Electrical Transmission and Substation Structures
Electrical Transmission and Substation Structures
Electrical Transmission and Substation Structures
SPONSORED BY
EDITED BY
Preface
These proceedings contain papers presented at the 2009 Electrical Transmission and
Substations Structures Conference sponsored by the Structural Engineering Institute
of the American Society of Civil Engineers, held in Fort Worth, Texas, November 812, 2009. This is the fourth specialty conference on utility structures, and begins an
attempt to hold the conference on a three-year cycle. The conference originated to
provide opportunities for utility engineers to gain further understanding of the
industry through technical papers, workshops, plant tours, and networking with other
utility professionals.
The electric utility industry faces the impending retirement of approximately half its
workforce in the next 10 years. The planning committee designed every aspect of this
conference to provide value and knowledge transfer to those in the transmission and
substation structures sector of our industry. We specifically focused on the needs of
the newer, less-experienced engineers and technicians, hence the title, Technology
for the Next Generation. However, there was also plenty of value for seasoned
veterans.
The conference included formal sessions on loadings, analysis and design,
foundations, substation design, applied technology, and case studies. The session
chairs and editor reviewed all papers included in these proceedings. All papers are
eligible for discussion in the Journal of Structural Engineering. The papers are also
eligible for ASCE awards.
Marlon W. Vogt, PE, Editor
Acknowledgments
Preparation of the 2009 Electrical Transmission Conference required time, effort, and
dedication by members of the conference steering committee and staff at ASCE/SEI.
Therefore, much of the success of the conference is a reflection of the level of effort
by this group of volunteers.
The committee would also like to acknowledge the critical support of the sponsors,
exhibitors, and participants who contributed to the success of the conference through
their participation.
SEI Staff
James Rossberg
Debbie Smith
Barbara Hickman
Donna Dickert
Sen Scully
Joanna Colbourne
Sponsors
Platinum
American Electric Power
Oncor Electric Delivery
Premium
FWT, Inc.
Power Line Systems, Inc.
Thomas & Betts Corporation
Valmont Newmark
Deluxe Plus
Electrical Consultants, Inc.
Sabre Tubular Structures
Deluxe
Black & Veatch
LOCWELD
POWER Engineers
SAE Towers
Basic
DIS-TRAN
Fabrimet, Inc.
Falcon Steel Company
KEC International Ltd.
Underground Devices, Inc.
Exhibitors
3M High Capacity Conductor
AIR2
Birmingham Fastener Manufacturing
COMENSA
Commonwealth Associates, Inc.
Copper Care Wood Preservatives, Inc.
Crux Subsurface, Inc.
DiGioia, Gray & Associates LLC
DIS-TRAN
Electrical Consultants, Inc.
Fabrimet, Inc.
Table of Contents
Loading I
Are Wood Poles Getting Weaker?
Brent Baker and Michael D. Brown
The Chaotic Confusion Surrounding Wood Equivalent Non-Wood Poles
Wesley J. Oliphant
Estimation of Containment Loads on a 230kV Steel Transmission Line Using Finite
Element Model
Asim Haldar, Maria Veitch, and Trevor Andrews
Common Sag and Tension Errors: Its Time for Template Technology to be Put in the
Drawer Forever
Otto J. Lynch
Foundations
Integration of Optimum, High Voltage Transmission Line Foundations
Freeman Thompson, Nick Salisbury, Anwar Khattak, Aaron Hastings, and Michael Foster
Geotechnical Investigations for a Transmission Line Are More Than Drilled Borings
Kristofer Johnson
Transmission Pole Foundations: Alternate Design Methods for Direct-Embedded Round,
Wood Transmission Poles
Cassie McNames and Sivapalan Gajan
Assessment and Repair of Steel Tower and Steel Pole Foundations
Nelson G. Bingel III and Kevin D. Niles
Substation Design
Substation Bus Design: Current Methods Compared with Field Results
Keith Malmedal and John Wendelburg
3D Automated Design of Substations
D. R. Lemelin, C. Bgel, C. Benavides, C. Lacroix, and Y. Drolet
Seismic Design of Substation Structures
E. Bazn-Zurita, J. Bielak, A. M. DiGioia Jr., and S. Jarenprasert
Analytical Techniques to Reduce Magnetic Force from High Fault Current on Rigid Bus
T. A. Amundsen, J. L. Oster, and K. C. Malten
Loading II
Wind Loading: Uncertainties and Honesty Suggest Simplification
Alain H. Peyrot
Review of Span and Gust Factors for Transmission Line Design
Roberto H. Behncke and T. C. Eric Ho
Wind Load Methodologies for Transmission Line Towers and Conductors
Leon Kempner Jr.
The Effects of Ice Shedding on a 500 kV Line
Alan B. Peabody and Ron Carrington
Case StudiesFoundations
Golden Pass LNG 230kV Double Circuit: Foundations
Ernest C. Williamson Jr. and Ronald L. Rowland
Deepwater Transmission Line Foundations Meet Trophy Bass Lake Environment
Applied Technologies
Vegetation Management Through LiDAR Derived CADD Models: Compliance with NERC
Reliability Standard FAC-003-1
N. Ferguson, S. Ryder, S. Verth, P. Richardson, and R. Ray
Sequential Mechanical Testing of Conductor Designs
Bruce Freimark, Bruce Vaughn, Jean-Marie Asselin, Craig Pon, and Zsolt Peter
H2S Entrapment and Corrosion on Direct Embedded Galvanized Steel Transmission Poles
Ajay Mallik and Andy Cooper
Line Rating: Its All about the Temperature!
Ryan Bliss
Posters
Dynamic Wind Analyses of Transmission Line Structures
Ferawati Gani, Frederic Legeron, and Mathieu Ashby
Temporary Support of Lattice Steel Transmission Towers
Scott Monroe, Rodger Calvert, and John D. Kile
Power Restoration Solution after Major Cyclone: Gonu Hit Oman in June 2007
Deepak Lakhapati
Analysis Challenges in the Evaluation of Existing Aluminum Towers
Rengaswamy Shanmugasundaram, Jason W. Kilgore, and Roger Calvert
Electrical and Mechanical Considerations in 765kV (Insulator) Hardware Assembly
Design
John Kuffel, Ziqin Li, Bruce Freimark, and Teja Rao
Strength of Steel Angles Subjected to Short Duration Axial Loads
Jared J. Perez, Wendelin H. Mueller III, and Leon Kempner Jr.
Residual Capacity in Compression of Corroded Steel Angle Members
L. -V. Beaulieu, F. Lgeron, S. Langlois, and S. Prudhomme
Steel Structure Surface Preparation for Below Grade Coatings
Edward M. Jacobs
Impact of Alternative Galloping Criteria on Transmission Line Design
D. Boddy and J. Rice
Civil Engineer, The Empire District Electric Co., 602 S. Joplin Ave., Joplin MO
64802, PH (417) 540-6028; email: [email protected]
2
Civil Engineer, Allgeier, Martin and Associates, Inc., P.O. Box 2627, Joplin MO
64803, PH (417) 499-1009; email: [email protected]
ABSTRACT
This paper follows the evolution of wood pole design over the last 80 years and an
example is used to demonstrate how the relative safety factor has increased. A brief
review of the changes in the National Electrical Safety Code (NESC) strength
requirements from 1927 to 2007 is presented to show how little it has changed.
Technological advances in calculators, computers and computer software make it
possible to analyze complex problems. Equations that were once simplified to
facilitate computation are now expanded to include components previously ignored.
However, the NESC has not adapted to incorporate these changes, resulting in more
conservative designs.
INTRODUCTION
Have you noticed that both transmission and distribution wood pole line designs
dont seem to be what they were? Span lengths seem shorter and pole classes seem
larger than they used to be? Have the codes changed that much or are the poles just
weaker than they were? We submit for your consideration that the methods used for
design today are much different than they have been over the last 80 years, resulting
in more conservative structure design. The purpose of this article is to walk you
through the evolution of wood pole design and demonstrate how the relative factor of
safety has increased over the years.
While there have been changes in the National Electrical Safety Code (NESC), the
strength requirements have changed very little between 1941 and 2007. In 1927 there
were three grades of construction, A, B, and C with overload factors (safety factors)
of 3, 2 and 1.33 respectively. The Heavy Load case was an 8 psf wind with radial
ice. In 1941 the grades were changed to Grade B, C and C at Crossings with
overload factors increased to 4, 2 and 2.67 respectively. The Heavy Load case was
changed to 4 psf wind with radial ice and the k factor was introduced (NESC
Rule 250B). In 1977 the extreme wind case was introduced (NESC Rule 250C). In
1990 an alternate method was introduced as an option that a designer could use. The
alternate, Method B, was an attempt to acquaint line designers with a Load and
Resistance Factored Design (LRFD) method as the first of several steps towards
completely revising the strength requirements for wood poles in future editions of the
NESC. The 1997 edition made the LRFD method the preferred method, but left the
overload factors as the alternate method. The 2002 edition revised the extreme wind
case requirements. The 2007 edition introduced an extreme ice with concurrent wind
load case (NESC Rule 250D). The 2007 edition also eliminates the alternate overload
factor method in 2010, leaving the LRFD method as the only method after that time.
Although sweeping changes were proposed for the 2007 NESC, the majority were not
approved. It appears that change continues to generate a lot of controversy and
opposition within the industry. Work is now underway on the 2012 NESC. Change
proposals were due by July 2008 and a preprint of proposed changes was due in
September 2009. The Public Comment period will end on May 1, 2010. It is
anticipated that many changes will be proposed for the strength and loading
requirements of Part 2 of the NESC. It is also anticipated that the controversy and
resistance to change will continue.
Of interest from a historical perspective is that the ultimate fiber stress used for poles
was increased sometime between 1961 and 1972, but pole circumferences six feet
from the butt were reduced, resulting in about the same groundline moment capacity.
EXAMPLE 1
For illustrative purposes a simple single pole, single circuit 69kV structure with 4/06/1 ASCR conductors and a 3/8 HSS shield wire has been used. The pole top
assembly, or crossarm configuration, is a TS-1 assembly as shown in Figure 1. The
pole and wire data used in all of the calculations are set forth in Tables 1 and 2
This assembly has been a standard Rural Utilities Service (RUS), formerly the Rural
Electrification Administration (REA), assembly for over 60 years. This assembly
was selected as thousands of miles of this type of construction has been used by both
investor owned utilities and rural electric cooperatives. Many of those lines have
been rebuilt or upgraded in some way over the years, but many are still in service.
Some would consider this a standing testimony to the reliability of this type of
construction and the original design methods.
Table 1.
POLE DATA
Species (ultimate bending stress = 8,000 psi)
Height - Class
Top Circumference (in)
Groundline Circumference (in)
Moment Capacity at A (Ft.-Lb)
Moment Due to Wind on Pole (Ft.-Lb.)
SYP
55-1
27
45.9
204,200
3,985
SYP
70-1
27
49.9
261,900
6,835
Table 2.
CONDUCTOR LOADS, UNFACTORED
NESC HEAVY LOADING
VERTICAL
TRANSVERSE
3/8 HSS (LB/FT)
0.8077
0.4533
4/0 6/1 ACSR (LB/FT)
0.9520
0.5210
Moment
21.26HS
22.92HS
39.60HS
83.78HS
The maximum horizontal span, HS, is found from the following equation:
204,200 ft.-lbs / 4 = 83.78HS
therefore HS = 609 feet
METHOD 2. The second design method is the same as Method 1except that wind
on the pole is considered. This method was built into the chart used in the 1950 REA
Transmission Line Manual, but no example or explanation was included. Both the
1961 and 1972 REA Bulletin 62-1 used this method as well. For a 55-1 pole the
maximum horizontal span, HS, is found from the following equation:
204,200 ft.-lbs / 4 = 83.78HS + 3,985 ft.-lbs.
therefore HS = 562 feet
METHOD 3. The third design method refined Method 2 by including the effects of
the unbalanced vertical loads due to the unsymmetrical placement of wires. RUS
Bulletin 62-1 (1980) indicated that this could impact span length by 2 to10 percent,
but left it to the engineers judgment if it was significant enough to include. Again
for a 55-1 pole, the maximum horizontal span (HS) is found from the following
equation assuming that the vertical or weight span is equal to the horizontal span
(which is assumed for all the following methods):
204,200 ft.-lbs / 4 = 83.78HS + 3.5HS + 3,985 ft.-lbs.
therefore HS = 539 feet
METHOD 4. This method is the same as method 3, except that it was done with the
aid of a computer program, PLS-POLE (2000). It was done as a basis of comparison,
as the rest of the methods addressed by this paper were done using PLS-POLE
(2000).
The resulting HS = 533 for the 55-1 pole.
It should be noted that all previous methods ignored the axial stress component (total
vertical load divided by the cross-sectional area of the pole), but the program does
account for it. The difference between methods 3 and 4 is attributed, in part, to that
axial stress component and to minor differences in the calculation of wind load and
pole circumferences. The difference of 1.1% is note worthy, but it certainly is minor
when compared to the differences obtained with the other methods.
METHOD 5. This method is identical to Method 4 except that the moment capacity
is checked at five-foot increments up the pole as well as at the groundline.
The resulting HS = 533 for the 55-1 pole, indicates that the maximum moment
occurred very near the ground line. A second example using a 70 pole, which will
be discussed later, had a lower horizontal span indicating that the maximum stress did
occur above the groundline.
METHOD 6. This method takes into account the P-Delta () or secondary moment
effects. Horizontal loads cause the pole to deflect a distance of delta (). The
resultant vertical load P, times the distance results in an additional moment
component or secondary moment. This secondary moment causes additional
deflection with an additional moment component. This non-linear behavior increases
the moment the pole must resist and greatly increases the math the engineer must deal
with. Hence the use of the computer program for the rest of the methods.
An equation for pole deflection appeared in the REA Bulletin 62-1 (1961), but the
secondary moment was not included in span length calculations until much later.
REA Bulletin 62-1 (1980) indicates that secondary moments are usually not
accounted for in wood pole designs stating: The high overload factor of four for
heavy, medium and light district loadings is intended to keep the design simple for
low height structures and in line with known strength, foundation response and
loading conditions. For tall single pole structures, the designer may want to increase
the OCF for NESC district loadings and high wind loadings in order to account for
the additional moment due to deflection.
RUS Bulletin 1724E-200 (1992) provides example calculations with and without
consideration of the secondary moments, that are linear and non-linear calculations.
The manual further indicates that many designers incorporate secondary moments in
their design, but that manual did not provide recommendations either way on it.
A general equation for determining the maximum groundline moment is:
MA (OCF)Mwp + (OCF)Mwc + (OCF) Mvo + (OCF) Mp
where:
MA = the ultimate moment capacity of the pole at groundline
Mwp = the groundline moment due to wind on the pole
Mwc = the groundline moment due to wind on the wires
Mvo = the groundline moment due to unbalanced vertical loads
Mp = the groundline moment due to secondary moments
OCF = the overload capacity factor
Note that the overload capacity factor, as it was called at that time, is applied after the
secondary moment is calculated and that the OCF could be applied to either side of
the equation. While the axial stress component (P/A) is not shown (it is not
considered in any of the RUS Manuals), it is incorporated in the computer analysis
results provided in this paper.
The resulting HS = 497
METHOD 7. This method has been included to show what would happen if one
used Method 6, but placed the entire safety factor on the load side of the equation
before calculating the secondary moment. In other words, instead of calculating pole
deflection based on a transverse load of T, one used 4T to get pole deflection.
The resulting HS = 414
METHOD 8. In 1990 the NESC introduced an alternate method, called Method B,
which broke the traditional safety factor or Overload Factor into a strength factor
component and a load factor component. In accordance with the 2007 NESC, the
LRFD method is the only method to be used after July 2010. This LRFD method is
fully embraced by the RUS Bulletin 1724E-200 (2005). By the time that manual had
been published, RUS was requiring that secondary moments be considered as part of
the design on most single pole structures. However, even the 2007 NESC does not
require secondary moments be included. NESC Rule 260.A.1 states that allowance
may be made for deflection when the effects can be evaluated.
A general equation for determining the maximum groundline moment is:
MA (LF)Mwp + (LF)Mwc + (LF)Mvo + (LF)Mp
where:
= strength factor, 0.65 for wood poles
LF = load factor, varies with the type of load (2.5 for wind loads, 1.5
for applied vertical loads and 1.0 for dead load or self weights
Again note that the load factor is applied after calculating the secondary moment,
which is how RUS intended it to be. NESC Rule 260.A.1 (2007) states that Such
deformation, deflection, or displacement should be calculated using Rule 250 loads
prior to application of the load factors in Rule 253. Clearly the intent is to base
deflection on an unfactored load, however that is easier said than done. Unless pole
strength calculations are done using approximations to simplify manual or
spreadsheet calculations, most programs are written to apply the load factor to the
loads before calculating deflection. In most other standards and codes, the intent is to
apply the load factor first since the load factor accounts for the uncertainty in
predicting the load in various return periods. Most other codes have load factors of
1.6 or less instead of the 2.5 required for wind loads when using Grade B
construction. Methods 8 through 11 below have the load factor applied before
calculating deflection since PLS-POLE (2000) was used for the calculations. The
results are therefore more conservative than what the NESC would require or that
RUS would recommend.
It should be noted that Methods 1 through 7 used a traditional safety factor of 4,
regardless of what side of the equation it was applied, with the exception of dead
load. The dead load or self-weight of the pole and structure was not even considered
in Method 1 thru 3 and Methods 4 thru 7 used a load factor of 1.0. Methods 8
through 11 are based on the LRFD method and the equivalent safety factor is 3.86 for
wind loads (2.5/0.65) and 2.3 (1.5/0.65) for applied vertical loads. Method 8 uses a
load factor of 1.0 for dead load for comparison to the previous methods.
The resulting HS = 534 is slightly higher than Method 6 due to the difference in
overall safety factors.
METHOD 9. This method is just like Method 8 except that the Load Factor for the
dead load has been increased from 1.0 to 1.5. The 1.5 load factor is in compliance
with the NESC and thus this method probably better represents what most designers
that use a non-linear analysis are doing.
The resulting HS = 507
METHOD 10. ANSI 05.1 (2002) incorporated a new requirement in Section 9 that
recommends the ultimate fiber bending stress be reduced with pole height. 100% of
the published fiber stress can be used at groundline and the fiber stress should be
reduced linearly to 75% of the published value at half the pole height above ground.
Previously this recommendation had been included in a non-binding appendix of
ANSI 05.1 (2002). A major effort was made to adopt this provision in the NESC C2
(2007), however that effort was not approved and the NESC remains based on the
ANSI 05.1 (1992). It is anticipated that this issue will be a proposed change for the
next NESC. The reduction recommended by ANSI 05.1 (2002) is incorporated in
Method 10.
The resulting HS = 430. Remember in Method 4 and 5 that the maximum stress
occurred at the groundline. However this example demonstrates that the reduction in
maximum fiber stress would require designers to check the portion of the pole above
ground and will probably impact most designs.
METHOD 11. As discussed previously, the advent of the computer has allowed or
enabled us to do things that we normally would not be able to do. This method builds
on the previous method by adding in the weight of the crossarms, braces and
insulators and includes the wind load on all of those components. We can even apply
the correct drag factors for the flat surfaces of the crossarm and the round surfaces of
the insulators.
The resulting HS = 423 is 1.6% less than the Method 10 results and is about the
same order of magnitude as that observed in Method 4 over Method 3.
EXAMPLE 2
The previously described methods are used on the same TS-1 structure with a 70-1
pole. Logic would indicate that taller poles will deflect more and therefore will be
impacted more by methods that include secondary moments. The results are shown
in Table 4 and bear out that logic.
0.65
0.65
10
0.65**
11
0.65**
Relative Factor
Of Safety
1
1
1
1
1
1
4.0
V*
2.5W
1.5V
2.5W
1.5V
2.5W
1.5V
2.5W
Horizontal Span
0.25
0.25
0.25
0.25
0.25
0.25
1.00
Relative Factor
Of Safety
N
N
N
N
N
Y
Y
70-1
Horizontal Span
Non-Linear
Y
Y
Y
Y
N
N
N
Load Factor
Groundline Check
Only
1
2
3
4
5
6
7
Strength Factor
Method
55-1
609
562
539
533
533
497
414
3.7
4.0
4.2
4.2
4.2
4.5
5.4
590
528
512
501
493
447
354
3.6
4.0
4.1
4.2
4.3
4.7
6.0
534
4.2
477
4.4
507
4.4
437
4.8
430
5.2
352
6.0
423
5.3
345
6.1
* Load Factor of 1.0 on dead load and 1.5 on applied vertical loads
** Fiber Stress reduced in accordance with ANSI 05.1(2002)
CONCLUSIONS
While wood poles may not be getting weaker, it certainly gives that appearance from
the results presented herein. Based on these results the following conclusions can be
drawn regarding the design of single pole wood structures:
the NESC heavy loading has changed very little
the method of calculation has changed drastically
the method of calculation has a major impact
secondary moments have a significant impact
the application of strength and load factors has a major impact on secondary
moments
ANSI 05.1 fiber stress reduction has a major impact
wood poles are not getting weaker
The methods presented by this paper are not intended to represent all methods that
have or could be used to determine the maximum horizontal span. Other methods or
different assumptions will result in different relative factors of safety. Other load
cases or the use of Grade C construction will also result in different relative factors of
safety. For example, structures controlled by extreme wind load cases may display a
much narrower range of relative factors of safety since a load factor of 1.6 or less is
normally used instead of 2.5, thereby giving more realistic deflections and secondary
moments.
CLOSING THOUGHTS
If thousands of miles of a structure design have been installed and those lines
performed well structurally for over 50 years, logic would indicated that the design
was satisfactory or perhaps even over designed. Yet all but one of the design
methods considered in this paper resulted in a more conservative design. Is it time to
re-evaluate our design method? Should we adjust the strength and load factors to be
consistent with the design method?
The advent of the computer and the ingenuity of software developers have made it
possible to analyze complex problems that could not be done previously. The
engineer must decide how best to use that technology to manage our resources while
protecting the safety and well being of the public. It needs to be recognized that
more conservative designs may not be in the best interest of the public. They may
just be more expensive and a waste of resources. The wide variety of design methods
in use and the wide margin in relative factors of safety indicate the need for design
standards. Perhaps the proposed ASCE Manual of Practice Design of Wood Utility
Pole Structures will be that standard.
Perhaps Dr. A.R. Dykes summed it up well when he said Structural Engineering is
the Art of moulding materials we do not wholly understand into shapes we cannot
precisely analyze, so as to withstand forces we cannot really assess, in such a way
that the community at large has no reason to suspect the extent of our ignorance.
ACKNOWLEDGEMENTS
A special thanks to Donald Heald of Rural Utilities Services for his assistance in
providing older versions of the REA Transmission Line Manual. The authors also
extend their gratitude to their employers, The Empire District Electric Company and
Allgeier, Martin and Associates, Inc. for the resources they provided in support of
this work.
REFERENCES
ANSI C2(2007), National Electrical Safety Code, Institute of Electrical and
Electronics
Engineers, Inc., New York, NY
ANSI O5.1(2002), American National Standard for Wood Poles Specifications and
Dimensions, Alliance for Telecommunications Industry Solutions,
Washington, DC
PLS-POLE(2000), Power Line Systems - Analysis and Design of Pole Structures,
Power Line Systems, Inc., Madison, WI
U.S. Dept. of Agriculture, Rural Electrification Administration, Transmission Line
Manual Physical Design Consideration, Government Printing Office,
Washington, D.C
U.S. Dept. of Agriculture, Rural Electrification Administration, Bulletin 62-1(1961),
Transmission Line Manual, Government Printing Office, Washington, D.C
U.S. Dept. of Agriculture, Rural Electrification Administration, Bulletin 62-1(1972),
Transmission Line Manual, Government Printing Office, Washington, D.C
U.S. Dept. of Agriculture, Rural Electrification Administration, Bulletin 1724E200(1980), Design Manual for High Voltage Transmission Lines,
Government Printing Office, Washington, D.C.
U.S. Dept. of Agriculture, Rural Utilities Services, Bulletin 1724E-200(1992), Design
Manual for High Voltage Transmission Lines, Government Printing Office
Washington, D.C.
U.S. Dept. of Agriculture, Rural Utilities Services, Bulletin 1724E-200(2005), Design
Manual for High Voltage Transmission Lines, Government Printing Office,
Washington, D.C.
10
11
12
species of wood poles. What might be less known and even confusing, is that during
the process of selecting a wood pole class, particularly for larger transmission poles, a
number of strength adjustments typically need to be made to account for
characteristics and processes influencing pole strength and stiffness. One of these
adjustments is the height adjustment. On poles taller than 50 ft. the maximum
cantilever stress will typically occur some distance above the groundline, not at the
groundline as most assume. It is also important to note that the fiber stress values
shown in ANSI 05.1 represent consensus values (a combination of clear specimen
tests, full size pole tests and engineering judgment).
All of this is important to keep in mind as we make very specific strength
comparisons between wood poles and steel, concrete and fiber reinforced polymer
poles. Keep in mind, the ANSI loads below are intended to be 50 percentile or
mean strength values.
13
14
Table 2 below shows the strength factors as they apply to structures from table 2611A of the NESC.
2007 NESC Strength Factors for Structures
Wood & Reinforced Concrete
(not prestressed)
Steel
Concrete
(prestressed)
FRP
0.65
1.0
1.0
1.0
0.85
1.0
1.0
1.0
0.75
1.0
1.0
1.0
Table 2 - Strength Factors for Structures (excerpted from Table 261-1A of the
2007 NESC)
Steel, prestressed concrete and fiber reinforced polymer materials used for structures
can use the same strength factor, irrespective of the NESC loading rules. Wood poles
and reinforced concrete poles (not prestressed) on the other hand, have variable
strength factors depending on the NESC load condition. And to what nominal
strength do these factors apply 1%, 5%, 50% (mean), LEL values? Why is it that
one pole material is variable, and all other pole materials are fixed? If this is an
attempt to inject differing levels of reliability into the design of structures, there are
better ways of dealing with this. But currently, we have confusion, and chaos.
And of course the confusion and chaos is amplified when we try to compare a
standardized steel, prestressed concrete or FRP pole load capacity to a wood pole load
class derived from such a crazy methodology. If the pole is originally designed as a
steel pole to begin with, we will use a 1.0 strength factor no matter what the loading
rule. If the pole is originally sized as a wood pole, varying strength factors would
have been used depending on the loading condition considered. Does it really make
good structural engineering sense that we can utilize 50% more of the steel poles
capacity than a wood pole when considering NESC rule 250B, grade B loadings, but
only 18% more capacity than a wood pole under NESC rule 250B, grade C loadings?
I propose this does not make much sense.
One can argue about the coefficient of variation (COV) being much larger with wood
poles than with the other non-wood pole materials. With that COV difference, a
larger number of poles, because of the higher variability, will exceed the required
loading, assuming that all poles are based on the same nominal strength rating. But
that doesnt make much sense either. I believe the goal should be to have the same
reliability against failure at a specific load no matter what the material used. We set
the loads, and we design poles to withstand those loads. We want to maintain a
certain confidence that every pole will hold the required load and we can do that if we
standardize the way we define nominal strength across all materials.
Most materials define strength as a 5% lower exclusion limit with a 50% confidence
level. We should do that with all poles no matter the material and we will gain
consistency in how we discuss nominal strength. ANSI 05.1 2008 is now providing
the tools to calculate the strength of wood poles at a 5% lower exclusion limit with a
50% confidence. We need to standardize on this for wood poles and as well all other
pole materials.
The Rural Utilities Service (RUS) has provided wood pole equivalency guidelines
and established standard classes of both steel poles and prestressed concrete poles.
The table below shows the standard class steel poles. The standard class prestressed
concrete poles have the same strength requirements as does the steel pole standard
classes. In addition to the tip load requirements, RUS also has several other
requirements that are important considerations. These requirements are a defined
Point-of-Fixity (POF) that is a distance from the pole bottom that equals 7% of the
pole length, and a minimum strength specification at 5 feet from the top of the pole.
Both are excellent requirements. But a catalog of non-wood poles designed to meet
the RUS requirements of these tables, would meet requirements that are not typically
considered with the selection of wood poles. Until we require wood poles to meet
the same criteria, there will continue to be confusion of wood pole equivalency
comparisons when these tables are utilized.
15
Durability:
Durability and strength are related. When structures deteriorate for whatever reason,
they can lose strength. As in real estate, it is location, location, location. A
woodpecker hole can be relatively benign structurally unless our friend decides to
make a full fledged condo out of the top 20 feet of a pole. The same principle also
applies to corrosion on a steel pole. A small bare spot mid way up the pole does not
16
necessarily reduce strength below that required of the loads. But, losing more than
33% of the wall thickness at groundline will definitely need to be considered
significant and analyzed with respect to the expected load.
But how is durability treated with regard to wood pole equivalency? NESC Rule
250B allows for the deterioration of wood poles up to 67% of its original installed
strength. However, non-wood poles must maintain their original required
strength throughout their design life. In Rule 250C & 250D deterioration of wood
poles is allowed up to 75% of its original installed strength. Again, non-Wood Poles
must maintain the original required strength throughout their design life. If a
calculated strength is required to achieve a given design life, then why do we allow
reduced strength allowances for wood poles, but not for steel, prestressed concrete,
and/or FRP poles? Does that make sense?
Stiffness:
In my experience, the issue of stiffness is the least understood structural characteristic
and most likely ignored when making wood pole equivalent comparisons.
Obviously there are pole top deflection issues.
Wood poles, and steel poles poles deflect in a typical linear fashion. Apply a load
and the tip of the pole will deflect a certain distance. Apply 10 times that load and
the tip will deflect roughly 10 times the original distance. Depending on the height of
the pole, wood poles will experience pole top deflections typically in the 10%-15% of
the pole height at rupture. Steel poles can deflect 10%-20% of the pole height at
ultimate load. There are stiffness differences.
Prestressed concrete poles have a variable modulus of elasticity. This is due to the
prestressing effects and the composite nature of the combined concrete and steel.
Concrete poles are very stiff and exhibit very minimum deflection, typically 2%5% of the poles height for up to 50% of the poles loading range. Beyond
approximately 50% of the ultimate loading, a prestressed concrete pole will become
very elastic and pole top deflections of 10%-15% of the pole height will not be
uncommon.
FRP poles can also be variable in stiffness depending on the nature of the composite.
Stiffness is a key design consideration for FRP poles. Especially when considering
FRP poles as wood pole equivalents special consideration should be given to
limiting the deflection to that of the wood pole.
So all poles deflect differently. But lets also discuss the other stiffness related issues
that I believe are even more critical. Substituting wood pole equivalent steel and
concrete structures for existing wood pole guyed structures and multipole structures
such as H-Frames, 3-Pole structures, etc should be done with great care. There is a
stiffness interaction in these types of structures that may cause guy forces to change
17
18
factors dependent on load conditions. Make the load factors variable (50 year
recurring events, 25 years recurring events, etc.) if differing levels of
reliability are desired.
4. Alternatively, NESC should just consider deferring structural design standards
altogether to organizations such as the Structural Engineering Institute of
ASCE.
5. Include in ANSI 05.1 the effects of strength degradation over time due to
durability issues, or work to exclude from NESC the currently allowed
strength reductions that apply only to wood poles.
REFERENCES:
1. Institute of Electrical and Electronic Engineers, Inc., National Electric Safety
Code, C2-2007, IEEE, New York, NY
2. American National Standards Institute, ANSI 05.1-2008, American National
Standard for Wood Poles, Specifications and Dimensions ANSI, New York,
NY.
3. Wolfe, Ronald W.; Bodig, Jozsef; Lebow, Patricia. 2001. Derivation of
Nominal Strength for Wood Utility Poles. Gen. Tech. Rep. FPL-GTR-128
rev.). Madison, WI: U.S. Department of Agriculture, Forest Service, Forest
Products Laboratory. 11 p.
4. Rollins, H.M. A Discussion of Wood Equivalent Poles, A technical
bulletin prepared for the North-American Wood Pole Coalition, January 2007
5. United States Department of Agriculture, Rural Utilities Service, Bulletin
1724E-214, Guide Specification for Standard Class Steel Transmission
Poles.
19
20
This figure also shows the location of the dead end towers on the two parallel lines
(Line 1 and Line 2) considered in this study and the many occurrences of the line
failures during the past forty years. It is seen that a large section of the two steel lines
does not have an adequate number of dead end towers to provide security against
cascade failures. Note that the steel transmission line system on the Avalon Peninsula
was upgraded during 1999-2004, considering cascade failure protection (Haldar,
2006). The dead end towers on Line 1 and Line 2 were not upgraded.
Figure 1 Location of transmission lines Figure 2 Glaze ice sample (1984 ice storm)
Lines 1 and 2 were designed for (1) 25.4 mm radial glaze ice, (2) 13 mm
radial glaze with 117 km/h wind and (3) 160 km/h wind respectively. An overload
factor of 1.33 was used. The conductor is 795 MCM 26/7 ACSR and the ruling span
in the long section is 430 m. The final tension under everyday condition at 40 C is
27.8 kN. The ratios of span to sag and span to insulator length are 32 and 200
respectively.
EXISTING TOWER CAPACITY ENVELOP DIAGRAM
The existing suspension tower was designed for a longitudinal load of 17.8
kN with a 62.0 kN vertical load on each phase or their combinations. Analysis of a
typical suspension tower shows that with no ice, the actual longitudinal capacity
( Rn ) under static loading is approximately 30.0 kN per phase or their combinations.
This capacity does not include any strength factor ( =1.0). Figure 4 depicts an
interaction diagram of vertical and longitudinal limit loads on the tower. The static
capacity can be increased further under short duration dynamic load (Perez et al,
2008) and this issue will be discussed later.
SCOPE OF THE PAPER
The objective of this paper is to estimate the length of a cascade failure zone
due to a broken conductor scenario. The length is estimated based on a finite element
21
22
model of a typical 230 kV line section where both transient (peak dynamic) and
steady state (quasi static residual) loads are evaluated. The numerical model considers
the following cases: (1) first tower next to the break is intact and (2) first tower fails
and the force redistribution in the remaining line system is assessed.
Next, a scenario with more than one tower failure is considered. Based on the
location of the first surviving tower after a failure, the most probable extent of the
failure zone is estimated and the mitigation to strengthen the tower to withstand the
required containment load is proposed.
BAY D'ESPOIR
DAMAGED AREA
(ONE OCCURANCE - 1970)
SUNNYSIDE
DAMAGED AREAS
(TWO OCCURANCES 1970 & 1984)
TERMINAL STATION
LOCATION OF DEADEND
TOWERS ON LINES 1 AND 2
230kV WOOD
POLE LINE
LEGEND:
230kV STEEL
TOWER LINE
(AREA OF INTEREST)
DAMAGED AREAS
(FOUR OCCURANCES 1984, 1988, 1990 &1994)
DAMAGED AREAS
(BETWEEN 1970 TO 1994)
DAMAGED AREAS
(ONE OCCURANCE - 1984)
ST. JOHN'S
Figure 3 Location of the Dead End Towers on the Existing Two Lines
80000
60000
40000
20000
0
0
10000
20000
30000
Longitudinal Load (N)
40000
Figure 4 Static Tower Capacity Envelope Diagram under Vertical and Longitudinal
Loads
23
the peak load measured on the first tower adjacent to the broken conductor span, in
Figure 6a (T1C). Kempner (1997) showed that this impact factor decreased as the
number of failed towers increased. If the first surviving tower withstands the peak
impact load after a limited number of tower failures, then the cascade will be
prevented and the extent of the failure zone can be estimated assuming symmetry on
both sides of the initial failure point. The variation of impact factor with number of
failed towers will be shown later.
24
McClure and her colleagues (2006) at McGill University have used ADINA
extensively to predict the dynamic load on overhead lines due to conductor breakage
as well as to study ice-shedding problems. The detailed methodology for modeling
overhead lines using ADINA is given in CEATI report (McClure, 2006) and in
Tucker (2007). The latter guide was used to develop and validate a numerical model
with full scale test data to study the failure of a suspension insulator string and its
effect on adjacent towers under dynamic loading condition (Tucker and Haldar,
2007).
METHODOLOGY- FINITE ELEMENT MODEL (FEM) IN ADINA
The general outline for transient response analysis in ADINA requires the
development of a finite element (FE) model of the line followed by a static run to
ensure that the initial position of the cable geometry under gravity load (self weight)
is captured. A routine check on the sag and the tension provides the basis for
assuming that the model is correct with respect to initial tension, axial rigidity (EA)
and the boundary conditions. Since the ADINA software used for this study had a
restriction of 900 nodes, it was important that the model be kept simple to provide
insight into the problem for determining whether the development of a more complex
model is warranted.
The following provides the highlights of the procedure that was used in the
present study to develop an ADINA model.
The conductor is modeled as truss elements with initial pretension.
The conductor material can be modeled as elastic (final modulus) or non
linear elastic (initial modulus).
The tower masts and the bridge are modeled as linear elastic beam elements
with appropriate stiffness while the guys are modelled as cable elements with
axial rigidity and initial tension.
The load is defined as mass proportional.
The static model is run first; the result is saved and used as the restart
condition for the transient dynamic analysis.
The transient dynamic analysis is initialized at 1.0 second and the cable
death element (element removed from the model to simulate a broken
conductor) is activated at 1.002 second to simulate the conductor break.
Once the static analysis is completed, the transient response analysis is carried
out by invoking the death element option in ADINA which allows the simulation of
a conductor rupture at any location. If this option is used, the corresponding element
group automatically becomes nonlinear. When the conductor element dies, the
insulators at the break point, as well as insulators in the other spans, are free to swing
25
fully. A step by step time step integration ( t = 0.002 second) of the equations of
motion of the discrete system provides the time history responses of various
parameters such as displacement, conductor tension, force in the insulator string and
on the tower, etc.
VALIDATION OF IMPACT FACTORS - A TWO SPAN MODEL
In order to validate the results, a simple two span model (Figures 7 and 8) is
used to understand the accuracy of the results with respect to number of elements
considered and its impact on various line parameters such as sag, tension and
frequency. Figure 9 compares the impact factor values for conductor tension obtained
from the analysis with those determined from the full scale EPRI tests (Peyrot et al,
1980). Two impact factors defined by Peyrot et al (1980) are used here for
comparison. These are Initial Impact Factor (IIF) and Final Impact Factor (FIF). The
IIF is defined as the ratio of the peak dynamic load to the conductor tension prior to
the break while the FIF is defined as the ratio of the peak load to the residual load.
ADINA model runs with 2% damping compares well with the test results. Figure 10
presents a typical time history plot of conductor tension at the tower adjacent to the
broken conductor span. The typical insulator swing was less than 800 degree under
steady state condition while under transient condition, this was more than 900 degree.
This section presents the methodology to assess the extent of the failure zone
due to conductor breakage. A section of the line is modeled with five equal level
spans under every day conductor tension. The failure is considered either in any one
single phase or their combinations. This provides five different failure scenarios: (1)
outside left phase (2) middle phase (3) outside and middle phases (4) two outside
phases and (5) all three phases. Analysis is first carried out assuming that the first
tower near the break is intact (conventional model) and later, allowing the first tower
to fail. Analysis also considers subsequent tower failures (Figure 6b).
26
Im p act F acto rs
7
6
IIF
FIF
4
3
2
1
0
No
1%
2%
Damping Damping Damping
ADINA
Model
ADINA
Model
ADINA
Model
Test 1
Test 2
Peyrot et Peyrot et
al.
al.
Figure 9 Comparison of Model and Full scale Conductor Tension Impact Factors
As indicated in the previous section, a line section of the steel line (Figure 3)
with five equal spans is modelled with cable elements for the conductor and the tower
guys, beam elements for the two tower masts and the bridge and a number of truss
elements for the insulator strings. Based on a sensitivity study, twenty cable elements
were selected for each span.
27
The model developed in the previous section is used to study the effects of the
loss of a phase conductor or their combinations on the intact towers (Figure 6a). The
ADINA model predicted that the longitudinal force on each tower decreases as the
towers are located farther from the failure location. The ratio of the longitudinal force
on each tower ( Fi ) divided by the force on the first tower next to the failure location
( F1 ) is shown in Figure 11. The longitudinal force on the towers decreases
significantly from tower 1 to tower 2 and continues to decrease at smaller amounts in
the remaining towers, conforming to a power trend. This trend was also noted by
Kempner (1997) in a small scale model test and is valid in general for broken
conductor case. If the tower closest to the failure location survives then it can be
assumed that all remaining towers will also survive.
ADINA model runs were made for an initial static tower capacity of 27.6 kN
( =0.9, Figure 4) under everyday loading condition. This capacity does not consider
any amplification factor under short duration dynamic load (Perez et al, 2008). Later
this static capacity was increased to 35.6 kN and 44.5 kN to reduce the extent of the
failure zone (lesser number of tower failures) and to assess the type of mitigation
(structural modification) that may be required to withstand the large peak forces. The
tower failure in ADINA is referred as the failure of the section of the bridge that
would be affected by the peak force from a conductor failure. This implies that a
single phase failure on one side of the bridge would only affect that side. Therefore,
the peak load in excess of the tower capacity imposed on the first tower near the
break will cause a tower failure and a significant amount of this longitudinal load will
be transferred to the next tower.
28
Fi / F1
0.8
0.6
0.4
2
Tower Number
Figure 12 Impact Factors on the Closest Surviving Tower (Middle Phase Failure)
However, the force on the tower is still larger than its static capacity (27.6 kN)
and therefore, an extrapolation is done to determine the acceptable number of tower
failures to ensure that the imposed force on the surviving tower is less than the static
capacity. In this case it is estimated that the failure of five towers on one side will
contain the peak load. However the line model could be enlarged to include more
spans to see the decrement of this peak force. The comparison plots with the results
29
30
presented by Kempner (1997) are also shown in Figures 12 and 13 for similar span to
sag and span to insulator ratios. The comparison is reasonable.
1.0
Fi / F1
0.8
0.6
0.4
2
Tower Number
One Phase
Failure
Side
One Phase
Failure Middle
10
8
2
10
10
8
Two Phase
Failure
Side and
Middle
10
8
6
Two Phase
Failure
Two Sides
Three Phase
Failure
10
8
6
10
8
6
31
One Phase
Failure
Side
One Phase
Failure Middle
Two Phase
Failure
Side and
Middle
Two Phase
Failure
Two Sides
Three Phase
Failure
A separate model run was also made by increasing the 44.5 kN static tower
capacity to 53.4 kN by including a dynamic capacity factor of 1.2 (Perez et al, 2008). This
was done to consider the effect of short duration peak load. Under this scenario,
ADINA model runs show that the line section will lose only one tower on each side of
the failure location (Table 2). It shows that the number of tower failures under cascade
situation can be reduced by strengthening the tower for increased static capacity and
taking into consideration the additional strength due to the effect of short duration force
on the tower (Perez et al, 2008). More information on the tower capacity under short
term dynamic loading is needed to model the event appropriately.
MEMBER FAILURE UNDER
40.0 kN and 50.0 kN LOAD
MEMBER FAILURE UNDER
50.0 kN LOAD
SECTION A
SECTION B
SECTION C
SIDE VIEW
BOTTOM VIEW
robust and is equally applicable with some ice loading on the conductor where the
initial tension and the ice weight of the cable would be larger. Reduction of the load
on the tower can be achieved further if one chooses to use a load reduction device
(Haldar and Tucker, 2008). This part of the work is still being studied.
REFERENCES
32
Vice President, Power Line Systems, Inc., 956 Marynell, Nixa, MO 65714
PH (417) 724-8292; FAX (417) 724-8492; email: [email protected]
Abstract
The calculation of proper sags and tensions of conductors and other wires are
the most vital aspect of transmission line design; structure heights and placements are
determined by the sags, structure designs are determined by the tensions, and lines are
constructed using both of these methods. Mostly gone are the days of plastic sag
templates. Or are they? Many use computer software programs thinking that they are
using Technology for the Next Generation, yet they merely duplicate the archaic
plastic template technology and thus their designs have sags and tensions that can
have significant error. Additionally, their lines are built using these same assumptions
and thus sags can be lower than anticipated or tensions can be higher than anticipated.
This paper will discuss common errors made in the industry and suggest
methods for more accurate sag and tension calculations for better structure and line
designs as well as construction. Its time that we truly use Technology for the Next
Generation and not continue to use these Electronic Sag Templates that merely
duplicate what weve always done before.
Introduction
The concept of using an equivalent span for a series of spans has been around
for a long time. The method now known as the Ruling Span Method was originally
developed by E.S. Thayer (Thayer) in 1924.
What is a Ruling Span?
Section 9.7.1 in the Design Manual for High Voltage Transmission Lines
(RUS Bulletin 1724E-200) defines the Ruling Span as follows:
A ruling span is an assumed uniform design span which approximately
portrays the mechanical performance of a section of line between its deadend
supports. The ruling span is used in the design and construction of a line to provide a
uniform span length which is representative of the various lengths of spans between
33
deadends. This uniform span length allows sags and clearances to be readily
calculated for structure spotting and conductor stringing.
Use of a ruling span in the design of a line assumes that flexing of the
structure and/or insulator string deflection at the intermediate supporting structures
will allow for the equalization of tension in the conductor between adjacent spans to
the ruling span tension.
The Report of the IEEE Bare Conductor Sag at High Temperature Task
Force entitled Limitations of the Ruling Span Method for Overhead Line
Conductors at High Operating Temperatures (IEEE Limitations of the Ruling Span
Method) states:
It (Ruling Span) is a fictitious span with a rate of slack equal to the
average rate of slack of the line section.
How is a Ruling Span Calculated?
The ruling span is traditionally calculated by the following equation:
RS
S
S
Where:
RS = Ruling Span
Si = horizontal length of individual spans in a Ruling Span section
It is important to note that the ruling span is NOT an average of spans; it is
weighted towards the longer spans.
How is a Ruling Span Used
Once the Ruling Span for a given section of line is calculated, a catenary
curve can be developed for its uniform span that will predict the sags of every span
within that Ruling Span section. Traditionally, this was made into a plastic template
that an engineer could then use for determining the layout of a line design. Line
design software has since been developed that can apply these same principles
electronically.
What is Ruling Span Error?
As described above, the Ruling Span concept assumes that the conductor
attachments are free to move in order for the tension to equalize. As most conductors
are usually clipped to the insulators, this can only be true at the time (and thus for a
given conductor temperature) of construction of the line when the longitudinal
tensions are equal. Unless all spans in a Ruling Span section are level and equal, this
34
assumption has varying degrees of error. Spans are usually unequal in length and as
the conductor temperature changes; shorter spans will expand and contract less than
longer spans. Depending on the structure configurations, this can create various
effects:
-
35
differentials at each structure properly accounted for in the structure and insulator
design.
Inclined Spans
The calculation of a Ruling Span in inclined spans such as in mountainous
construction and some river crossings may not be what it appears to be at first glance.
To illustrate this, lets first cover some basic concepts of sag and tension.
The Southwire Overhead Conductor Manual (Southwire) provides that for a
level single span of wire, the conductor length can be calculated by the following
equation:
2
sinh
2
Where:
L = total conductor length in a span
H = horizontal component of conductor tension
w = conductor weight per unit length
S = span length
Further, Southwire provides that for a level single span of wire, the sag can be
calculated by the following equation:
cosh
Where:
36
simply use a more effective ruling span that will accurately predict the expected sag
and tension in the incline as expected on a level span. The immediate solution is
usually counterintuitive; simply using a Ruling Span of the chord length does not
yield the expected sag and tension. The Ruling Span must actually be shorter than
the actual span so that the expected behavior of the actual span will match. It is
important that the slack be calculated properly as the sensitivity of this slack is critical
when calculating the sags, particularly when considering temperature and loading
effects.
The slack of an inclined span is simply equal to the slack of a level span
having the same horizontal projection multiplied by the cosine of the slope of the
span chord. Thus, the effective span to use, i.e. the Ruling Span, of an inclined span
is simply the horizontal span length multiplied by the cosine of the slope of the span
chord. This will result in a theoretical inclined span length the same as the horizontal
span length. In our example above, even though the span is 1000, the Ruling Span
would be 866.03 feet.
For a series of spans, the traditional Ruling Span can be slightly modified to
account for the individual inclined spans and an Improved Ruling Span can be
developed as follows:
S
C
C
RS
Where:
RS = Ruling Span
Si = horizontal length of individual spans in a Ruling Span section
Ci = chord length of individual spans in a Ruling Span section
To prove this concept, we will further explore the example span above using a
traditional Ruling Span, an Improved Ruling Span, and exact Finite Element Model
(FEM) analyses for a series of inclines, considering temperature and loading effects,
and compare those results. The 1000 foot span with four inclines were considered;
0, 15, 30, and 45. The traditional Ruling Span is of course 1000 foot, and the
Improved Ruling Spans are 1000, 966, 866, and 707 feet respectively. The spans are
subjected to high (212 F) and low (-20 F) temperatures as well as a loaded (1 inch
ice at 30 F) condition. The sags, horizontal tensions, and maximum tensions are
summarized in Tables 1 through 3. Numbers in parentheses, when indicated, are
percent errors relative to the FEM values.
37
38
15
30
45
31.15
0%
32.25
-1%
35.95
-6%
43.99
-14%
31.15
0%
32.78
0%
38.56
1%
52.74
4%
31.15
32.72
38.2
50.94
16.25
0%
16.83
1%
18.76
4%
22.97
7%
16.25
0%
16.57
0%
17.61
-2%
19.79
-8%
16.25
16.65
18.04
21.48
29.12
0%
30.14
-2%
33.59
-7%
41.09
-17%
29.12
0%
30.45
-1%
35.11
-3%
45.84
-8%
29.12
30.67
36.22
49.69
15
30
45
4384
0%
4384
1%
4384
6%
4384
16%
4384
0%
4314
0%
4088
-1%
3657
-3%
4384
4321
4126
3787
8411
0%
8411
-1%
8411
-4%
8411
-6%
8411
0%
8539
0%
8959
2%
9761
9%
8411
8502
8748
8993
15934
0%
15934
2%
15934
8%
15934
21%
15934
0%
15769
1%
15248
3%
14289
8%
15934
15658
14783
13184
15
30
45
4418
0%
4719
1%
5413
6%
6786
14%
4418
0%
4647
0%
5074
-1%
5766
-3%
4418
4655
5117
5947
8429
0%
8872
-1%
10047
-4%
12463
-6%
8429
0%
9004
0%
10678
2%
14369
8%
8429
8966
10435
13284
16042
0%
17102
2%
19583
7%
24513
19%
16042
0%
16932
1%
18796
3%
22202
8%
16042
16818
18263
20652
It can be seen that the Improved Ruling Span method gives much better
results relative to the more exact FE results than the traditional Ruling Span. While
most lines do not have the extreme inclines used to illustrate this theory, there are still
differences at the lower inclines and in rare cases of extreme inclines, these
differences can become quite significant. It is obvious that Finite Element modeling
yields the absolute correct results to design or analyze an overhead line, but should
the Ruling Span concept be the method of choice, then the Improved Ruling Span
method should be used.
Weight Span Errors
Another common mistake made in the days of plastic sag templates was in the
calculation of the weight span on each structure. Sag templates were traditionally
made with the maximum sag condition under temperature and/or ice loading for the
proper spotting of structures to maintain proper clearance to the ground. If a shield
wire was used, there was most likely a minimum sag condition for that wire used for
checking against uplift on a structure and/or hardware binding.
While there are many other inherent problems with using the simplified Wind
and Weight span concept for structure analysis, those are beyond the scope of this
paper. Without going into depth on the theory of Wind and Weight Span methods and
in keeping with the theme of this paper, a brief discussion of this process for
understanding its impact will follow. Traditionally, hatch marks were made on the
drawings at the low point, or the belly, of the maximum sags drawn with the sag
template. The distance between these two low points was then measured and
multiplied by the unit weight of the wire with ice and/or wind and any load factors to
obtain the maximum vertical load on each structure. A variant of this was to reverse
calculate the maximum vertical load a structure could support and then divide that by
the unit weight of the wire with ice and/or wind and any load factors to obtain the
maximum Weight Span the structure can support.
The problem with this concept is that the maximum weight span does not
occur under the maximum design temperature condition. In reality, other conditions
such as cold temperatures, ice, and wind can create a higher vertical load on the
structure. See Figure 1. It is nearly impossible to simulate with a plastic template or
computerized methods that merely duplicate plastic template technology the effects
that wind has on inclined spans as described in detail in Re-Engineering The
Transmission Line Design Process; this effect alone can increase the weight spans by
50% or more in average terrains.
Figure 1 represents a section of an actual line where the Weight Span under
what was assumed to be the maximum sag condition of Maximum Operating
condition is 990 feet. The structure was designed for a maximum of 1000 foot
Weight Span. Under the Cold condition, the weight span increases to 1230 feet and
even more critical, it increases to 1357 feet under the Extreme Wind condition. This
39
ice span is actually slight shorter in this instance, but the vertical load is actually
greater due to the weight of the ice in addition to the weight of the conductor.
40
spotting and another different ruling span, and this would occur differently for each
Ruling Span section of a line.
In the days before templates could easily be made with software and a local
print shop, many engineers would simply say close enough within some certain
range of the actual Ruling Span and the one used for spotting. Some utilities have
even gone so far as to require only certain increments (i.e. 300, 600, 900, 1200, etc.)
for the Ruling Span and in one case that the author is aware of, a utility designs every
line in their system with a single Ruling Span. But just exactly what is close
enough?
The RUS Bulletin 1724E-200 addresses this in Section 9.7.4 and Table 9-4:
9.7.4 Effects of the "Wrong" Ruling Span: It is important that the actual
ruling span be reasonably close to the ruling span value that is used to spot the line.
If this is not the case, there may be significant differences between the predicted
conductor tensions and clearances and the actual values. There have been instances
where sags were greater than predicted, resulting in clearance problems, because the
wrong ruling span was assumed. Table 9-4 will be of use in determining how
conductor sags differ from the predicted value when there are differences between
actual and assumed ruling span. Note that tension variation is opposite of that of the
sags. Thus, increased sags mean decreased tension and vice versa.
Our Drake conductor example above will be used to illustrate these exact
effects on various spans and their magnitude. As a basis of comparison for various
Ruling Spans and spans, the same horizontal tension of 6300 lbs at 60 F at the
Design Ruling Span as used above will remain the constant across the analysis and all
spans are assumed to be level.
At the stringing condition of 60 F, there is no change in the sag regardless
of what Design Ruling Span was assumed (see Table 4). In essence, any errors in the
Design Ruling Span versus the Actual Ruling Span will not be apparent during the
construction phase.
41
42
Tension
500
600
700
800
900
1000
1100
1200
1300
1400
1500
(lbs)
500
5.43
7.81
10.63
13.89
17.58
21.71
26.27
31.27
36.71
42.58
48.89
6300
600
5.43
7.82
10.64
13.9
17.6
21.73
26.3
31.3
36.74
42.62
48.93
6300
700
5.42
7.81
10.63
13.89
17.58
21.71
26.27
31.26
36.7
42.57
48.88
6300
800
5.43
7.82
10.64
13.9
17.6
21.73
26.29
31.3
36.73
42.61
48.92
6300
900
5.43
7.82
10.64
13.9
17.59
21.72
26.28
31.28
36.72
42.59
48.91
6300
1000
5.43
7.82
10.64
13.9
17.59
21.72
26.28
31.28
36.72
42.59
48.91
6300
1100
5.43
7.82
10.64
13.9
17.59
21.72
26.28
31.29
36.72
42.6
48.91
6300
1200
5.43
7.81
10.64
13.89
17.59
21.71
26.28
31.28
36.71
42.59
48.9
6300
1300
5.43
7.81
10.64
13.9
17.59
21.72
26.28
31.28
36.72
42.59
48.9
6300
1400
5.43
7.82
10.64
13.9
17.59
21.72
26.28
31.29
36.72
42.6
48.91
6300
1500
5.43
7.81
10.64
13.9
17.59
21.72
26.28
31.28
36.72
42.59
48.9
6300
The problem occurs when the line is heated up. From our discussion on
slack in the Inclined Spans section above, the calculation of the final sags under the
maximum operating temperature is a function of the slack. If the slack is calculated
using the wrong Ruling Span, then the wrong final sag will be estimated. To illustrate
this concept, the sag of the conductor at 212 F has been calculated for a 1000 span
(i.e. the Actual Ruling Span) using various Design Ruling Spans (see Table 5).
Table 5. Maximum Operating Condition
Sag (ft)
Tension
500
600
700
800
900
1000
1100
1200
1300
1400
1500
(lbs)
500
10.78
15.53
21.15
27.63
34.99
43.21
52.32
62.3
73.16
84.9
97.53
3172
600
9.73
14.01
19.08
24.93
31.56
38.98
47.18
56.18
65.97
76.55
87.92
3515
700
8.94
12.87
17.52
22.9
28.99
35.8
43.33
51.59
60.57
70.28
80.71
3827
800
8.35
12.03
16.37
21.39
27.08
33.44
40.47
48.18
56.57
65.64
75.38
4096
900
7.9
11.38
15.5
20.25
25.63
31.65
38.31
45.6
53.54
62.11
71.33
4327
1000
7.55
10.87
14.8
19.34
24.48
30.23
36.59
43.55
51.13
59.32
68.12
4529
1100
7.27
10.47
14.25
18.61
23.56
29.1
35.21
41.92
49.21
57.09
65.56
4705
1200
7.04
10.13
13.79
18.02
22.81
28.17
34.09
40.58
47.63
55.26
63.46
4860
1300
6.85
9.86
13.42
17.54
22.2
27.41
33.17
39.49
46.35
53.77
61.75
4994
1400
6.69
9.63
13.11
17.13
21.68
26.77
32.4
38.57
45.27
52.52
60.31
5113
1500
6.55
9.44
12.85
16.78
21.24
26.23
31.75
37.79
44.36
51.46
59.09
5218
43
Examining this table for our 1000 foot example level span, if a Design Ruling
Span of 1000 foot is used on the 1000 foot Actual Ruling Span, the sag would be
30.23 feet. However, if an 1100 foot Design Ruling Span was used, the sag would be
29.1 feet. Translating this in the design practices, if an 1100 foot Design Ruling Span
was used to spot the structures and sag the wire and the Actual Ruling Span ended up
being 1000 feet and the line designer said close enough, the structures spotted
would be 1.13 feet lower than the designer intended. While this probably wont
create clearance issues in every span due to line grading, the only way to verify this is
to redraw the entire line as spotted with the Actual Ruling Span and make changes
where this is necessary.
Figure 2 shows a relationship of Ruling Span difference expressed as a
percentage of the design span versus the actual span sag error at 212 degrees,
expressed as a percentage.
60.0%
40.0%
Error in Sag
20.0%
0.0%
-100%
-50%
0%
50%
100%
150%
200%
-20.0%
-40.0%
-60.0%
-80.0%
Design Ruling Span Error to Actual Ruling Span
500
600
700
800
900
1100
1200
1300
1400
1500
1000
250%
44
15.0%
10.0%
Error in Sag
5.0%
0.0%
-50%
-30%
-10%
10%
30%
50%
-5.0%
-10.0%
-15.0%
-20.0%
Design Ruling Span Error to Actual Ruling Span
500
600
700
800
900
1100
1200
1300
1400
1500
1000
45
1.5
0.5
-20%
0
-10%
0%
10%
20%
30%
40%
50%
-0.5
-1
-1.5
-2
Design Ruling Span Error to Actual Ruling Span
500
600
700
800
900
1100
1200
1300
1400
1500
1000
above parameters should be within 40 feet of the Actual Ruling Span. However, it
would simplest, and most accurate, to just use the Actual Ruling Span on the line.
Even better, Finite Element analysis should simply be used as it is the most accurate
method available today.
Construction with the Wrong Ruling Span
Even further compounding this problem is even if the most prudent line
design possible has been done, the line is constructed with traditional Stringing
Charts where the Ruling Span has been rounded can lead to these same errors.
If the conductor is strung at the same temperature as the reference
temperature, the use of the wrong ruling span is insignificant. However, if the
temperature varies from this basis temperature, errors will be introduced. If a
stringing chart from a shorter ruling span than the actual ruling span is used and the
ambient temperature is lower than the reference 60 F, the resulting sag will be less
than expected, and if the ambient temperature is higher than 60 F, the resulting sag
will be higher. If a stringing chart from a longer ruling span than the actual ruling
span is used and the ambient temperature is lower than the reference 60 F, the
resulting sag will be higher than expected, and if the ambient temperature is higher
than 60 F, the resulting sag will be lower.
In the example above, the basis for the comparison is to assume that the
tension at 60 F is 6300 lbs across all Design Ruling Spans. In an Actual Ruling Span
section of 1000 feet, if the stringing chart for a 900 foot Ruling Span is used, the sag
at 212 F (Initial) will be 31.65 feet versus the 30.23 feet expected by the design. For
this reason, it is critical that the stringing charts be used for the Actual Ruling Span.
Further, this is not just a transmission problem; distribution lines can be just as
susceptible if not more sensitive to these errors. One cannot force a span or a series
of spans to behave different than physics dictates.
Conclusion
By not considering the actual line layout in the calculation of the Ruling Span,
the wrong Ruling Span can be selected and the results can be error, particularly in the
cases of extreme temperature and loading events. Further, simply using a computer
program to calculate the behavior on a level span and then applying those results to
an entire line design can lead to an incorrect design and/or analysis. While the Ruling
Span certainly served its purpose for approximately designing overhead lines with
little temperature changes anticipated and with technology available at that time,
modern overhead line designs with the maximum operating temperatures that they are
expected to operate at in the 21st century should simply use modern Technology for
the Next Generation. With the tools readily available to the overhead line engineers
46
47
Robert Whapham 3
P.E., SM.IEEE
Principal Engineer, American Electric Power (AEP), 700 Morrison Road, Gahanna,
OH 43230, Phone: (614) 552-1944, email: [email protected]
Senior Engineer, American Electric Power (AEP), 700 Morrison Road, Gahanna,
OH 43230, Phone: (614) 552-1821, email: [email protected]
Global Market Manager - Transmission, Preformed Line Products (PLP),
660 Beta Drive, Mayfield Village, Ohio 44143, Phone: (440) 473-9202,
email: [email protected]
ABSTRACT
Over the past number of years AEP and other utilities have been experiencing fatigue
failures of steel davit arms in our Texas region, especially when used in conjunction
with concrete poles. It was assumed that these failures were the result of conductor
motion, either as a result of high wind gusts (wind sway) or smooth moderate winds
(aeolian vibration).
This paper describes the results of a field vibration study that was conducted on one
line which experienced davit arm failures to establish the level of aeolian vibration
versus wind velocity and vibration frequency. Also described are the results of
extensive laboratory testing that were conducted on a vibration span to determine the
relationship between conductor motion, the davit arm motion, and the fatigue stress
on the arms. Further laboratory tests were also conducted to determine if dampers
used to control aeolian vibration on the conductors would have a mitigating effect on
the davit arm failures.
Further long-term field measurements are planned on a line where davit arm failures
have occurred to collect data on a full scale span and to verify the preliminary
findings of the laboratory testing.
FIELD VIBRATION STUDY
A two week field vibration study was conducted on a 138KV line in Corpus Christi,
Texas in the span adjacent to structures where davit arms have failed (Figure 1).
The line is built with a vertical bundle (12 spacing) of 795 kcmil 26/7 ACSR
(Drake) conductor on concrete poles. The ruling span length is approximately 880
feet and the surrounding terrain is very flat farmland (Figure 2).
Page 1 of 11
48
49
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
50
1
1
1
1
1
3
3
3
3
1
1
2
4
7
5
8
4
7
7
16
18
2
3
1
3
3
4
2
4
10
12
14
1
1
1
1
1
4
6
5
9
8
11
8
9
2
1
5
3
5
7
4
10
7
4
12
2
3
3
8
5
3
6
7
9
12
4
2
4
2
2
3
4
2
8
8
20
22
24
26
2
3
1
1
5
6
8
9
1
1
1
3
1
3
4
3
9
5
7
5
3
5
9
7
2
2
5
7
3
1
3
2
1
28
30
32
34
36
38
40
42
1
1
1
1
1
1
1
44
46
48
Frequency (Hertz)
Page 3 of 11
Accelerometers were attached to the davit arms as required to monitor both vertical
and horizontal dynamic displacements and frequencies of the arms during the tests
(Figure 10).
Figure 8
Dual Axis Strain Gauges
Figure 9
Strain Gauge on Side
Page 4 of 11
Figure 10
Accelerometers
51
52
As shown in Table 1, the vertical fundamental frequencies for the three arm designs
were very similar, but there was a marked difference in the horizontal fundamental
frequencies.
Before the fundamental frequency tests were conducted, the same test setup was used
to determine the static stress on the davit arms adjacent to the weld, and the vertical
deflection at the end of the arm. Of significant interest was the level of static stress
on the arms as determined from the strain gauges. These values, as measured by
strain gauges near the welds, were 2 to 3 times higher than a linear calculation of
stresses using the arm geometry, material properties, and the position of the load.
However, the strain gauge positioned 6 inches from the weld on the mounting plate
gave results consistent with the linear calculations.
Possible explanations for this large difference are that there exist concentrated
stresses adjacent to the weld due to the shape of the arm and the type (and quality) of
the weld. Further proof requires a non-linear, finite element, analysis of the arm
shape and welds. However, it is known that higher static stresses on the arm will
result in a lower fatigue stress endurance limit within the arms life cycle as long as
the fatigue stress is larger than the fatigue threshold.
Table 1 Fundamental Frequency vs. Load
Load
0#
500#
1,000#
1,500#
2,000#
Arm #1
24.6
4.9
3.6
2.9
2.4
Fundamental Frequencies
Vertical
Horizontal
Arm #2
Arm #3
Arm #1
Arm #2
Arm #3
25.2
24.5
N/A
7.4
8.6
4.6
4.8
N/A
6.3
9.0
3.3
3.4
N/A
6.9
7.5
2.5
2.7
N/A
6.0
7.5
2.1
2.4
8.5
5.8
7.5
Page 6 of 11
Figure 12
Forced Response Test
53
It was hoped that from this testing that higher order resonant frequencies of the arms
could be found so that this information could be used to determine which frequencies
to use in the span testing.
Unfortunately, only the fundamental vertical frequency was apparent when the
outputs of the accelerometers were closely reviewed.
SPAN TESTING
Figure 13 shows the basic setup for the span testing, where the davit arm with
insulator is connected to the center of a 90 span of 795 kcmil 26/7 ACSR (Drake)
conductor. Figure 14 shows a close-up of the suspension clamp and the weights
attached below it to simulate the weight of the conductors.
Page 7 of 11
54
55
The vibration shaker positioned near one end of the span is shown in Figure 15.
Figure 16 shows the levered weight basket arrangement at one end of the span used to
adjust and maintain the tension on the conductor (15 lbs of tension for each 1 lbs of
weight in the basket).
Similar span testing was conducted for each of the three davit arm designs. At a
number of discrete conductor tensions (ranging from 10% to 15% of the conductor
breaking strength) the span was tuned and held at six or seven resonant frequencies
(see Table 2 for example). Therefore, for each arm, as many as 34 different
conductor vibration frequencies were investigated.
Table 2 Lab Span Resonant Frequencies
Tension
4726#
3976#
3600#
3376#
3150#
Freq 1
9.94
9.11
8.79
8.49
8.25
Freq 2
14.81
13.57
12.95
12.70
12.39
Freq 3
19.76
17.88
17.36
16.84
16.33
Freq 4
24.86
22.84
21.98
21.42
20.86
Freq 5
30.27
27.72
26.74
26.10
25.43
Freq 6
36.52
33.59
32.42
31.66
30.98
Freq 7
40.06
38.70
36.98
36.30
For each resonant conductor frequency the peak-to-peak amplitude of the anti-node
of one of the free standing wave loops in the span was held at 0.5.
By recording the dynamic stress at various points on the arm (with strain gauges), and
the vertical and horizontal movement of the arm tip (using accelerometers) it was
possible to correlate aeolian vibration activity in the span to movement (vertical or
horizontal, or a combination) of the arm tip.
The most compelling test data was at a conductor frequency of 28.35 hertz for one of
the arms. As shown in Figure 3, the amplitude of conductor vibration was one of the
highest around this frequency during the field vibration study.
The movement of the arm at the 28.35 hertz frequency produced the highest dynamic
strain levels. The movement was vertical, and even though the frequency of the span
Page 8 of 11
56
was 28.35 hertz, the arm was moving at the vertical fundamental frequency of the
arm/insulator/weight combination (2.4 hertz), as clearly shown in Figure 17. The
measured maximum strain is 100 microstrain, which occurs at the top of the arm.
0.5
0.5
0.5
0.5
0.5
0.5
1121
1123
1125
1137
1141
1143
Shaker
Power
(W)
1.3
6.2
18.7
23.6
27.0
41.2
66.1
56.7
48.0
58.0
67.9
62.8
59.0
Shaker
Velocity
(m/s)
0.0632
0.1593
0.2716
0.2801
0.2947
0.3605
0.4426
Shaker
Force
(N)
69.1
181.2
342.1
374.1
393.6
461.5
540.8
Phase
Angle
(deg)
128
115.4
113.7
116.8
117.8
119.7
123.5
0.4776
0.4472
0.4557
0.4468
0.4291
0.415
561.5
540.5
566.4
525.6
514.4
504.2
115
113.4
116.7
125.3
124.7
124.3
Comments
During the two-week field vibration study, measurements were also made on phases
that had dampers (one per span) installed. Figure 18 shows the standard IEEE1
formatted results of the testing comparing the undamped vs. damped vibration
activity. This clearly shows that the amplitude of vibration was reduced below the
threshold for arm movement.
Page 9 of 11
1.2
0.8
0.4
0.0
0
100
200
R1 - Undamped
300
Microstrain p-p
R2 - Damped
400
500
R3 - Undamped
Page 10 of 11
57
the aeolian vibration of the conductors, which may be affecting the davit arm such as
high wind gusts.
By also studying phases with and without dampers with the planned system, the
effectiveness of the dampers in mitigating the arm movements can be verified in the
field.
A follow up field vibration study is also planned which will allow correlation
between the movements measured on the arms with actual vibration data on the span.
CONCLUSIONS
The field vibration study showed that there is moderately severe vibration
activity on spans without dampers.
Static stress measurements on the davit arms adjacent to the weld (with strain
gauges) were considerably higher (2 to 3 times) than the stress calculated
using standard linear analysis.
Span testing in the laboratory revealed that vertical (aeolian) vibration of the
conductor can produce both vertical and horizontal (or combination)
movements of the tip of the arms.
Further testing showed that there was a threshold value of the free loop antinode amplitude, below which the arm did not move. This supports the use of
dampers to control the level of the vibration in the field to mitigate arm
movements and subsequent fatigue failures.
References
1
Page 11 of 11
58
Georgia Transmission Corporation, 2100 East Exchange Place, Tucker, GA 30084; PH (770) 2707699; FAX: 770-270-7775; Email: [email protected]
59
FIGURE 1
After the extensive studies, it was concluded that an upgraded existing APC delta
configuration tower family is the desired 500kV standard structures going forward.
The new family of structures will have following improvements to the existing
structures studied:
- Provide adequate clearance required for hot line maintenance
- Provide adequate phase spacing between shield wires and conductors required
for helicopter maintenance
- Provide engineered buzzard shield for protection of insulators from buzzard
contamination
- Design an additional angle structure to provide efficient line design with a more
completed family of structures
- Accommodate more effectively with 150 feet rights-of-way width under the
various loading conditions encountered
- Use more readily available angle sizes and provide interchangeable angles
between AISC and metric angles in design and detailing
- Provide better electric behaviors
2
60
Based on the assessment of work involved in developing the new family of 500kV
structures, a working team consisting of members from Southern Company and GTC
was formed. The team was responsible for coordinating design requirements of
Southern Company and GTC, setting design parameters, determining bidding process,
selecting winning bidder and reviewing design and details submitted by the bidder.
BIDDING PROCESS
In 2005, the working team started the pre-bidding process. It began with initial
contact with consulting firms in the United States which have known experience in
design and analysis of 500kV transmission lines. After review of the information
collected in the initial contact, considering the scope of the project and capacity of the
consulting firms, the following four (4) firms were selected for further consideration:
- Black & Veatch
- Commonwealth Associates Inc.
- Burns & McDonald
- Power Engineers
The working team conducted face to face interviews with each of above consulting
firms. The interview started with the consultants presentation on experience of
latticed steel tower design, recent 500kV transmission line work, and design software,
particularly PLS-TOWER (TOWER). Then detailed discussions on project scope,
schedule, quality control process, tower detailing and tests, tower fabrication, tower
erection and other related topics were carried out. It was intended initially to have a
consulting firm perform all engineering work including design, analysis and detailing
of the family of tower. However during the interview and discussions, it was found
out that the consultants have either very limited or no capacity at all for detailing of
latticed steel tower. After much deliberation, it was decided that the project would be
better carried out in two steps. First, the team would solicit proposals from the
consulting firms to develop design specifications for the latticed steel towers. Then a
tower supplier would be selected to provide services in design, analysis, detailing and
testing of the towers based on the design specifications developed. The tower supplier
selected would also supply 157 steel towers for GTCs thirty-eight (38) miles of
Thomson Warthen 500kV Transmission Line Project.
After proposals for developing the design specifications from the consulting firms
were evaluated, Black & Veatch was selected. In order to incorporate requirements
for hot line and helicopter maintenance in design specifications, a large meeting was
called for soliciting advice and recommendations from personnel of helicopter
companies specialized in transmission line maintenance, transmission line
construction contractors with extensive experience in 500kV transmission line
construction, suppliers of hotline maintenance tools, and Southern Company and
GTC maintenance. A detailed design specification was developed for use to solicit
proposals from tower suppliers for design, analysis, detailing, testing and supply of
latticed steel towers. Following suppliers were contacted initially for pre-qualification:
- Fort Worth Tower Inc. USA
- Locweld Inc. Canada
- Falcon Steel Company USA
- Fabrimet Canada
- Thomas & Betts USA
- SAE Towers Mexico
61
Kalpataru India
Mitas Turkey
62
tower design. It is common practice in the United States to use only one diameter and
grade of bolt in any single tower design. ASTM A394 Type 0 and Type 1 bolts are
the most commonly used bolts for transmission line towers in North America.
However, considering more and more transmission line towers are designed and/or
supplied by overseas fabricators, it may be desirable to specify ASTM A307 and
A325 bolts for transmission line towers. Because ASTM A307 and A325 bolts are
widely used in steel structural joints of buildings and bridges, the ASTM A307 and
A325 specifications are better known internationally. These bolts are more readily
available and generally have better quality assurance and quality control process.
Transmission tower connections are designed as bearing type connections. Decision on
the design of connections with threads included or excluded from the shear plane could
have significant impact on costs. If the connections are designed with threads included,
then all threaded bolts could be used to provide ease of construction. It would also
eliminate concerns about possible mistake of installing bolt threads in the shear plane
which could lead to significant strength reduction of the connection when the design
called for excluding threads in the shear plane. However, connections designed with
threads included would require more bolts in the connection and increase number of
gusset plates used in the connections. Gusset plates are always expensive, and
sometimes they will increase the eccentricities in the connections. In addition to the
cost of furnishing, fabricating, handling and erecting the additional piece, each gusset
plate usually requires a few extra bolts and fills or washers which contribute more
adverse economic comparison (EPRI, 1990). It should be a major goal in design and
detailing to minimize the number of gusset plates used in a tower.
It was required in the design specifications that the transmission line structures to be
designed in accordance with the ASCE 10-97. The loads, configuration and electrical
clearances were provided to the bidders in loading and configuration drawings. The
structures furnished under the design specifications included a family of following
latticed steel towers:
DSS-T
50 125 ft, self-supporting tangent structure, 0-5
DSS-HT
50 135 ft, self-supporting heavy tangent structure, 0-5
DSS-SA
50 120 ft, self-supporting small angle structure, 5-15
DSS-LA
50 120 ft, self-supporting large angle structure, 15-30
DSS-DE1
65 125 ft, self-supporting deadend structure, 0-45
DSS-DE2
65 125 ft, self-supporting deadend structure, 45-115
DGM-T
50 125 ft, guyed mast tangent structure, 0-5
DGM-HT
50 135 ft, guyed mast heavy tangent structure, 0-5
After reviewed the preliminary designs submitted by the bidders, it was found that the
structure DSS-DE2 was extremely heavy because it was designed to accommodate a
very large line design angle. Past experience in Southern Company and GTC 500kV
transmission line designs shows it is very rare in line design that any structure would
have line angle greater than 90. Therefore it was decided to reduce design line angle of
the structure DSS-DE2 to 90 and add another design of DSS-DE3 structure with line
angle 90-115. This would provide more economic utilization of all structures.
5
63
64
It was expected that all bidders would put sufficient effects on designing the lightest
towers possible in their proposals. However, it is extremely important for the owner to
keep in mind that an economical tower design is a compromise between tower costs
and foundation costs. The basic guidance is that a narrower base will lead to a lighter
tower, but larger foundation reactions. While a wider base will in general reduce the
foundation reactions, but may require longer, more costly tower bracing members and
lead to a more expensive tower.
After initial evaluation of design proposals submitted by the bidders, it was found that
further design optimization would be necessary taking into consideration of foundation
reactions and costs in the tower designs. Request for additional design information was
made to the two short listed bidders. The two bidders redesigned three towers selected
by GTC to provide information on tower weights and foundation reactions under
various tower base widths. After receiving the information on the revised tower base
width and foundation reactions, foundation designs using EPRIs CUFAD
(Compression and Uplift Foundation Analysis and Design) were performed. A cost
comparison for different widths of tower base taking into consideration of both tower
and foundation cost was then carried out. The data used is summarized below.
- The foundation designs were based on two soil borings obtained in a project area
where two 500kV transmission lines will be built in the near future. The first soil
boring was categorized as "Good Soil" with a blow count N = 22 34, unit soil
weight = 120 125 lb/ft3, soil friction angle = 32 37 degrees and no water table.
The second soil boring was categorized as "Average/Poor Soil". The top fortyseven feet soil has blow count N = 8 35, dry unit soil weight = 115 120 lb/ft3,
soil friction angle = 30 32, water table depth at 10 feet.
- Tower base widths considered:
Tangent tower: 36 ft, 40 ft and 45 ft
Large angle tower: 40 ft and 48 ft
Deadend tower: 46 ft, 50 ft and 52 ft
- Tower costs:
Tangent tower steel: $0.843 / lb
Deadend tower steel: $0.951 / lb
Large angle tower steel: $0.910 / lb
Steel erection cost: $0.900 / lb
- Concrete pier foundation cost was assumed to be $600 per cubic yard, including
concrete, reinforcement and contractor labor and materials.
- Uplift and compression foundations were designed differently for angle and
deadend towers. The minimum foundation diameter considered was 3 feet and
maximum burial depth was 40 feet. Under same loading and soil condition, a
smaller diameter and deeper foundation is in general cheaper than a larger diameter
and shallower foundation.
The cost comparison results were summarized in the table below.
Table 1: Cost Comparison for Different Widths of Tower Base
Tower Type
Base Width Increase (ft)
Cost ($)
Steel & Erection
Good Soil
Foundation
Total Cost Difference ($)
Cost ($)
Steel & Erection
Avg./Poor Soil
Foundation
Total Cost Difference ($)
Tangent
36 to 40
36 to 45
1339 ()
4468 ()
-720 ()
-1680 ()
619 ()
2788 ()
1339 ()
4468 ()
-1680 ()
-7680 ()
-341 ()
-3212 ()
Large Angle
40 to 48
3244 ()
-11160 ()
-7916 ()
3244 ()
-25800 ()
-22556 ()
Deadend
46 to 50
46 to 52
2701 ()
4081 ()
-11160 () -20400 ()
-8459 () -16319 ()
2701 ()
4081 ()
-23040 () -33840 ()
-20339 () -29759 ()
The cost comparison shows that increasing tower base width to reduce foundation
reactions will result in total savings except for tangent tower in good soil area. The
savings could be quite significant for large angle and deadend towers located in
average/poor soil areas. It was found that the deadend tower with 46 feet base width
would require the uplift foundations in average/poor soil area to be designed with a
diameter of 10-11 feet and a burial depth of over 40 feet, which would be a very
expensive and impractical construction.
A guidance to determine the width of tower base could be to set the slope of the post
legs of tower so the theoretical intersection of their extensions is at the elevation close
to the centroid of the horizontal loads of the worst and the second worst design loading
conditions. This design would provide following advantages (EPRI 1990):
- The post legs of the tower body form almost a true A frame and the horizontal
shear bracing will be minimal.
- Forces in the post legs will be fairly uniform from ground line to tower waist.
- Foundation reactions will be fairly constant from the shortest to the tallest tower.
TOWER MODELING
It is a standard practice to model latticed towers as ideal elastic three-dimensional
trusses with pin connection at joints. Tower members are considered as axially loaded
tension or compression truss elements. Moment due to eccentricities is assumed to be
small and is not calculated in the analysis.
PLS-TOWER is a specialized computer
program for analysis of latticed
transmission towers and is widely used
by utilities around the world because of
its compatibility to transmission line
design software PLS-CADD. All
Figure 2 Beam Elements (Thick Lines)
computer programs used in analysis of
latticed transmission towers almost exclusively modeled the tower members as truss
elements until recently, PLS-TOWER now recommends the use of beam elements to
stabilize planar joints and mechanisms. PLS-TOWER recommends that all tower
members be modeled as beam elements, except diagonals and single horizontal struts
as shown in Figure 2 (PLS-TOWER, 2008).
PLS-TOWER emphasized that the main purpose of using beam elements is to
stabilize planar joints and mechanisms. Beam elements, in addition to axial tension
and compression can carry shears and moments, but they are not intended to be loaded
in bending and there is no design check for moments and no moment report, except for
the moment results table that can be generated using Model/ Results/ Moments for
Angles Modeled as Beams. For design purposes, the beam members are still
assumed to be loaded axially. The beam elements can be used to stabilize the model,
but the model should still be triangulated (PLS-TOWER, 2008).
65
Tower Body
66
or any other message was given in the output. However, when the tower was tested in
early 2008, both members MN and PQ failed prematurely.
The misuse of beam element shown in the examples of Detail A and Detail B are
pretty obvious for someone with experience in tower design. But sometimes it can be
quite intricate. In View A-A of Figure 3, horizontal and leg members were modeled
as beam elements except members P10P11, X-bracing and V-bracing were modeled as
truss elements. At the attachment point P8, two V-bracing at inside and outside panel
was formed to provide support at all direction. The model was properly triangulated.
No obvious flaw or violation to the recommendations and guidelines provided by the
TOWER was found. However the result of modeling horizontal and leg members as
beam elements is significantly different to the result when these members are modeled
as truss elements. Table 2 shows the result of member force in term of its capacity.
Table 2 Member Capacity Comparison: Beam Elements vs. Truss Elements
Member Force as % of Its Capacity
P1P2 P1P3 P3P5 P5P7 P7P10 P1P4 P3P6 P5P8 P7P11 P7P8 P10P11
Beam
P2P4 P4P6 P6P9 P9P11 P2P3 P4P5 P6P8 P9P10 P8P9
Element
87
48
36
57
97
83
91
33
21
57
33
Model
P1P2 P1P3 P3P5 P5P7 P7P10 P1P4 P3P6 P5P8 P7P11 P7P8 P10P11
Truss
P2P4 P4P6 P6P9 P9P11 P2P3 P4P5 P6P8 P9P10 P8P9
Element
238
112
100
80
140
220
213
65
93
179
192
Model
The result in Table 2 shows that member design is within its allowable capacity in
beam element model that simulates a rigid frame panel, whereas in truss element model
majority of members are significantly overstressed. When the bracing is insufficiently
stiff to provide proper support, the members modeled as beam elements shown carried
moment and shear, but no design check for shear and moment is performed in TOWER.
This had raised concerns about using beam elements in tower models, especially when
the tower was modeled by someone with limited experience in modeling. The concerns
were expressed to PLS in early 2008. Suggestions were made to PLS for adding a
warning or a pop-up message when TOWER detects significant moments in members
modeled as beam elements. The TOWER program released after April 2008 has an
additional warning now listed when beam elements modeled are carrying significant
moment. A graphic warning was also added in 2008. The graphic warning is a very
helpful tool to ensure that the beam elements were used properly in TOWER. The
tower designer should exercise extra caution and heed any warnings issued when beam
elements are used in the tower model.
DETAILING
The tower designer is responsible for ensuring structural integrity when the drawings
are created by detailers. The drawings used for fabrication of tower should be
consistent with the model. Moments due to framing eccentricities are generally not
calculated in the design model. The detailer should minimize the framing
eccentricities in detailing processes so that no significant moments would be
generated. When framing eccentricities are unavoidable at some connections, proper
design checks and calculations shall be performed to adjust the model and member
selection accordingly.
67
10
68
69
failure. Over a dozen premature failures occurred in the tests of next four towers,
which were more than it would expect normally. All failures can be attributed to
either detailing or modeling errors. Detailing errors occurred when double angles
were used in the analysis, but not properly detailed contributed to several failures.
Figure 5
Figure 5 shows one failure where V-bracing was designed with back-back double
angles, but the detail in View A-A did not properly design the sufficient bolts at
connections between the V-bracing and the gusset plates. There was also a premature
buckling failure when double angles were used in X-bracing, but only a single angle
was shown on the drawings. Detailers should pay special attention when double
angles are used. The tower designer should work closely with the detailers and
perform a thorough review of all details when they are completed.
Detail A
Failure at Connection B
Detail AA
Figure 6
Figure 6 shows two failures occurred at two different connections when testing a
deadend tower. The failed bolts in Connection A showed clear cut-off surfaces at the
shear plane, which indicates the connection was significantly under designed. The
connection was designed using AISC method for eccentric loads on fastener groups
assuming each bolt group would withstand the force that connected to the member.
Since the leg member was not continuous at the waist (see Detail A), the above
assumption was not accurate. The combined forces from different members on each
bolt group should have been considered. Detail AA shows an improved detailing with
leg continuous above the tower waist.
11
The failure occurred in Connection B was quite different. The failed bolts showed
clear necking at the failed surface. Deformation at shear planes was observed, which
could be due to lack of hardness of the bolts used, or the shear planes slipped
significantly when the test loads applied. The failure appeared to be tension failure
due to prying action. Tower connections are designed as bearing type connection. To
avoid prying action at connection, it would be a good practice to use high strength
bolts in connections that connect high strength steel and are heavily loaded.
SUMMARY
The proliferation of demands for reinforcing extra high voltage transmission line
networks in recent years and for complying with updated codes and standards
requires many utilities in the United States to develop new families of latticed steel
structures. Design of latticed steel transmission towers is a challenging task which
involves lengthy modeling and detailing of significant amount of steel members and
connections. There are very limited resources in the United States capable of carrying
out the detailing and design of latticed steel towers. Selecting experienced designers
and detailers are critical for the success and economy of the project.
An economical tower design is a compromise of tower and foundation costs. The
foundation costs should be considered when developing tower designs. It is extremely
important to thoroughly understand the capability of the software used in the design
and modeling of the structures, and to have knowledge of structural behavior. It is the
responsibility of tower designer and detailer to ensure structural integrity of the
towers designed. Full scale tower tests provide an indispensable tool to validate the
adequacy of the structures designed, particularly when the structures were designed
by someone with limited experience in the subject.
REFERENCES
1. AISC, (2005), Steel Construction Manual, Thirteenth Edition, American
Institute of Steel Construction Inc.,
2. ANSI/ASCE 10-97, (2000), Design of Latticed Steel Transmission
Structures, American Society of Civil Engineering, Reston, Virginia.
3. ASTM A394-07, (2007), Standard Specification for Steel Transmission
Tower Bolts, Zinc-Coated and Bare, ASTM International, West
Conshohocken, Pennsylvania.
4. EPRI, (1990), Considerations for the Structural Design of a Lattice Steel
Electrical Transmission Tower, Research Report RLMRC-89-R6, Haslet,
Texas.
5. EPRI, (2003), Foundation Analysis and Design (FAD), Haslet, Texas.
6. PLS-CADD, (2008), Power Line Systems Computer Aided Design and
Drafting, Power Line Systems, Inc., Madison, Wisconsin.
7. TOWER, (2008), Analysis and Design of Steel Latticed Towers Used in
Transmission and Communication Facilities, Power Line Systems, Inc.,
Madison, Wisconsin.
8. RCSC, (2000), Specification for Structural Joints Using ASTM A325 or
A490 Bolts, Research Council on Structural Connections, Chicago, Illinois.
12
70
71
72
by its own personnel. The scope of each subcontract is defined as clearly as possible
to reduce the likelihood of change orders for the Design/Build contractor. The final
project cost from the Design/Build contractor to the owner will include fees and
expenses for providing the overall project management, accepting and managing
project risks, and recovering the substantial cost of preparing the initial Design/Build
proposals to the owner. In some cases, the markups and the fee can make the
Design/Build contract more expensive than the Design/Bid/Build approach.
Advantages of Design/Build Approach
1. The project cost and schedule can be defined very early in the project.
2. If the scope is well defined and the owner identifies its requirements in the
Design/Build contract, there should be few, if any, change orders.
3. The owner can obtain guarantees on cost, schedule, and performance from a
single responsible party.
4. Owners with limited staff to dedicate to the project can rely on the
Design/Build contractor to coordinate each phase of the project. The owner
can use its limited staff to monitor the contractor to confirm that the project
meets the requirements of the contract.
5. The project schedule can be reduced due to close coordination of the design,
procurement and construction activities by the Design/Build contractor
allowing the owner to better define the project prior to issuing EPC bid
documents.
6. Design/Build contractors may have innovative approaches that are less costly
than the owners standard practices and may meet the owners requirements at
a reduced cost.
7. All contractual disputes between the Design/Build contractor and the
subcontractors are the direct responsibility of the Design/Build contractor.
8. Relatively large penalties in the form of liquidated damages can be established
to motivate the Design/Build contractor to complete the project on schedule.
9. Lower material and equipment costs may be passed though to the owner
during the bidding process due to the volume buying power of the
Design/Build contractor.
Disadvantages of Design/Build Approach
1. The Design/Build contractor receives an additional fee for managing and
accepting project risks and higher exposure to risks lead to higher pricing.
2. The contingency included in a Design/Build price depends on the level of
detail in the bid documents, project schedule, difficulty of construction and
unknowns. To limit the amount of contingency in the Design/Build price most
of the unknowns, such as subsurface or environmental risks, need to be
defined prior to issuing the bid documents. Well-defined subsurface
conditions also reduce owner exposure to change orders.
73
3. The owner generally does not select the equipment supplier. The Design/Build
contractor will generally solicit bids from owner-specified vendors and select
equipment with the lowest initial cost. Performance specifications that
consider life cycle costs must be part of the initial bid package. Some owners
may take bids on specialized equipment and then assign those contracts to the
Design/Build contractor. This allows the owner to select this primary
equipment and still obtain overall system guarantees.
4. Generally the owner is not involved in the detailed design decisions and has
very limited time, if any, to review design drawings.
5. Scope and schedule need to be clearly defined very early in the project.
Changes in scope or schedule after award will most likely cause the
Design/Build contractor to revise the price of the project.
6. The owner will receive the contractor's standard drawing formats unless
otherwise specified in the bid documents. The owner may also not receive
detailed documentation, calculations, or manuals, since the engineer typically
only produces those documents necessary for construction of the project.
Program Management Approach
The Program Management approach provides for meeting all an owners project
delivery needs related to a large and complex effort involving the construction of
multiple facilities over a several year period. The owner may choose to use internal
resources for program management or, if the owner has limited resources, the owner
may hire an outside resource as program manager. The program manager can be a
construction general contractor or a consulting engineer, in which case, the program
manager is commonly referred to as the owners engineer. The program manager is an
agent of the owner and serves as a single point of management for the entire process
of completing the project(s).
The program manager provides detailed managerial support and added technical value
to the owner and is normally involved in the earliest stages of a project. A program
manager offers current planning methodologies, public involvement and testimony
capabilities, design expertise, knowledge of construction methods and pricing, an
understanding of competitive market conditions, and effective scheduling and cost
control systems. Program management is successful when the projects planning,
permitting, design and construction phases are effectively integrated into a single
process. The program manager can deliver a project(s) with consistently successful
results in several areas of measurement over the course of several years by developing
an overall master plan for the project(s). The master plan might include such tasks as:
74
Describe the relationships between the program management staff and the
owner as well as future consultants, contractors, construction managers,
facilities users, government agencies, interveners and any other groups.
Describe the key program management tasks and establish the procedural
basis for executing program management functions.
Establish state-of-the-art management controls to effectively manage time,
quality, costs, and resources to achieve a high level of performance.
Establish communications procedures, information document flow and data
base management procedures.
Define program-wide standards, policies, and procedures.
Define program public information and coordination requirements.
Identify the approvals and timing required.
Program management still requires that designs are completed, materials are procured
and transmission lines and substations are built. The program manager may
recommend either the Design/Bid/Build approach or the Design/Build approach for
each individual project within the construction program.
Advantages of Program Management Approach
1. This approach provides comprehensiveness and continuity of services for the
owner with staff or a third party overseeing the entire process from project
planning to project closeout.
2. The owner has more control of the design and construction processes
throughout the project. The program manager provides timely cost and
schedule information to the owner allowing informed decisions on cash flow
and accurate information to executives, the public and regulators.
3. The owner can avoid hiring additional staff for specific project-related
functions. The program manager can easily increase or decrease resources on
the program applying specific resources when and where they are needed to
maintain project objectives.
4. The program manager provides current knowledge of design and construction
resources, material vendors and best practices. Program managers typically
are involved in the largest, most complex projects in the country and the
world. They bring a unique perspective and understanding to the owner.
5. Bid packages for engineering, materials and contracting can be staged based
on project needs such as outage windows, financial objectives, seasonal
weather, permitting requirements and environmental restrictions.
6. The program manager can be provided with penalties or incentives to maintain
schedule and budget.
Disadvantages of Program Management Approach
1. The program manager assumes no contractual responsibilities for design or
construction. Contracts for engineering, materials and construction labor are
held and maintained by the owner. Contract terms and conditions must contain
strong enforcement language that allows the program manager to maintain
budgets and schedules.
75
2. A strong working relationship between the owner and the program manager is
paramount to success. Early involvement in the program develops trust and
confidence between the program manager staff and the owners staff. A keen
understanding of the critical internal and external issues related to the project
is fundamental for success.
3. The program manager must rely on the owner for key decisions that impact
the project budget and schedule. Regular, documented communication is
required.
Current Mega Project in the United States
A mega project, or program, is defined by a capital cost over $1 Billion and an
implementation schedule of several years. Several programs meet these parameters to
include: the Middletown-Norwalk Project, the New England East-West Solution
Project, the TrAIL Project, the Susquehanna to Roseland Project, the Tehachapi
Project, the Gateway West Project, the PATH Project, the Maine Reliability Project
and the Baseload Project. Programs are underway or recently completed using the
following delivery methods:
A.
Northeast Utilities (http://www.nu.com/) Middletown-Norwalk Project
The Middletown-Norwalk Transmission Project involves construction of nearly 70
miles of new overhead and underground 345-kilovolt transmission line.
The project is among the largest transmission capital projects ever constructed and is
the first project to include a program management role in the power delivery industry.
It will use more 345-kilovolt cross-linked polyethylene (XLPE) underground
conductor cable than has previously been installed on any other single project in the
country.
This project is using Program Management under a Program Manager of Choice
contract with Northeast Utilities. The successful PMOC was Burns & McDonnell
(www.burnsmcd.com). Burns & McDonnell is responsible for permitting, design,
procurement, construction management, right-of-way acquisition, environmental
issues, community relations, and document control. The project is energized and won
the Utility Automation/ T&D 2008 Project of the Year and Engineering News
Records Energy Project of the Year.
B.
Northeast Utilities New England East-West Solution Project
New England East-West Solution (NEEWS) is four related transmission projects
developed by a working group of planners from Northeast Utilities, National Grid and
ISO-New England.
The problems were identified by ISO-New England (http://www.iso-ne.com/) in its
regional planning process. They are:
Limitations to east-west movement of electricity on the New England power
grid
76
Each of these problems poses a threat to the reliability of electric power in southern
New England and the region overall. These problems also negatively affect our ability
to benefit from access to cleaner, competitively priced electric power.
This project is using Program Management under a Program Manager of Choice
contract with Northeast Utilities. The successful PMOC is Burns & McDonnell.
C.
Allegheny Powers TrAIL Project (http://www.aptrailinfo.com/)
The PJM board (http://www.pjm.com) has approved a five-year program designed to
maintain the reliability of the transmission grid in the Mid-Atlantic region. The plan,
includes construction of approximately 210 miles of 500-kV transmission lines within
Alleghenys service territory.
Specifically, the plan calls for construction of a new 500-kilovolt (kV) line extending
from southwestern Pennsylvania to existing substations at Mt. Storm, W.Va., and
Meadow Brook, Va., and continuing east to Dominion Virginia Powers Loudoun
Substation. The plan incorporates portions of Alleghenys proposal that was
submitted to PJM as the Trans-Allegheny Interstate Line.
A preliminary estimate of Alleghenys portion of the expansion is in excess of $850
million.
Line mileages by state: Maryland10 miles, Pennsylvania40 miles, Virginia40
miles and West Virginia120 miles.
The Trans-Allegheny Interstate Line (TrAIL) (http://www.aptrailinfo.com/) will
extend from Southwestern Pennsylvania (37 miles) to West Virginia (114 miles) to
Northern Virginia (28 miles). Dominion Virginia Power will build the line from an
interconnection point east of Alleghenys Meadow Brook substation near
Middletown, Va. to its Loudon Substation. About 36 miles of the 500-kV line in
Southwestern Pennsylvania are needed to serve local demand in Greene and
Washington counties.
This project is using an open book EPC delivery method with a management fee paid
on milestones achieved through the life of the project. Successful contractor is Kinney
Construction.
77
D.
PPLs and PSE&Gs Susquehanna to Roseland Project
In June 2007, PSE&G (www.pseg.com) announced plans to invest nearly $1 billion to
construct new transmission lines to keep up with increasing demand and to ensure
continued reliability. The Susquehanna to Roseland line is the first of three proposed
PSE&G projects to be approved by PJM. This new 500,000-volt line will run from
PPLs Susquehanna substation near Berwick, Pa., to PSE&Gs substation in
Roseland, NJ. The entire line will be about 130 miles long. The New Jersey portion of
the line will be 40-50 miles long.
On August 5, PSE&G selected the preferred route for the new Susquehanna-Roseland
power line. This decision was based in part on a detailed study performed by a New
Jersey-based siting contractor with national experience on projects such as this (read
the routing report). PPL Electric Utilities (PPL) also announced the selection of this
route as the best pathway for this new line. In addition to the placement of new towers
and lines, PSE&G also will construct two new substations -- one in Jefferson
Township, Morris County and another at its existing property in the Roseland/East
Hanover area.
Based on a number of critical factors, PSE&G has determined that this route is the
best choice for the project because it:
Would be constructed entirely within an existing 230,000-volt (230kV)
transmission right-of-way, which already contains transmission structures and
wires, for its entire length in New Jersey. This minimizes the impact to the
environment, requiring no construction on virgin right-of-way and minimal
clearing of vegetation.
Crosses the least amount of wooded wetlands and forested lands, and has the
least potential to permanently alter these important habitats.
Has the least impact on aesthetics where the line will be built and provides the
least environmental, engineering and construction challenges.
On January 12, PSE&G submitted its application to the New Jersey Board of Public
Utilities (http://www.state.nj.us/bpu/) to construct the Susquehanna-Roseland power
line. PSE&Gs partner in this project, PPL Electric Utilities, also filed an application
for
approval
with
the
Pennsylvania
Public
Utilities
Commission
(http://www.puc.state.pa.us/). We expect that the BPU will issue a decision regarding
this project later this year. The BPU process will include public meetings and
hearings.
During the review period, PSE&G will conduct final design work to determine the
location of each of the new structures and temporary access roads, as well as other
construction details.
In addition to the BPU review process, the project will undergo environmental review
by all government agencies with jurisdiction.
78
The line is expected to cost between $1.2 billion and $1.3 billion. The New Jersey
portion will cost approximately $750 million, which includes construction of new
substations in Northern New Jersey. The cost will be shared by the 51 million electric
customers in the PJM region, which includes 13 states and the District of Columbia.
The Pennsylvania portion of this project is using Program Management under a
contract with PPL. The successful Program Manager is Burns & McDonnell. The
New Jersey portion of this project is proposed for Program Management and PSE&G
is in the process of evaluating proposals.
E.
Southern California Edisons Tehachapi Project
Southern California Edison Company (SCE) (www.sce.com) is proposing to
construct
the
Tehachapi
Renewable
Transmission
Project
(TRTP)
(www.sce.com/PowerandEnvironment/Renewables/Wind/) to provide the electrical
facilities necessary to interconnect new wind turbine based electrical generation in
excess of 700 megawatts (MW) and up to approximately 4,500 MW from the
Tehachapi Wind Resource Area (TWRA). The TWRA has been identified as the
richest wind resource area in California. The proposed TRTP would consist of a series
of new and upgraded high-voltage transmission lines (T/L) and substation facilities to
deliver electricity from new wind farms in eastern Kern County, California, to the Los
Angeles Basin.
The Tehachapi Renewable Transmission Project (TRTP) Segments 1 to 3 is a series
of new and updated electric transmission lines and substations that will deliver
electricity from new wind farms in the Tehachapi area to SCE customers and the
California transmission grid. TRTP is a vital part of meeting Californias renewable
energy goals. The California Public Utilities Commission (CPUC) approved project
Segments 1 to 3 in March 2007 and construction is now underway.
The proposed TRTP consists of eight segments enumerated as Segment 4 through
Segment 11. Proposed Segments 4, 5, and 10 that would involve upgrading and
expanding SCEs transmission system north of SCEs Vincent Substation in order to
integrate TWRA wind generation to SCEs electric system. Proposed Segments 6, 7,
8, and 11 would involve upgrading and expanding SCEs transmission system south
of SCEs Vincent Substation in order to deliver TWRA wind generation to SCEs
load centers in the Los Angeles Basin. Segment 9 would involve building a new
substation (Whirlwind Substation in Kern County), expanding two existing
substations (Antelope and Vincent substations), and upgrading three substations
(Gould, Mesa, and Mira Loma substations) (refer to Figure ES-1).
Southern California Edison (SCE) has chosen to use a modified form of Program
Management called Owners Agent for the TRTP 123 and TRTP 4-11. Unlike
Program Management, the SCE version of Owners Agent does not allow for right-ofway acquisition, the procurement major equipment or the performance of detailed
design work. SCE has a significant staff and the services of several large A/E firms
to support this effort. The successful Owners Agent is Burns & McDonnell
79
F.
Idaho Powers Gateway West Project
Idaho Powers (www.idahopower.com) Gateway West project will supply present and
future needs of customers and improve electric system reliability in the service
territories of both companies. In addition, the project will enable delivery of new
generating resources, including wind, to more customers in the region. Known as
Gateway West, line segments are scheduled to be completed in 2014. At certain times
of high customer demand, transmission lines reach full capacity. Idaho Power and
Rocky Mountain Power are committed to ensuring reliable electric service to
customers and respond to others requesting use of the transmission system. Work has
begun to complete this project in the next five to seven years.
The companies are proposing to construct, operate and maintain approximately 1,150
miles of new 230 kilovolt (kV) and 500 kV electric transmission lines between the
planned Windstar Substation near Glenrock, Wyoming to the planned Hemingway
Substation near Murphy, Idaho.
This project is using the EPC method of project delivery. Power Engineers has been
hired as the Owners Engineer to prepare the EPC bid package.
G.
AEPs PATH 765kV Project
AEPs (www.aep.com) PATH project will employ advanced transmission
technologies to make it the most reliable, efficient and environmentally sensitive
project possible. Extra-high voltage transmission inherently costs less and requires
less land to carry the same amount of power than other transmission designs, and new
765-kV designs reduce line losses 40 percent over older designs. That means the
PATH line will transport electricity more efficiently and reduce the power generation
necessary to meet electricity demand. Additional advancements in 765-kV and 500kV conductor designs and new control technologies in the project's substations also
will enhance the reliability and efficiency of the PATH lines compared with other
transmission lines.
The PATH project includes approximately 244 miles of 765-kV extra-high voltage
transmission from AEP's Amos substation near St. Albans, West Virginia, to
Allegheny's Bedington substation, northeast of Martinsburg, West Virginia. Another
46 miles of twin-circuit 500-kV transmission will be constructed from Bedington to a
new substation to be built near Kemptown, southeast of Frederick, Maryland. The
total project is estimated to cost approximately $1.8 billion. AEP's estimated share of
the project is approximately $600 million.
This project is using the design/bid/build method of project delivery. AEP is using
internal design resources on the 765kV transmission and will contract for outside
construction services once they obtain the needed permits to construct.
80
H.
Maine Power Reliability Project
The Maine Power Reliability Program (www.mainepower.com) includes a proposal
to build a new, 345-kilovolt (kV) transmission line from Orrington, Maine (15 miles
south of Bangor), to Newington, New Hampshire. The line will follow existing
transmission corridors through nearly 80 Maine towns, including Detroit, Benton,
Windsor, Lewiston, Yarmouth, Gorham, and Eliot. The program includes investments
in new substations, upgrades to existing substations, and improvements to the 115kilovolt (kV) electric system in central Maine.
The Maine Power Reliability Program began in January of 2007 with a study to
project the region's future needs for electricity service. The first phase of the study,
which was completed last summer, found that serious problems would emerge as
early as 2012 without significant changes in demand patterns, transmission capacity,
or new supply. The proposal includes transmission investments and recommendations
to encourage alternatives to transmission, such as new generation, or programs to
manage the growth in peak electricity demand.
CMP prepared its plan in conjunction with neighboring utilities in Maine and New
Hampshire with oversight by ISO New England (www.iso-ne.com), the organization
responsible for managing electricity supply and transmission for the New England
states.
The preferred route calls for the construction of an additional 345kV transmission line
following existing transmission corridors from Orrington to the Maine-New
Hampshire interconnect in Eliot (Orrington-Detroit-Benton-Windsor-LewistonGorham-Eliot-Newington, NH). CMP is also planning to build several new 115kV
lines and upgrade others on existing corridors.
This project is using the Program Management approach. Central Maine Power
(http://www.cmpco.com/) has very limited staff and has selected Burns & McDonnell
as the successful Program Manager.
I.
Progress Energy Floridas Baseload Project
Progress Energy Florida (www.progress-energy.com) is considering adding about 200
miles of transmission lines across nine counties to maintain reliability and to move
energy efficiently to customers throughout the region and state. The lines would be
230 and 500 kilovolts (kV).
Transmission lines are planned in three main segments: from Levy County south to
Hernando County, from Pinellas County east to Kathleen in Polk County, and from
Levy County east to the Wildwood/Leesburg area. These proposed transmission lines
could affect the counties of Citrus, Hernando, Hillsborough, Lake, Levy, Marion,
Pinellas, Polk, and Sumter.
The transmission lines are expected to be built on two types of structures. Typically,
the single steel poles are between 90 and 165 feet tall; the H-frame (two-pole)
81
structures about 120 feet tall. The structures would be spaced 700 to 1,300 feet apart,
depending on the structure and terrain.
This project is using the EPC and Program Management approach. Progress has hired
Patrick Engineering as an Owners Engineer to prepare 500-kV standards and EPC
bid packages. In addition, Progress plans to hire a Program Manager to handle
construction management, inspection, budgeting, scheduling, public relations and
more. The current plan is to allow the future Program Manager the opportunity to bid
on the EPC work.
82
ABSTRACT
When TXU Electric Delivery (now known as Oncor Electric Delivery) filed a CCN
application with the Public Utility Commission of Texas (PUCT) in 2006 it planned
to construct a new 345 kV line approximately 6.5 miles long on single pole
structures. During the PUCTs review of the application the City of Dallas intervened
in the process and requested that a portion of the line be constructed using UG
technology.
The new line was necessary to provide additional capacity and reliability to the
Dallas Central Business District. The circuit needed to be rated at 3400 Amperes.
After much deliberation, the PUCT ruled that Oncor should construct a 0.7 mile
segment of the line underground.
Thus began the first 345 kV underground project on the Oncor system. There are a
number of 345 kV underground projects in service throughout the United States but
none with a dynamic rating capability near the 3400 Ampere range required for this
project. Selecting the correct technology and materials and developing a design to
properly implement them has presented engineering (and political) challenges that
Oncor will not soon forget.
The line is currently under design and will be constructed Spring through Fall of
2009.
Background
In March, 2006 Oncor Electric Delivery filed an application for a Certificate of
Convenience and Necessity (CCN) with the Public Utility Commission of Texas
(PUCT) to construct the W. Levee-Norwood 345 kV line. The line would consist of
two 345 kV circuits, with one circuit installed initially and the second as system
needs dictated. The line would connect two existing major switching stations and was
necessary to provide additional capacity and reliability to the area near downtown
Dallas. As required in the filing Oncor had selected preferred and 10 alternate routes
between the two existing stations.
During the PUCT review of Oncors application a number of parties exercised their
right to intervene and become a part of the regulatory process, including the City of
Dallas. The City desired to modify the route of the line and also requested the PUCT
to direct Oncor to construct a portion of the line using underground technology
instead of overhead. The City contended that an underground line in the area near
downtown would be more aesthetically pleasing and that a portion of the route was
83
84
subject to near term development plans due to the Citys proposed Trinity River
Project which had been approved by voters in a May 1998 bond election. The City of
Dallas also offered to reimburse Oncor for 25% of the difference between cost of the
installed underground line and what the overhead alternative would have cost.
The PUCT agreed with the Citys proposal and directed Oncor by a 2-1 vote to
construct a 0.7 mile section of the line underground. The remainder of the line would
be constructed utilizing over methods but a large portion of the line would be built
within the median of Irving Blvd., a major arterial street.
Thus began a process that is currently entering the construction phase and is
scheduled to be complete by December 2009.
Underground
West Levee
5
Figure 1. Project Route
In order to keep from having to de-rate the load transfer capability of the new circuit,
the new underground circuits would need to provide the ability transfer the same
load, 3400 Amperes.
Oncor has approximately 60 circuit miles of underground transmission line in its
system, all at 138 kV. Industry contacts at other utilities and other sources identified a
number of applications at 345 kV both in the United States and abroad but none at the
3400 Ampere capacity required for this project.
Oncor decided to seek outside engineering expertise to assist in the design and
construction management areas of the project. Oncor developed a Request for
Proposals from multiple firms qualified in this area of design. The RFP included an
evaluation of the two most prevalent underground cable technologies, extruded cross
link polyethylene (XLPE) solid dielectric cable systems and conventional high
pressure fluid filled (HPFF) pipe type systems.
After a detailed and thorough evaluation of the proposals submitted by the design
firms, Oncor selected Power Engineers, Inc. to provide design, material evaluation,
construction management and support services for the project. The decision was also
made to proceed with a conventional HPFF underground system. The decision was
based primarily on the following items:
The suitability and robust nature of the HPFF underground system for use on
a high capacity (3400 A.) through flow transmission line like this project.
Total installed cost analysis for this project, including associated
pressurization systems and cathodic protection systems.
Oncors operating experience with existing HPFF underground systems on its
transmission system.
Materials
The various materials required for use on the underground line were evaluated with
respect to the project design requirements. The major items selected for use on this
project are listed below with a brief discussion of reasons for their selection:
Conductor: 3-3500 kcm Copper, Compact Segmental PPP Laminate Insulated
cables per phase
Pipe: 10 ID steel with flared ends ASTM A523
Coatings: Liberty Coating Pritrec 10/60 Extruded Polyethylene
Dielectric Fluid: Polybutene Insulating Fluid ASTM D 1819 84
Thermal Backfill: Fluidized Backfill RHO of 60 thermal ohms or less
Design
The route of the underground section of the line was determined by the PUCT. It
followed a path North from the W. Levee Sw. Station crossing under Singleton Ave.,
a major arterial street leading into downtown Dallas, and continued under the
pavement of the South/West bound lanes of Canada Dr. The final design of the
underground line section turned out to be approximately 0.9 mile in length.
85
This route posed two immediate problems (opportunities) for the designers. The first
issue was to determine how to cross Singleton Ave. After consultations with the City
it was determined that Singleton Ave. could not be closed continuously for term of
construction activities in this area. It was decided that the use of directional drilling
technology would be the best solution to install the UG circuits across this street.
The second issue was to determine how to install the circuits under Canada Dr. The
City agreed that portions of Canada Dr. could be closed for periods of time adequate
for conventional open-trench construction. This would allow the use of more readily
available construction equipment and would also reduce the construction window
significantly for that portion of the line.
Each cable trench is designed with a number of specific features. The HPFF steel
cable pipes are located parallel and approximately 3 ft. on center. Each pipe is placed
near the bottom of the trench on a sand bag cushion at a depth of approximately 4-4.5
ft. Each HPFF pipe has a 1 PVC temperature monitoring duct attached to the pipe
surface. The trench also includes a 4 diameter oil supply line pipe and a 4 PVC
communications duct. A 1 ft. protective layer of low compression strength Fluidized
Thermal Backfill (FTB) material is placed around and over the steel pipes and ducts
to aid in the transmission of heat away from the HPFF pipes. The FTB material
contains a red dye to aid in the prevention of dig-ins after the underground lines are
in-service. A 6 layer of 3000 psi concrete is then placed on top of the FTB material
to further protect the underground system. The trench is then backfilled with layer of
compacted native material to the elevation of the street sub-grade. A typical crosssection is shown below:
86
87
The proposed line has also been designed with a Dynamic Thermal Monitoring
System (DTS) with the capability to continuously monitor the thermal characteristics
of the entire line segment via fiber optic cable. This system provides Oncor the
capability to dynamically rate the capacity of the line on a real time basis based on
the changing thermal conditions of the line.
The line is also equipped with a leak detection system with the capability to
continuously monitor and detect leaks as small 3 gallons per hour within a 24 hour
period. The system also includes flow rate indicators, cumulative flow indicators,
temperature data from the cable system and the earth, and load information from the
substation feeder.
Construction
The design of the underground section of the line was completed in September 2008
and Oncor issued a Request for Proposals from two potential contractors to construct
the line and provide many of the required material including:
UG Cable
Pumping Plant and Pressurization Control System
In a move to ensure the availability of the specialized HPFF steel pipe required for
this project Oncor had previously purchased the pipe and made arrangements to have
the coating applied. Oncor also made plans to acquire the 18 overhead to
underground terminations from G&W Electric Company due to the long lead time
(30 week) for this product.
In October 2008 Oncor selected Prysmian Power Cables and Systems, LLC as the
primary construction contractor for the project. Prysmian had a long working history
with Oncor on many HPFF underground systems installed on its transmission system.
Construction began on the underground section of the line in February 2009 with the
installation of six (6) directional drill bores, each approximately 600 ft. in length in
the section under Singleton Ave. Construction of the open-trench section under
Canada Dr. began in late April 2009 and continues at this time with a proposed
completion date of October 2009.
Figure 3.
Directional Drill Section - 10 Pipe being
lifted and aligned in preparation for back
pull through bore hole.
F
Figure 4.
Directional Drill Section - 10 pipe in final
position. City of Dallas skyline in
background.
Figure 6.
Directional Drill Section 10 HPFF
and 4 Oil Supply Line Yoked for back
pull through bore hole.
Figure 7.
Directional Drill Section Jetted rotating
mandrel begins back pull through bore hole.
88
Figure 8.
Transition Section Directional Drill to
Pipe in Trench Section
Figure 9.
Transition Section FTB Placement
Figure 10.
Pipe in Trench Section FTB Placement
Figure 11.
Trifurcating Manholes at Topeka
89
Figure 12.
Trifurcating Manholes at Topeka
Figure 13.
345 kV Double Circuit
Tangent Structure
Figure 14.
345 kV Tangent Structure
w/138 kV Underbuilt Circuit
90
Due to the line route, structure heights and conductor loadings, galvanized tubular
steel structures with base plates and anchor bolts were selected for use on this line.
Line Design Issues
The route selected by the PUCT posed a number of challenges to the design and
successful construction of the overhead section of the project:
An approximate 2.1 mile segment of the Irving Blvd. route was designated as
State Highway 356 and was under the control of the Texas Department of
Transportation (TXDOT). TXDOT prohibits the placement of transmission
line structures inside highway right-of-way.
Approximately 1.2 miles of the line lie within the levee system of the Trinity
River. This segment required permits from the US Army Corps of Engineers
(USACE). These line segments also required an easement from the City of
Dallas since the area inside the levees are designated as recreation area by the
City.
The entire 4.9 mile segment inside the median of Irving Blvd. required
approval from the City for any modifications to the existing medians and
street required to create a space to place the proposed transmission poles.
The City of Dallas, TXDOT and the North Texas Tollway Authority (NTTA)
were in the design phase of a new Sylvan Ave. bridge across the Trinity River
and the new Trinity Parkway toll road parallel to the river. The new bridge is
to be located approximately 160 ft. west and parallel to the proposed route of
the new transmission line. The bridge construction will not take place until
2010-2012 and therefore required close coordination between Oncor, the City,
TXDOT and NTTA to ensure proper structure placement to eliminate
conflicts with the construction of the future bridge and parkway.
There are three 138 kV lines that the new line is required to cross. One of
these is located within the levee footprint and the existing structure can not be
relocated and changed out or lowered to make a crossing easier to achieve.
Although the vast majority of the overhead section of the line is located on
public property, the route crosses four pieces of private property and will
require the acquisition of easements from these owners.
The route crosses one railroad and thus requires the acquisition of a crossing
permit from the railroad.
A number of the pending issues were not of a design nature and Oncor was
responsible for coordinating with the City of Dallas, TXDOT, NTTA and the USACE
to reach a favorable resolution of those issues.
First, the City of Dallas and TXDOT began negotiations to transfer the ownership and
thus the operating and maintenance responsibilities for the Irving Blvd. section of SH
356 to the City of Dallas. These negotiations also were expanded to include the City
of Irving regarding a section of SH 356 that was located within the city limits of
Irving. As it turned out when TXDOT and the cities reached agreement it was
91
discovered that the transfer would also require the approval of the Texas Attorney
General and finally the Governor. These negotiations were successfully consummated
in January 2009 when Texas Governor Rick Perry signed an order finalizing the
transaction.
Oncor was required to prepare and submit a crossing permit to the US Army Corps of
Engineers (USACE) for two crossings of the East Fork of the Trinity River. At these
locations, all structures were designed to be outside of the USACEs zone of
concern with respect to the levees. The USACE has heighted the review of
structures that penetrate the footprint of the levees since Hurricane Katrina in 2005.
At present time, the USACEs review of Oncors submittal is in progress.
Oncor was also required to obtain an easement from the City of Dallas for the same
Trinity River crossing since the City is the owner of the property and operates the
Trammell Crow Park and Lake. Oncor prepared an easement submittal package and
submitted it to the Dallas Park Board for their action. The Park Board held a public
meeting and approved and forwarded their recommendation to the Dallas City
Council for final action. The City Council approved the Park Boards
recommendation and authorized granting the easement in June 2008.
Overhead Line Design
Oncor engaged the services of M&S Engineering, Inc. to perform the line design,
assist in the material evaluation, assist Oncor in the preparation of permit applications
for submittal to the USACE, City of Dallas and the applicable railroad, design street
modifications and assist in the construction management activities on this project.
M&S Engineering began to immediately address the remaining issues affecting the
design of the line. They began working with the LIDAR data and began preliminary
structure spotting followed by site visits to resolve conflicts with the numerous
overhead utility crossings.
After preliminary structure locations were developed, meetings were scheduled with
the City of Dallas to resolve issues where modifications to the road median would be
required to enable the placement of a structure. The Public Works & Transportation
Department of the City of Dallas worked with Oncor to finalize the design of the
required street modifications. Final construction plans for the street modifications
were approved by the City in June and construction is scheduled to begin in early
July 2009.
The crossing of the three existing 138 kV lines was the next issue that had to be
resolved. Two of the crossings were resolved by changing out existing double circuit
structures with multiple shorter single circuit concrete poles and removing the shield
wires in the crossing span. The third crossing was more of a problem. As stated
previously, the existing structure was located within the levee template and could not
be replaced by a new shorter structure. It was decided to design a replacement for the
bottom conductor arm to support two phases and lower the top phase and shield wire
92
to the next lower arm positions. This work enabled the new 345 kV poles crossing the
138 kV line to be 20 ft. shorter and avoided the FAA requirement to install aviation
lighting on structures.
The remaining line design issues were completed in early June 2009 and a preconstruction meeting for the project was held on June 17, 2009.
Construction
The design of the overhead section of the line was completed in March 2009 and
Oncor issued a Request for Proposal from potential vendors to provide the major
items of material required for the line. The following is a list of the vendors selected
and the items they are providing:
FWT, Inc. - Tubular Steel Poles
Southwire, Inc. ACSR Conductor
MacLean Power Systems, Inc. Insulator Assemblies
Hubbell Power Systems, Inc. Insulator Assemblies
Brugg Cables, LLC Fiber Optic Cable
Oncor began geotechnical investigations for the structures located outside of the
levee system in May 2009. The geotechnical investigations were conducted by
TEAM Consultants, Inc. The USACE also approved the boring plan for the structures
located inside the levees in late May. The geotechnical investigations at all accessible
sites were completed in June 2009.
Oncor completed the design of the structure foundations for all accessible poles on
June 24, 2009.
Oncor selected Chapman Construction Company as the primary construction
contractor to construct the overhead segment of this line. Chapman began structure
foundation construction in July 2009 and plans to begin the installation of structures
during the Summer and conductor during the Fall 2009.
The project has a scheduled completion date of December 15, 2009.
93
94
INTRODUCTION
The paths of transmission lines include varying geographies: varying in
terrain, subsurface conditions, environmental sensitivity, and accessibility.
Approximately one third of the cost of each transmission tower/structure is beneath
the ground surface and is never seen. The value of these foundations is of
tremendous importance because remediation of a failed structure foundation incurs
far greater cost than any other type of single-structure failure. Conventional deep
foundations, which are cast-in-place drilled piers, are still the most economical
solution for many structures, but when certain obstacles are encountered, micropile
foundations become the most economical solution. These high costs, high values,
and different foundation options elicit a closer look at this invisible portion of the
structure.
OBSTACLES
Efficient foundation design and construction are the major topics that will be
discussed in this paper. One key to efficient design of proper foundations is the
acquisition of accurate and adequate geotechnical and geological information of the
tower sites, as this information greatly influences the foundation design requirements
for each tower. An accurate geological survey requires quality survey data and
supporting geotechnical information. A proper geotechnical investigation requires,
among many things, access to critical locations on the site with small investigatory
drills and associated equipment. Gaining such access is often an obstacle that can be
very difficult and involve much more than will be discussed in this paper. It should
be noted, however, that such access and permitting challenges will be similar or more
stringent in the construction phase of the project, and the foundation design team
should be aware of such obstacles.
Building transmission lines in remote areas brings about many
constructability issues. Limited access can cause problems with getting equipment
and material to the sites, which can increase time and cost. Some areas can be
restricted such that no vehicular traffic is allowed to the structure sites at all. This
can result in the need for hand digging foundations and/or blasting with explosives.
This is where safety concerns increase tremendously in traditional construction
techniques. Traditional foundations for deadend structures can extend to depths of 20
to 30 feet below grade. This introduces many safety concerns for hand-dug
foundations. A typical 500-kV single circuit deadend structure can require up to 15
yards of concrete per leg. This requires approximately 60 helicopter trips with a
single-yard concrete bucket. Clearly, this is both lengthy and expensive.
INVESTIGATION
Geological and geotechnical investigation programs involve the study of the
local terrain for stability and drainage, insitu tests such as standard penetration tests
(SPTs), collection of disturbed and undisturbed soil samples, and pertinent laboratory
95
tests. Foundation design parameters are then estimated from the collected test data.
The types of foundations chosen depend upon these geologic and geotechnical
conditions identified, the tower loading, and various other factors, such as access
constraints and environmental restrictions.
A limitation of typical geotechnical investigation programs is the number of
locations that can be practically and economically explored during the design phase
of the project. Typically, geotechnical borings are obtained at deadend locations and
at least one per mile. This seems to be a commonly accepted practice, but leads to
interpretation and conservatism in geotechnical properties to account for possible
variations between boring locations.
After proper geological and geotechnical investigation programs have been
performed and the respective reports and associated recommendations have been
conveyed to the foundation designer, the final foundation design process begins. The
final design considers the many different requirements and limitations, and finally
produces foundations that satisfy all such controlling factors.
CONVENTIONAL FOUNDATIONS
The design of conventional foundations (also called traditional foundations)
often involves drilled-shaft-type deep foundations of sufficient size to support the
specified structure loading in the worst expected geologic and geotechnical
conditions identified or suspected in the geological and/or geotechnical reports and
recommendations. These types of foundations are referred to as conventional (or
traditional) foundations for good reason: they have been effectively used for
transmission structure foundations for the last century, and will continue to do so.
Where both access is reasonable and subsurface conditions are favorable,
conventional deep foundations are the optimum choice. An excepted limitation of a
traditional foundation is that the loads supported by the foundation are usually too
large to economically perform insitu testing to verify adequacy of the design. This
causes additional factors of safety to be applied to the design to insure an adequate
foundation. It also impedes the ability to streamline the design if soil conditions are
better than expected.
INTRODUCTION TO MICROPILES
Before discussing the benefits of micropile transmission structure
foundations, the following is a brief introduction to micropile technology, which is
still not well-understood by many and thus continues to be labeled as an emerging
technology.
Micropiles were conceived in Italy in the early 1950s, by Dr. Fernando Lizzi
(Armour, T., et al., 2000). They were introduced into the eastern United States nearly
20 years later and have continued to grow in applicability and versatility for the last
30 years. A micropile is a small (about 4- to 12-inch-diameter) replacement pile -
96
versus a displacement pile, such as a driven pile. It is typically installed using a small
drill (weighing 15- to 40-thousand pounds) and can be installed in virtually any
geotechnical condition, from soft silts and clays to hard, igneous bedrock. During the
drilling, the hole is stabilized (when necessary) using high grade steel tubing (casing),
which is often extended behind the drill bit to the bottom of the hole. The drill steel
and bit are then completely withdrawn from the fully-cased hole, a string of
continuously-threaded rebar is inserted which extends the full length of the hole, and
the hole is then tremie-filled with neat cement grout. The casing is withdrawn,
exposing a portion of the micropile grout to the soil (this portion is called the bond
zone), and leaving a portion of the casing in place around the upper portion of the
micropile for connections and flexural performance. Figure 1 is an illustration of a
typical composite micropile.
A typical composite micropile consists of these three main elements: a single
strand of continuously-threaded rebar (bar) that extends the full length of the
micropile in the center of the micropile; a design-specified length of steel tubing
(casing) around the circumference that extends from the top of the micropile to a
critical depth (about half the length of the micropile in Figure 1); and neat cement
grout, encasing the bar, and bonding the micropile components together and to the
geotechnical strata. Other ancillary components of the micropile include items such
as couplers, which connect segments of bar to form a continuous element, and pvc
centralizers, which position the bar in the center of the drill hole. Typical sizes of bar
(nominal diameters) range from 1 to 3 inches, and typical strengths of such bar are
75,000 to 120,000 psi, minimum yield. Typical sizes of casing (outer diameter by
wall thickness) range from 5 inches by 0.375 inches to 8-5/8 inches by 0.500 inches,
with minimum yield strengths ranging from 45,000 to over 100,000 psi. Typical neat
cement grout is composed of two primary elements: Portland cement and water, with
28-day unconfined compressive strengths ranging from 3,000 psi to more than 5,000
psi.
A single typical micropile can exhibit a capacity ranging from as little as 25
kips up to and exceeding 500 kips. While the lateral capacity of a single micropile is
understandably limited, groups of appropriately arranged micropiles can support
tremendous loading - both axially and laterally. The size, length, and configuration
of micropiles are designed to accommodate the magnitudes and proportions of the
different loading (axial loading, base shear, and overturning moment) on a projectspecific basis.
Micropiles are not inexpensive foundation elements, nor is their cost as
volatile as that of traditional foundations when limitations are imposed on access
and/or construction procedures. Such limitations highlight the clearest benefit that
micropiles can provide. The construction techniques for micropile installation are
refined and controlled enough to satisfy many environmental concerns and still retain
flexibility to meet higher, project-specific construction demands or limitations.
97
MICROPILE FOUNDATIONS
When factors such as access constraints and environmental restrictions
become substantial obstacles, the advantages micropile foundations have over all
other deep foundations, which are mostly construction-related, have significant
effects on schedule and cost.
Micropiles can be installed through virtually any subsurface condition, from
soft sediments and clays to hard, competent rock. This may seem counterintuitive to
many, but in harder and more competent rock, the penetration rate for micropile
installation typically becomes noticeably faster than in granular soils or less
competent rock formations. Different subsurface conditions pose different obstacles,
which will be highlighted in a few case histories.
98
Drill rigs used for micropile installation are small and light-weight (typically
15- to 40-thousand pounds). Specialized rigs are modular and sometimes even
smaller, so mobilization of these rigs to limited access locations with medium-lift
helicopters has become standard practice.
When subsurface conditions vary from those expected, the installation
procedures for micropiles are flexible enough and have become streamlined to
accommodate such variance by modifying the lengths, number, and/or configuration
of the micropiles at each site. Such modification is part of design contingencies
outlined in the construction documents in something called a decision tree or flow
chart. With such a guidance tool, drill operators and inspectors have clear directives
to follow in the event of such variant subsurface conditions.
Another unique quality of micropiles, because of their size and geometry, is
their ability to be economically tested to validate their capacity. This ability reduces
the need for additional conservatism and increased Factors of Safety - which are used
to accommodate uncertainties in foundations that are not tested.
Health and safety are concerns in any type of construction, but the hazards
present during micropile construction, while they do exist, are minor and are
managed through regular, simple safety training protocols.
Other benefits of micropiles (which are not typically important in this
industry) are the compact sizes of many micropile drills for use in restricted access
and head space, as well as low noise and vibration emission.
Again, micropile foundations are not an inexpensive system, so they are not
beneficial when a traditional foundation can be easily constructed, but the cost of
micropile foundations do not escalate as rapidly as traditional foundations when
various constraints and restrictions are imposed or adverse ground conditions are
encountered.
MICROPILES IN THIS INDUSTRY
The first micropile foundations for transmission structures in this industry
involved triangular and rectangular micropile configurations with steel or concrete
pile caps. The advent of radial micropile arrays has revealed numerous design and
construction advantages. Radial-arrayed micropile configurations, installed with
specialized equipment, can be completed with surprising accuracy and efficiency. In
completely helicopter-supported site conditions, micropile foundation elements for all
four legs of a 500-kV suspension tower have been installed in less than a day. In the
same conditions, the same accomplishment for a 500-kV deadend tower has been
achieved in under three days.
99
Both steel and concrete pile caps have been utilized for non-lattice structures.
Concrete pile caps are currently the preferred choice for lattice towers, but
advancements are continually being investigated and implemented for improved
design and construction efficiency.
Even today, as constraints and restrictions grow and impede project
advancement, micropile design and construction techniques continue to evolve to
accommodate these increasing limitations. Some of these interesting evolutions and
solutions will be illustrated in a few case histories.
The current state of practice for micropile foundations in the electric
transmission industry is still developing as the industry leaders gain familiarity and
comfort with this technology.
INTEGRATION OF THE DIFFERENT FOUNDATIONS
Kangley-Echo Lake Transmission Line (500-kV). One of the first major attempts
to implement micropiles as lattice transmission tower foundations was in part of the
Kangley-Echo Lake 500-kV transmission line that extends through a protected
watershed east of Seattle, Washington (Mathieson, W.L., et al., 2004). At that time
(15 years ago), micropile construction techniques and efficient environmentally
sensitive practices were still advancing, and only one tower was constructed using
micropiles on this project.
Swan-Tyee Intertie (138-kV). This 138-kV line connects the Four Dam Pool Power
Agencys hydroelectric facilities at Swan Lake and Lake Tyee. This 57-mile segment
of line includes more than 250 structure locations with more than 350 foundations, all
of which are supported by micropile foundations with steel pile caps. These
structures consist of guyed single-shaft structures, guyed and un-guyed Y-structures,
guyed H-structures, guyed 3-pole structures, and guyed A-frame structures for major
water crossings. Lateral structure loading was supported via battered foundation
micropiles and guy wires. Figure 2 is an example of one of these micropile
foundations. This was the first micropile transmission line foundation project of this
size, and micropiles were chosen because of the remote, mountainous terrain of the
structure sites. This project was completely helicopter-supported. Another major
variable encountered from site-to-site was the subsurface condition, ranging from 3blow-count silt to 25,000-psi rock. With a foundation construction schedule of
directions to follow, dependent on structure type and subsurface conditions
encountered, micropile foundations were the only feasible approach to complete this
project within the narrow seasonal window available.
100
101
102
Ebey Slough (230-kV). The rebuild of this 2.5-mile section of transmission line
consisted of the replacement of 80 old wood poles with 15 new steel, single-shaft,
self-supporting structures, carrying two 230-kV circuits. Of these 15 structures, the
foundation construction for 10 was possible only with micropiles. These 10
structures are located in a sensitive estuary, having strict environmental restrictions,
and silty, organic geotechnical material that simply could not support more than
about 500 psf on the surface. Blow counts (N-values) for the first 50 to 75 feet of this
material were between about 0 and 5 blows per foot. Directly beneath this material
was the bearing unit of dense granular material. Because of these rare subsurface
conditions, micropile lengths ranged from 70 to 120 feet. Special transportation
vehicles, drilling equipment, and construction techniques were utilized to
accommodate these unique challenges. High-capacity marsh buggies were assembled
and used for transporting equipment, material, and crews to the structure sites;
unique construction platforms were designed, assembled, and placed; and modular
drill rigs were customized to drill these micropile foundations in a controlled, precise,
and efficient manner. The suspension structures required 10 to 16 micropiles per
structure, with a 9- to 10-foot-diameter concrete cap, and the one deadend structure in
this section required 36 micropiles with an 18-foot-diameter concrete cap. This
deadend foundation was designed to support a 15,000-kip-foot moment, and to
exhibit less than three degrees of rotation. Figure 5 is a photo of this deadend
foundation.
103
104
CONCLUSION
In this industry, traditional foundations are practical, trusted, and often the
most economical choice for transmission structures; however, when varying
obstacles and restrictions present themselves, threatening budget and schedule,
micropile foundations can provide solutions that are being realized by the industry.
REFERENCES
Armour, T., Groneck, P., Keeley, J., and Sharma, S. (2000). Micropile Design and
Construction Guidelines Implementation Manual, 1.5-1.6.
Mathieson, W. L., Mitchell, C. L. B., Robertson, C. A., and Rustvold, J. (2004).
Micropiles to Support Electric Power Transmission Towers. Proceedings of
the 29th Annual Conference on Deep Foundations, 2004, Vancouver, British
Columbia, Canada, 123-136.
105
106
Abstract
With increasing need for design support on transmission lines, engineers need to be
comfortable in evaluating limited geotechnical foundation design data spread out over
long distances.
This report will show that through merging engineering geology and geotechnical
engineering evaluations, data can be grouped into sections comprised of different
design parameters along the length of a transmission line. The goal of this merger is
to create a better characterization of the foundation design data along an entire
alignment. Some tools and methods used by others are presented for consideration.
When done well an engineer can better understand the nature of the foundation
conditions along the alignment and why/where the geotechnical design parameters
should be used and changed as a result of how they are grouped around similarities in
geologic conditions.
Introduction
To understand the needs and considerations necessary to design foundations for a
transmission line both the engineering geology and overall geotechnical conditions
must be combined. Often there is too much of an emphasis on drilled borings located
at a predetermined distance and drilled to specified depths. Instead it is important to
look for trends expressed through changes in geology. Such changes then need to be
confirmed by geotechnical sampling at a frequency that matches the needs of a
project. These trends in data must then be presented so that they may be reviewed
and implemented into the design process.
107
Geophysical investigation
Testing of soil and rock for strength and index tests
Geotechnical design recommendations
Power Administration
Power Delivery
Design Firm
Utility Company
Entity
Required
Boring
Depth
50', 70' if
blow counts
are 15 or
less at 50'
Drilled Boring
Locations
Every tower
with one boring
1
*
At angle
points
> 3.99
degrees
30'
1 boring every
1/2 mile or
angle points,
dead ends,
changes in
geology, and
roads
30'-45'
dependent
on structure
type
1 boring every
mile or angle
points, dead
ends, 3 degree
change in
alignment,
changes in
geology
Required
Depth to
Core Into
Bedrock
Sampling
Intervals
Within a
Boring
Complete
full depth to
50'
10'
20'
Environmental
Sampling
Requirement
Unique
Requirement
Sample 2',5',7'
and 10'
intervals using
Mod. Calif.
2
Sampler *
Full suite of
hazardous
material tests
at 2', 5' & 7'.
Requires use of
Design
Summary
Report
Sample every
2.5' in upper
10', then every
5'.
None
None
None
On boring log
plot graph of
SPT blow
counts as
representation
of penetration
resistance for
soil and
weathered
rock.
Continuous
coring using
HQ3 wire line
system for
rock.
108
As shown each entity has a different way to obtain the geotechnical data needed for
designing the foundations for a transmission line. Each of these entities may use their
guidelines in the same state so there is no real geographical reason for the differences.
Instead each is defined by their desired level of risk with some influence from the
planning and permitting process. Although each acknowledges the need for
description of geologic information in their guidelines the main topic by far is
obtaining subsurface geotechnical engineering data.
These specifications often put a lot of weight on an 8 to 10 inch diameter boring that
represents one location that is extrapolated to represent a one half mile to 1 mile of
geotechnical conditions. Since lab data or boring log data by itself is risky to use, the
engineer then has to evaluate the geologic conditions to fill in the gaps of
information. This should be done in a way that associates key geotechnical changes
with changes in the engineering geologic conditions along the alignment.
109
Often what one is looking for are trends in geologic and geotechnical conditions that
are similar so that foundations can be constructed using the same general parameters
to keep design and construction consistent over long line lengths. In Figure 1 and 2
the boring logs indicate very similar conditions as reflected by the SPT blow count
graph where near surface conditions having blow counts from 10 to 20/ft and below
10 feet blow counts are above or greater than 50/foot. This example represents a
segment of a power line that covers 6 miles. The similarity of SPT blow counts is
matched to the same geologic formation present through out that distance. Such
similarities allow the engineer to group the entire distance under the same foundation
design recommendations. Outside of the grouped sections there is a need to focus on
those few locations where there are significant differences in either geology or line
loads.
SPT blows/foot
Figure 1
SPT blows/foot
Figure 2
110
SPT blows/foot
111
SPT blows/foot
112
113
soil where higher modulus values can reduce the foundation size and
associated project costs.
Benefits of How
Geotechnical Data is
Presented
When delivering geotechnical design
recommendations for a transmission
line project, consultants need to think
about how that data is to be presented
and will be used. There are a number
of report formats used that aid in
communicating the geotechnical
Figure 8 Geologic Strip Map
conditions more clearly and help make
the information more supportive to the
designers needs. Often a power line project is starting to be constructed at the same
time the
geotechnical
report is being
finished. The
data needs to be
delivered
SPT blows/foot
quickly and in
an easy to use
format that
communicates
to the design
engineers where
the concerns
and
constructability
issues are
located. Using
a geotechnical
report format
generally used
for a building
would make the
information
hard to find and
too time
consuming to
easily utilize.
Some formats
for data
presentation are
Figure 9 WAPA Boring Log Sheet
8
114
115
allows for the addition of subsurface descriptions, lab test results where obtained, and
moisture content. The SPT log helps determine where the blow counts transition
from soil into bedrock. One can also see that the bedrock has much lower blow
counts than might be anticipated for rock where the deepest shale is encountered.
The presence of blow counts in bedrock show that it is highly weathered and has the
consistency of dense sand or clay. Hard bedrock would be cored and blow counts
would indicate refusal conditions.
Another format used with Southern California Edison makes use of a Geotechnical
Design Summary Sheet shown in Figure 10. This form recognizes the importance of
active faults throughout Southern California on a foundation design. With this form a
geotechnical engineer can give design parameters on a per tower basis. This data can
then be easily considered in preparation of LPILE input parameters and other shaft
parameters necessary for the design of drilled pier foundations. The form even allows
for recognizing corrosion potential of subsurface conditions, surface depths to ignore
and caving potential. This type of reporting format may seem time consuming
initially but readily provides data necessary for foundation design that is often
discussed in several pages of a standard geotechnical report.
Design
Recommendations
Taking into account the
design methods to be used
during foundation
engineering is important
when grouping engineering
geology and geotechnical
data. Drilled piers are the
predominant foundation type
for lattice towers and
monopoles. Foundations
designers can use
geotechnical data in a
number of different design
methodologies. One
common design method used
for the drilled piers on
transmission lines is LPILE
(Ensoft Inc.).
LPILE is a method that
analyses pile movement due
to lateral loads. The software
models soil behavior with
data from p-y curves (Figure
10
116
11) that are internally generated for standard soil types. It takes into account the
lateral modulus of subgrade reaction (k) to evaluate pier dimensions.
Given the similarity of the subsurface conditions as represented in Figures 1 and 2
earlier, design data was prepared and grouped as shown in Example #1 and Table 1.
The data represents a 6 mile span of distance along a transmission line. Looking for
such opportunities to group data simplify the presentation of the geotechnical design
parameters for a project.
Example #1- Engineering Parameters (Geotechnical L Pile Design
Parameters Tower Locations 70/4 to 76/1 Older Alluvial Soil)
Older Uplifted Alluvial Soil Deposits. The soil is loose to very dense,
comprised mostly of silts and sandy silts that are locally clayey. At the
surface, these soils are dry and very loose.
These granular soil deposits may be susceptible to sloughing. The contractor
should be prepared to case these holes if needed during pier drilling.
Recommend that the strength contribution from the upper five feet of soil be
neglected. Vertical pressure due to weight of this layer on underlying layers
need not be neglected.
Design for lateral loads should be for a primarily non-cohesive soil.
TABLE 1
Layer
No.
1
2
3
Layer
Thickness
(ft)
Effective
Weight of
Soil (pcf)
Friction
Angle
(degrees)
Cohesion
(psf)
Ultimate
Bearing
Capacity
(psf)
Ultimate Skin
Friction
Compression
(psf)
Ultimate
Skin
Friction
Uplift (psf)
5
105
10
120
36
12,000
750
550
30
120
40
20,000
1,500
1,150
Note: Cells with no data indicate zero strength or zero capacity for the soil
layer.
Conclusion
Towers 70/4 to 76/1
Figure 12
11
117
project alignment. When done well an engineer can better understand the nature of
the foundation conditions along the alignment by grouping the engineering
parameters around these similarities and changes in geologic conditions.
References
Black & Veatch Corporation, (2004). Geotechnical Engineering Services
Specification, Geotechnical Services, Exploration Borings and Laboratory Tests.
Kleinfelder, (January 2004), Summary Geotechnical Investigation Report, Path 15
500 Kv Transmission Line, Gates Los Banos, California.
Southern California Edison (SCE). (2008). Guideline for Geotechnical Investigation
For Transmission Line Projects, Transmission and Distribution Business Unit
Civil/Geotechnical Engineering Group, Rosemead, California.
Kondziolka, R., E., and Rojas-Gonzales, L., F., (1991), Case History of a Cost
Effective Design Approach for Laterally Loaded Drilled Shafts, IEEE Xplore.
Western Area Power Administration (WAPA). (2008). Geotechnical Investigation
Statement of Work, Sacramento Voltage Support.
Dembowski, Scheffield and Smith, A Foundation Cost Comparison: Alternate
Methods of Specifying Loads for the Design of Rigid Base Tower Foundations.
12
Ulteig Engineers, Inc., 3350 38th Avenue South, Fargo, ND 58104; PH: (701) 2808636; FAX: (701) 280-8739; email: [email protected]
2
North Dakota State University, Department of Civil Engineering, 1410 14th Avenue
North, Fargo, ND 58105; PH: (701) 231-5648; email: [email protected].
ABSTRACT
Although close attention is paid to the design of transmission poles, many
times the design of the foundation is left up to a rule of thumb method. Reliable
and established methods for analyzing laterally loaded piles have been incorporated
into the current study to calculate the required depth of embedment for transmission
poles, and the results are compared to the current methods. The results show that the
current methods for determining embedment depths for direct-embedded transmission
poles significantly underestimate or overestimate the required embedment depth
depending on the soil conditions. Recommendations are made for alternate design
methods for transmission pole foundations. A case study is presented where directembedded wood transmission poles fell over due to inadequate foundation design.
The results of the case study are compared to the methods presented in this paper. It
is concluded that the alternate design methods will generate more accurate results
than the current methods.
INTRODUCTION
Tangent transmission structures endure lateral loads due to wind pressures on
the conductors and the structure itself. The lateral loads are resisted by the soil along
the depth of the pole below grade. The industry standard is to embed the pole a depth
of 10% of the total pole height plus 0.61 m (2 ft) below grade, which does not take
into account the pole or soil properties. In 2002, Mehran Keshavarzian studied the
current methods for foundation design and found that no equation in the form of a
rule of thumb or a single chart is likely to be found to yield consistently reliable pole
foundations in all soil types and for all possible pole classes, lengths, species, and
pole loading scenarios (Keshavarzian, 2002, p. 155). Although few studies have
been performed on transmission poles specifically, several researchers have studied
lateral loaded piles. These concepts have been adopted in this literature to determine
more accurate embedment depths for round, wood transmission poles. Using the
information presented in this study, the designer will be able to generate more reliable
foundation designs for round, wood transmission structures.
118
119
MATERIALS
Wood transmission poles are classified based on the American National
Standards Institute (ANSI) O5.1 (ANSI, 1987). The poles are classified by a
predetermined ultimate load that the pole must support for a given pole diameter
(ANSI, 1987). The pole classifications investigated in this paper include class 2, 1,
H1, H2, and H3 with heights ranging from 12.2 m (40 ft) to 36.6 m (120 ft). The
species of wood used in this study is Western Red Cedar. Since the applied load on a
transmission pole will vary greatly between projects, the loading used to determine
the required embedment depths was based on the ultimate loads found in the ANSI
O5.1 standard (ANSI, 1987).
The soils analyzed include homogeneous cohesive and cohesionless soils.
Undrained shear strengths, cu, ranging from 5 kN/m2 (100 psf) to 192 kN/m2 (4000
psf) were used to calculate embedment depths in cohesive soils. Based on research
by the authors, it was determined that the undrained shear strength of the soil plays a
large role in determining the required embedment depths for cohesive soils, especially
for poles placed in poor soils. As the undrained shear strength increases, the
embedment depth required decreases. Therefore, the authors assembled a soil
classification system based on the results of this study for cohesive soils which is
listed in Table 1.
Table 1. Soil Classification Summary for Cohesive Soils (Adapted from McNames,
2008)
Soil Classification
cu, kN/m2 (psf)
Very Poor Clays
5-23 (100- 499)
Poor Clays
24-47 (500-999)
Average Clays
48-95 (1000-1999)
Good Clays
>96 (2000)
Both saturated and unsaturated conditions were analyzed for the case of
cohesionless soils. Friction angles, , varying from 20o to 40o were selected for the
analysis of poles placed in cohesionless soils. The passive earth pressure coefficient,
Kp, was determined using Rankines method where K p = tan 2 ( 45 + / 2) . Small
differences in the effective unit weight, 3.1 kN/m3 (20 pcf), do not have a significant
impact on the embedment depth required, which is how the classification system
shown in Table 2 was derived. Average values for the unit weight of cohesionless
soils were taken as 17.3 kN/m3 (110 pcf) for the moist unit weight and 7.4 kN/m3 (47
pcf) in the case of fully saturated soils.
Table 2. Soil Classification Summary for Cohesionless Soils (Adapted from
McNames, 2008)
Soil Classification
, degrees
Poor Sands
20-29
Average Sands
30-39
Good Sands
>40
Page 2
120
0.61 m
P
1.5D
f
g
2cUD
3D
9cUD
9cUD
FLEMING ET AL.'S METHOD
BROMS' METHOD
COHESIVE SOILS
0.61 m
0.61 m
0.61 m
L
z
L
z
3DzKp
BROMS' METHOD
z
2
Kp Dz
FLEMING ET AL.'S METHOD
10
(1.3tan+0.3)
10
(1.3tan+0.3)
0.6x
1.02x
COHESIONLESS SOILS
Figure 1. Soil pressure distribution for cohesive and cohesionless soils.(Adapted from
McNames, 2008)
Page 3
121
Cohesive Soils. The pressure distribution for Broms (1964b) method for cohesive
soils is shown in Figure 1. Based on this pressure distribution, the required
embedment depth, z, is determined by solving the following equation for y(z) = 0,
Papp
Papp ( L 0.61 z + 1.5D +
)
Papp
9cu D
y ( z ) = 1.5D +
+
z , where Papp is the
9cu D
2.25cu D
applied lateral load, L is the length of the pole, and D is the groundline diameter of
the pole as determined by ANSI (McNames, 2008). Fleming et al.s (1992) assumed
pressure distribution is also shown in Figure 1. The required embedment depth is
determined from this distribution by solving the following equation for y(z) = 0,
y ( z ) = 1.2(4.5cu Dz 2 10.5cu D 2 z + 10.5cu D 3 Papp ( L 0.61)) (McNames, 2008).
Cohesionless Soils. Using the pressure distribution shown in Figure 1 for Broms
(1964a) method, the required embedment depth is calculated as
2 Papp ( L 0.61)
z=3
, where is the unit weight of the soil ( when water table is
K p D
below the pole and where the water table is present along the length of the pole)
(McNames, 2008). Fleming et al. (1992) suggest that the pressure distribution is
linear as Broms proposed in 1964, however, the pressure is directly proportional to
Kp2 rather than Kp based on tests performed by Barton in 1982 (as cited in Fleming et
6 Papp ( L 0.61)
al., 1992). The required embedment depth is calculated as z = 3
K p2 D
(McNames, 2008). Based on the distribution proposed by Prasad and Chari (1999),
the
embedment
depth
is
determined
to
be
equal
to
Papp
z = 1.588x
, where x is the distance from the
(1.3 tan + 0.3)
0.408 [10
] D x
groundline to the point of rotation as shown in Figure 1 (McNames, 2008).
CURRENT DESIGN METHODS
The 2004 RUS Bulletin 1724E-200 recommends a rule of thumb for
determining the required embedment depth for most wood pole structures in good
soils and not subjected to heavy loadings (RUS, 2004, p. 12-3) as
z = 0.1L + 0.61 (rule of thumb), where the constant 0.61 is in meters. The RUS
(2004) also includes an equation found in the Wood Preserving News (as cited in
RUS, 2004) for determining appropriate embedment depths for round, wood poles as
S e z 3.75
P=
(RUS equation), where Se is the soil constant: 140 for good
L 0.61 0.662 z
soils, 70 for average soils, and 35 for poor soils (RUS, 2004).
Page 4
122
Cu = 5 kN/m2 (100psf)
Cu = 48 kN/m2 (1000psf)
Rule of Thumb (10%+0.61m)
RUS (Average Soils)
Cu = 24 kN/m2 (500psf)
Cu = 96 kN/m2 (2000psf)
RUS (Poor Soils)
RUS (Good Soils)
14
12
10
8
6
4
2
0
Pole Height (m)
Figure 2. Required embedment depths for Broms method for cohesive soils (class
H3 poles). (Adapted from McNames, 2008)
Page 5
Cu = 5 kN/m2 (100psf)
Cu = 48 kN/m2 (1000psf)
Rule of Thumb (10%+0.61m)
RUS (Average Soils)
12
123
Cu = 24 kN/m2 (500psf)
Cu = 96 kN/m2 (2000psf)
RUS (Poor Soils)
RUS (Good Soils)
10
8
6
4
2
0
Pole Height (m)
Figure 3. Required embedment depths for Fleming et al.s method for cohesive soils
(class H3 poles). (Adapted from McNames, 2008)
20 degrees
40 degrees
RUS (Poor Soils)
RUS (Good Soils)
30 degrees
Rule of Thumb (10%+0.61m)
RUS (Average Soils)
0
Pole Height (m)
Figure 4. Required embedment depths for Broms method for moist cohesionless
soils (class H3 poles). (Adapted from McNames, 2008)
Page 6
20 degrees
40 degrees
RUS (Poor Soils)
RUS (Good Soils)
124
30 degrees
Rule of Thumb (10%+0.61m)
RUS (Average Soils)
0
Pole Height (m)
Figure 5. Required embedment depths for Fleming et al.s method for moist
cohesionless soils (class H3 poles). (Adapted from McNames, 2008)
20 degrees
30 degrees
40 degrees
Rule of Thumb (10%+0.61m)
RUS (Poor Soils)
RUS (Average Soils)
RUS (Good Soils)
8
0
Pole Height (m)
Figure 6. Required embedment depths for Prasad et al.s method for moist
cohesionless soils (class H3 poles). (Adapted from McNames, 2008)
Page 7
125
Linear Curve Fitting for Cohesive Soils. It was determined that the results for
Broms (1964a) and Fleming et al.s (1999) methods tend to follow a linear curve.
Therefore, linear equations for each method were generated using the soil
classifications listed in Table 1. The purpose of providing the linear equations is to
give the designer a method similar to the traditional rule of thumb. To simply the
amount of equations, the results for class H3 poles and the lower limit of the
undrained shear strength (Table 1) were used to generate the linear equations. As was
noted earlier, embedment depths for class H3 poles are typically 21%, 16%, 10% and
5% greater than class 2, 1, H1, and H2 respectively. Therefore, the linear equations
presented will provide conservative results for pole classes smaller than H3. The
linear equations and the coefficient of reliability, R2, associated with each equation
are listed in Table 3 for cohesive soils.
Table 3. Summary of Linear Equations for Cohesive Soils (Adapted from
McNames, 2008)
Method
Soil Classification
Linear Equation
R2 Value
Very Poor Clay
z (m) = 0.18*L + 6.7 m
0.990
Poor Clay
z (m) = 0.09*L + 3.4 m
0.972
Broms
Average Clay
z (m) = 0.06*L + 2.7 m
0.966
Good Clay
z (m) = 0.05*L + 2.1 m
0.964
Very Poor Clay
z (m) = 0.15*L + 6.7 m
0.999
Poor Clay
z (m) = 0.07*L + 3.4 m
0.979
Fleming et al.
Average Clay
z (m) = 0.05*L + 2.4 m
0.953
Good Clay
z (m) = 0.04*L + 1.8 m
0.989
It was revealed that embedment depths for transmission poles placed in
cohesive soils can easily determined from normalized plots using the quantity
cuDL2/Mapp as shown in Figure 7. This quantity was chosen because all of these
properties play a large role in determining the required embedment depth, and are
parameters that can be easily obtained. This is also the normalization that gives the
best fit for the data collected for each method. Some advantages of using the
normalized plots rather than the linear equations include: more accurate embedment
depths because all of the data gathered in the research is incorporated to the
normalized plots, the applied loads can be tailored to fit a specific pole scenario, and
the factor of safety is left up to the discretion of the designer.
It should also be noted that both methods provide unstable results for values
of cuDL2/Mapp near 3.2. This occurs when heavily loaded poles are placed in soils
with very low undrained shear strengths. To account for this, a lower limit of 3.2 has
been set for both of the methods, and the data points below this limit have been
removed to generate the curves shown in Figure 7. The upper and lower limits of the
rule of thumb are also plotted along with the power curves to illustrate the
limitations of the current method. Based on the data shown in Figure 7, it can be
concluded that the rule of thumb will most often generate unconservative
embedment depths.
Page 8
126
0.6
0.5
Broms' Method
0.4
0.3
0.2
0.1
Fleming et al.'s
Method
0
cu DL2 /Mapp
Figure 7. Normalized summary plots for laterally loaded transmission poles
embedded in cohesive soils. (Adapted from McNames, 2008)
Linear Curve Fitting for Cohesionless Soils. Broms (1964a), Fleming et al.s
(1992), and Prasad and Charis (1999) methods for short-rigid laterally loaded piles
were adopted for use in determining the required embedment depths for direct
embedded transmission poles. It was determined that for nearly all pole classes, pole
heights, and soil types; the embedment depths calculated by the existing methods
provide unconservative results.
Based on the research, it was discovered that when plotted the data trends are
nearly linear for all three methods studied. Therefore, linear equations were fit to
each method based on the soil classifications listed in Table 2 using the lower limits
of the friction angle for each classification and the data gathered for class H3 poles.
The linear equations are shown in Tables 4 and 5 for moist and fully saturated
cohesionless soils respectively. The embedment depths obtained using these
equations will be conservative for class 2, 1, H1, and H2 poles. As was noted earlier,
the embedment depths for class H3 poles are 15%, 11%, 7%, and 3% greater than
class 2, 1, H1, and H2 poles respectively.
Page 9
127
Table 4. Summary of Linear Equations for Moist Cohesionless Soils (Adapted from
McNames, 2008)
Method
Soil Classification
Linear Equation
R2 Value
Poor Sand
z (m) = 0.04*L + 3.4 m
0.974
Broms
Average Sand
z (m) = 0.03*L + 3.0 m
0.981
Good Sand
z (m) = 0.03*L + 2.4 m
0.953
Poor Sand
z (m) = 0.05*L + 3.7 m
0.988
Fleming et al. Average Sand
z (m) = 0.03*L + 3.0 m
0.984
Good Sand
z (m) = 0.03*L + 2.1 m
0.968
Poor Sand
z (m) = 0.06*L + 4.6 m
0.986
Prasad and
Average Sand
z (m) = 0.05*L + 3.7 m
0.974
Chari
Good Sand
z (m) = 0.03*L + 3.0 m
0.944
Table 5. Summary of Linear Equations for Fully Saturated Cohesionless Soils
(Adapted from McNames, 2008)
Method
Soil Classification
Linear Equation
R2 Value
Saturated Poor Sand
z (m) = 0.06*L + 4.3 m
0.962
Broms
Saturated Average Sand
z (m) = 0.05*L + 3.7 m
0.976
Saturated Good Sand
z (m) = 0.04*L + 3.4 m
0.990
Saturated Poor Sand
z (m) = 0.06*L + 4.9 m
0.990
Fleming et al. Saturated Average Sand
z (m) = 0.05*L + 3.7 m
0.979
Saturated Good Sand
z (m) = 0.04*L + 2.7 m
0.988
Saturated Poor Sand
z (m) = 0.09*L + 5.8 m
0.992
Prasad and
Saturated Average Sand
z (m) = 0.07*L + 4.9 m
0.978
Chari
Saturated Good Sand
z (m) = 0.06*L + 3.7 m
0.970
The methods presented in this paper for cohesionless soils can also be
normalized based on the quantity (tan)1.5DL3/Mapp. Each of these parameters plays
a role in determining the embedment depth and is a parameter that can easily be
obtained by the designer. The data for each method has been fit to a power trendline
with the results shown in Figure 8 for cohesionless soils. These results are compared
to the rule of thumb, and it can be seen that the rule of thumb is unconservative
for most conditions when compared to the alternative methods.
Page 10
0.6
0.5
0.4
Broms' Method
128
0.3
0.2
0.1
(tan)1.5DL3/Mapp
Figure 8. Normalized summary plots for laterally loaded transmission poles
embedded in cohesionless soils.(Adapted from McNames, 2008)
CASE STUDY
Hancock 34.5 kV Windfarm Line (Wen, 2004). In early 2004, several round, wood
structures used to support a double circuit 34.5 kV transmission line fell over in
Hancock County, Iowa after a storm left the ground saturated (Wen, 2004). Henry
Wen was asked to investigate the cause of the failure, and found that the foundations
were undersized when analyzed for the design loads using the soils found at the site.
The poles used on the line were Western Red Cedar, 16.8 m (55 ft), class H1 poles.
The embedment depth used in the original design was based on the rule of thumb,
such that the poles were embedded in the ground 2.3 m (7.5 ft) (Wen, 2004).
Based on soil borings performed in the area, Wen (2004) determined that the
poles should have been embedded in the ground up to 6.1 m in some areas (Wen,
2004). The PLS-Caisson (2002) program distributed by Power Line Systems (PLS,
2002) was used by Wen (2004) to determine the proper embedment depths for the
poles. The PLS-Caisson (2002) program uses a modified version of Broms method
(1964a&b) to determine proper embedment depths of transmission poles embedded in
layered soils (PLS, 2002).
Page 11
129
Comparison with Alternate Design Methods. Using the design loads and soil
borings found in the report assembled by Henry Wen (2004), the alternative design
methods included in this paper were compared to the PLS Cassion (2002) results
determined by Wen. Since the soils found at the site consisted of layered cohesive
soils, a weighted average for the undrained shear strength (cu*) was calculated along
the depth of embedment. (Additional information regarding the soils found at the site
can be found in McNames, 2008.)
Wen (2004) determined that the poles placed near Boring #4 should have been
embedded in the ground a total of 5.9 m (19.37 ft) with the PLS-Caisson (2002)
program using an average diameter of 0.44 m (1.45 ft), pole height of 16.8 m (55 ft),
applied moment of 257.7 kN-m (189.5 k-ft), and weighted average undrained shear
strength of 12.2 kN/m2 (255 psf) (Wen, 2004). The linear equations presented in this
paper for Broms (1964b) and Fleming et al.s (1992) methods predict required
embedment depths of 9.7 m (31.9 ft) and 9.2 m (30.25 ft) respectively. Wens
(2004) results are also compared to the normalized plots presented in this paper. The
predicted embedment depths are 5.9 m (19.25 ft) and 5.2 m (17.1 ft) for Broms
(1964b) and Fleming et al.s (1992) methods respectively.
The design parameters Wen (2004) used for Boring #5 are as follows: 0.44 m
(1.45 ft) pole diameter, 16.8 m (55 ft) pole height, 257.7 kN-m (189.5 k-ft) applied
moment, and an average undrained shear strength of 29.1 kN/m2 (607 psf) (Wen,
2004). Based on these parameters, Wen (2004) determined that the embedment depth
required is 3.8 m (12.37 ft ) from the PLS-Caisson (2002) program (Wen, 2004).
When these parameters are used with the linear equations presented in this paper, the
predicted embedment depth is 4.9 m (15.95 ft) for Broms (1964b) method and 4.5 m
(14.85 ft) for Fleming et al.s (1992) method. The normalized methods produce
embedment depths of 4.0 m (13.2 ft) and 3.5 m (11.55 ft) for Broms (1964b) and
Fleming et al.s (1992) methods.
Conclusion. When compared to the existing rule of thumb, the simplified linear
equations generated conservative values for the embedment depth. This is due to the
fact that the linear equations presented are generated using the ultimate moment
capacity and geometry of class H3 poles, and are based on the lower limits of the soil
parameters presented in Table 1.
As can be seen in the previous examples, the normalized plots provide
embedment depths within 12% of those calculated by the PLS-Caisson (2002)
program. The rule of thumb that was used in the original design of the foundations in
question underestimated the embedment depth by as much as 61% in some locations.
The normalized plots predict a much improved and more reliable design compared to
the traditional rule of thumb method. The method of using a weighted average for
layered cohesive soils proved to be fairly accurate, although it should be tested with
additional soil types.
Page 12
130
Page 13
131
Page 14
132
Abstract
During the course of their service-life steel pole and lattice tower foundations are
subjected to significant environmental forces that can have a detrimental impact on
their service life. Foundations are critical components of all transmission systems and
therefore are crucial to the national power grid.
Deterioration identified in these foundations during regular inspections and
assessment should be categorized into specific levels of structural severity. Each
should be addressed accordingly by a varied regimen of mitigation and repair options.
Coatings and anode installation are most commonly used to help prevent
deterioration. Repair options include but are not limited to: bolted steel repairs,
welded steel repairs, reinforced concrete encasement of concrete foundations and
direct member replacement.
This paper will address assessment and repairs specific to steel tower and steel pole
foundations inclusive of structural durability, corrosion issues and standard repair
methods.
133
Guyed Steel Lattice Towers similar to the self supporting steel lattice
structures described above, guyed towers are distinctly different because of
their reliance on multiple guys to provide structural stability.
Guyed Steel Poles similar to self-supporting steel poles, guyed steel poles
rely on multiple guy wires to provide stability.
Foundation Types
All steel structure types require an adequate foundation to support the loads imposed
on the structure. Because of the varying types of environments and soil conditions
where these structures may be located, it is necessary to utilize different types of
foundations in order to provide a stable platform for the structure. These include:
Direct-embedment, which can be utilized for both steel poles and steel
grillage, involves the excavation of a foundation hole in the earth by means of
a drill rig or similar equipment. Excavations are made to accommodate the
tower leg and grillage, which are set into the excavation and backfilled with
either native soil, compacted gravel, crushed stone or concrete.
Direct embedment of steel poles and steel grillages may sometimes be
preferred over other foundation options due to their typically lower cost and
ease of installation. However, they are usually not used where shallow rock
and ledge are present or in areas where soil stability is a concern due to sandy
soil conditions or in soil weakened by the influence of water.
134
One of the most significant advantages of reinforced concrete foundations is that the
structural steel members are protected from the environment by the concrete, which
limits their exposure to corrosion and significantly increases their service life.
Base Plates, Anchor Bolts and Stub Angles
When utilizing drilled shaft or pier slab foundations it is necessary to have a reliable
interface connection between the steel and concrete.
Three of the most common connections are
Base Plate and Anchor Bolts Lattice towers and steel poles have used both
base plates and anchors bolts of various designs to anchor the upper support
structure to the foundation.
Stub Angles usually a short piece of angle steel with shear connectors,
which is set directly into the concrete.
Pin Bases several designs have been engineered to create a pin connection
between a tower and its foundations. These are normally used with guyed
structures such as guyed vees.
135
Reasons to Inspect
It is important to understand that with all structures there is a need for periodic
inspection to ensure structural integrity and public safety. Steel structures and their
foundations are no exception, as these structures are sometimes placed in
environments that are less than ideal relative to the longevity of their structural
components.
Environmental conditions can have a dramatic impact on the structural integrity of
steel and concrete over relatively short periods of time. Detrimental influences such
as agricultural activity, changes in soil elevation from construction and emissions
from manufacturing facilities can all contribute to the degradation of structural
components. These effects can be compounded by unseen defects in structural
members, coatings and concrete at the time of construction.
Often these environmental influences combine to create active corrosion cells which
are detrimental to steel structural components.
Definition of a Corrosion Cell
The basic electrochemical corrosion cell consists of four components; a cathode, an
anode, an electrolyte and a metallic path. Corrosion within the cell exists because of
the flow of electrons and ions between these components.
Anode the point where corrosion takes place within the cell due to the flow
of positively charged ions away from it.
Cathode negatively charged ions migrate from the cathode, which creates a
polarization helping to protect the cathode from corrosion.
Metallic path conducts negatively charged electrons from the anode to the
cathode.
With a corrosion cell on steel towers and steel poles, the electrolyte is represented by
the soil the structure is buried in. The metallic path is the metal of the structure and
the cathode and anode can either be between the different legs of a steel tower or
along the surface of a steel pole where there is a difference of corrosion potential
between two points.
136
Excavation because steel towers and steel poles may be buried in soil it is
often necessary to make shallow inspection holes. The excavation extends the
visual assessment and physical measurements through the initial transition
area of the steel to soil interface. Usually inspection excavations are limited to
a depth of 18 24 but can occasionally be deeper if necessary to determine
the extent of visible corrosion.
137
138
Half Cell Measurements measure the structure to soil potential within the
environment. Typically, this is done utilizing a digital or analog potential
meter. The more negative the measurement is (< -.850 mV), the more likely it
is that the structure will have a reduced potential for corrosion. The less
negative the measurement (> -.400 mV), the more likely it is to corrode.
Corrosion Rating
The factors previously discussed are only indicative when considered alone. They
must be evaluated as to how they relate to each other in a specific environment. To do
this an algorithm was developed to measure these factors together in order to
determine a structures potential for corrosion or its corrosion rating.
These ratings are divided into specific categories quantifying each structures
potential for future corrosion, as listed below:
Low the potential for corrosion is low and no further action is necessary at
this time. Typically the recommendation for these structures is to re-inspect in
another ten years.
Mild the potential for corrosion is slightly elevated and, while it may not be
visually evident, it is possible that minor corrosion activity is present. Similar
to the structures rated as low, it is recommended to re-inspect these structures
in another ten years
Severe these towers are significantly corroded and show signs of metal loss
such as large pitting, edge loss and thinning. In severe cases large perforation
of structural members are present.
Mitigation Alternatives
Similar to inspection methods, corrosion mitigation methods have been used in gas,
petroleum and other industries for many years. As such, many of these products can
be readily adapted for use in similar applications.
7
139
The most common and practical mitigation efforts available are the application of
protective coatings. Coatings protect the steel from exposure to its surrounding
environment, providing a barrier against moisture and other corrosive mechanisms.
A good coating application program involves an extensive excavation and a thorough
surface cleaning of the structure prior to application of the coatings. These surface
preparations are critical to the overall performance of the coating system and are
covered by a number of ASTM and NACE standards. Coatings are available in
several different types for a variety of applications. It is very important that the
correct material is used with the recommended application procedure to achieve the
desired level of protection.
When coatings alone are not sufficient to protect structures from the effects of
corrosion, an additional measure of protection that can be applied is sacrificial
anodes.
Sacrificial galvanic anodes give up of themselves to protect steel structures from
corrosion by means of the galvanic reaction that takes place between the steel and the
anode. Because the anode is higher in the galvanic series than steel, when it is put in
direct contact with the steel structure the loss of negatively charged ions is transferred
from the steel to the anode so that the anode corrodes in place of the steel.
Anodes are usually applied by placing them into trenches or augured holes in close
proximity to the structure. The anodes are then connected directly to the steel
structure by means of a conductive wire.
Because of the varying amount of steel exposed in the ground on various structures,
protective anode systems, or beds, are designed specifically for the structure they are
to protect. A misapplication of anodes will usually result in inadequate protection for
the structure and will likely lead to an early failure of the anode system.
General Repairs
Structures that have deteriorated beyond the protective capabilities of coatings and
anodes usually are significantly weakened by section loss of their supporting
structural members or foundations. In most cases, in-place repairs can be individually
engineered for these structures in order to avoid the high cost of replacing the entire
structure.
Concrete Foundation Repairs
Concrete foundation repairs can usually be achieved by simply encasing the degraded
foundation in new concrete. This usually involves excavating around the foundations
to an adequate depth to remove any loose or broken concrete and then placing forms
140
in place to accommodate the new concrete. Once the concrete is poured and hardened
sufficiently the forms are removed and the new foundation is backfilled.
Steel Repairs
Weakened structural steel members must either be replaced or reinforced to continue
to provide adequate support to the structure. There are a number of different ways for
this to be done:
Bolted Splice Repair this repair leaves the weakened member in place
similar to the encasement repair. However, in this repair the weakened
member is sandwiched between two similarly sized pieces of steel, which are
drilled and bolted in place effectively transferring the load of the weakened
member to the two new pieces of steel.
141
Summary
Because of the aging infrastructure in the United States relative to the national power
grid, it is important to bring to light practices and methodologies to the electric
industry that help to improve practical economic decisions concerning its long term
durability.
In this paper, some of the most common assessment and repair techniques and
procedures available specific to steel tower and steel pole foundations were
addressed.
It is hoped that this information generates further investigation into practical
economic solutions that will maintain the national power grid through the responsible
use of appropriate maintenance programs.
10
142
References
P.R. Roberge, Scope of language and Corrosion Corrosion Basics: An Introduction,
Second Edition NACE (2006): pp. 1519.
H.H. Ulig, The Cost of Corrosion in the United States, Chemical and Engineering
News 27 (1949): pg 2764.
G.H. Koch, M.P.H. Brongers, N.G. Thompson, Y.P. Virmani, J.H Payer, Costs of
Corrosion and Preventative Strategies in the United States, National Technical
information Services, FHWA-RD-01-156, Springfield, VA, Sept. 30, 2001.
NACE Publication 3C149, Economics of Corrosion (Houston, TX: NACE, 1994).
R.D. Gundry, Corrosion/93 Plenary & Keynote Lectures (Houston, TX: NACE
International, 1993).
W. Lynes, Some Historic Developments Relating to Corrosion, Journal of the
Electrochemical Society 98 (1951): pp. 3C 10C.
S.D. Cramer, B.S. Covino, G.R. Holcomb, M. Ziomek-Moroz, Information Sources
and Databases for the Corrosionist, in vol. 13A, Corrosion: Fundamentals, Testing,
and Protection, eds. D.S. Cramer, B, S, Covino (Metals Park, OH: ASM
International, 2003), pp.999-1001
11
143
144
145
0
1.5
2
3.5
8
1,000
1.5
2.5
4
8
2,000
1.5
2.5
5
8
3,000
1.5
3
5
8
4,000
1.5
3.5
5
>8
5,000
1.5
3.5
6
>8
6,000
2
3.5
6
>8
8,000
2
4
8
>8
10,000
2
4
8
>8
15,000
2.5
5
8
>8
20,000
2.5
5
8
>8
Where:
Wi = Ice weight (lb/ft)
r = radial ice thickness (in)
d = bus diameter (in)
[1]
146
Moment
of Inertia
(in4)
0.1056
0.3912
0.8679
1.924
3.894
6.281
9.611
20.67
40.501
105.743
i2
D
147
[2]
Where:
Fsc = Short Circuit Forces (lb/ft)
i = Short Circuit Current (Amps rms)
D = Centerline-Centerline spacing of buses (in.)
Wind loading may be calculated using pressures derived from ASCE 7-05 or
from IEEE Std. C2. When ASCE 7-05 is used there are three loading cases that must
be considered and two wind values are needed. The three loading cases are:
1.
2.
3.
The first load case requires determining the fastest possible wind for the
location in question. The second load case requires the value of wind shown
coincidental with the worst icing conditions. The wind force is then calculated using
the first value on a bus diameter without ice and the second value on a bus diameter
including ice. The third case is a vertical load only. The largest resultant force is
then used in the design. If IEEE Std. C2 is used, only one calculation is needed
(unless the bus is over 60 feet above ground level) based upon the value for wind
pressure given for the loading district (light, meduim or heavy) at the location. In
either case Equation 3 is then used to calculate the force per unit length on the bus.
Fw = 0.083PW d
[3]
Where:
Fw = Wind loading (lb/ft)
Pw = Wind pressure (lb/ft2) from ASCE 07 or IEEE Std. C2
d = Outside diameter of the conductor including ice if
needed (in.)
The total bus loading FT, in lb/ft may then be determined using Equation 4
where WC is the sum of the bus and the dampening conductor weight in lb/ft.
FT =
(FSC + FW )2 + (WC + Wi )2
[4]
If wind loads were computed using ASCE 7, then Equation 4 must be solved three
times. The first case will include wind without ice in which case Wi is not included.
The second case will be coincident wind with ice. The third case is ice only and Fw is
not included. In all cases short circuit forces Fsc should be included. The worst total
FT will then be used to determine the maximum bus support spacing. If wind loads
148
are computed using IEEE Std. C2, then Equation 4 applies both wind and ice loads
simultaneously for the approriate loading district. However, IEEE Std. 605 suggests
applying laterial and gravity loads individually and calculating FT first with only
maximum wind and then with only maximum ice and then using the largest value of
FT. If the bus height were to exceed 60 ft. then IEEE Std. C2 would also require an
additional computation for the extreme wind condition in which case Equation 4 must
be solved again using the extreme wind pressure but no ice load. RUS 1724E-300
takes a more conservative approach and suggests extreme wind be considered for all
bus designs no matter their height. So, while the design standards are not completely
consistent, it would seem prudent to consider the following load cases for Fw and Wi
when solving Equation 4.
1.
2.
3.
4.
[5]
Where:
LM = Maximum bus support spacing (ft)
KSE = Multiplying factor from Table 5
FB = Maximum allowable fiber stress (lb/in2)
28,000 lb/in2 for 6061-T6
20,000 lb/in2 for 6063-T6
S = Section modulus of the bus (in3) from Table 3.
FT = Total conductor loading (lb/ft)
Table 5: Values of KSE and KDE to be used in Equations 5, 6, 7 and 8.
Bus System
KSE
Fixed Both Ends
1.0
Fixed-Simply Supported
0.82
Simply Supported Both Ends (single span)
0.82
Simply Support
0.82
(two equal spans)
Simply Supported (three or more equal
0.88
spans)
KDE
4.5
9.34
22.5
9.34
11.9
The RUS gives Equation 6 for calculating vertical bus deflection and
recommends limiting deflection to L/200 including the effects of ice. IEEE Std. 605
however, suggests limiting deflection to a maximum of L/150 while not including ice
in the calculation.
y = K DE
Where:
y=
KDE =
E=
I=
LD =
(WC + Wi )L4D
EI
[6]
The maximum allowable length (ft) between supports using IEEE Std. 605 deflection
limits would then be calculated from Equation 7.
L D = 92.8
I
K DE WC
[7]
The maximum allowable length (ft) between supports using RUS deflection limits
would be calculated using Equation 8.
L D = 84.3
I
K DE (WC + Wi )
[8]
The calculation of deflection requires that the fixity of both ends of the span in
question be determined. No end conditions will produce either completely fixed or
simply supported, and the condition becomes more difficult to determine when some
bus work is welded at the supports and some is not. One common condition is that a
bus span is continuous on one side and supported by several supports over a long
length, and simply supported on the other side. IEEE Std. 605 suggests a bus span of
this type should be considered fixed on the continuous end and pinned on the other
end. The RUS suggests, as may be seen in constant KDE from Table 5, that the
continuous end should not be considered entirely fixed.
The final factor limiting the allowable distance between bus supports is the
strength of the insulators supporting the bus. The RUS recommends Equation 9 for
determining the maximum allowable bus support spacing governed by the support
insulator cantilever strength. This equation includes a suggested safety factor of 2.5
so the insulators working load will be approximately 40% of the rated cantilever
strength.
149
LS =
WS
2.5(FSC + FW )
[9]
Where:
LS = The average of the two adjacent span lengths, or the span length
if equal spans (ft)
WS = Rated insulator cantilever strength (lbs)
The maximum bus span length can now be determined by finding the smaller of LS,
LD, and LM.
DESIGN EXAMPLE
In a recent switchyard design a critical long span (60 ft) was included which
was needed to cross a service road. This span was critical because it was highly
visible and for aesthetic reasons the owner wanted a strict L/150 limitation on the bus
deflection. The span in question was supported as shown in Figure 1. It was
continuous on one side and supported by more than three supports to which the bus
was welded, and simply supported on the other side where it made a connection to an
expansion fitting and a switch. Six-inch schedule 80 bus was used. A problem was
encountered when the calculated deflection was compared to the actual results after
construction was completed.
150
151
choose which assumptions best fit their design. Special consideration should be
given to deflection calculations. Deflection in a substation bus design is typically not
critical from a performance or structural standpoint but may become important for
aesthetic considerations. However, when deflection does become critical, the fixity
assumptions suggested in some standards were found to be somewhat in error and in
the case of those in IEEE Std. 605 were non-conservative. In the case considered
here, with a continuous bus on one side of the span and a pinned support on the other
side of the span, the results using the assumptions for multiple simply-supported
spans given in the RUS documents produced the most accurate results when
compared with actual measured deflections, and the suggestions contained in IEEE
Std. 605 resulted in considerable error when compared to field measured results.
REFERENCES
IEEE Std. 605-1998. IEEE Guide for Design of Substation Rigid-Bus Structures.
IEEE , New York, NY.
IEEE Std. 1427-2006. IEEE Guide for Recommended Electrical Clearances and
Insulation Levels in Air Insulated Electrical Power Stations, IEEE New
York, NY.
IEEE Std. C2-2007. National Electrical Safety Code. IEEE New York, NY.
RUS Bulletin 1724E-300. (2001). Design Guide for Rural Substations. Rural Utilities
Service, U.S. Department of Agriculture. Washington D.C.
ASCE Standard No. 7-05. (2006). Minimum Design Loads for Buildings and Other
Structures, American Society of Civil Engineers, Reston, VA.
152
ABSTRACT
This paper reveals a new approach to the design of civil engineering components
of a substation. A design pattern is presented that enables the engineer to design and
generate 3D structures, foundations, oil containment devices and soil cut and fill of a
substation. The process is automated for minimum user operations and the final
product is a 3D model of the entire substation. For the client, the outcome is cost
reduction, better quality control, overall lifecycle project time reduction and a virtual
realistic representation of the substation that can be looked at from different angles
and directions. Quantities for cost estimation by the project engineer are readily
available from the 3D model. This paper also portrays the 3D model of a substation
as a virtual tool that embraces all aspects of the lifecycle of a substation project.
INTRODUCTION
Numerous 3D computational tools are available to calculate the many aspects of
civil engineering of substations. However, the engineering process of a substation
project still relies on 2D information. The electrical department prepares 2D
drawings showing plan and elevation views by voltage levels. These drawings are
then submitted to the civil engineer who produces its own 2D drawings showing
infrastructures with drainage, foundations layout and details, steel structures layout
and details as well as oil spill containment devices and fire/noise wall protection.
This process is cumbersome and needs constant interaction between the disciplines to
ensure that changes are taken into account from both sides. A new approach based on
a 3D design pattern is presented below where advantages are foreseen in the lifecycle
of a project from design to delivery to the client as shown in Figure 1. Catia V5 and
Smarteam from Dassault Systemes company are respectively the 3D CAD
(Computer-Aided Design) and PLM (Product Lifecycle Management) database
softwares used to develop the automated solutions described in this paper.
153
154
PROJECT
DESIGN
ESTIMATION
FEASIBILITY
STUDY
CONSTRUCTION
CLIENT
MAINTENANCE
AND
EXPLOITATION
155
ELECTRICAL EQUIPMENT
STEEL SUPPORT
LATTICED
STEEL
STRUCTURES
156
157
1-
2-
3-
158
Step 1:
characteristics of
required
foundation are
extracted from our
calculation tool
Step 2:
characteristics
are imported in
our Excel tool
developed to
convert the
collected data to
corresponding
realistic 3D
model of the
foundation
Step 3: CatiaV5
is the design tool
used to generate
the foundation in
3D. A 2D
drawing and
steel report are
also available.
They are
automatically
generated from
the 3D model
Figure 5. Automated foundations 3D design process for substations.
Finished ground
surface
Existing ground
surface
159
CONCLUSION
The approach of design pattern using Catia V5 to generate a 3D model of a
substation is very well suited with engineering activities that evolve starting from the
feasibility study and ending with the final design project. Bill of quantities are
generated automatically and adjusted when changes appear. The benefits of the
design pattern approach with Catia V5 also include :
-
REFERENCES
Alexander C. (1977). A Pattern Language: Towns, Buildings, Construction.
Oxford University Press, USA.
Judson S.R. et al. (2003).A Metamodeling Approach to Model Refactoring. IEEE
Computer Society.
160
DiGioia, Gray and Associates, LLC, 570 Beatty Road, Monroeville, PA 15146; PH: 412372-4500; FAX: 412-372-1972; [email protected]
2
Department of Civil and Environmental Engineering, Carnegie Mellon University,
Pittsburgh, PA 15213; PH: 412-268-2958; FAX: 412-268-7813; [email protected]
ABSTRACT
The 2008 Edition of ASCE Manual 113, Substation Structure Design Guide, devotes
several sections to criteria and procedures for the seismic design of structures and
components of electric substations. This paper describes concepts that interpret and
supplement the seismic provisions of Manual 113. Among the issues discussed are:
definition of seismic design spectra, simultaneous occurrence of ground motion in three
mutually perpendicular directions, incorporation of soil-structure interaction effects and
estimation of deflections under seismic loading. Numerical examples are presented to
illustrate some concepts, and recommendations are given to implement them in practice.
INTRODUCTION
Electrical substation structures support sensitive and costly equipment that could be
damaged by earthquakes. In the United States, the reference document for the design of
substation structures and components is ASCE Manual 113, Substation Structure Design
Guide. The 2008 edition of this document devotes Sections 3.1.7, 5.6.2.1 and 6.7 to
criteria and procedures for seismic design. In this paper we review, interpret and
supplement the seismic provisions of Manual 113. To this end, we have conducted
parametric studies and have reviewed related documents including IEEE 693-2005,
ASCE 7-05, and ASCE 4-98 to identify appropriate procedures and guidelines.
SEISMIC INPUT
The seismic demand can be defined as one of the following options:
A seismic coefficient that applied to the weights of the structural members and
the supported equipment provides static lateral forces.
Design spectrum, i.e., a table of maximum pseudo-acceleration versus period
or frequency, as input in a modal spectral analysis.
Acceleration time histories for linear or nonlinear step-by-step analyses.
Seismic coefficients and design spectra are provided by local or national codes, or can be
developed via site specific seismic hazard analyses. In the United States, elastic design
spectra are prescribed in Standard ASCE 7-05 and are based in seismic hazard maps
developed and published by the USGS, most recently in 2008. This information is used
to calculate design spectra with a shape similar to that given in Fig. 1.
Section 3.1.7 of Manual 113 defines the seismic design spectrum in its Eqs. 3-8 and 3-9,
based on the USGS spectra which correspond to 5% of critical damping. To convert a
161
162
5%-spectrum to other damping ratios, Manual 113 provides Fig. 5.2 with normalized
spectra for damping ratios of 2%, 5% and 10%, reproduced in Fig. 1. The USGS spectra
correspond to rock sites. Thus, to account for flexible subsurface layers, Tables 3-12 and
3-13 of Manual 113 adopt the site modification factors from the ASCE 7-05 Standard.
IEEE 693, 2%
3.2
Normalized Acceleration
IEEE 693, 5%
2.8
ASCE 7-05, 5%
2.4
2.0
1.6
1.2
0.8
0.4
0.1
10
100
Frequency (Hz)
163
function of frequency, R(f), that varies smoothly between 1 at say 33 Hz and remains
constant for frequencies smaller than or equal to a characteristic value fR, for instance:
R(f) = R
R(f) = 1 + (R 1) [(33 f)/25]2.6
R(f) = 1,
f = natural frequency in Hz
for f 8 Hz
for 8 f 33 Hz
for f > 33 Hz
(1)
(2)
(3)
These equations correspond to fR = 8Hz. Resulting reduced spectra are shown in Fig. 2 for a
damping ratio of 5%.
Table 1. Structure Response Modification Factor, R.
Structure or Component Type
R for USD
Moment-resisting steel frame
3.0
Trussed tower
3.0
Cantilever support structures
2.0
Tubular pole
1.5
Steel and aluminum bus supports
2.0
Station post insulators
1.0
Rigid bus (aluminum and copper)
2.0
Structures with natural frequency 25 Hz
1.3
R can also vary with the direction of motion because different structural systems can be
used to resist lateral forces in different directions. In addition, the systems that resist
vertical seismic accelerations are frequently different than those resisting horizontal
forces. Gravity can lead to asymmetric hysteresis loops for vertical vibratory loads,
diminishing the ability of the structure to dissipate inelastic energy and thus decreasing R.
SEISMIC ANALYSES OF SUBSTATION STRUCTURES
Seismic analyses provide estimates of stresses and deflections, for use in load
combinations that include seismic effects in the verification of the adequacy of members
or components. Current seismic analysis methodologies are classified as static, modal
response spectrum analysis, and time history (step-by-step) analysis.
Static method
The static method is applicable to structures or components with a dominant mode, and
uses as seismic coefficient the spectral value corresponding to the fundamental
frequency. Following Standard IEEE 693-2005, Manual 113 requires that the seismic
coefficient be multiplied by 1.5 when higher vibration modes could contribute
significantly to the seismic response. Static seismic forces on each component or
equipment are calculated by multiplying their mass by the seismic coefficient in each
direction of motion, and are applied at their centers of mass.
Since seismic vertical and horizontal ground motions occur at the same time, codes
require simultaneous consideration of seismic forces in both directions. Considering that
spectra are maximum responses that are unlikely to occur simultaneously in the three
164
directions, Section 5.6.2.1 of Manual 113 allows estimation of the combined response as
the square root of the sum of squares (SRSS) of the responses to each direction of ground
motion. Another rule combines 100% of the response to one direction of motion with
40% of the responses to the motion in the other two directions (100, 40, 40 rule). ASCE
4-98 Standard describes the use of these rules.
R=1
Noirmalized Acceleration
2.5
R = 1.5
R=2
2.0
R=4
1.5
1.0
0.5
0.0
0.1
10
100
Frequency (Hz)
This reflects past practice in the nuclear industry that considered no amplification of the
PGA for frequencies equal to or higher than 33 Hz. Recent ground motion models for
Eastern North America (Atkinson and Boore, 2006), however, indicate that the seismic
response at frequencies higher than 33 Hz can be significantly above PGA. The issue is
less relevant when nonlinear behavior is anticipated since the reduced spectra are flatter.
Flexible soil layers modify the ground motion. Typically, these site-effects are quantified
with one-dimensional wave propagation analyses that account for the stiffness and
damping properties of the soil/rock layers. ASCE 7-05 adopts a simpler approach where
sites are assigned a Class A to F based on shear wave velocity, standard penetration
blowcounts, and/or undrained shear strength. Then, two modification site coefficients are
tabulated for each site class in terms of the spectral amplitude at rock. This approach is
adopted by Manual 113 and is sufficient for most substation projects. If necessary, wave
propagation analysis can be conducted with linearized procedures (Schnabel et al., 1972).
The most important task is the determination of shear modulus and viscous damping that
properly represent the dynamic nonlinear response of the soil layers.
For combining modal responses, Section 5.6.2.1 of Manual 113 allows the SRSS method
when the relevant natural frequencies differ more than 10 percent. Otherwise, a complete
quadratic combination (CQC) method must be used. Since CQC rules can be easily
implemented in computer codes, there is no reason to use the SRSS method.
Three modal analyses are conducted with spectra for three orthogonal directions of
ground motion. In each analysis, the total response is a CQC combination of modal
responses. These three responses are combined with the SRSS or the 100-40-40 rules.
Time history (step-by-step) analysis
The time history analysis consists of numerical integration of the differential equations of
motion in small time steps. For linear systems, the time history is simplified because the
mass, damping and stiffness matrices remain unchanged. The analysis must consider
only the modes with larger participation, and the input consists of three accelerograms
matching the design spectra in each direction of ground motion. Uncorrelated records
can be input simultaneously, considering all combinations of positive and negative signs
to capture the most unfavorable effects on each component. For correlated records, the
analysis is performed independently for each ground motion direction and the results are
combined with the SRSS or the 100-40-40 rules. Linear time history analyses also have
to consider nonlinear effects. One possibility is to develop time histories that match the
reduced spectra, but it is preferable to perform the actual nonlinear analyses. Linear stepby-step analysis constitutes a convenient option for systems with non-proportional
damping that have complex modes and frequencies; otherwise, the modal response
spectrum approach is sufficiently accurate.
Spectral matching must satisfy several requirements. In addition to enveloping the
design spectrum, a time history must contain sufficient energy in all relevant frequencies.
Typical frequency content, duration and matching requirements are given in Appendix A
of IEEE 693. Time histories also must reflect site conditions, usually by means of one-
165
166
dimensional wave propagation analysis (Schnabel et al., 1972); however, more accurate
two and three-dimensional formulations have been developed by Bielak et al. (2003).
Time history analysis is unavoidable if one needs to take into consideration nonlinearities
in components or equipment. Appropriate constitutive models must be identified to
capture the main nonlinear structural characteristics under cyclic loading. Since the
nonlinear response can vary significantly under different input time histories, sufficient
earthquake records must be selected to represent the potential seismic events at the site.
An added complication is the need to incorporate all combinations of signs and directions
of seismic input. Owing to these difficulties, nonlinear analyses are rarely conducted,
except for seismic qualification of equipment excessively large to be qualified by testing.
EFFECTS OF DAMPING RATIO ON DESIGN SPECTRA
As prescribed in current codes, R values are applicable to structures with 5 percent of
critical damping, usually appropriate for buildings. Different damping ratios, say 2%,
may be adequate for substation components. The design seismic coefficient or spectrum
can be adjusted using Fig. 1. However, based on the IEEE 693 equations, the spectrum
for a prescribed percent damping ratio, d, is obtained by multiplying the 5-percent
spectrum by Fd, as follows:
3.5
2%
Normalized Acceleration
3.0
5%
10%
2.5
Fd, 2%
Fd, 10%
2.0
1.5
1.0
0.5
0.0
0.1
1.0
Frequency (Hz)
10.0
100.0
for 0 f 8 Hz
for 8 f 33 Hz
for f > 33 Hz
(4)
(5)
(6)
(7)
167
percent. Fig. 3 shows that the average spectra (continuous lines) resemble design spectra
and indicates an excellent agreement with spectra calculated with Eqs. 4 through Eq. 7.
We have developed a similar approach where the 5-percent spectrum is multiplied by a
factor , as follows:
= 0.3 + 0.01d
(8)
(9)
(10)
= (5/ d )
= 1 + ( 1) exp(0.05 f)
3.5
exact 2%
Normalized Acceleration
3.0
exact 5%
exact 10%
2.5
approx 2%
2.0
approx 10%
1.5
1.0
0.5
0.0
0
10
100
Frequency (Hz)
(11)
168
3.0
exact 2%
Normalized Acceleration
2.5
exact 5%
exact 10%
2.0
approx 2%
approx 10%
1.5
`
1.0
0.5
0.0
0.1
1.0
10.0
100.0
Frequency (Hz)
1.6
exact 2%
Normalized Acceleration
1.4
exact 5%
exact 10%
1.2
approx 2%
1.0
approx 10%
0.8
0.6
0.4
0.2
0.0
0.1
1.0
10.0
100.0
Frequency (Hz)
169
exact 2%
0.9
Normalized Acceleration
exact 5%
0.8
exact 10%
0.7
approx 2%
0.6
approx 10%
0.5
0.4
0.3
0.2
0.1
0.0
0.1
1.0
10.0
100.0
Frequency (Hz)
should be kept in mind that piles could experience excessive lateral displacements and
bending moments in liquefied soils. A possible solution is the used of drilled shafts
embedded into deeper non liquefiable soils or in rock.
DISPLACEMENTS
Section 5.2 of Manual 113 highlights the importance of accurately estimating
displacements in substation components, warning that A structure designed for strength
may have excessive deflections. By contrast, Section 3.1.11.3 of the Manual states that
loads from earthquakes should not be considered in deflection analysis. It is argued that
nonlinear dynamic analyses can be difficult and unreliable. However, seismic
regulations, such ASCE 7-05, accept that linear analyses along with structural-response
modification factors, is sufficiently reliable for estimating stresses and displacements
under seismic loads. Most of the uncertainty in seismic response estimates resides in the
assessment of seismic activity, rather than in structural analyses methodologies,
The static and the modal spectral seismic analyses produce estimates of the displacements
at any point of a structure. Since the linear analyses use forces reduced by a factor R, on
account of nonlinear behavior, the ensuing displacements are also reduced and have to be
increased to estimate the inelastic deformations. To this end, seismic regulations
stipulate amplifications factors for R-reduced displacements. For instance, ASCE 7-05
requires that displacements be amplified by a factor Cd tabulated along with R. Manual
113 does not provide values for Cd. Upon examination of Table 12.2.1 of ASCE 7-05,
we propose that Cd be equal to R.
EXAMPLES
The following examples are based on the Dead-End Structure example of Section 3.6.1
and Fig. 3.5 of Manual 113. Cantilever support poles constitute the structural system in
one horizontal direction. We understand that 5 percent damping was considered in this
example. Manual 113 considers a Site Class D and provides the following information:
Fa = 1.33, SS = 0.590, Fv = 2.06, S1 = 0.186
SDS = 2/3 Fa SS = 2/3 1.33 0.590 = 0.523, assumed to control.
Manual 113 assumes R=2 and importance factor IFE=1.25. Thus, the seismic design force,
FE, is equal to (Sa/2) 1.25W = 0.33W. For the other direction, we assume that the
structural system is a moment-resisting steel frame. Now R=4 and FE = (Sa/4) 1.25 W
= 0.17 W, if all other data remain the same. In the vertical direction, lacking any other
guidance, we would use FEV equal to 0.8 times the larger horizontal force, i.e., FEV =
0.80.33 W = 0.26 W.
Now let us consider that a damping ratio of 2 (rather than 5) percent is appropriate.
Using Figure 5.2 of Manual 113, or Eqs. 6 and Eq. 7, the amplification in the flat region
of the design spectra (say at a frequency of 7 Hz), the factor equals 3.25/2.5 = 1.3.
However, using equations 9 and 11, with = R = 2, and d = 2, we have:
170
171
Standard IEEE 693, adopted in Manual 113, is adequate for modifying elastic spectra.
However, substation structures are designed anticipating inelastic behavior with energy
dissipation occurring via viscous as well as hysteretic damping. Using inelastic nonlinear
analyses of single-degree-of-freedom systems we have verified that viscous damping is
less effective when higher hysteretic energy dissipation occurs. This implies that the
required modification of the design spectrum due to a change in the damping ratio
decreases when the target ductility demand increases. Equations 9, 10 and 11 have been
developed herein to calculate spectral adjustment factors for damping ratios between 2
and 10 percent, considering average ductility demands between 1 and 4. These factors
correctly approach unity at high frequencies.
REFERENCES
Atkinson, G.M. and Boore, D.M. (2006). Earthquake Ground-Motion Prediction for
Eastern North America. Bulletin of the Seismological Society of America, 96(6):
2181-2205.
ASCE (2000). Standard ASCE 4-98: Seismic Analysis of Safety-Related Nuclear
Structures. Reston, VA.
ASCE (2006). Standard ASCE/SEI 7-05: Minimum Design Loads for Buildings and
Other Structures. Reston, VA.
ASCE (2008). Manual 113: Substation Design Guide. Edited by L. Kempner Jr.
Reston, VA.
Bielak, J., Loukakis, K., Hisada, Y., and Yoshimura, C. (2003). Domain reduction
method for three-dimensional earthquake modeling in localized regions part I:
Theory. Bulletin of the Seismological Society of America, 93(2): 817-824.
USGS
(2008).
United
States
National
Seismic
Hazard
Maps.
http://earthquake.usgs.gov/research/hazmaps/
Gazetas, G. (1991). Chapter 15: Foundation Vibrations. Foundation Engineering
Handbook, Edited by H-Y Fang, Van Nostrand, New York.
Idriss, I.M. and Boulanger, R.W. (2008). Soil Liquefaction During Earthquakes.
Earthquake Engineering Research Institute, Oakland, CA.
IEEE (2006). IEEE Standard 693: Recommended Practice for Seismic Design of
Substations, American National Standards Institute.
Jennings, P.C. and Bielak, J. (1973). Dynamics of building-soil interaction. Bulletin
of the Seismological Society of America, V. 63, No. 1, pp. 9-48.
Prakash, S. and Sharma, H.D. (1990), Pile Foundations in Engineering Practice, John
Wiley and Sons, New York.
Puri, V. K. and Prakash, S. (2007). Foundations under Seismic Loads. 4th
International Conference on Earthquake Geotechnical Engineering. Paper No. 1118.
Thessaloniki, Grece.
Schnabel. P.B., Lysmer, J. and Seed, H.B. (1972). SHAKE: A computer Program for
Earthquake Response Analysis of Horizontally Layered Sites. Report UCB/EERC
72/12. University of California, Berkeley, CA.
172
PE, SE, Member ASCE, Manager, Sargent & Lundy LLC, 55 E Monroe, Chicago IL
60603, Email: [email protected], Phone: 312-269-2637.
2
EIT, Associate Member ASCE, Associate, Sargent & Lundy LLC, 55 E Monroe,
Chicago IL 60603, Email: [email protected], Phone: 312-269-2623.
3
PE, SE, Fellow ASCE, Senior Manager, Sargent & Lundy LLC, 55 E Monroe,
Chicago IL 60603, Email: [email protected], Phone: 312-2696486.
ABSTRACT
Fault currents are increasing on existing and new substations due to system
additions and modifications. Bus designs are typically based on the circuit breaker
rating of 40kA to 80kA. Phase spacing is no greater than that required by insulation
level, thus for configurations in the 115kV to 138kV range, magnetic forces over 100
pounds per foot are not uncommon when computed per IEEE 605 methods.
The magnetic force calculation in IEEE 605 is proportional to the decrement
factor squared and 1.6 is the default decrement factor. The first reduction strategy is
to calculate the actual decrement factor. The decrement factor is a function of fault
clearing time, system reactance and system resistance. Typical system reactance to
resistance ratios (X/R) can reduce magnetic force 12 to 18 per cent for a two cycle
breaker. More reduction is available with greater clearing times.
The second force reduction strategy is to compare the natural frequency of the
bus to the forcing function frequency. This paper presents a simplified generalized
coordinate method to determine the frequency based response of the bus to the
magnetic field.
The paper will also review the cost savings as a result of applying these
techniques.
DECREMENT FACTOR
Decrement factor defined. The purpose of the decrement factor is to account for the
momentary peak effect of the AC and decreasing DC components of the short-circuit
current during the first half-cycle of the fault, where the DC component is at a
maximum. A full explanation of the theory behind the decrement factor is outside the
scope of this document see section 10.2 of IEEE Std. 605-1998 for further
explanation.
173
174
Influence of the decrement factor. The following equations show that the shortcircuit force is proportional to the square of the decrement factor (See Equation 12
IEEE Std. 605-1998).
)2
C Df 2 ISC
FSC = Kf
D
(1)
where
FSC = short - circuit force, lbf/ft
Kf = mounting - structure flexibility factor usually taken as unity
C = 5.4 x 10-7 for USCS units
Df = decrement factor as given in the equation shown below
ISC = symmetrical short - circuit current, A
D = conductor phase spacing center - to - center, in
Df =
1+
Ta
tf
2 tf
Ta
1 exp
(2)
where
Ta =
R 2 f
(3)
and
tf = fault current duration, sec
X = system reactance
R = system resistance
f = 60 Hz
Figure 1 plots the decrement factor for a common 5 cycle clearing time, tf =
0.083 sec, at a fault current of 80 kA over a range of (X/R) ratios with a common 138
kV phase spacing of 8-0. Note that even for very high X/R ratios the calculated
values approach but do not reach the default maximum decrement factor of 1.6.
Figure 2 plots the corresponding short-circuit force using Equation 1. Note
that only for very high X/R ratios do the calculated values approach 160 lbf/ft
maximum short-circuit force using the default 1.6-decrement factor.
175
1.6
Decrement Factor
1.5
1.4
1.3
1.2
1.1
1
10
20
30
40
50
60
70
80
90
100
X/R
180
160
140
120
100
80
60
40
10
20
30
40
50
60
70
80
90
100
X/R
as a function of X/R
using default decrement factor of 1.6
requires two additional electrical design inputs - the X/R ratio and the fault clearing
time.
DYNAMIC RESPONSE OF BUS
Introduction. The short-circuit force is dynamic in nature. Consequently, the
dynamic response of a given span of rigid bus is highly dependent on the natural
frequency of that span in relation to the frequency of the forcing function. The
standard method of calculating the short-circuit force does not account for the
dynamic response of the rigid bus to the short-circuit-forcing function and often leads
to overly conservative results.
IEEE recognizes this and even states that for certain natural frequencies the
rigid bus will have fault current forces of less than one half the calculated first
quarter-cycle force when the conductor span reaches full deflection (IEEE Std. 6051998). In fact, the standard equation computes the equivalent static force on the bus
assuming the bus has fully deflected at the same time the forcing function reverses
direction. However, such a condition is extremely unlikely considering that the
natural frequency of a common rigid bus span is significantly different from the
frequency of the forcing function (60 Hz) see Figure 3 below.
Most utilities account for the conservatism with an overload factor around 1.0
applied to short-circuit forces for insulator design, compared to the much higher
factor applied to wind and dead loads of approximately 2.0. Note that the smaller
overload factor for insulator design no longer applies with a more accurate estimate
of the short-circuit load; an overload factor similar to that for wind is instead
appropriate.
The following sections explain the effect of the natural frequency on the
dynamic response of the bus to the short-circuit forcing function. Also presented is a
simplified method for determining the dynamic response of rigid bus to the shortcircuit forcing function.
Assumptions. The following common rigid bus assumptions and materials provide
the basis for the dynamic analysis method for reducing the short-circuit force.
176
177
Note that one can apply this simplified method for rigid bus arrangements
with rigid connections between the high and low bus using structural modeling
software capable of dynamic analysis.
Common natural frequencies. This section computes the natural frequencies for a
range of realistic rigid bus spans using the aforementioned assumptions and materials.
The standard equation for the natural frequency of a fixed-fixed beam is:
fn =
3.52
wL4
384EI
where
fn = natural frequency of the rigid bus span, Hz
w = self weight of bus, kip/in
L = span length, in
E = modulus of elasticity, ksi
I = moment of inertia, in4
Plotting the above function for the previously defined range of span lengths
produces Figure 3 shown below.
70
Natural frequency, Hz
60
50
40
30
20
10
0
15
20
25
30
35
40
178
The results shown in Figure 3 indicate that, even for very short and stiff rigid
bus spans, the natural frequency of a rigid bus span is significantly less than that of
the forcing function. For example, the highest natural frequency shown in Figure 3
(15 span) is 39.6 Hz which is only 36.9/60 = 0.66 times the forcing function
frequency. Thus, the natural frequency of the rigid bus is usually significantly less
than that of the forcing function, which corresponds to a similarly reduced dynamic
response and equivalent static short-circuit force than that calculated using the default
equation.
Overview of method. A discussion of the structural response of beams subjected to
uniformly distributed load is outside the scope of this document. However, the
method presented below follows the distributed mass undamped system response to
a uniformly distributed harmonic loading using the generalized single-degree-offreedom methods presented in sections 2-7 and 4-1 of the Dynamics of Structures
textbook (Clough 1975). The required steps and simplified equations are as follows:
(5)
where
5. Calculate the natural circular frequency of the beam (rad/sec) with the
following equation with the generalized mass and stiffness defined above.
n =
kG
mG
179
(6)
6. Calculate the generalized short-circuit force for the shape function with:
L
FG = FSC (x)dx
(7)
where
FG = generalized short-circuit force for the shape function, lbf/in
FSC = short-circuit force calculated using Equation 1 converted to units
of lbf/in from lbf/ft
(x) = shape function previously defined
7. Noting that the deflection everywhere in the beam is at a maximum at the
same time the deflection at midpoint is a maximum; use the general
solution of the displacement at midspan with the following equation.
FG 1
(sin t sin n t ) m
2
kG 1
m (t ) =
(8)
where
m(t) = general solution of deflection at midspan vs. time, in
FG = generalized short-circuit force for the shape function, lbf/in
= circular frequency of the forcing function, rad/sec
t = time (ranging from 0 to the fault clearing time, tf), sec
= ratio of the forcing frequency to the natural frequency = /n
kG = generalized beam stiffness, lb/ (in-sec2)
m = is the deflection value of the shape function evaluated at
midspan = (x=L/2)
8. Plot the general solution for midspan deflection over the duration of the
fault (from t = 0 to t = tf) and locate the maximum midspan deflection
(max).
9. Calculate the static component of the maximum midspan deflection using:
st =
FG
m
kG
where
st = static component of the maximum midspan deflection, in
FG = generalized short-circuit force for the shape function, lbf/in
m = is the deflection value of the shape function at midspan
kG = generalized beam stiffness, lb/ (in-sec2)
(9)
180
10. Calculate the dynamic maximum response (DMF) of the rigid bus span
to the forcing function as shown below.
DMF =
max
st
(10)
where
max = maximum total deflection at midspan from step 8, in
st = static component of the maximum deflection at midspan from
Equation 9, in
11. Finally, compute the equivalent static short-circuit force applied to the
bus with:
(11)
where,
FSC = short-circuit force calculated using Equation 1, lbf/ft or lbf/in
DMF = dynamic maximum response calculated above
12. Design the rigid bus and insulators the usual way with the equivalent
static short-circuit force and an insulator overload factor of 2.0 to 2.5
applied to the short-circuit load.
Note that it is conservative to apply the resulting reduced short-circuit load,
FSC_EQ, for the stiffest span to all rigid bus spans.
Force reduction. The amount of short-circuit force reduction is proportional to the
dynamic maximum response factor, DMF, which will be less than 1.0 for normal
rigid bus spans. The actual reduction depends on the rigid bus material, span length,
support conditions, other loads (such as wind and ice) and the utilitys overload
factors for insulator design. The results of the hypothetical case study (see following
section) provide reasonable estimates of the force reductions one can expect for
common 138 kV rigid bus spans.
COST SAVINGS
The cost savings associated with both the decrement factor and dynamic
analysis methods to reduce the short-circuit load are dependent on many variables
unique to each substation such as the rigid bus materials, insulator strengths, other
rigid bus loads, span lengths, support conditions and the insulator overload factors.
Therefore, this section of the paper uses the results from a hypothetical case study to
provide reasonable estimates of the force reductions and cost savings one can expect
for the typical 138 kV substation.
181
Design inputs. The following design inputs define the case study rigid bus
geometry, support conditions, loads, fault current data and insulator overload factors.
Reduced /
Default
62.4
1.00
0.0
Df reduced to 1.27
39.2
0.63
37.2
1 to 3
14.4
0.23
76.9
10
0.14
85.6
Combining methods
Percent
See
reduction Equation(s)
Total lateral force reductions. The cost savings associated with reducing the shortcircuit force is related to the total reduction in lateral load (wind and short-circuit). In
addition, the insulator strength usually controls the allowable span length between
support rather than bus stresses and deflection. Therefore, the total lateral load
including insulator overload factors is the most important criterion when determining
182
cost savings. Table 2 below presents the effect of the both the decrement factor and
dynamic analysis methods of short-circuit force reduction as well as a case combining
both methods.
Table 2. Total lateral loads including overload factors with corresponding
reductions
K2Fsc*
(lbf/ft)
K1Fw *
(lbf/ft)
99.8
24
124
1.00
0.0
Df reduced to 1.27
62.8
24
87
0.70
29.9
36
24
60
0.48
51.5
22.5
24
46.5
0.38
62.4
FT ** Reduced / Percent
(lbf/ft)
Default reduction
* K1 = 2.5 and K2 = 1.6 or 2.5 for short-circuit load without and with dynamic
reduction, respectively
** FT = K1Fw + K2FSC
Estimated cost savings. As mentioned earlier, the insulator strength generally limits
the allowable rigid bus span lengths. Therefore, one can estimate the cost savings
based on the ratio of default to reduced total factored lateral loads because the ratio
applies to the number of bus supports required for a given bus run. For example,
Table 2 above shows that, compared to using the default short-circuit force, one
would only need approximately 38% as many bus supports (including insulators,
fittings, support structures and foundations) after combining both reduction methods.
Hence, the percent reduction column in Table 2 above represents the cost savings in
insulators, fittings, support structures and foundations as a percentage of the default
costs.
CONCLUSIONS
One can expect the following cost savings for a typical 138 kV rigid bus
design including insulators supports and foundations
-
BIBLIOGRAPHY
Institute of Electrical and Electronics Engineers (IEEE). (IEEE Std. 605-1998). IEEE
Guide for Design of Substation Rigid-Bus Structures, New York, New York.
Clough, R.W. (1975). Dynamics of Structures. McGraw-Hill, USA
183
184
Aspects, or NNA) as long as the general nomenclature is respected. This is what most
of them do.
This writer is in a unique position to comment on the recent wind provisions as: 1) he
has conducted academic research and taught graduate level and professional classes in
wind engineering, reliability-based design and transmission line design, 2) he has been
a member of some of the CIGRE (CIGRE, 1990) and ASCE (ASCE Manual 74, 1984,
1991, 2008) committees that developed the recent wind provisions, 3) he has worked
as a transmission line consultant for many years and investigated line failures, and 4)
he has participated in the implementation of the wind provisions of many international
codes and standards in his companys design software (PLS-CADD & TOWER) that
is used in over 100 countries. In doing so he has sometimes uncovered provisions that
were simply untested, incomplete and confusing. Being confronted with so many
ways of approaching the common problem of designing safe lines economically and
practically, has led this writer to reflect on what we are doing, what is really justified
and to suggest some simplifications. This is the rationale for writing this paper and
hopefully convincing future generations that: 1) there is nothing sacred behind the
current code procedures, 2) the complexity of the problem is not amenable to its
description by fancy equations, and 3) simpler formulations should be considered.
Refining the wind loading equations has been part of the larger goal of improving
design by what is now commonly known as Reliability Based Design (RBD). There
was even some hope in the early days of RBD development, that given a sufficient
amount of research and data collection, our industry would someday be able to
quantify the probability of failure of a line (for example statistically determine the
extent and number of failures of a line over a period of time). However, in part
because of the uncertainties described in this paper, other uncertainties related to the
size of the storms and the corresponding number of exposed structures, we will never
be able to achieve this lofty goal. In spite of this shortcoming (inability to quantify the
probability of failure), RBD is still a very valuable guide to develop consistent design
procedures covering various combinations of loading events and materials (CIGRE,
1990 and 2006; ASCE Manual 111, 2006; Ghannoum, 2002; Mozer et al, 1984; Peyrot
et al, 1984). Some concepts of RBD and simplicity are not incompatible.
2. TYPES OF WIND STORMS
For the proper understanding of the various engineering approaches to the
determination of wind loads on transmission lines, it is imperative to have some
knowledge of the various types of wind storms that may be damaging to our lines.
Recommended reading on this subject are CIGRE Brochures 256 (2004), 344 (2008)
and 350 (2008). Here we will limit our discussion to winds with gusts in the range of
40 to 60 m/s (about 90 to 140 mph) that cover at least two spans. Tornadic winds,
while extremely violent, generally cover less than two spans and their peak velocity
values are generally not considered in the statistics that form the basis of the Basic/
Reference Wind maps produced around the world. While tornadic winds can be
considered in design, for example as explained in ASCE Manual 74 (1991 and later),
they will not be discussed in this paper.
185
186
187
V[z,t]
2 OR 3 SEC.
value
Vf[z,t]
(the
10 MIN. OR
1 HOUR
turbulence) as shown in Fig. 1. The fastest wind value is the peak gust Vg[z].
Actually, given the precision of everything related to wind, the peak gust or the often
quoted 2-sec or 3-sec gusts are equivalent and will simply be referred to as the peak
gust or simply gust in this paper. The gust is the maximum wind that a structure or
a very short span experiences. While Fig. 1 depicts the variation over time of the wind
velocity at one point, it could also be interpreted as the spatial variation of the wind
velocity along the centerline of a transmission line at one instant of time.
2.5.1 The mean value
In the context of a winter storm, it can be argued that if you are far enough above the
ground (above the gradient height), there is very little turbulence and the wind velocity
is not affected by the retarding effect of the roughness of the ground (i.e., when z is
higher than the gradient height, Vf[z,t] is equal to zero and Vm[z] is constant and
equal to the gradient speed). Based on experimental observations and considering the
need for simplification, an academic wind assumes that, below the gradient height, the
average wind velocity decreases following a well defined profile. Below the gradient
height, the rougher the surface of the ground (upstream from the location where we are
interested in the wind speed), the slower the average wind speed Vm[z] will be. Fig. 2
shows typical academic wind profiles, completely defined by the assumed reference
wind speed VRef at 10 m above the ground and the ground roughness.
Profile equations proposed by three major international codes for Open Country or
Reference category (called Category C by ASCE, Category B by IEC, and Category 2
by CENELEC) are:
ASCE/ NESC
1/9.5
Vm[z] /Vm[10] = 1.42(z/275)
IEC
CENELEC
0.16
188
Eq. 1
Eq. 2
Eq. 3
HEIGHT
At 40 m above ground, the
Z
GRADIENT
above equations give increases
GRADIENT
SPEED
HEIGHTS
in velocity of 1.16, 1.25 and
URBAN
1.27, respectively (actually,
> 200 m (700 ft)
Eq. 1 is applied to the gust
SUBURBAN
even though its source implies
an established wind over
OPEN ( REFERENCE )
several minutes). Given that
W ATER SURFACE
pressures are proportional to
ZONE OF INTEREST
the squares of velocities, the
pressure increase at 40 m over
10 m ( 33 ft )
Vm [z]
that at 10 m are 35%, 56% and
V
REF
61% respectively. Nobody can 0
really say that one of the
Figure 2 Typical profiles
prescribed increases is better
than the other, as in a real storm the values are quite variable. One of the few full
scale tests conducted on power lines (Houle et al., 1991) during Winter Storms
conditions, show a very high coefficient of variation (from 30 to 42%) for the
exponent of the power laws in Eq. 1 and 2. These observations, coupled with the fact
that terrain categories are somewhat arbitrary and difficult to assign in real situations,
suggest that there is significant uncertainty in the assumed shapes of the profiles. In
fact, as described in CIGRE Brochure 350 (CIGRE, 2008), even the general shapes of
the profiles shown in Fig. 2 are totally inappropriate for local storms.
While the profiles described by Eqs. 1 to 3 are supposed to be valid for the mean
velocity and extend to very high elevations (the gradient height), actual profiles of
interest are generally limited to less than 60 m (except for river crossing towers).
2.5.2 The fluctuating value
The fluctuating value Vf[z,t] is assumed to be a stationary random process with
invariant statistical properties over the averaging time period Tav. This assumes that
during at least 10 minutes the storm properties do not change (stationarity). One can
immediately see that this assumption is invalid for all local storms. Due to the
retarding effect from the roughness of the ground, the turbulence increases from the
gradient height down to the ground level, and is greater for rougher terrains (i.e.
increasing from terrain Category D to Category A according to the ASCE
189
-2/3
Eq. 5
190
paper between 2 sec. and 10 min., but it is being proposed to determine Gust Factors.
ASCE Manual 74 uses a curve developed by Durst (Durst, 1960) to estimate GFs.
CENELEC uses the following equation for the Gust Factor at height z to relate the
peak gust Vg[z] to the 10-min average for Open Country:
GF[z] = 1 + 2.28/ Log[z/0.05]
Eq. 6
This author developed some teaching material for an advanced graduate course on the
dynamics of structures subjected to random loads such as wind and earthquakes and he
used a well-developed textbook on the subject (Ghiocel et al, 1975). Based on this
experience, it was concluded that some advanced knowledge of calculus, statistics and
random processes, and several class hours were needed simply to understand the
concepts and limitations behind the academic winds, their power spectra and crossspectra, and the corresponding structural responses. The complexity of the
formulations has intimidated many engineers and prevented them from questioning
their validity.
In summary, the academic winds assume the following:
At the somewhat arbitrary gradient height there is a hypothetical wind with no
191
turbulence that remains constant over at least 10 minutes. However, due to the
roughness of the ground, the mean wind is slowed-down by friction against the ground
and turbulence is generated by that friction. The amount of turbulence depends on the
roughness of the ground that is somewhat arbitrarily defined for 3 to 5 categories:
smooth water surfaces such as coastal areas (ASCE Category D), slightly rough
surfaces such as open country (ASCE Category C), rougher surfaces such as suburban
areas (ASCE Category B), etc. Estimated gusts associated with academic winds are
very approximate statistical outgrowth of the turbulence and are not a measured
quantity.
Unfortunately, because of the impossibility of characterizing the almost infinite ways
in which real winds can vary during a storm, their reduction to academic winds as
described above (1960 and 1970 vintage models) has gained credibility due to: 1) the
absence of alternates, 2) the lack of understanding of the limiting assumptions, 3) the
elegance of the formulations, and 4) the status acquired by their appearance in print.
Our industry was grasping for something to quantify a basically unknowable
phenomenon. Therefore, like with other parts of the wind loading equations that are
discussed later in this paper, the academic winds spread like unchecked viruses in
many codes, are often accepted as facts, and are unnecessarily complicating our
design processes.
3. EFFECT OF TOPOGRAPHY
It is well known that the presence of hills and valleys locally affects the wind velocity.
The wind blowing over a hill ridge and perpendicular to that ridge has a higher
velocity and different turbulence characteristics than the wind undisturbed by the hill.
There are proposed solutions to that problem (Armitt et al., 1975; ASCE Manual 74,
2008), but it is generally not taken into account in transmission line design, probably
because of the wide variation of topographic features along a line route and the need
for standardization. However this is an important concept for building design where
the topography is well known, the structure is custom designed and the need for
precision is higher.
4. FROM WIND VELOCITY TO STRUCTURAL RESPONSE
If a transmission structure or a span were just a small rigid body at a distance z above
the ground, the maximum wind force on it would simply be:
F = qg[z] Cd A
Eq. 7
where qg[z] is the stagnation pressure (also called the dynamic wind pressure) equal to
2
0.5 x mass density of air x Vg[z] (where Vg[z] is the gust velocity at height z), Cd is
the drag coefficient of the body, and A is its exposed area (perpendicular to the wind).
4.2.1 Span Factor and Gust Response Factor
Unfortunately our transmission structures and our spans may not be small in size and,
192
as some have suggested, they may have some kind of resonant dynamic response to
the wind. We will first deal with the resonant response concern and then with the
size effect.
By resonant dynamic response, we are not talking about the pseudo static response of a
line component that simply follows slowly increasing or decreasing winds, or the
aeolian vibration (perpendicular to the wind velocity) that some spans and structural
components experience under laminar wind, or the galloping of spans. We are talking
about the possibility of some along-wind resonance, i.e. a resonant response in the
direction of the wind due to the fact that the wind spectrum could have significant
energy close to some natural frequencies of the system. Fortunately for us, along-wind
resonance of transmission lines is NOT a concern. The majority of the wind loads
come from the spans and when a span is subjected to extreme wind (blown out with
non-uniform wind along its length) there is no conceivable mechanism or identifiable
pendulum-type natural frequency that could be excited dynamically by the wind. As
to some dynamic response of latticed towers, they have high natural frequencies at
which the wind does not have energy. This is unlike some tall buildings that have
much lower frequencies. Steel and concrete poles, if wires were not attached to them,
could have some resonant along-wind response. However, with the wires attached, it
is not possible for these structures to vibrate as they would do alone, both because of
the restraint from the wires and the damping that they provide. Therefore, any attempt
to include a possible resonant response factor in our transmission line design practice
should be resisted vigorously as unfounded. This is similar to the attempt made by
building designers to force our industry to include an earthquake loading case for
transmission lines. Transmission lines, as shown during real earthquakes or
theoretically, do not respond significantly to earthquakes as substation or building
structures do.
Now let us discuss the size effect, also referred to as aerodynamic admittance. First
consider a 40 m/s gust. In 3 seconds it will have covered a space of 120 m. Therefore,
such a gust will envelope an entire transmission structure (except probably a river
crossing tower) at once and also a very short span (certainly the span of a distribution
line). Therefore, there should not be any size effect concern for most of our
transmission structures if the velocity of the gust is the reference design value.
However, for spans, it is very likely that when a gust hits a portion of a span, the rest
of it is subjected to lesser wind velocities. This is because wind velocities are not fully
correlated over longer spans. Therefore, this size effect can be taken advantage of
by allowing a reduction of the unit wire design loads for long spans. This is usually
handled by adding a Span Factor (SF) to Eq. 7 as shown in Eq. 8:
F = qg[z] SF Cd A
Eq. 8
Codes that use Span Factors as shown in Eq. 8 normally neglect the size effect on
structures and neglect all dynamic effects. Only the lack of correlation of the wind
velocities along the span are considered.
193
Another approach to the handling of the size effect and the possible resonant
effect is the Gust Response Factor (GRF) approach. The GRF approach consists of
specifying a force F, which, if applied statically, would cause the system to reach its
expected peak response. With the GRF approach, Eq. 8 is replaced by Eq. 9 where the
2
stagnation pressure qm[z] is now equal to 0.5 x mass density of air x Vm[z] (where
Vm[z)] is the 10-min average, or longer, wind velocity):
F = qm[z] GRF Cd A
Eq. 9
Codes that use Gust Response Factors traditionally consider both the size effect and
the resonant effect, not only on spans but also on structures. The ASCE Standard 705 (ASCE, 2006) for buildings includes the resonant effect in its gust factors, but this
standard is certainly not applicable to transmission lines.
There is sometimes some confusion in the naming of the factors: for example,
CENELEC (CENELEC, 2001) in Art. 4.2.2.3 calls Gust Response Factor what is
really the square of the Gust Factor (Gust Factor = ratio of peak wind to 10-min
average) and it calls Structural Resonance Factor what is really as Span Factor. We
also know of big mistakes that have been made when using 10-min mean wind values
as input to equations such as Eq. 8 or gust values as input to equations such as Eq. 9.
Depending on the formulation, SF and GRF may be a function of the height above
ground z.
The derivation of the Gust Response Factor proposed by Davenport (Davenport, 1978
& 1979) that eventually found its way into the ASCE/ NESC equations is very
elaborate and based on many assumptions, one of which is quite arbitrary. This
arbitrary assumption simply says that the peak response of a span or a structure is
equal to its response to the mean wind value (10-min average) plus a certain number
(statistical factor) of standard deviations of the response. The statistical factor is
suggested to be a number between 3.5 and 4. The standard deviation is calculated as
the area under the power spectrum of the response. There are also approximations in
the mathematics. For any wind that does not follow exactly the stationarity
assumption and the power spectrum of the academic wind (as so many winds do), one
has to question the validity of the proposed GRF.
Because Eq. 9 starts with the mean wind, the Davenport approach should end up with
GRFs for small structures and short spans that are larger than 1. However, because
the NESC/ASCE procedures are not using an average reference wind but a gust
reference wind, qg[z] replaced qm[z] in Eq. 9 and the GRFs were decreased by the
square of the gust factor.
4.2.2 Wind direction
Whether one uses a design equation such as Eq. 8 or Eq. 9 for determining the wind
force on a span, it is almost universally assumed, as a worst case hypothesis, that the
wind blows perpendicular to the spans. However, if the wind were to blow at a certain
incidence angle from the normal to a span, the force would decrease by a factor that is
equal to the square of the cosine of the incidence angle. For example, if the wind is at
45 degrees, the force goes down by a factor of 2. For any incidence larger than 18
degrees, the force will go down by more than 10%, i.e. for winds that are equally
likely to come from any direction, there is an 80% chance that the force will be 10%
smaller (and sometimes much less) than assumed by a design equation that applies the
wind normal to the span. Since most design wind maps are statistics of wind
velocities that do not consider wind direction, assuming that these wind will occur
perpendicular to the spans as most codes require is quite conservative. For directional
winds such as Downslope Winds, some lines will be much more vulnerable than
others.
The wind direction effect is one of the major uncertainties when trying to estimate
wire loads on a probabilistic basis. Some building and communication tower codes
even include a wind directionality factor to account for the reduced probability of
maximum wind coming from any direction.
4.2.3 Center of pressure
There is the question about what value of z should be used in Eqs. 8 and 9. For
structures, the NESC suggests using a single value at 2/3 the height of the structure
and then using the resulting pressure over the entire height. Other codes expect the
designer to break down the structure into sections at different heights and to compute a
different pressure at each height based on the center of gravity of the area of the
section. For spans, some codes require the use of the center of pressure for the
conductor, which is approximated as 1/3 of the sag below the attachment points. One
issue is whether to have a different z for each wire or to use some kind of average.
Some codes suggest using the average attachment height of all wires or that of the
highest wire. Calculated loads are certainly sensitive to these assumptions.
4.2.4 Wire tensions
Wire tensions affect the transverse loads on all angle structures and all dead ends.
They also affect the vertical loads on all structures with non-horizontal adjacent spans.
So, the tension is not really due to the effect of the wind on the two spans adjacent to a
structure, but the uncorrelated wind on all the spans in the tension sections (from dead
end to dead end) where the structure is located. If this tension section includes many
spans, no one will ever know what length to use for the calculation of the SF or the
GRF, or even if the equations are valid for such a calculation.
4.2.5 Drag Coefficient
ASCE Manual 74 (1984 and 1991) had quite a bit of information on published values
of drag coefficients for cables, structural members and some assemblies.
While these ASCE documents and IEC suggest a drag coefficient of 1 for all wires,
194
195
international practice regarding this varies widely. Houle et al (1991) and ASCE
Manual 74 show that, even for a given Reynolds number, there a wide scatter of
conductor drag coefficients.
For poles, drag coefficients close to 1 are generally recommended.
Regarding latticed towers, the NESC simply requires a combined drag coefficient of
3.2 (1.6 for the members of the front face plus 1.6 for the members on the back face)
to be applied to the exposed area of the front face and a constant design pressure
calculated at 2/3 the height of the tower. However for rectangular cross section
towers, the IEC, CENELEC and ASCE all use the same formula for the combination
of drag coefficient and exposed area. The formula is:
2
2
2
(1+0.2 Sin [2a]) (Ct At cos [a] + Cl Al sin [a])
Eq. 10
where a is the direction of the wind relative to the tower transverse axis, At is the
area of the tower face exposed to pure transverse wind, Al is the area of the tower face
exposed to pure longitudinal wind and Ct, Cl are drag coefficients based on the
solidity ratios of the faces (these coefficients can vary from 4 to slightly less than 2).
This is an example of a formula that spread like a virus from one code to the other as
there is no real knowledge of what happens to a full size tower in a real storm. The
formula is of no help to the designers of towers that have non-rectangular cross
sections such as shown in Fig. 4 and is almost impossible to automate in a tower
design computer program.
The limitation of the formula was determined
experimentally (De Oliveira et al, 2006).
The limited use of Eq. 10
prompted the latest ASCE
Manual 74 (2008) to
include a universal wind
on members procedure as
an alternate to the wind on
face approach of Eq. 10.
This alternate procedure
conservatively
ignores
shielding
(which
is
impossible to know for any
configuration
different
from a perfectly rectangular
section) and determines the
wind force on each member
independently, based on the
relative orientation of the
wind and the axis of that
member. In the comparisons
of Section 6.2, we will refer to the ASCE 2008 method that uses the face as ASCE
2008 F and that that handles each member separately as ASCE 2008 M.
In addition to its lack of universality, the formula of Eq. 10 requires that a tower be
divided into sections, each section having its own solidity ratio and height above
ground. The combination of having to micro-manage the solidity ratio together with
the section height is another complexity (fortunately not required by the NESC) which
cannot be justified by the uncertainties involved.
Micro-managing the wind load on a tower can be pushed to extremes as demonstrated
by the UK-NNA (2004). This specification considers the probability that the
maximum conductor loads are not fully synchronized with the maximum load on the
tower itself, and therefore the two are not simply additive. Since the portion of the
resulting force in a member due to the loads from the wires as opposed to wind loads
directly on the structure varies depending on the type of member and its location (for
example main leg member vs. diagonal), there is a factor that adjusts wind loads based
on the relative sensitivity of the members to wire and tower loads.
4.2.6 Wind on insulators
Some codes require that wind on insulators be calculated, including the effect of
gustiness and increase of wind velocity with height. This is probably the epitome of
un-necessary complexity as no-one knows the shielding effect of the structures on
insulators on one side of a structure and the exposed area/ drag coefficient of an
inclined insulator under extreme wind. Given the small contribution of the insulator
load to the total structure load, this is certainly something that can be neglected as
rightly done in the US.
5. CODES, STANDARDS AND GUIDES SELECTED FOR COMPARISONS
Three design methods that have received international attention are those of the
ASCE/ NESC, CENELEC, and IEC. We will add to the list a fourth method, the UK
NNA method, which is one of the many variants of CENELEC, but is based on a one
hour average academic wind. All are loosely based on academic wind models
developed more than 30 years ago by the few individuals cited in Section 1. Very few
engineers on the current code committees even know the assumptions and complexity
of the original mathematics. However, over the last 30 years, the three design methods
have evolved somewhat independently due to the political and adaptation processes
that eventually led to their official adoption since 2000.
All four methods start with the selection of a basic 50-year Reference Wind measured
at the Reference Height of 10 m. Reference Winds are normally available from
national wind maps. However, the type of Reference Winds and their averaging time
are totally different. In IEC document, it is a 10-minute average wind. In CENELEC
(depending on the NNA or method chosen), it can be a 2-sec gust, a 10-min average or
even a 1-hour average as used in the UK NNA. In ASCE/ NESC, it used to be a
fastest mile (which is close to a 1-min average), but it is now a 3-sec gust. Since gusts
196
are the velocities of importance in design, collecting and using gust data as the
starting point (Reference Wind) is certainly the right thing to do, rather than counting
on some lose statistical relationships between a gust or a peak structural response and
the corresponding 10-min or 1-hour reference values. So we can certainly say that the
adoption by the ASCE/ NESC of a 3-sec Reference Wind is a major move in the right
direction. One of the common-sense features of the ASCE wind map is that measured
gust data for the non-hurricane zones of the US were assembled from a number of
stations in state-sized areas to decrease sampling error (Peterka, 1998). Then, based
on the insufficient variations over the Eastern 3/4 of the lower 48 States to justify
contours, a single zone of 40 m/s (90 mph) was adopted. This laudable simplification
is in contrast to some unbelievable micro-zoning maps included in some other national
codes. But it is understood that this simplification (harmonizing all winds into larger
zones) has its inherent inaccuracies and has been challenged (Simiu, 2003).
One thing is clear though with the ASCE/ NESC maps: they do mix together data from
large scale wind storms and local storms, as well as those from tropical cyclones, but
they exclude winds from tornadoes. They are wind estimates at a point. However, the
50-year reference wind that would be needed for a better reliability estimation of a line
should not include maximum winds at a point but maximum winds over the space of
the line: but those data are rarely available.
All four methods use wind profiles (effect of height), and some combination of gust
factors with span factors, or gust response factors, all of which are essentially only
valid for the academic winds. In fact, the ASCE/ NESC procedure, which is using
the best basis for the Reference Wind (3-sec), is still borrowing internally the
mathematics of an academic wind. Now, let's think for a minute about what it is
doing. It starts with a 3-sec gust which is related to measured data, but the academic
wind theory says that we need an established wind over at least 10-minute for the
theory behind the profiles and the GRFs to be valid. So the ASCE/ NESC procedure
takes a good design value (the 3-second gust) and it reduces that value internally by a
somewhat arbitrary number to get the corresponding 10-min wind so that it can
salvage the academic wind basis of the profile and the vintage GRF. It makes very
little sense to say that a measured gust of say 40 m/s is really the peak wind of an
assumed 10-min wind of 28 m/s or an assumed 1-hour wind of 25 m/s.
This writer understands the huge amount of sincere efforts behind the development of
the code procedures mentioned in this section. The negative comments throughout
this paper are only meant to emphasize the underlying uncertainties, the unnecessary
complexity and the lack of consensus.
For the design of buildings or other isolated structures as described in the ASCE
Standard 7-05 (ASCE, 2006) one can possibly justify some refinements in the
description of the nearby terrain roughness and topography, as well as the dynamic
and shape characteristics of these structures. But transmission lines are totally
different systems from buildings that are normally custom designed for a site.
Buildings have volumes for which there are very important issues of local force
coefficients that affect local pressures around the building and possible dynamic
197
behavior. Transmission lines and their supports, on the other hand, have line like
components with insignificant along-wind resonant dynamics and where local force
coefficients are irrelevant, but instead drag coefficients are needed. Transmission
lines also benefit from some level of standardization and they can cover a wide range
of terrain characteristics: their most important loads are span loads for which ground
roughness, span length and wind incidence introduce considerable uncertainties.
Therefore, design rules for transmission lines should be unique and not be imposed by
regulating bodies or academic-types that deal with building structures.
6. SOME COMPARISONS
Very simple examples are presented in this section to demonstrate the combined effect
of height and span length on wire loads and the combined effect of height and drag
coefficients on structure loads with the goal of appreciating the sensitivity of these
loads to some code assumptions and the corresponding lack of consensus.
6.1 Wire loads
The wire load examples all
consider a Drake conductor
sagged at 20% of ultimate
after creep for various
combinations of attachment
heights, span lengths and
code methods. For the five
attachment
heights
considered, the span lengths
were maximized so that they
would have an 8 m clearance
above ground at 115 deg C
after creep as shown in Fig.
Figure 5 Span examples
5 (a vertical to horizontal
scale ratio of 10/1 was selected for the display of the PLS-CADD models in Fig. 5 for
clarity of presentation).
As a reference for our wire load examples, a 1 m long rod having the same diameter as
the Drake conductor, located 10 m above the ground and subjected to a gust of 40 m/s
would be subjected to a force of .613 x 40 x 40 x 0.0281 = 27.6 N/m. In the rest of
this section, we will report the load per unit length of that same Drake conductor in the
models of Fig. 5 when subjected to combinations of heights, span lengths and code
assumptions as the rounded ratio (Wire Load Ratio) of the calculated code load
divided by the Reference Load of 27.6 N/m. The ratio is rounded to two digits, as any
other precision for our demonstration purpose would be meaningless.
For comparison with the old traditional US design methods, the ratios are also
calculated for NESC Rule 250B (including the 2.5 load factor for wind and the ice
thickness, if any). For example, the ratio is 0.93 (25.6 N/m divided by 27.6) for the
198
Heavy Loading District, 0.71 for the Medium District and 1.10 for the Light District.
For the various codes considered, the Reference Winds at 10 m above the ground were
taken as follows:
NESC 2007
ASCE 2008
IEC 2003
CENELEC 2001
UK NNA 2004
For the Reference Winds described above, Table 1 shows the Wire Load Ratios for the
Open Terrain categories of various codes. For completeness, these categories are
described below (the wording of the documents is purposely included to draw
attention to the sensitivity of the answers to some imprecise definitions):
NESC 2007
ASCE 2008
IEC 2003
CENELEC 2001
UK NNA 2004:
In order to show how loads change when going from one loosely defined terrain
category to the next rougher one, Table 2 shows Wire Load Ratios similar to those in
Table 1 for the categories defined below:
NESC 2007
ASCE 2008
IEC 2003
CENELEC 2001
UK NNA 2004:
199
Table 1
200
NESC 2007
ASCE 2008
IEC 2003
CENELEC 2001
UK NNA 2004
0.80
0.80
0.94
0.92
0.85
0.78
0.78
1.03
0.99
0.81
0.80
0.80
1.05
1.05
0.84
0.83
0.83
1.06
1.10
0.86
0.85
0.85
1.07
1.13
0.88
0.93
0.71
1.10
NESC 2007
ASCE 2008
IEC 2003
CENELEC 2001
UK NNA 2004
0.64
0.76
0.85
0.69
0.62
0.85
0.72
0.75
Not used
0.65
0.88
0.77
0.78
0.68
0.89
0.82
0.81
0.70
0.90
0.85
0.83
the span (the size effect), thus the relative insensitivity (a term we will use later in
this paper) of the wire loads.
But looking at numbers in identical locations in Table 1 and in Table 2 will show
much larger differences. For example, the fact that the number of trees or buildings
may vary from a few to numerous will lower the IEC span loads by over 17%,
which is more than the 14% variation due to the full range of heights and span lengths
considered. This is a very significant contribution to the uncertainty of the calculated
numbers when one considers that a line will most likely traverse terrains where the
roughness and topography of the upwind surface varies.
6.2 Structure loads
To illustrate some of the complexities,
variations and uncertainties of some code
tower wind loads, we purposely selected a
very tall tower (86 m) as shown in Fig. 6.
That tower has a fairly simple geometry,
so that Eq. 10 can be used on the 7 main
sections of the tower whose important
properties are summarized in Table 3.
The arms at each of the three levels were
modeled as three separate sections with
transverse wind areas of 4.5, 4.5 and 5.8
2
m , respectively.
Table 3
Sect.
#
1
2
3
4
5
6
7
Tower Sections
Average
Height
(m)
Face
Area
2
(m )
Solidity
Ratio
(%)
81
70
57
46
35
21
8
6.2
9.4
8.2
6.1
9.1
8.9
10.4
37.1
39.6
23.7
14.4
9.4
8.3
6.1
201
202
considered. For the Reference F case, the reference pressure of 981 Pa was applied
to the entire face of the tower assuming a drag coefficient of 3.2 for that face (similar
to that required by the NESC): this is a wind-on-face approach. For the Reference
M case, the reference wind of 40 m/s was applied on each individual member of the
tower, assuming a drag coefficient of 1.6 for each member and letting the computer
determine the wind load on each member based on the relative orientation of the wind
velocity and the member axis: this is a wind on members approach.
For each of the two Reference Cases and for some codes, the overturning moments at
the tower base were calculated as shown in Table 4. The overturning moment is the
simplest and most important measure of the wind load and its demand on the tower.
All the calculations in Table 4 assumed an Open Country (as was done for the wires in
Table 1). Since ASCE 2008 allows two alternate methods (a wind-on-face approach
and a wind-on-members approach), the two methods are included in the table and
are designated as ASCE 2008 F and ASCE 2008 M, respectively. All other code
methods use the wind on face approach. Where appropriate, Table 4 also includes
the pressures and drag coefficients for 3 of the 7 sections.
Table 4
Sect. 1
Press. CD
(Pa)
Sect. 3
Press. CD
(Pa)
Sect. 6
Press. CD
(Pa)
Reference F
NESC 2007
ASCE 2008 F
IEC 2003
CENELEC 2001
UK NNA 2004
10,300
11,700
9,200
10,500
13,000
11,300
981
1110
1190
1310
1700
1560
981
1110
1110
1300
1580
1390
981
1110
901
1090
1260
1000
ASCE 2008 M
11,800
Reference M
10,500
3.2
3.2
2.2
2.3
2.3
2.3
3.2
3.2
2.9
2.8
2.8
2.8
3.2
3.2
3.7
3.5
3.5
3.5
varies the drag coefficient based on the solidity ratio of the face of the section.
The IEC, CENELEC and UK NNA formulae do not specifically isolate a GRF for the
tower (even though the TOWER CENELEC calculations include the recommended
structural resonance factor of 1.05), but they require varying pressures with height
and varying drag coefficients.
The large differences in base moments when comparing the IEC, CENELEC and UK
NNA (there are much greater variations with other European codes that we have not
summarized in this example) were surprising given the supposed cross-pollination
between these specifications. This is just one example of the lack of consensus which
we have observed all over Europe.
Comparing the base moment for Reference M (which ignores the increase of wind
velocity with height and the complexity of the wind on face approach and associated
solidity ratio issues) to the moments from the other codes, suggests that the
conservativeness of the wind on members approach somehow makes up for ignoring
the other factors.
Although not shown here, there would be a substantial drop in the tower loads if the
terrain category was changed to the next rougher category.
As with the wire loads, there is some relative insensitivity of the final tower load
(measured by the base moment) to the tower height and other parameters. This is
because the increase of velocity with height may be tempered by a built-in code size
effect and section drag coefficients that are substantially smaller in the high portion
of a tower than near its base. Another contributor to the insensitivity is the fact that
short towers are often used on top of a hill where the wind velocity may be higher as
opposed to taller towers used at lower elevations. Given that: 1) some relative
insensitivity is observed for some code procedures, 2) tower wind loads are much
smaller than the sum of the wire loads, 3) tower wind loads are not fully synchronized
with the wire loads (lack of spatial correlation), and 4) large uncertainties related to the
use of academic winds and terrain categories, one should wonder if the minutia of
dividing a tower into sections is necessary.
7. NEED FOR SIMPLIFICATION
It has been suggested that, given that most lines are now designed by computers where
complex code formulas are automated (PLS-CADD, TOWER, Etc.), complexity is not
an issue. This writer totally disagrees with this argument for the two reasons discussed
below.
7.1 Reason 1 - Honesty
Given that damaging winds can come from a wide variety of storms and given all the
uncertainties discussed previously, it is basically dishonest to pretend that our wind
designs will be better (better balance between costs and reliability) if one fine-tunes
203
the contribution of each of the many factors affecting the problem with some
questionable equations. The multitude of the current factors that have to be accounted
for increases the chance of errors and provides additional litigation opportunities to
ignorant parties that will focus on minute irrelevant details rather than understanding
what is important. As engineers, we should always favor common sense.
7.2 Reason 2 - Simplify the life of the designer.
One extremely useful concept often used in line design considers that a family of
supports, for given supported wires and code criteria, has some allowable wind and
weight spans. This concept is immediately invalidated by making wind loads
dependent on structure height, conductor heights and span lengths. Another useful
concept is that of designing a standard family of supports with the same top geometry
but with different heights. For example, the upper portion of a tower and its shortest
body is common to the entire family, with body and leg extensions taking care of the
need for varying heights. Other issues that can plague the engineer when the wind
loads vary with height and span length are the corresponding calculations of the wire
tensions. Therefore, there are very good practical arguments for eliminating the
dependence of wind loads on height and span length for most common design
situations.
7.3 Suggested simplifications
7.3.1 Utilize the 3-sec gust as the Reference Wind
This simplification does not apply to the US where NESC/ ASCE already use gust as
the reference wind speed. CENELEC also allows as an option the use of the gust.
Since gust winds are those that destroy lines, starting the design process with the gust
wind (Reference Wind) eliminates the large inherent uncertainty in some codes of
going from mean wind velocity to gust through the Gust Factor or going from mean
structure response to peak value through the Gust Response Factor.
The pressure caused by the Reference Wind is the Reference Pressure qRef. Some
codes specify that pressure as the starting point of their wind calculations instead of
the corresponding gust velocity.
7.3.2 Eliminate terrain categories
Since terrain categories have such an uncertain effect on wind gusts and are not
amenable to clear definitions along a line, they should be eliminated. The reference
Open Country category is the only one to keep. This is currently done by the NESC.
A coastal or lake increase factor of about 20% might be appropriate for such exposures
(current ASCE Category D or IEC Category A).
7.3.3 Abandon the concept of Gust Response Factor
204
As its name indicates, a Gust Response Factor is the ratio of a peak structure response
divided by the average response due to the mean wind. As mentioned in Section 4.2.1,
the GRF normally accounts for a possible resonant dynamic effect and a size
effect (lack of correlation of wind gusts at distant points). But the dynamic effect is
not a factor in transmission lines. The size effect is significant for wires, but not for
their supporting structures. Therefore, the only contribution from the GRF should be a
reduction of wire loads for longer spans, which is exactly what the concept of a Span
Factor does.
Because: 1) there is no average structure response to apply a GRF to if one starts with
a reference wind which is a gust, 2) there is no identifiable resonant response in
transmission lines, 3) there is no significant size effect on transmission structures,
and 4) the only components that can benefit from the size effect are the wires, the
concept of GRF is inappropriate and should be replaced by a simpler Span Factor
where there is a need to reduce unit loads on very long spans. This turns out to be the
alternate empirical approach of CENELEC.
7.3.4 Eliminate height and span length as variables for the majority of designs
For all structures with maximum height below a cutoff value (to be determined but
certainly above 50 m) use the Reference Pressure qRef over the entire height of the
structure and use a reduced pressure (suggested to be around 0.90 qRef), for all wires
attached to the structures. The reduction factor accounts for the low probability of
having winds perfectly perpendicular to the spans and having the span loads perfectly
synchronized with the structure loads. Using a constant design value accounts for
some of the relative insensitivities discussed in Sections 6.1 and 6.2.
For unusual situations (river crossing or very long spans), an increase of wind velocity
with height and a Span Factor may be considered.
An obvious and even more justified extension of this recommendation would be to
simply use the Reference Pressure qRef on all structures and wires within a substation
(ASCE Manual 113, 2008).
7.3.5 Use the same wire wind load for the determination of the lateral wire load
transmitted to the supporting structure and for the determination of the wire tensions
This was discussed in Section 4.2.4, and if followed, this simplification will generally
result in conservative values of mechanical tensions. However, since tension affects
the loads on angle and dead-end structures and since these structures are normally
designed to a higher reliability level than regular tangent structures, conservative
tensions are desirable.
7.3.6 Offer an alternative to the solidity ratio-based drag coefficients for towers
205
206
207
Kluge, R.O. (2005). Change Proposal CP2718 for Rule 250C of the 2007 Edition of
the National Electric Safety Code, on behalf of Wisconsin Utilities
Association.
Manuzio, C. and Paris, L., (1964). Statistical Determination of Wind Loading Effects
on Overhead Conductors, CIGRE Report 231, Paris, France.
Mozer, J.D., Peyrot, A.H. and DiGioia, A.M. (1984). "Probabilistic Design of
Transmission Lines", J.of Structural Engineering, ASCE, Vol. 110, No. 10,
Oct., 2513- 2528.
NESC/IEEE (2007). National Electrical Safety Code , Standard C2-2007, IEEE, New
York.
Peterka, J. and Shahid, S., (1998). Design Gust Speeds for the United States, J. Of
Structural Engineering, ASCE, 124 (2), 207-214.
Peterka, J. (2005). Comment in Support of Change Proposal CP2718 for Rule 250C
of the 2007 Edition of the National Electric Safety Code, Cermak Peterka
Petersen, Inc., Fort Collins, CO 80524
Peyrot, A.H. and Dagher, H.J. (1984). "Reliability Based Design of Transmission Line
Structures", J. of Structural Engineering, ASCE, Vol. 110, No. 11,Nov., 2758
2777.
PLS-CADD, A Computer Program for the Analysis and Design of Overhead Electric
Lines, Power Line Systems, Madison, WI, USA.
Saul, W.E., Jayachandran P.M. and Peyrot, A.H. (1976). Response to Stochastic
Wind of N-Degree Tall Buildings, J. of the Structural Division, ASCE, Vol.
102, May.
Simiu, E. (1973). Gust Factors and Along-Wind Pressure Correlations, J. of the
Structural Division, ASCE, April, 773-783.
Simiu, E. et al. (2003). Wind Speeds in ASCE 7 Standard Peak-Gust Map:
Assessment, J. of Structural Engineering, ASCE, April.
TOWER, A Computer Program for the Analysis and Design of Steel Latticed Towers,
Power Line Systems, Madison, WI, USA.
UK NNA (2004). UK Normative Aspects (UK NNA), European Standard EN
50341-13-9, Issue 3, CENELEC, European Committee for Electrotechnical
Standardization, Brussels
208
Senior Project Engineer, POWER Engineers, Inc., Meridian, Idaho 83642; Tel: (208) 288-6499;
E-mail: [email protected].
Director, The Boundary Layer Wind Tunnel Laboratory, University Of Western Ontario, London,
Ontario N6A 5B9; Tel: (519) 661-3338; E-mail:[email protected].
ABSTRACT
Current criteria for the calculation of synoptic wind load effects on transmission line
components are based on Davenports gust response factors, which take into account
the different response of structures and long spans of wires to turbulent wind. The
gust factors include resonant components that imply possible dynamic amplifications
of the wind effect on the structures and wires, although the wire resonant response is
generally ignored because of high aerodynamic damping.
Although the gust response factor equations were derived for 1-hour mean wind
records, the factors are being applied for winds with different sampling intervals but
with reference to 10-minute winds. The conversion is based on Dursts velocity ratio
curve, which was not derived from records of extreme winds but from low velocity
storms. Due to these differences, and also other simplifications, there could be errors
in calculated line or component reliability of at least one order of magnitude.
While line design weather loading conditions are derived from synoptic wind records,
it is well known within the industry that most line or structure failures are induced by
non-synoptic storms, such as downdrafts or tornadoes generated by thunderstorms.
Field evidence from these storms indicate that the 1-hour mean factors, even when
extrapolated to very short-interval gusts, do not explain the observed effects and
therefore new span and gust factors are needed for design.
The paper reviews the basic concepts and history of span and gust factors and the
impact of some of the current assumptions and simplifications. It also focuses on
recent research on the actual effects of non-synoptic storms and the new factors that
are derived from these experiences.
INTRODUCTION
Current transmission line design codes such as ASCE (1991, 2006), NESC (2007),
IEC (2003) and BSI (1986) are based on winds measured within the atmospheric
boundary layer by anemometers mounted on masts, usually at airports and airfields.
These are large-scale air movements or pressure gradients generated by global
atmospheric instability as determined by seasonal patterns in temperate climates. The
10-minute or 1-hour or fastest-mile mean wind records are then normalized to a
common surface roughness and height above ground and, using statistics of extremes,
209
design wind speeds are established usually for a 50-year return period. In the U.S. 50year, 3-second gust design wind speeds outside of the hurricane regions are in the
order of 85 to 90mph (~ 40m/s.)
As the wind speed at any point within the turbulent boundary layer is not uniform but
a random function of time and space, the instantaneous velocity and spatial variations
of wind can be described only in statistical terms involving mean values, turbulence
intensity (root mean square fluctuations about the mean), spectrum of turbulence and
correlation functions. The resistance or response of structures within a flow of
turbulent wind, on the other hand, includes drag effects, inertial reactions due to the
acceleration of the structure, and structural damping.
In the case of a transmission system, the probability that a high-speed wind gust or
front can envelope the line and act simultaneously on all points of a medium to long
span of wires is very low, but at the same time most points of a structure are likely to
be affected by a single gust. The response of towers and conductors to turbulent wind
is also different, as they have different along-wind dimensions, shapes and inertial
properties.
In recognition of this most line design standards have usually incorporated factors
that adjust the effects of the turbulent wind on the different system components,
primarily wires and structures. These factors are known in the industry as span, gust
or, more recently, gust response factors.
The first reference that the authors could find on the subject of wire response as a
function of span is a CIGR paper by Manuzio and Paris (1964), which introduces an
analytical approach for the determination of wind effects on conductors based on
statistical theory.
Subsequently, Castanheta (1970) defines the ground roughness and the span length as
the most important factors influencing the conductor response, and introduces a
spatial coefficient of reduction for wind forces on long cables. However, the paper
focused more on the effect of anemometer response, as it represented the bulk of the
uncertainties at the time when local meteorological data were unreliable and the
measurement methods were not standardized.
Armitt et.al. (1975) summarizes the research work carried out by a European working
group for the determination of design loads on transmission systems. Armitts
description of the span factor is based on local anemometer and terrain conditions
with a 2-second averaging time. Three sets of curves were given for the three terrain
conditions studied, but with no specific instructions on the available options from
them.
All these early investigations made reference to the statistical description of natural
winds by Davenport (1960, 1961) and its relationship to the wind loading on
structures, Davenport (1969.) The same statistical description of the wind was used in
210
the development of the gust loading factor for line-like structures by Davenport (1967)
suitable for general code use.
GUST RESPONSE FACTORS
Based on the statistical methods discussed above, Davenport (1979) developed
simplified gust response factors (GRF) for transmission line structures and wires. The
concept was to develop design loadings from turbulent wind acting simultaneously on
the wires and the structure, recognizing firstly that high speed gusts would not
envelope a large span of conductors, and secondly that wind fluctuations at
frequencies near the critical values could result in dynamic amplification and
potential damage due to structure resonant response. The development of the GRF
equations and their application is well documented, Davenport (1979), ASCE (1991,
2006), NESC (2007). Only the basic concepts will be briefly reviewed here.
The peak response R of a structural system to fluctuating winds, in terms of bending,
shear or displacements, can be separated into a constant mean component R and a
dynamic component R that fluctuates randomly about the mean. The fluctuating
component can be subdivided into a background response that varies slowly in time,
and a sinusoidal resonant component cycling at the natural frequency of the structure.
Following Davenport (1979), the gust response factor is developed based on the peak
response or load effect R in the tower, given by the following:
R = R + g R
(1)
211
212
simple loading model the variability of the other parameters may be ignored. Solving
Equation (1) for the above conditions and the external parameters, the peak tower
response R can then be expressed as follows:
R = tc pc GRFc + tt pt GRFt
(2)
where the terms are influence functions for the tower response due to the
conductors and tower loads; terms p are the mean wind forces acting at the effective
height of conductors and tower, and the GRF terms are the gust response factors of
conductors and tower, both adjusted for effective height and ground surface
roughness. As the aerodynamic damping in the wires is significant, the dynamic
amplification Rc due to resonance in the wires can be ignored. The reduction in wind
loading in the conductor gust response factor is then due only to the lack of spatial
correlation of the wind gust along the line. The gust response factors are as indicated
below.
GRFc = 1+gE(Bc )
(3)
GRFt = 1+gE(Bt + Rt )
(4)
in which is a load coincidence coefficient ranging from 0.7 to 1.0 that accounts for
the non-simultaneous occurrence of peak tower and conductor loads, and E is an
exposure factor equal to two times the turbulence intensity Iu. The gust factor given
above for each line component is the ratio of the peak response, given by the sum of
the mean and the fluctuating responses, with respect to the mean system response.
DISCUSSION ON GRFs
This approach proposed by Davenport (1979) is now part of most line design codes,
especially ASCE (1991, 2006) and NESC (2007.) GRFs for 1-hour mean and other
sampling intervals have been developed for approximately equivalent wind speeds
and are included in Table 1. The comparison is made for a 125ft (38m) tall structure
in open country terrain, with conductors and tower effective heights of 115ft (35m)
and 85ft (26m), respectively. The tower GRF includes a 1% damping coefficient and
a critical frequency of 2 Hz for Rt.
Table 1: GRFs for different sampling intervals
Sampling
Interval
Equivalent Wind
Speed (mph)
GRFc
Peak Response
Conductor (psf)
GRFt
Peak Response
Tower (psf)
1-hour
Davenport (1979)
Fastest Mile
ASCE (1991)
3-second
ASCE (2006)
60.0
75.0
90.0
1.34
0.80
0.56
17.8
16.6
15.6
1.97
1.20
0.87
24.2
22.8
22.1
These results show that the calculated wind pressures are consistent for all the
sampling intervals. The conductor GRF for each reference velocity has a mean plus
dynamic proportion of approximately (1+0.3), and the tower GRF has a proportion of
approximately (1+1.0), thus indicating that the fluctuating wind adds a load effect
equivalent to 30% of the mean wind to the conductors, and doubles the mean wind
effect on the tower.
The square root of the conductor GRF can also be interpreted as a wind gust effect
factor that adjusts the reference velocity of each sampling interval to an equivalent 2to 5-minute wind, thus slow enough to affect the entire span length. Similarly, the
square root of the structure GRF can be interpreted as a gust factor that adjusts the
reference velocity to approximately an equivalent 10-second wind gust that is large
enough to envelope a complete structure.
There has been a tendency to disregard the tower dynamic response in design, as
lower voltage structures would have little resonant response. The use of a 3-second
gust dynamic pressure without any reduction or amplification, with approximately a
10-second equivalent gust response as discussed above, would be sufficient for these
shorter towers. However, the structure dynamic response factor Rt should not be
systematically ignored for taller towers, as in very high voltage lines or river
crossings.
Although comprehensive, the gust response factor approach contains several
assumptions and simplifications. For example the statistical factor g for 1-hour mean
winds in Equations (3) and (4) is in the range of 3.5 to 4.0, or a 15% variation. The
load coincidence factor is estimated at a range of 0.7 to 1.0, depending on the
relative span to structure contribution to the load effect, thus with a 40% variation.
For the application of the GRF approach to different sampling intervals the GRFs
have to be adjusted by the ratio of velocities kV squared. Normally this is done using
the conversion curve proposed by Durst (1960), who based his analysis on wind
speeds of no more than 40-50mph. Sissenwine et.al. (1972) has shown that the
velocity ratio, or gust factor, kV is not the same for low and for high wind speeds.
Further, the authors note that the GRF equations in ASCE (1991, 2006) and NESC
(2007) estimate the kV gust factor with respect to a 10-minute reference velocity, and
not 1-hour mean as in the original GRF development by Davenport (1979.) This
systematically introduces a 15% difference in the GRF calculations using these codes.
The load variations due to the above discrepancies may have a minor impact on the
calculation of wire or structure loads. But these differences are important if what is
needed is an estimation of the reliability of the system. As we know, the doubling of
the return period for winds with a COV of 15-20% adds approximately 15% to the
calculated loads. Even if a reference 50-year wind speed could be established with
reasonable approximation for a point in a line, the introduction of any of the variances
described above, or their combination, would result in an error in the reliability
213
estimation of at least one order of magnitude, which invalidates any reliability based
theory for transmission line design. Changes in line reliability can still be calculated,
but only relative to a reference level, as suggested by Behncke et.al. (1994.)
GUST AND SPAN FACTORS
A different approach that has been used in the past for the determination of wind
loads on line components consists of two separate factors, namely a gust factor that
addresses the time variations of the wind, and a span, or gust extent factor that
accounts for the space variations of gusts. In this case the conductor dynamic
response can be omitted.
The gust factor is calculated as the ratio of the gust wind speed to the mean wind
speed, usually 5-second to fastest mile or 1-hour wind. The gust factor is then
adjusted for height above ground, as the wind turbulence decreases with height. Using
the same kV velocity ratio based on Durst (1960), the following expression for the
reference gust factor is obtained for fastest mile records, VFM, after EPRI (1987.)
GFo = 1.13+5.5/VFM
The total multiplier to calculate the design peak velocity at the effective height zh is as
follows:
GFh = GFo (z0 /zh )1/
where z0 and zh are the reference and effective heights respectively, is the power
load exponent as a function of ground surface roughness, and the total design gust
wind is thus (VFM GFh ). This concept is used in a report by IEEE (1977), where a
reference gust factor GFo of 1.3 is suggested.
The span factor is defined as the ratio of the effective gust wind load on the wires to
the fully correlated gust wind load on the entire span. While the background
component of the wire load represents the reduction due to spatial correlation, the
relationship between the span factor and the gust response factor can be considered as
an equivalent span factor, as follows:
Equivalent Span Factor = (1+2gIuBc )/(1+2gIu ) (5)
where the denominator is the fully correlated loading relative to the mean velocity,
and Iu is the turbulence intensity. This span factor equation is the same as that
proposed by Holmes (2008) except the load coincidence factor . Holmes considered
the spatial correlation in the wind, but omitted the non-coincident occurrence of the
gust effects of the wind loads from the tower and the wires. It can be shown that, for
the case when the response ratio is in the order 0.1, i.e. the load effect from the tower
load is 10% of the load effect from the wire load, the coincidence factor for a typical
span length can be as large as 0.8. The omission of the load coincidence factor may
lead to a conservative specification by about 10%. A value of 0.75 is recommended
for the load coincidence factor .
214
Holmes (2001, 2008) proposed the following expression for synoptic wind span
reduction factor (SRF) valid both for open country and semi-urban ground roughness,
in which L is the span length in ft:
SRF = 0.59+0.41 e -(L/690)
(6)
Figure 1 shows the comparison of the span factor SRF for synoptic winds from
Equation (6) as proposed by Holmes (2008) and that implied by the specification of
the conductor GRF, calculated from Equation (5) for open country terrain and
conductor effective height of 115ft (35m.) It can be seen that they are equivalent
quantities.
215
(7)
Following the basic formulation of Davenports GRF described in this paper, the
effect of the high spatial correlation in a downburst can be represented by a large
integral length scale LS, which approximates the size of the correlated wind gust.
Figure 2 shows a comparison of Holmes suggested span factor in Equation (7) with
the equivalent span factor derived from the GRF formulation in Equation (5), with an
integral length scale LS=1,500ft and for a conservative load coincidence factor =1.0.
The higher coincidence factor recognizes the more dominant effect of the wires due
to the higher spatial correlation in downdraft winds.
216
Behncke and White (1984), Behncke et.al. (1994) and Ishac and White (1995) among
others have studied the effects of strong tornado winds on transmission lines. It was
observed that the windward masts of guyed structures were invariably overwhelmed
by bending and P- effects generated by the increased wind pressure. In selfsupporting structures the primary failure due to a tornado wind occurs in the lateral
bracings of the structure, which then induce the collapse of the main chords as a
secondary effect. As tornado winds are highly localized, extreme wind speeds along
the tornado path are often accompanied with little or no wind a short distance away,
thus suggesting that the failure is related to a more or less direct impact on the tower,
as the load effect on wires would not be significant.
Based on the statistics mentioned above it was initially proposed that a load
equivalent to F2 wind in the F-scale be applied on the towers alone, with both the
velocity pressure exposure coefficient and the structure gust factor equal to 1.0, which
reflected the lack of knowledge on the vertical profile of HIWs. The F2 wind
intensity was not based on any risk or return period calculation, but solely on the
basis that it represents the majority of extreme winds expected to intercept a line, it
requires small additions of strength to the structures, and the modeled behavior of
towers under the full range of F2 load as a minimum closely resembles the observed
field effects.
Recognizing that even a narrow-front tornado would have some effect on the wires,
Behncke and White (2006) subsequently proposed a span reduction factor for tornado
wind, given by the following expression:
SRF = WG (1 - 0.25WG / L)/L
where L is the average span length and WG is the tornado gust width. Figure 3 shows
the proposed span ratio for tornadoes, assuming a 300ft gust front.
217
apply since it assumes that the entire span would receive mean and background
dynamic wind loading.
CONCLUSION
The development and evolution of span and gust factors for transmission line design
has been examined. The initial approach of two separate factors, namely a wind
velocity ratio and a span reduction factor applied to conductors, has been replaced in
design codes by a single tower peak response concept that takes into account both the
temporal and spatial variations of wind gusts, as well as the dynamic responses of the
line components at their characteristic frequencies and according to their inertial
properties.
These factors are applicable to the near-the-ground turbulence generated by synoptic
winds, in the order of 90mph (40m/s) in the U.S., which are well documented from
records at typical weather stations. Although this approach provides a good basis for
line component design, experiences with long transmission lines in many countries
have demonstrated that structures usually fail due to non-synoptic winds. For example,
a CIGR (1996) survey shows that nearly 65% of the reported line failures had been
triggered by high intensity winds in the form of downdrafts or tornadoes generated by
local atmospheric thermal instability.
Although the wind speed in convective downdrafts is in the range of F0 or F1 in the
F-scale, their large spatial correlation over the typical length of line spans results in
greater loads on the conductors. Results from field investigations on downdrafts show
that span reduction factors for these storms should be between 40 and 50% greater
than those for synoptic winds, as shown in Figures 1 and 2. In view of the
uncertainties related to proper identification and analysis of thunderstorm winds in
the meteorological data, it would be also prudent to take a more conservative
approach in specifying the corresponding gust response factor for the conductors.
In the case of tornadoes, experience shows that failures caused by direct wind impact
on the tower, excluding projectiles, can be reasonably modeled by the application of a
narrow-front wind with intensity F2 in the F-scale, as this intensity represents the
majority of observed tornadoes. The proposed smaller span factor in Figure 3
corresponds to the reduced gust effect on the conductors. The designer should select
the maximum wind velocity within the F2 range based on the local indications of
tornado occurrence.
In considering the design wind loads for transmission line systems, wind loads due to
different types of storms can be considered independently. Although the general
specification of the design wind speeds inherently includes the convective wind
storms such as thunderstorms and downbursts, in many of the areas the synoptic
winds would be the dominant component in these combined wind speeds. Applying
the higher GRF or span factor for thunderstorm or downburst winds with these wind
speeds would be overly conservative.
218
An appropriate design procedure may include the use of the specified combined
(synoptic and thunderstorm) wind speed with the GRF or span factor for synoptic
winds and then perform a design check using the higher GRF or span factor with a
separately specified thunderstorm wind map that needs to be developed. For tornado
winds, the design wind speed is dependent on the probability of occurrence per unit
area. Design wind maps should be developed based on this probability as well as
the overall length of the transmission line system under consideration.
REFERENCES
Armitt, J., Cojan, M., Manuzio, C. and Nicolini, P. (1975) Calculation of Wind
Loadings on Components of Overhead Lines, Proc. IEE, Vol. 122, No 11.
ASCE (1991) Guidelines for Electrical Transmission Line Structural Loading, ASCE
Manuals and Reports on Engineering Practice No. 74, American Society of Civil
Engineers, New York, 139 pp.
ASCE (2006) Guidelines for Electrical Transmission Line Structural Loading, ASCE
Manual 74, Peer Review Draft, ASCE Manual 74 Committee.
Banik, S.S., Hong, H.P. and Kopp, G.A. (2008) Assessment of tornado hazard for
spatially distributed systems in southern Ontario, Journal of Wind Engineering
and Industrial Aerodynamics, 96. pp 1376-1389.
Behncke, R.H. and White, H.B. (1984) The Alicur 500kV Transmission System,
Paper 22-02, CIGR International Session, Paris.
Behncke, R.H., Milford, R.V and White, H.B. (1994) High Intensity Wind and
Relative Reliability Based Loads for Transmission Line Design, Paper 22-205,
CIGR International Session, Paris.
Behncke, R.H. and White, H. B. (2006) Applying Gust Loadings to Your Lines,
Proceedings of the 9th International Conference on Overhead Lines. Fort Collins,
Colorado.
BSI (1986) BS 8100-1 Lattice Towers and Masts Part 1: Code of Practice for
Loading, British Standards Institution.
Castanheta, M.N. (1970) Dynamic behaviour of overhead power lines subject to the
action of the wind, Paper 22-08, CIGR International Session, Paris.
CIGR (1996) Review of IEC 826: Loading and Strength of Overhead Lines. Part 3:
Analysis of Recent Transmission Line Failures, SC-22 WG22-06.
Davenport, A.G. (1960) Rationale for determining design wind velocities, Journal of
the Structural Division, Proceedings of the American Society of Civil Engineers,
Vol. 86, ST5, pp 39-68.
Davenport, A.G. (1961) The spectrum of horizontal gustiness near the ground in high
winds, Quarterly Journal of the Royal Meteorological Society, Vol. 87, No. 372,
pp194-211.
Davenport, A.G. (1967) Gust loading factors, Journal of the Structural Division.
Proceedings of the American Society of Civil Engineers, Vol. 93, ST3, pp 11-34.
Davenport, A.G. (1969) The application of statistical concepts to the wind loading of
structures, Proc. Inst. Civil Engrs. 19, pp. 449-472.
219
Davenport, A.G. (1979) Gust response factors for transmission line loading, Wind
Engineering, Proceedings of the 5th International Conference on Wind
Engineering. Colorado State University, Ed. J.E. Cermak. Pergamon Press.
Durst, C.S. (1960) Wind Speeds over Short Periods of Time, Meteorological
Magazine, Vol. 89, pp. 181-186.
EPRI (1987) Reliability Based Design of Transmission Line Structures: Final Report,
Volume I: Methods, Electric Power Research Institute, Project 1352-2.
Fujita, T.T. and Pearson, A.D. (1973) Results of FPP Classification of 1971 and 1972
Tornadoes, 8th Conference of Severe Local Storms, American Meteorological
Society, Boston, Mass. pp. 142-145.
Holmes, J.D. (2001) Wind Loading of Structures, Spon Press, London, U.K.
Holmes, J.D. (2008) Recent Developments in the Specification of Wind Loads on
Transmission Lines, Journal of Wind and Engineering of the Indian Society of
Wind Engineering, Vol. 5 No.1.
Holmes, J.D., Hangan, H.M., Schroeder, J.L. and Letchford, C.W. and Orwig, K.D.
(2008) A forensic study of the Lubbock-Reese downdraft of 2002, Wind and
Structures, Vol. 11, No. 2, pp 137-152.
IEC (2003) Design Criteria for Overhead Transmission Lines, International
Standard IEC 60826, International Electrotechnical Commission, 3rd Edition
2003-10.
IEEE (1977) Loading and Strength of Transmission Line Systems, Part 2: Designing
for Wind Loads, Task Group of Line Loading and Strength of Transmission Line
Structures, Paper PES A 77 229-8.
Ishac, M. and White, H.B. (1995) Effect of Tornado Loads on Transmission Lines,
IEEE, Transactions on Power Delivery.
Manuzio, C. and Paris, L. (1964). Statistical determination of wind load effects on
overhead line conductors. CIGRE, Paper 231.
NESC (2007) National Electrical Safety Code, National Bureau of Standards,
Washington, D.C.
Sissenwine, N., Tattleman, P., Granthan, D.D. and Gringorten, I.I. (1972) Extreme
Wind Speeds, Gustiness and Variations with Height, Air Force Cambridge
Research Laboratories, Technical Report 73-0560, Aeronomic Laboratory, Project
8624.
220
221
222
223
Response Factor, discussed below, takes into account the structures response to the
dynamic characteristic of the wind. Gust Factors have values greater than 1.0.
The Gust Response Factor (GRF ) is the ratio of the peak gust load effect on
the structure or cables to the mean load effect corresponding to the mean wind speed.
The gust response factor accounts for the additional load effects due to wind
turbulence and dynamic amplification of flexible structures and cables. The gust
response factor represents the cumulative effect or integration of the gusts and lulls of
the wind over the range of span lengths of typical transmission lines, as well as the
effect of the wind on the supporting structures. The IEC, ASCE 74, and NESC
provide equations for determining the Structure and Conductor gust response factors.
The IEC refers to this factor as the Combine Wind Factor. ASCE-7 calls this
parameter the Gust Effect Factor. The Gust Response Factors have values that are
typically 1.0 or less, depending on the basic wind averaging period.
It should be noted that the recommended Gust Response Factor in ASCE 74 is
modified from that presented in Davenport, 1979. The original Davenport equation
accounted for the structure/conductor background and dynamic wind response.
The recommended ASCE 74 Gust Response Factor does not account for the dynamic
component. It was decided by the original ASCE 74 committee that the dynamic
component did not represent the actual behavior of a transmission system to applied
winds.
The Velocity Exposure Coefficient (kz) accounts for the increase wind with
height, including the effects of the terrain category. The basic wind speed values are
typically measured at 33 ft [10 m] above the ground. The wind speed will increase
with height up to the gradient wind height. The gradient height is where the wind
speed is assumed to be constant above the ground surface. The variation of wind
speed with height is caused by ground friction that varies with ground roughness. The
height of the gradient wind varies with terrain category. An exception to increased
wind speed with height has been noted during certain high intensity wind (HIW)
conditions.
Two equations are used to represent the wind distribution with height: Power
Law [V(z) = Vh (z / zh)1/] and the Logarithmic Law [V(z) = (1/k) V* ln(z /z0)].
Where: V(z) = Velocity at height z above the ground, Vh = Measured wind speed at a
reference height zh, zh = Measured reference height, = Power Law Coefficient, V* =
Shear velocity or friction velocity = (0 / ), 0 = stress of the wind at ground level,
= air density, k = von Karman constant, approx. 0.4, and z0 = roughness parameter of
the ground. The parameters used in these equations are determined by the terrain
exposure categories. ASCE 74, NESC, and IEC use the power law equation to
increase the wind speed with height (see Figure 2a).
Wind effects are influenced by the terrain roughness. The greater the
roughness, the more turbulent and slower the wind is at lower ground levels. There
are four basic terrain exposure categories shown in Table 2. It should be noted that
ASCE 74 and IEC 60826 reverse the categories. The NESC only uses one category,
C. Wind characteristic parameters for the different terrain exposure categories are
listed in Table 3. It should be noted that the used in the Power Law changes
based on the averaging period of the basic wind speed.
224
225
(a)
(b)
(c)
226
Number (Re), have significant variations, Cd ranges from 0.7 to 1.4, between Re
values of 3(10)3 to 3(10)5. The conductor drag is typically assumed to be 1.0.
227
along the spans. The Span Factor is applied to the wire wind force. The wind speed
varies not only with time and distance above ground level, but also horizontally over
a front normal to the wind flow, Figure 4. The actual wind pattern is of a turbulent
nature where the wind speed is non-uniformly distributed over the transmission line
span. It is generally accepted that in most instances a wind gust will not impact more
than a fraction of the transmission line span.
228
guidelines may provide a wind load adjustment Spatial Factor to account for the
vulnerability of a transmission line based transmission line length.
COMPARISON OF ASCE 74, NESC, AND IEC 60826
The Canadian Electric Association Technology International (CEATI), Inc.,
Montreal, Canada, completed a study of the comparison of the ASCE 74 (2005
Draft), IEC 60826 (2003), and NESC (Rule 250C, 2002) wind load methods. The
study, Report T053700-3324 was commissioned by the CEATI Overhead Design
Issues and Wind & Ice Storm Mitigation Interest Group (WISMIG). The principal
investigator was Mr. Elias Ghannoum. The report also included a comparison with
the European Standard CENELEC EN 50341-1-2001 and CSA C22.3 No. 1-01, but
these results are not presented herein.
The purpose of this study was to identify the similarities and difference
between the wind load methods. Key wind load parameters, assumptions, and
limitation were presented. For comparison of the calculated wind loads a simple
structure with varying parameters was used. The tower heights used were 82, 98, 115,
131, and 164 ft. The unique heights listed are the result of the conversion from meters
to feet. The report used SI units. The towers were assumed to be lattice structures
with an average solidity ratio of 0.40, and the conductors at one level, flat
configuration.
In general, the ASCE 74 and NESC methods are similar, whereas the IEC
60826 and CSA C22.3 methods are similar. For the purposes of this paper,
comparisons between ASCE 74, NESC, and IEC 60826 for normal wind to the tower
and conductors will be discussed.
Some of the major differences between the ASCE, NESC, and IEC documents
are presented. The first is the basic wind speed averaging period. ASCE uses the 3second gust wind speed. The IEC uses the 10-minute average wind speed. ASCE uses
equations to determine the structure and conductor Gust Response Factors, where IEC
uses a graph to determine the Combine Wind Factors (Gust Response Factors). Both
ASCE and IEC use the 50-year return basic wind speed as the design level. Other
recommended return periods are different in these two documents. ASCE used 50,
100, 200, and 400 year return periods, whereas IEC uses 50, 150, 500 year return
periods. Both these documents use the power law to vary wind speed with height. The
wind parameters use in the power law equations are different because of the different
wind speed averaging periods used by the two documents. The IEC has a separate
factor to account for the Span Factor (GL). The author has been told that the
Davenport Gust Response Factor accounts for the span factor affect. There are some
other differences in the determination and application of the wind loads and these will
not be discussed here. The following comparisons were made using a terrain category
of C (ASCE/NESC) and equivalent B (IEC). The basic wind speed was converted to
be equivalent for the method used.
Table 6 show the comparison of the effective wind pressure (psf) [N/m2] on
conductors. This is the pressure that includes all the adjustment factors due to height,
gust response factor, span factor, etc. These values show that ASCE/NESC values are
20% less than that obtained from the IEC method.
229
230
Table 8, Relative variation of wind pressure with span length (Const. Ht = 98.4 ft)
Span (ft)
IEC (psf)
IEC
ASCE/NESC (psf) ASCE/NESC
656
25.9
100%
20.8
100%
820
25.6
99%
20.3
98%
984
25.2
97%
19.9
96%
1148
24.8
96%
19.6
94%
1312
24.2
94%
19.3
93%
1476
24.0
93%
19.1
92%
1640
23.6
91%
18.9
91%
[1 ft = 0.3048 m, 1 psf = 47.9 N/m2]
FUTURE NEEDS FOR TRANSMISSION LINE WIND LOAD METHODS
This author believes that the advancements in wind load methodology applied
to transmission line systems are necessary so that design engineers understands the
characteristic parameters and limitations used to determine tower and conductor wind
forces. This knowledge allows the engineer to make decisions on wind loads that
directly affect the reliability of the transmission line. The advanced information is
also important when an engineer is conducting a failure investigation of a tower after
an extreme event wind storm. Also, the utility engineer can use the advanced methods
to calibrate a simpler in-house procedure. Site specific evaluation of a transmission
line tower requires the information from an advanced method to determine the
adequacy of the structure.
The difference between the IEC and ASCE 74 are more readily acceptable
when detailed information is presented on their development. The design engineer
should remember that the wind characteristic parameters and extreme wind speeds are
not exact representations of what the actual transmission line may experience.
Utilities need to decide the complexity level of the wind load methodology that
provides adequate transmission line performance. It should be remembered that short
term performance, such as less than 50 years experience, by its self, does not justify
acceptable performance. The advanced wind load methodologies will help the Utility
engineer to make the best decisions with respect to the appropriate wind load method,
be it a simple method or a more complex method.
The IEC and ASCE 74 committees should work together to present a common
and consistent wind load methodology. This would benefit the US and World
transmission design communities. The new research in HIW will soon provide an
additional wind methodology that may control in some regions future tower designs.
The transmission line engineer should be fully engaged in this research effort. Both
CIGRE (www.cigre.org) and CEATI (www.ceati.com) are active in wind load
research.
REFERENCES:
ANSI/TIA 222 G (2005), Structural Standard for Antenna Supporting Structures and
Antennas, Telecommunication Industry Association, Arlington, VA
231
232
233
Jamaleddine et al. 1993; Klmn et al. 2007; Roshan Fekr 1995; Roshan Fekr and
McClure 1996; Roshan Fekr and McClure 1998; Roshan Fekr et al. 1998).
Oertli, Freitag, Morgan and Swift and PTI all assumed that the worst case jump
height occurred with one loaded span surrounded by bare spans. Jamaleddines
model tests showed that this is not the case. In his two span model, the jump height
was higher when the neighboring span was loaded.
Roshan-Fekrs study using the ADINA software showed that changes in elevation
(within the bounds of his study) had a minor impact on the jump height of the cables.
Ice shedding in the center 1/3 to of the span is more important than ice shedding
from near the supports. None of the tests and studies addressed the question of the
affect of multiple loaded spans adjacent to the span shedding.
Greisser, Oertli and Jakse all indicated that in areas of heavy icing, double circuit
structures need, not only horizontal offsets between the middle phase and the top and
bottom phases, but also horizontal offsets between the top and bottom phases.
For tangent sections of line with no wind, the cables jump almost completely
vertically with practically no transverse motion. This was observed and reported for
the tests performed by Oertli, Morgan and Swift, and PTI. Oertlis tests showed that
substantial lateral movement can result when ice sheds from spans ending in I-string
running angle insulators. Oertlis tests also showed that wind may be a factor in
causing substantial transverse motion when a cable sheds ice.
The tests with suspension insulators all were carried out on lattice steel towers which
are very stiff compared to tubular steel structures. The flexibility of the supports,
particularly in the transverse direction, had not been adequately studied to date.
The tests and simulations noted above were typically made on lower voltage lines
with relatively stiff lattice steel towers. This study extended the range of previous
studies to a 500 kV line using relatively flexible tubular steel structures.
There are relatively few reported field observations of ice shedding. Several of the
papers reviewed here have noted the tendency of ice to slide towards the low point of
sag and accumulate thereeven to the extent of damaging suspension clamps in sidehill spans. Laforte et al. observed this sliding assisting in the removal of ice by
collisions between chunks of ice. The significant patterns studied have included full
spans shedding, and ice concentrated near the low point of sag shedding.
Finite Element Models
Both 2D and 3D models were used. The 2D models were used to explore the effect
of the span length and number of loaded spans on the jump of the conductor after ice
shedding. They were also used to examine the effect of various patterns of ice
loading on the height of jump.
234
The 2D models used single Bluebird conductor. In the 3D model, the 2-bundle
Bluebird ACSR conductor was modeled as a single conductor with the mass,
stiffness, loads and damping equivalent to the twin bundle. This assumption reduces
the number of nodes and elements by approximately one-half which reduces the
amount of time to perform the analysis. The only limitation imposed by using an
equivalent single conductor is that shedding from only one of the two sub-conductors
cannot be modeled. The 3D model also used inch EHS shield wire. Catalog
properties were used with moduli of elasticity of 9,020,000 psi and 25,000,000 psi,
for the conductor and shield wire respectively.
The wires were divided into pin-connected truss elements 10 to 15 ft long. The
conductor and shield wire were modeled using horizontal tensions at 32 F final
tension including creep. Table 1 shows the horizontal tensions.
Table 1 Conductor and Shield Wire Horizontal Tensions
Ruling Span
900
1200
1500
Conductor
(lb)
11,390
11,716
11,934
Shield Wire
(lb)
3,642
The internal damping of the wires was modeled as the axial viscous damping of a rod.
The critical axial viscous damping for a rod is given in the following equation where
ccr is the critical damping constant, A the area of the rod, E the modulus of elasticity
and m the unit mass. Axial damping of 0.5% of critical was used.
c cr = 2 A E m
Aerodynamic damping is due to the motion of the conductor relative to the air. In
still air, the aerodynamic damping force is given by Equation 2 (Dyrbye and Hansen
1996, p. 76).
Fd =
1
Vr2 C d A p
2
Where Fd is the damping force, is the air density and Vr is the velocity relative to
the air, Cd is the drag coefficient and Ap is the projected area. Cd depends on the
Reynolds number (following equation) where dw is the conductor diameter and is
the viscosity of air.
Re =
Vr d w
235
For smooth circular cylinders, Cd is between approximately 0.9 and 1.2 for Reynolds
numbers between 200 and 100,000 (Binder 1973; Eisner 1931). As the relative
velocity is reduced below a Reynolds number of 200, Cd steadily increases to over 50
at a Reynolds number of 0.1. A Cd of 1.25 was used for the calculations as relatively
low velocities were expected.
Fig. 1 shows a 2D conductor model. When 3D models are used, additional horizontal
dampers are added at each node to damp the transverse motion of the conductor.
With no wind, the aerodynamic dampers are identical in placement and number as the
vertical dampers.
236
237
Component
Pole
Shield Wire Arms
Upper and Middle
Xarms
Lower Xarm
Base Diameter
(in Flat to Flat)
69
12
End
Diameter
(in Flat to
Flat)
24
6
Length
(ft)
180
12
No of
Flats
12
6
Wall
Thickness
(in)
Table 3-1
0.250
30
0.375
22
12
32
0.375
22
12
Wall
Thickness
(in)
.4375
.3750
.2500
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
238
Figures 4(a) and (b) represent the last span loaded in a section shedding its ice from
the full span. Figures 4(c) and (d) show a span in the center of a fully loaded section
that sheds all of its ice at once. Figures 4(e) and (f) show only the 500 ft in the center
of a span shedding its ice in a section with all the spans fully loaded. Figures 4(g) and
(h) show the center 500 ft shedding its ice in an otherwise fully loaded section. This
would typically take place after the center of the span has already shed its ice and the
ice in the sides of the span slides down to reload the center of the span.
Figure 5 compares the effect of the different loading patterns shown in Figure 4. All
the curves are based on the center span of nine 1500 ft spans with an ice load of 1.0
radial glaze ice equivalent.
Table 4 summarizes the results of the analyses of the 2D nine span line sections
simulated. The middle 500 ft of the center span was loaded for all of the partial
lower span loads. The upper span sag is the ruling span sag for the span and radial
ice thickness listed. The vertical gap and overlaps are based on a phase spacing of
37.5 ft.
239
240
Span
(ft)
900
1,200
1,500
900
1,200
1,500
900
1,200
1,500
900
1,200
1,500
Radial
Glaze
Equiv.
(in)
0.50
0.50
0.50
1.00
1.00
1.00
0.50
0.50
0.50
1.00
1.00
1.00
Lower
Span
Loading
Partial
Partial
Partial
Partial
Partial
Partial
Full
Full
Full
Full
Full
Full
Upper
Phase
Sag
(ft)
24.3
40.9
61.8
26.6
43.8
65.1
24.3
40.9
61.8
26.6
43.8
65.1
Lower
Phase
Starting
Sag
(ft)
22.8
36.9
54.5
24.0
36.7
52.1
24.3
40.9
61.8
26.6
43.8
65.1
Lower
Phase
Jump
Sag
(ft)
10.0
19.2
32.4
0.7
4.9
11.9
8.7
15.4
25.1
-1.6
0.0
1.3
Jump
(ft)
12.8
17.7
22.1
23.3
31.8
40.2
15.6
25.5
36.7
28.2
43.8
63.8
Vertical
Gap
(ft)
23.2
15.8
8.1
11.6
Overlap
(ft)
1.4
15.7
21.9
12.0
0.8
9.3
6.3
26.3
120
100
80
60
40
10
15
20
25
30
35
40
Time (s)
Top Phase
Bottom Phase
(a)
(b)
Fig. 6 3D Model Results
Figure 6b compares the time histories of the vertical positions of the center points of
the top and bottom phases. With this long span, there is substantial overlap
continuing for the entire 30 seconds simulated.
241
Fig. 7a compares the time histories of the horizontal positions of the center of the
span at the mid point between the two subconductors. Note that the horizontal
displacements of both phases are increasing at the end of the 30 second simulation.
The transverse motion of the conductor is caused by the motion of the tubular steel
structures when the ice sheds. Before the ice sheds, the structures deflect towards the
side with two phases due to their being balanced by only one phase on the other side
of the structure. When the ice sheds, the bending load is reduced and the structures
rebound swaying back the other way inducing transverse vibrations in the structure
which are coupled to the conductors through the insulator strings.
Fig. 7b is the difference between the time histories shown in Fig. 7a. In addition, the
difference has been reduced by 1.5 ft to take into account an18 inch bundle spacing.
This figure shows that the sideways motion causes the horizontal paths to also
overlap slightly based on the 2.4 ft static offset of the structures modeled.
2
20
19
18
17
16
10
15
20
25
Time (s)
30
35
40
21
1.5
0.5
0.5
CL Top Bundle
CL Bottom Bundle
10
15
20
25
30
35
40
Time (s)
(a)
(b)
assumptions must be made. In this paper, the assumed ice distributions have been
reduced to a few simple patterns. The lines studied have also been simplified to
sections with even spans and wire attachment elevations. The results of the study
cannot provide definitive guidance, but mainly serve to enhance our engineering
judgment when designing lines in icing areas.
When ice and snow begin falling from a line, distortions of the wire system will
occur. When a span is more heavily loaded than the spans around it, its sag increases
(see Fig. 5 curve a) in comparison with the calculated sag (based on evenly loaded
line sections). Modern line design software has some capability for determining the
static position of the wires for different assumed loading patterns while also
considering the movement of the insulators and supporting structures. Large
distortions of the wire system have been observed in Alaska after snow has shed
partially or fully from some spans while others remained more heavily loaded. In
areas with very localized in-cloud icing, similar distortions can be expected when
spans loaded with rime are adjacent to spans that are not. When ice sheds suddenly
from a lower phase, there may already be an existing imbalance in ice loads that has
reduced the phase to phase vertical clearance.
When designing a new line, areas where icing can be expected should be identified
and the type and amount of ice estimated. A static analysis considering uneven
loading, insulator movement and structure deflection should be considered when
phases are placed one above the other.
Structures are typically designed with clearances for switching surges in everyday
conditions and during moderate winds. An additional criterion may be added to
maintain 60 Hz clearances during extreme winds. Similarly to extreme wind events,
60 Hz clearances should be maintained during ice shedding to avoid flashovers.
The studies and analyses described in this report show that the risk of flashover due to
vertical jump during ice shedding is relatively small for phases directly above one
another (37.5 ft spacing) for spans under 900 ft. The risk of flashover increases both
with span length and with the amount and pattern of ice. For spans longer than 900 to
1000 ft, offsetting the adjacent phases horizontally is recommended when rigid
towers, e.g. wide base, 4-legged lattice steel towers, are used. If flexible structures are
used, e.g. tubular steel pole steel structures, the horizontal offset should be increased
to account for the lateral motion of the conductor due to the motion of the tower.
Acknowledgements
The authors are grateful for Southern California Edisons support of this work.
References
Binder, R. C. (1973). Fluid Mechanics, Fifth Edition, Prentice-Hall, Inc., Englwood
Cliffs, NJ.
242
Dyrbye, C., and Hansen, S. O. (1996). Wind Loads on Structures, John Wiley &
Sons, Ltd, Chichester, GB.
Eisner, F. (1931). Das Widerstandsproblem. In Proceedings of the 3rd International
Congress for Applied Mechanics, Stockholm 24-29 August 1930, edited by C.
W. Oseen and W. Weibull. AB. Sveriges Litografiska Tryckerier. Stockholm.
23-42.
Freitag, A., and Brandt, E. (1981). "Dynamische Beanspruchungen von
Mittelspannungs-freileitungen beim Abwurf von Eislasten (Dynamic Loads
During Release of Ice Loads from Medium Voltage Overhead Lines) in
German." Elektrizittswirtschaft, 80(19), 668-676.
Fuheng, S., and Shixiong, J. (1988) "Icing on Overhead Transmission Lines in Cold
Mountainous District of Southwest China and its Protection." Proceedings of
the Fourth International Workshop on the Atmospheric Icing of Structures,
Paris, France, 354-357.
Greisser, V. H. (1913). "Effects of Ice Loading on Transmission Lines." Transactions
of the AIEE?? Check(June 27, 1913), 1829-1844.
Jakse, J., Al Harash, M., and McClure, G. (2001) "Numerical modelling of snowshedding effects on a 110 kV overhead power line in Slovenia." ISOPE-2001:
Eleventh (2001) International Offshore and Polar Engineering Conference,
Stavenger; Norway, 690-694.
Jamaleddine, A. (1994). Effets Statiques et Dynamiques du Dlestage de la Glace sur
les Lignes Ariennes de Transport D'nergie lectrique (Static and Dynamic
Effects of Ice Shedding from Overhead Electric Transmission Lines) in
French, PhD, Gnie Mecanique (Mechanical Engineering), Universit de
Montral, cole Polytechnique, Montreal.
Jamaleddine, A., Beauchemin, R., Rousselet, J., and McClure, G. (1996) "WeightDropping Simulation of Ice-Shedding effects on an Overhead Transmission
Line Model." Proceedings of Seventh International Workshop on Atmospheric
Icing of Structures, SaguenayLac-Saint-Jean, Qubec, Canada, 44-48.
Jamaleddine, A., McClure, G., Rousselet, J., and Beauchemin, R. (1993). "Simulation
of Ice-shedding on Electrical Transmission Lines Using ADINA." Computers
and Structures, 47(4, 5), 523-536.
Klmn, T., M., F., and McClure, G. (2007). "Numerical analysis of the dynamic
effects of shock-load-induced ice shedding on overhead ground wires." Computers
and Structures, 85, 375-384.
McClure, G. (1989). Interactions Dynamiques dans les Lignes Ariennes de
Transport d'lectricit Soumises aux Bris de Cbles (Dynamic Interactions in
Overhead Electrical Lines from Broken Wires, in French), Ph.D. Thesis,
Dpartment de Gnie Civil (Department of Civil Engineering), Ecole
Polytechnique de Montral, Montreal.
McClure, G., and Lapointe, M. (2003). "Modeling the Structural Dynamic Response
of Overhead Transmission Lines." Computers and Structures, 81(8-11), 825834.
McClure, G., and Tinawi, R. (1987). "Mathematical Modeling of the Transient
Response of Electric Transmission Lines Due to Conductor Breakage."
Computers and Structures, 26(12), 41-56.
243
McClure, G., and Tinawi, R. (1989). "Comportement dynamique des lignes ariennes
de transport d'lectricit d aux bris de cbles. II. Problmes numriques
associs la modlisation mathmatique. (Dynamic Behavior of Overhead
Electric Transmission Lines Due to Broken Cables. II Numerical Problems
Associated with the Mathematical Modeling)." Canadian Journal of Civil
Engineering, 16(3), 354-374.
Morgan, V. T., and Swift, D. A. (1964). "Jump Heigh of Overhead-line Conductors
after the Sudden Release of Ice Loads." Proceedings of the Institution of
Electrical Engineers, 111(10), 1736-1746.
Oertli, H. (1950). "Oscillations de Cbles lectriques Ariens aprs la Chute de
Surcharges (Vibration of Overhead Electric Cables after Dropping Surcharge
Loads, in French)." Bulletin de L'Association Suisse des Electriciens (Bulletin
of the Association of Swiss Electrical Engineers), 41(13), 501-511.
Peabody, A. B. (2004). Applying Shock Damping to the Problem of Transmission
Line Cascades, Ph.D. Thesis, Dept. of Civil Engineering and Applied
Mechanics, McGill University, Montreal.
Power Technologies, I. (1978). Transmission Line Reference Book 115-138 kV
Compact Line Design, Electric Power Research Institute, Palo Alto.
Roshan Fekr, M. (1995). Dynamic Response of Overhead Transmission Lines to Ice
Shedding, Master of Engineering, Dept. of Civil Engineering and Applied
Mechanics, McGill University, Montreal.
Roshan Fekr, M., and McClure, G. (1996) "Numerical Modelling of the Dynamic
Response of Ice Shedding on Electrical Transmission Lines." Proceedings of
Seventh International Workshop on Atmospheric Icing of Structures,
SaguenayLac-Saint-Jean, Qubec, Canada, 49-54.
Roshan Fekr, M., and McClure, G. (1998). "Numerical Modeling of the Dynamic
Reponse of Ice-Shedding on Electrical Transmission Lines." Atmospheric
Research, 46(1), 1-11.
Roshan Fekr, M., McClure, G., and Hartmann, D. (1998) "Investigation of
Transmission Line Failure Due to Ice Shedding Effects using Dynamic
Analysis." Proceedings of the Eighth Intenationa Workshop on Atmospheric
Icing of Structures, Reykjavk, Iceland, 11-16.
Schauer, H., and Hammerschimd, W. (1983) "Experience Concerning Ice Loads on
Overhead Lines in Austria." Proceedings of First International Workshop,
Atmospheric Icing of Structures, CRREL Special Report 83-17, Hanover, New
Hampshire, USA, 333-340.
Stewart, J. R. (1983) "Ice as an Influence on Compact Line Phase Spacing."
Proceedings of First International Workshop, Atmospheric Icing of
Structures, CRREL Special Report 83-17, Hanover, New Hampshire, USA,
77-82.
244
245
that a relatively high copper content could contribute to the formation of a dense
oxide coating (patina) at the surface of the material. The research results indicated
that, although these steels initially corrode at the same rate as plain carbon steel, they
soon exhibit a decreasing corrosion rate. After a few years, under favorable
environments, the continuation of corrosion is practically nonexistent. The adherent,
protective films on high-strength low-alloy steels would essentially seal the surface
against further penetration of water and delay the progression of metal corrosion.
This tough adherent oxide layer would stabilize and act as a tightly adhered barrier
coating, preventing future atmospheric corrosion.
Interestingly enough, the corrosion resistance aspect of this steel alloys
performance was almost unnoticed under the shadow of its high strength attributes.
Indeed, this aspect was largely ignored until the mid 1960s. At that time, with
maintenance costs rapidly escalating under the combined effects of inflationary labor
price, high cost exotic coating materials, and greater environmental awareness of
general public, the natural corrosion resistance properties of the material caught the
attention of some marketing personnel at major steel suppliers. Weathering steel was
portrayed as the final solution to all corrosion problems in advertising campaigns. It
was marketed for such diverse applications as office buildings, bridges, utility
structures, light poles, highway guide rails, and, unsurprisingly, a few unusual
architectural applications.
Metal Corrosion
Corrosion is the result of an interaction of metal with its surrounding environment
that adversely affects the metals properties. It is an undesirable deterioration of
metal, as it impacts the structural integrity.
Corrosion of metals is an electrochemical process. It involves both the flow of
electrons and chemical reactions. The basic electrochemical reaction that drives the
corrosion of metals is galvanic action. Galvanic corrosion normally occurs when two
dissimilar metals are placed in contact with each other in the presence of an
electrolyte or a conductor such as water, acids, or salts. A galvanic cell can also be
formed if the same metal is exposed to two different concentrations of electrolyte.
Consequently, a current is generated internally by physical and chemical reactions
occurring among the components of the cell. The resulting electrochemical potential
then develops an electric current that dissolves the less noble material.
Copper, brass, bronze, and nickel are more "noble" materials, and will act as auxiliary
cathodes to the steel and accelerate its anodic dissolution. Magnesium, aluminum,
zinc and zinc-base alloy are nearly always less noble, and tend to divert the attack.
They have a greater tendency to lose electrons.
From a thermodynamic perspective, the tendency to reduce internal potential energy
is the main reason for the corrosion of metals.
246
The actual corrosion process that takes place on a piece of bare mild steel is very
complex due to factors such as variations in the composition of the steel, the presence
of impurities, uneven internal stress, or exposure to a non-uniform environment. In
addition, the rate at which metals corrode is also controlled by numerous other factors
such as the electrical potential and resistance between anodic and cathodic areas, the
pH value of the electrolyte, temperature, and humidity. The combination effects can
be complicated and complex.
To simplify the discussion, four (4) elements contribute to the basis of corrosion and
corrosion prevention. They are the anode, cathode, electrolyte, and the return current
path. Removing any one of these elements will stop the current flow and corrosion
will not occur. The firmly adhered patina layer on weathering steels could essentially
seal the steel member surface against further penetration of water, effectively
removing the electrolyte, thus stopping the corrosion of the inner material.
Weathering (COR-TEN) Steel
The United States Steel Corporation (U.S. Steel) holds the registered trademark on
the name COR-TEN.
High strength steels can be produced under ASTM A-242, A-588, and A-871
specifications for atmospheric corrosion-resistant steels intended principally for
applications requiring durability. They reduce the overall structure weight; minimize
maintenance work, and lower initial costs.
The corrosion-retarding effect of the protective layer is produced by the alloying
elements. Several compositions of weathering steels have been developed for
atmospheric corrosion resistance properties. Relatively small amounts of copper,
chromium, nickel, and silica, will result in a tenacious oxide film that improves
atmospheric corrosion resistance if the surfaces are allowed to dry periodically in
normal service. Copper is the key alloying element, but others enhance the beneficial
effects of copper.
Even with the improved metallurgical compositions, the weathering characteristics
of these steels can only be developed in favorable environments. When used outside
of these suitable conditions, weathering steel is still vulnerable to extensive corrosion.
A better understanding of these limitations is needed to allow for appropriate usage of
this material.
To begin with, complete wet and dry cycles are essential for the development of a
uniform patina layer to function as an effective corrosion barrier. The rapidity at
which the protective film can be formed is highly dependent on the frequency with
which the surface is made wet by dew or rainfall and dried by sun or wind. The oxide
layer can only be formed through exposure to sufficient cycles of wetting and
247
complete drying of the steel surface. At first, the oxidation rate and its process is
similar to that of carbon steel. However, if the protective patina layer forms properly,
the process slows such that the corrosion of the steel is greatly diminished. In 1975,
R.L. Brockenbrough and R.J. Schmitt of U.S. Steel1 issued a paper establishing
certain guidelines for the design of structural steel transmission towers. The authors
state that Alternate wetting and drying is necessary for the formation of the
protective oxide film. Should the corrosion-resistant high-strength low-alloy steel
remain wet for prolonged periods, they will corrode at the same rate as carbon steel.
This caution is not only limited to the atmospheric wetting and drying of the general
surrounding for the entire structure. Localized details within the structure also need
to be addressed. Due to design flaws or improper maintenance, certain parts of a
structure can accumulate debris such as sand, dirt, pine needles, leaves, bird nests,
animal remains, etc., which, when wet, will remain constantly moist for a prolonged
duration of time. The resulting localized corrosion can become catastrophic. This
issue has also been discussed in the Brockenbrough report: The design of the
structure should minimize ledges, crevices, and other areas that can hold water or
collect debris. Details should be self cleaning if possible. Because it may not be
readily apparent that abnormal conditions which can accelerate corrosion of bare
COR-TEN steel structures are occurring, periodic inspections of towers should be
made to determine if such conditions exist. Attention should be given to determining
if 1) vegetation is causing corrosion damage, and 2) debris is present on ledges and
causing corrosion.
As stated, proper steps should be taken during the design stage to minimize the
retention and collection of moisture. Weathering steel is not rustproof in and of itself.
If water is allowed to accumulate in pockets, those areas will experience accelerated
corrosion rates. Structural details are a primary concern in order to achieve
satisfactory, long term performance.
Weathering steel structures that are enclosed or covered by vegetation have similar
corrosion concerns. Over-grown vegetation prevents air circulation and maintains a
high moisture environment around the steel. Structure members or portions that have
experienced constant damp surroundings will not be able to form a dense patina layer
and will have metal loss at a rate consistent with bare carbon steel. Obviously, there
is also the concern of deep pitting from the root systems of certain vegetation species.
Weathering steel is also sensitive to salt-laden air environments. These include
exposure to salt water, salt spray, or even salt fog. In such environments, it is
possible that the protective patina may not stabilize, but instead continue to corrode.
In the previously mentioned paper, the authors also warn that, It should be
recognized, however, that the steel is not capable of performing satisfactorily under
some conditions. One such condition is exposure to the seacoast where recurrent
wetting by salt spray can occur. Another condition is exposure to chemical
environments heavily contaminated with corrosive substances. The paper then goes
on to state, Where it is known that heavy localized chemical contamination or
248
249
250
Institute (AISI) and Steel Structures Painting Council (SSPC) to address technical
concerns and develop guidelines. Several reports and technical advisories were
issued.
In short, the results of these studies demonstrated that:
251
Weathering steel is not a type of material that engineers can simply put out there and
forget about it. Methodical inspection and maintenance activities are still needed.
Indeed, inspection and maintenance programs must be written specifically to address
the unique characteristic of weathering steel. A recent survey indicates that only
12.5% of the 56 companies have ever conducted corrosion surveys, such as plate
thickness measurements, on their structures. This number is considered to be too low
in authors opinion. Well-planned inspection techniques and maintenance programs
that prevent degradation of material should be implemented and could extend the
useful service life of the structure.
Reference
1
252
ABSTRACT
Transmission Line Catenary structures are steel cable structures from which
transmission circuit conductors are suspended. The objective is to keep the entire
support structure above locations of high avalanche risk where other structure types
have little chance of surviving. The worlds first Catenary was built in 1955 in British
Columbia, Canada. The authors have engineered and provided construction
management for the installation of the worlds second and slightly larger Catenary. It
was installed 1,000 meters north of the first. The authors have also installed a smaller
Catenary and done preliminary engineering for thirty four other Catenaries, all in
British Columbia. This paper discusses transmission line Catenaries in general and
will describe the 2007-2008 engineering and installation of Catenary 2 in detail.
INTRODUCTION TO CATENARIES
A cable of uniform weight along its length suspended between two points and subjected to
gravity forces forms a shape called a catenary. In the small community of people who have
engineered, built and maintained a transmission line structure made entirely of long cables,
this same word has become the identifying label for the structure type. Before the worlds
second large transmission line Catenary was built in 2008, the first Catenary was simply
known as the Catenary. They are now called Catenary 1 and Catenary 2 since they are both
on the same line. These names were quickly shortened to Cat 1 (also called the 1955
Catenary by those with a sense of history) and Cat 2.
Cat 1 was conceived and engineered by
H. Brian White in 1955 in response to the
wiping out of three structures on two
parallel single circuit lines that were
placed in the belly of a mountain cirque
on their way from a near sea-level river
valley to a high elevation Pass and on
into the adjoining valley. Brians design
used two 76 mm (3 inch) diameter
galvanized steel rope cables each 1,158 m
(3,800 ft) in length. The cables were
essentially parallel, about 100 meters
apart and oriented perpendicular to the
253
transmission line direction. Each cable was anchored into the high rock walls of the cirque
and suspended high above the two 300 kV transmission circuits below. The 122 m (400 ft)
middle segments of the cables over the circuits below were brought to a separation of 914
mm (3 ft) by yoke plates. Along the middle segment, a catwalk was suspended using the two
cables as handrails and six large suspension insulator assemblies were also suspended there
to carry the conductors that pass beneath the cables.
The ahead and back spans of the conductors suspended from the Catenary are 760 meters
(2,500 ft) and 1,220 meters (4,000 ft) respectively with a weight span of about 1,000 meters
(3,300 ft). The design load is 60 kg/m (40 lb/ft) and the conductor itself is a 3,364 kcmil
ACSR 108/37 Emu weighing 7.08 kg/m (4.76 lb/ft). Access to the catwalk section of the
Catenary is by helicopter to a small helipad mounted on the cables between the circuit
attachments. The helipad is 135 m (450 ft) above the ground below. The structures
dimensions and forces on it are large. It is a dramatic structure in a dramatic location. For 53
years, this structure stood alone as a proud achievement that some consider as the safest
structure on the 80 km (50 mile) transmission line.
In the 53 years since the Catenary was installed, the transmission line has been battered and
damaged in several locations at several different times by large avalanches. The most
problematic location was the next structure north of the Catenary on one of the two circuits
where the lines rise up towards the Pass. Here, the original 35 ton latticed steel tower was
harmed in a 1988 avalanche event and crumbled to the ground in late 1992 by a powder
avalanche event. Its identical replacement tower was wiped from the face of the earth by a
large, wet snow avalanche in March, 2007. These two circuits are a singular lifeline to the
viability of the owners facilities. These lines service was very expensive to reinstate
because the work occurs in the heart of the winter in BCs coastal mountains, by definition.
Revenue losses are extraordinary and the owners ongoing facility operations are put at
serious risk. To the owners question, Can we build a Catenary in this location and stop this
from happening again? we answered yes after a brief review of the sites terrain.
OTHER CATENARIES
In the years immediately after the 1992 tower failure, the author provided preliminary
engineering for a Catenary to be located a few miles away in the valley north of the abovementioned Pass. There, a tower had been lost in 1985 and a study by avalanche experts had
identified a whole series of at-risk towers along the valley floor in that area. The 1955 Catenary
replaced six original structures in the floor of the cirque. This proposed Catenary used a similar
two-cable/four anchor configuration. In the overall, it was to be much larger. The cables were
86 mm (3.375 inches) in diameter, over 1,500 meters (5,000 ft) long and the conductors
weight span was near 2,100 meters (7,000 ft). The helipad access on the cables was about 330
m (1,000 ft) above the ground below and the Catenary structure was to replace a total of nine
original structures on the two circuits below. This Catenary remains in the drawer, waiting to
see the light of day.
Concurrently, proactive action was taken at the site of the 1992 tower loss. Here, a single cable
Catenary was installed between the bedrock at the base of the tower that had failed and a point
high on the mountainside to the southeast. This single cable was a 44.5 mm (1.75 inch)
galvanized bridge strand 670 m (2,200 ft) in length. It was considered to be an insurance policy
against any future failure of that tower. It was put in place such that, upon the towers loss, the
circuit could be re-instated within weeks rather than in the nearly three months as was the case
after the 1992 event.
254
To ones delight/surprise/dismay depending on who you were, the 2007 failure of the same
tower put this small Catenary to work, as planned 13 years earlier. A very important lesson to
learn for locations as rugged and wild as high up in the BC Coastal Mountains is that Mother
Nature has a large bag of tricks and will surprise the engineer in you routinely. The temporary
insurance Catenary was used to support the circuit for a full year, summer 2007 to summer
2008. However, the 2007 avalanche
event was very different from the 1992
event and our well laid plans based on
the 1992 experience needed serious
modification. The damage rendered to
the circuit was far more extensive in
2007, even touching the 1955 Catenary
as it was the adjacent structure. With the
decision made to install a permanent
Catenary in the location, this small,
temporary Catenary became known as
the KitiKat in recognition of its relative
size to Cat 1 and Cat 2 and as a play on Figure2The single cable KitiKat in use for one year
words for the name of the nearest town,
Kitimat.
During the installation of Cat 2 in the summer of 2008, the owners of a proposed transmission
line elsewhere in the province visited the site. After a tour of Cat 1, Cat 2 under construction,
the KitiKat and the site of the still shelved Catenary for the valley north of the Pass, the authors
were asked to provide feasibility, preliminary designs and cost estimates for Catenaries as
needed for avalanche-prone sections of their proposed routes. On one route, seven Catenaries
were proposed in series to cover a 12 km (8 mile) long, deep, narrow and steep walled valley
without touching the valley floor. The lessons learned from the above-noted four Catenaries
three in existence and one engineered lead to a proposed design for these Catenaries that is
unique from all the rest.
CATENARY 2
The Design Constraints
The fundamental requirement of a Catenary structure is that it cannot be damaged by an
avalanche. Unless the cable anchors can be placed out of the reach of avalanches, or rendered
indestructible somehow, little is gained. Even though it can appear that the vast surface area of
a mountainside would offer endless anchor location possibilities, this is not the case. The slope
of the cables as they leave the anchor site out into the span can be as steep as 35 below the
horizon. Anchor sites must have a slope in front that is steeper than this. It would be best if
there was little avalanche threat above the site either by the absence of more elevation or by the
shape of the slope above such as might direct a flow elsewhere. Above all, the anchor locations
must allow cable load sharing that meets criteria; must be accessible for construction and must
allow positioning the cables above the conductors attachment points so that the line layout is
served. Anchor locating can be a challenge. Finally the anchorage assembly must be low
profile so as to not attract large impact forces when an avalanche does run over it.
If there are avalanches, there is snow. If there is snow, there is snow load. In this part of the
world, the design loads for snow or ice on wires are substantial. It made no sense in this case to
intentionally degrade the lines original design load case of 60 kg/m (40 lb/ft) of snow/ice
255
weight on the conductors with a stress limit of 100%. For reference, this oddly stated load
and stress limit combination translates into a span dependent and more typically stated load
of about 37 kg/m (25 lb/ft) of load at 60% of conductor breaking strength. Since the structure
itself is made of wires, it is impossible to rationalize not applying that load to the wire
structure as well. This is a departure from our normal transmission structure line of thinking.
With loads this large on the wire structure, it is difficult to find other load cases of concern
for its design. The two conditions that were monitored but in a near-qualitative manner were
the cables tensions for Aeolian vibration concerns and the bare wire, nice day tensions in
place when construction and maintenance people are on the cables. Cable vibration damage
must be mitigated and the factor of safety of the system components when personnel are on it
must be acceptable. We sought a factor of safety of at least 5 whenever possible for the latter
concern. This exceeds the factor of safety of a typical steel transmission support structure
under these conditions but it honors the intent with mechanical cable system standards, even
though it can be argued that a Catenary is not that.
We learned from the various events at this location that the Catenary can be asked to support
a broken wire condition of very high tension. Although the Cat 2 conductor insulator
assemblies are suspension types, they are also capable of carrying the breaking strength of
the attached conductors, as is the Catenary itself. With the large ice load case and long weight
span, it happens that the design load in suspension is nearly equal to the breaking strength of
the conductors. How convenient!
If the entire assembly is intended to carry suspension or conductor tension loads, it must also
be capable of significant articulation and large deflections that will occur as the forces on the
Catenary change from intact cases to broken wire cases. Similarly, a very long span of cables
will move in high wind. A very high wind over the Pass will strike Cat 2 transversely and
push it a meter or two along line. The conductors will not be inclined to follow and will
attempt to remain where they are. This will move the insulator strings out of plumb and they
could interfere with the catwalk. If this differential displacement is not dealt with in the
design details, conflict of parts and damage will occur.
The nature of the site defined the remaining design constraints. There is no road access
within five kilometers (3 miles) of the site and all access and storage locations were at least
that far away and 1.6 km (1 mile) below in elevation. In other words, everything and
everyone could only be brought to the site by helicopters. Therefore, all materials and
equipment had to transportable by
helicopter. History told us that the S-64
SkyCrane and the 234 Boeing Chinook
were generally available to us. We
designed with these heavy lift machines
in mind.
The Design Choices
Catenary 1 was constructed by laying its
large, heavy cables on the ground below
their final position and raising them into
place but pulling their ends towards the Figure3The Boeing Chinook put 114 tons of material onto
mountain site in 6.5 hours from 12 miles away and one
anchors with large winches set well below the
mile below. Here, three 6,000 lb cables reels go at once.
the anchor sites. Each of the two large
256
cables was made up of four segments up to 366 m (1,200 ft) long. The joints were poured zinc
socket sets. Our decision with Cat 2 was to string the cables in place above the two operating
circuits from one anchor site to its mate on the other side of the valley and to do this with jointless, single-length cable.
The cables of Catenary 1 and our general understanding of its design provided great initial
design guidance. We knew that our cables might be slightly longer and might need to be a bit
stronger. Its 3 inch steel rope cables are effectively made up of 6 one-inch cables helically
wrapped around the central seventh. It would be impossible to lift long lengths of 3 inch
cable to our work site and we had already envisioned the concept of assembling a cable in
situ from a collection of seven smaller cables.
However one-inch diameter cables that would form an unwound 3 inch cable were going to be
a bit heavy and require installation tensions larger than we wanted. Smaller cables were
attractive but assembling them into a larger cable would require a second layer of 12 cables
over the envisioned six for a total of 19 cables. Construction of that assembly was considered
too unwieldy. We settled on an arrangement of 14 cables, each inch (19 mm) in diameter and
placed in two parallel sets of seven cables each. In this arrangement, we had a total of 28
cables that we expected would be up to 1,340 m (4,400 ft) in length. At this length, their
on-reel weight would reach 2,700 kg (6,000 lbs) each and the stringing tension would be
about 2,300 kg (5,000 lbs) each. These were acceptable values.
The comparable weight and strength comparisons of this cable set choice to Cat 1 cables was
attractive. It was slightly heavier and slightly stronger allowing some freedom in establishing
its four yet-to-be determined anchor sites. There was some attraction to having a pair of parallel
58 mm (2.25 inch) cables representing each of the two main cables. Given that marker balls
would be required on the cables, a parallel set increased the construction cart configuration
options.
The Design Process and Tools
From a structural analysis and design perspective, the Cat 2 project offered a host of challenges
that required the unconventional use of several common software programs. During the
process of laying out the new Catenary, the objective was two-fold: to locate the best possible
anchor locations using aforementioned criteria, and to increase the wire clearance to ground.
This was done so that Cat 2 could be constructed well above the energized circuits, and upon
completion, the wires could be lifted (almost) vertically to the new attachment points. This
configuration also ensured that additional load would not be imparted on the existing
structures. From the early design phases, the idea was to cross the two sets of cables so that the
NE (A) anchors shared cables with the SW (B) anchors and the same with the SE and NW
anchors (C and D anchors,
respectively) - keeping the wires close
together in space to facilitate the
construction of the catwalk by linemen in
carts.
The design process began with a PLSCADD model of the entire line section;
extending from the southern dead-end
towers (parked safely out of avalanche
areas) to the dead-end towers at the top of
Figure4The PLS-CADD model of the line section containing
both Cat 1 and Cat2.
257
the Pass. The model also contained a detailed terrain model of the entire cirque. Using this
model to identify preliminary anchor locations, the Catenary system coordinates (east/west
yoke positions and assembly attachment points) and cable lengths were calculated using an
Excel spreadsheet. This information was passed into PLS-TOWER in order to construct the
Catenary structure. While TOWER is intended for 3-D lattice trusses with (4) anchors (legs) at
typical spacings of 20 60, our application was a bit different: our TOWER model consisted
of approximately 16,040 of cables and (8) anchor locations with a spacing of about 4000
between the east and west side anchors. The TOWER models interaction with PLS-CADD
required a coordinate transformation between a local structure coordinate system (TOWER)
and the global coordinate system (PLS-CADD). This transformation was more complicated
since the anchors sit almost 800 above and 2000 offset from the structures centerline location.
Once the preliminary model was built, it was moved around and adjusted within the PLSCADD file until the optimum anchor locations and conductor positions were achieved. Once
the anchor positions were surveyed and field-reviewed, they were constructed in the summer of
2007. Upon completion, the anchor sites and saddle locations were surveyed in and passed
into the model essential information to
determine the exact cable lengths. The
cable lengths were calculated taking into
account Cat 2s shape as well as the cables
pre-stress load and the offsets between the
anchor points and the saddle. Finally,
fourteen 3,673.80 ft reels were ordered for
the A-B cables and fourteen 4,372.45 ft
reels were ordered for the C-D cables.
While PLS-TOWER was useful and
accurate for determining cable lengths and
the structures deformed shape under
design loads, it could not provide all the
details necessary for the design and Figure5The spider web of cables within the cirque as
construction. Initially, we exported the modeled in SAPS: 71,340 of conductor and steel cable!
PLS-CADD file (with Cat 1 and Cat 2 structure models) into SAPS in order to satisfy
ourselves that the Catenary would withstand a broken wire condition the results of which are
more accurate when all the conductors and support cables play a role in the calculation.
This model provided a more accurate depiction of the insulator assemblies. One hardware
piece borrowed from Cat 1 was the suspension shoes and steel straps that connected the shoes
to an assembly yoke plate. However, the Cat 1 design has different departure angles than those
of Cat 2. Therefore, we had to adjust the strap lengths between the shoes and the yoke plates in
order to guarantee that neither shoe contacts the yoke plate and that the load is shared well
between the two shoes. This was done in SAPS by altering the uphill strap length accordingly
and examining the deformed shape and strap loads.
Finally, the SAPS model assisted in determining the required conductor lengths adjustments for
each phase and the attachment locations of each phase. The calculated values were verified in
the PLS-CADD file using finite element analyses, and we provided the contractor with wire
lengths to splice into each phase.
258
259
critical, operating circuits 240 m (800 ft) below the anchor sites.
The west side anchors were on the edge of the west side mountain top. They were about 150 m
(500 feet) apart. The entire area was generally flat and this was our working site where we
would store and stage materials and from which we would string cables to the east side. The
east side anchors were well down that mountain slope and work space was very restricted.
We flew a 3,280 m (10,000 ft) long, 34,000 kg (75,000 lb) breaking strength pulling rope
across the valley from west-side anchor B to east-side anchor A. There, it passed through a
turning block and we returned its end to the west side where the puller sat about 100 m (300
feet) from the tensioner that payed out the cables. The first cable was pulled across over the
circuits by this rope and pinned temporarily beyond its permanent anchor pins. This cable sat
about 50 feet above its final sag. We placed a trolley on that cable and used the pulling rope and
a haul-back rope attached to the trolley to pull the lead end of the 13 remaining cables across
the valley. The use of the trolley greatly reduced the risk to the circuits below if some
component of the pulling assemblage failed. The owner was adamant about that risk mitigation
for obvious reasons. The entire pulling operation was then repeated for the other 14 cables
between anchors D to C.
With four sets of seven cables strung across the valley in two parallel sets, the linemen in carts
began to bundle each set of seven cables to form a single cable unit. This was done with
specially sized Preform wraps placed at 17 m (50 ft) intervals. Between the wraps, the seven
cables were in contact but not bound to each other. It was expected that this loose arrangement
would give the finished product considerable self-damping characteristics against Aeolian
vibration. So far, this seems to the case.
The mid-section of the Catenary has a catwalk section over the circuit attachment area. At the
catwalk ends, the cables were brought together and held in place with large aluminum yoke
plates. Sets of cable clamps on the yoke plates gripped the cables against longitudinal slippage
and had to resist lateral splitting loads caused by the cables tensions and the angles turned by
the cables as they were re-directed
towards the anchorages. The yoke
plates were set in place as the cables
were rigged and pulled laterally by the
men working on them. The west yoke
plate was set first. With cables later
pulled together at the east yoke plate
site and now parallel between the yoke
plate locations, the 28 catwalk sections
each 4.88 m (16 ft) in length were set
in place from west to east. They were
brought to the men on the Catenary by
an A-Star helicopter, as was all of the
Figure8Crews in Carts bundled the small cables into
attaching hardware.
larger cables and rigged the parallel pairs towards each
other at yoke plate positions
260
261
September. Besides, the crew was beat. It was time to let them go home.
As of this writing, the 2009 plan for the relocation of the left circuit is underway and on
schedule. The entire installation will be completed by July 2009. Other that routine inspections
to ensure that the Catenary components behave as planned, there should be no more
monumental and expensive service recovery events at this location in the years ahead.
FUTURE CATENARIES
It is noted above that the lessons learned on each Catenary tackled have been insightful and
very useful is developing the next.
The assembling of seven small cables into a larger cable on site has proven to be a workable
idea with expected performance benefits. One of our Cat 2 anchorage designs was unique
because the rock shape did not allow our planned configuration. However, it later became
recognized as the better idea. As with any structures, whether they are buildings, bridges or
transmission towers, ideas and methods are made attractive by the local and present
circumstances. Certainly, this is the case with transmission line Catenaries. For example, the
combination of constraints on this project leads us to believe that green field Catenaries those
not fit into existing lines can be installed for half of the cost of this one, perhaps less.
The conditions that make a transmission line Catenary useful are rare but not unheard of.
Recognizing the existence and constructability of such structures should remind us that the
things that we think are genuine problems can be overcome when we apply our engineering
and construction skills and are not stopped by conventional practices.
References:
PLS-CADD, (1992), Power Line Systems Computer Aided Design and Drafting, Power Line Systems, Inc.,
Madison, Wisconsin
TOWER, (1984), Analysis and Design of Steel Latticed Towers Used in Transmission and Communication Facilities,
Power Line Systems, Inc., Madison, Wisconsin
SAPS, (1984), Structural Analysis of Power and communication Systems, Power Line Systems, Inc., Madison,
Wisconsin
262
American Electric Power, Transmission Line Projects Engineering, Director, 700 Morrison
Road, Gahanna, OH 43230; (614) 552-1690; email: [email protected]
2
American Electric Power, Project Management, P.O Box 2021, Roanoke, VA 24022; (540)
562-7055; email: [email protected]
3
American Electric Power, Transmission Line Projects Engineering, P.O Box 2021, Roanoke,
VA 24022; (540) 562-7295; email: [email protected]
ABSTRACT
Transmission line construction during the early 1900s was primarily dictated by
engineering issues with little devotion to mitigating visual impacts on the surrounding
environment. Beginning in the 1960s and 70s, a new environmental ethic began to
make its way into the planning process. As a result, heightened sensitivities to visual
impact issues are now the norm on new transmission projects, thus making them a
major objective on newer projects. This paper addresses the aesthetic challenges
confronting future expansion of transmission lines, such as overcoming public
objections, environmental stewardship and improving the appearance of transmission
facilities.
The paper uses illustrations, photographs and case studies to review current solutions
by American Electric Power (AEP) and other utility companies to reduce visual
impacts on transmission projects. Current best practices revolve around route
development, structure design, topographically sensitive siting techniques, off-site
mitigation, accurate visual simulations, stakeholder input, construction management,
and special structure and conductor finishes. Finally, the future of transmission line
siting is discussed along with areas of needed research. These issues challenge the
engineering team with the opportunity to provide innovative, creative solutions.
INTRODUCTION
Future expansion of transmission lines is vital. The existing system is aging and
insufficient to meet Americas long-term energy needs. Additionally, the existing
grid was not designed to transmit the twenty-first centurys renewable energy, such as
wind and solar power that may be generated far from where it is needed.
Nevertheless, the greatest challenge confronting future expansion of transmission
lines is aesthetics and associated issues, such as overcoming public objections, and
improving the appearance of transmission facilities.
The scenic landscape is a highly valued resource that enhances peoples lives and
benefits society it provides visual and sensory pleasure (USFS 1995). Aesthetic
mitigation focuses on protecting the existing scenic quality and beauty, especially
263
264
land cover; avian flyways; and endangered species. Though aesthetics is just one of
many environmental resources, a major objective of route development is to avoid
key and sensitive viewsheds while minimizing adverse impacts to other
environmental resources (e.g., natural and cultural). A complete discussion of the
environmental or aesthetic analyses employed by AEP is beyond the scope of this
paper, yet the following are highlights that have proven beneficial in addressing
aesthetic impacts during route development.
Scoping and Stakeholder Input
During the scoping process, it is critical to engage the public and identify their
concerns, key viewsheds, and preferences. Input can be gathered through phone
messaging (autodial) to reach citizens in immediate proximity (i.e., within 500 feet)
of a study segment, press releases that alert the public of opportunities to provide
input, public information workshops, written comments and direct interviews.
Additionally, homeowner organizations, clubs and other local organizations should
be contacted and a website established that includes general project information and a
commenting forum.
Further stakeholder input from federal, state and local officials is collected through
direct meetings and data request letters. These agencies should be involved early and
throughout route development; their input and expertise are essential (e.g.,
endangered species, local future land use, and cultural and recreational resources).
During field reconnaissance, input is validated and ground-truthed (e.g., critical
viewsheds) and engineering constraints are also recorded.
The route development process should be iterative and collaborative with multiple
opportunities for public input, agency review and participation, and route
modifications as needed. During a recent 765kV project, lack of initial involvement
from federal and state agencies hindered progress; however, once participation began
and agencies shared critical insights concerning their resources, potential routes
rapidly evolved. For example, by way of agency suggestion, a corridor developed
that crossed a sensitive trail near an existing disturbed land use (an interstate
highway). In the end, an acceptable route to all parties was approved.
Key Observation Points
Key observation points (KOPs), a U.S. Forest Service term (USDA 1995), are
representative view points with an associated viewshed and could include residences,
communities, parks, recreation, conservation areas, historical resources, trails, and
scenic travelways (see Figure 1). They are identified through the initial scoping,
public input, agency (federal, state and local) correspondence, websites, publications,
electronic data, and aerial photography. Impacts to these valued and sensitive
community viewsheds should be minimized to the extent possible during routing,
thus it is important to define the KOPs viewsheds. To accomplish this, a
geographical information system (GIS) viewshed analysis (ESRI ArcMap) is
typically used.
265
266
residences, and other community features) are identified and considered. When used
correctly, GIS viewshed analyses are valuable tools during route development.
Visualizations
Visual simulations (see Figures 3 and 4), computer terrain models, and line-of-sight
models illustrate degree of visual impacts, provide means to suggest mitigations,
verify GIS viewshed analyses, and compare alternatives. Additionally, they are one
of the best means of engaging decision makers, local officials and the public by
showing them what the project will look like and dispelling exaggerated visions of
the proposed project (IEA & LI 1995). For example, Figure 4 is a simulation used to
show local officials what a proposed 161 kV line would like; it was helpful since the
local officials believed the structure would be much larger. The project has
progressed to completion.
267
of a transmission line when it pre-dates their development; see Figure 7, the existing
500kV line did not deter new residential development as shown.
Additionally, utilizing existing farm, logging, jeep, and coal roads reduces access
road development, resulting in less landscape scarring and aesthetic impacts. Hence,
these land uses should be considered during routing.
Paralleling
Paralleling existing utility and non-scenic linear features such as natural gas lines,
utility lines, railroads, roads and fence rows (see Figures 8 and 9) condenses aesthetic
impacts to one corridor, reducing the before and after contrast as compared to a new
line through an uninterrupted forest canopy. Paralleling, however, does have
shortcomings. In Figure 8, the landscape could have absorbed one major line, but
three have transposed the previous rural character into an industrial character. Other
limitations include engineering code requirements and constraints (such as safety
clearances), and reliability issues.
Offset Mitigation
Offset mitigation is a creative way to reduce aesthetic impacts for the overall
landscape at critical areas. There are various offsetting mitigation methods. On a
new 765-kV transmission line, the removal of 17 miles of old transmission line and
ROW is being phased in over a period of time, to partially offset impact to U.S.
Forest Service lands. For another project (138kV line), a 69-kV line was retired and
removed from the communitys recreation area. Yet, another offset mitigation
method is property exchange; for example, an equivalent property is exchanged to
offset an impact of a proposed project. This type of mitigation is occasionally used
268
269
270
toward the sky making them iconic features carpeted with an uninterrupted forest
canopy due to the undevelopable slopes. To background views, the ROW clearing
and shadow line appear as an unnatural slice through the forest, creating a high level
of visual contrast (see Figures 14 and 15). Another advantage, the lattices longer,
higher spans provide the opportunity to minimize impacts on sensitive features like
riparian areas, forest, scenic roads (see Figure 16), water courses and cultural sites.
Low-Reflectivity Materials
Darkened or low-reflective materials (see Figures 17 and 18) for structures and
conductors reduce the visual impact of transmission facilities. This is of particular
benefit when the project is first constructed when it is most prominent; in contrast,
the traditional shiny, galvanized finish appears synthetic at first, especially when
contrasted with a natural landscape. The darkened effect is achieved by various
treatments to the material surface during the fabrication process. Furthermore, dark
and low reflectivity hardware, such as yoke plates, spacers, guy hardware and
insulators can also reduce aesthetic impacts. There is, however, a premium
associated with the low-reflectivity materials.
Conductor Configuration
New conductor configurations for extra high voltage (EHV) transmission lines
reduce the audible noise produced by the corona on the conductors. Noise reduction
reduces the transmission line presence and thus its visual attention. The new
configuration, six-conductor bundle was used on AEPs latest 765-kV line in West
Virginia and Virginia to decrease noise level, particularly at higher elevations. In
271
practical terms, the new design cuts the audible noise level by half for a person
standing 100 feet from the edge of the ROW. Furthermore, the use of a six-bundle
configuration, as compared to a four-bundle configuration, reduces the individual
conductor diameter (gauge), resulting in a lighter grey conductor and shadow and
slightly less visibility.
Maximizing the ROW
Aesthetic impacts can be further reduced by maximizing the ROW with efficient
designs. Modern EHV transmission lines have unparalleled performance and
efficiencies, resulting in less environmental impact since fewer lines and less ROW
need to be constructed. For example, one 765kV line requires 200 feet of ROW,
while an equivalent number of 345kV lines requires 900 feet ROW (see Figures 19
and 20).
272
of this is when the existing distribution ROW and the proposed transmission ROW
are co-located to minimize impact. Multi-circuiting is when transmission lines are
combined on the same structure (see Figures 13, 21, and 28). For example, single
lower-voltage regional circuits (230kV or less) can sometimes be combined onto one
larger and taller structure. In very limited cases, a single EHV circuit (765kV or
500kV) can be combined with a single lower voltage regional circuit (230 kV or
less); this, however, may require tower designs that have not been achieved to date.
Two bulk transmission lines, such as 765kV and 500kV, should not be combined on
the same structure due to reliability concerns.
CONSTRUCTION
After the transmission line is sited, approved, and engineered, aesthetic attention
should be integrated throughout construction. Much of this responsibility falls on the
construction management team. During and after construction, the team should
develop strategies to reduce construction impacts, monitor and address unforeseen
aesthetic impacts, verify restoration and revegetation, and confirm cleanup. The
following are mitigation examples during construction.
Vegetation Clearing Plan
As repeated throughout this paper, minimizing ROW clearing is an important
aesthetic mitigation. To this end, a vegetation clearing plan should be prepared to
direct the clearing operations and to avoid over-clearing. Generally, trees in valleys
and lower elevations, except for those locations where trees are tall enough to
threaten the safety and reliability of the line, should not be cleared. Frequently, lowgrowing species such as redbud and dogwood are compatible with transmission line
operations and left in the ROW in environmentally sensitive areas. Figures 22 and 23
are examples of selectively-cut ROWs. EHV lines are the exception due to their
national importance and clearing procedures are often inflexible.
Minimizing Disturbance
The construction management team should determine, mark, and enforce the limits of
construction activity to prevent unnecessary impacts, and where appropriate, include
mitigations in the construction specifications and contract(s). For example, all
construction vehicle movement outside the ROW should be restricted to pre-
273
Natural Shapes
Natural appearing shapes reduce visual contrast with the adjacent landscape. Hence,
during construction, vegetation clearing for ROW, access roads, and staging yards
should use curvilinear and natural appearing boundaries instead of straight lines to the
extent practical and grading should be performed in a manner that minimizes erosion
and conforms to the natural topography. Additionally, alignment of any access roads
or cross-country routes should follow the landform contours to minimize ground
disturbance and/or reduce scarring (visual contrast) of the landscape.
Restoration
Seeding and fertilizing (Figure 26) in areas of cut and fill disturbance speed
revegetation, provide screening, reduce erosion, promote and maintain wildlife
274
habitat, reduce invasion pressure by non-native plants, reduce bird nest parasitism
and predation, and restrict access by off-road vehicles. Material should be native,
perennial, and compatible in color to the adjacent landscape and installed as soon as
possible after grading.
POST-CONSTRUCTION
Mitigation
Occasionally, unpredicted aesthetic impacts occur during construction; these should
be addressed proactively and in a timely manner. Figure 27 is an example of an
unforeseen impact. At the request of residents in the immediate foreground views of
the transmission line, vegetation buffers should be considered where effective and
practical to limit views of the line. At critical visual areas (e.g., a greenway trail),
new trees should be strategically grouped and planted either at the view point or near
the transmission line, to break-up visibility of the line.
275
Most EMSs are built on the "plan, do, check, act" model. This model
leads to continual improvement based upon: Plan planning,
including identifying environmental aspects and establishing goals;
Do implementing, including training and operational controls;
Check checking, including monitoring and corrective action; and
Act reviewing, including progress reviews and acting to make
needed changes to the EMS. (USEPA 2009)
The EMS not only manages habitats, ecosystems, and vegetation but other issues
such as ROW usage. As mentioned earlier, the ROW clearings, especially on
mountains, can be a large impact to background views. Uses that encourage erosion
scarring on mountain slopes such as all-terrain vehicles (ATVs) should be
restricted barriers (access gates, woody material and tree slash, tank traps, and
ditches) should be installed at strategic locations to discourage and prevent
inappropriate use on the ROW. In contrast, for low voltage lines located in areas near
communities, consider public ROW access such as trails (see Figure 28), recreation
fields, or nature areas. These amenities benefit the community and offset impacts.
For lands not owned in-fee simple, the EMS must be created with landowner
cooperation and support. The final form of the EMS could include documents, plans,
GIS, maps, inventories, and so forth.
Furthermore, an EMS should build off of existing research and standards. For
example, the Electric Power Research Institute (EPRI) has co-sponsored (with
various utilities, academics and agencies) the development of integrated vegetation
management (IVM) performance standards to provide a system for assessing the
efficiency and effectiveness of the ROW vegetation management (EPRI 2006). IVM
is an operational-level component of an EMS (Novak et al 2008). Continuing
research, such as IVM standards, is needed to evaluate and determine the most
effective strategies.
CONCLUSION
New transmission lines are vital to update the current system and to make renewable
energy sources accessible. The electricity delivery system in the United States is
characterized by an aging infrastructure and largely reflects technology developed in
276
the 1950s or earlier (Edison Electric Institute 2009). Over seventy percent of
transmission lines are 25 years or older (Center for Smart Energy 2005), yet
Americans are using 13 times the electricity they used a half century ago (Tierney
2008). As a result, the system is being pushed to serve regional power markets and to
perform functions for which it was not originally designed.
Additionally, access to domestic energy resources is vital to our national security and
energy independence (Tierney 2008). An efficient, interstate transmission
superhighway is required today to access Americas clean, renewable generation.
AEPs chairman, president, and CEO Michael Morris states the Dakotas, Minnesota
and Iowa have some of the very best wind generation resources in the United States,
but the wind potential in this region cannot be developed unless we build a very
efficient transmission superhighway to bring this clean, renewable generation to
population and electricity load centers (Morris 2008). Transmissions role in
solving these national challenges [i.e., development of domestic energy resources] is
too often overlooked as the focus shines on power plants and demand-side measures.
But inadequate attention to enhancing the nations electric transmission system will
undermine if not prevent our ability to satisfy our national economic goals while
also addressing climate change and our needs for energy independence (Tierney
2008).
To address the challenge of future expansion, not only are aesthetic mitigations for
transmission projects needed, the industry will need to think out of the box. Bridges
that were once considered eyesores are now considered architectural assets and
engineering achievements. Perhaps, to a degree, the same can be true for
transmission lines. Imatran Voima Oy (IVO), the largest utility company in
Finland, is doing some motivating work with transmission structures. They are
incorporating artistic features into their structures to break up the monotony
perceived in transmission lines. This work has yielded a family of structures that are
both pleasing to the eye and may be mass-produced at reasonable cost (Nieminen
1996). In Korea, 300-foot 765 -kV structures were developed to span over trees to
completely eliminate ROW clearing and maintenance.
Because of the recent power outages to our transmission systems in California and
the Northeast blackout in 2003, Americans are beginning to understand that new
transmission lines are essential and necessary. Electricity is fundamental to our
quality of life (Tierney 2008). Few citizens, however, are tolerant of high-voltage
transmission lines, let alone the superhighway proposed for an extra high-voltage
system. This deep protective instinct in their valued viewsheds makes aesthetic
mitigation a formidable challenge. Nevertheless, through collaboration case
studies, modern designs, new approaches, and on-going research we will expand
and strengthen the nations electrical grid, access renewable energy resources, and
invent modern solutions to these aesthetic challenges.
277
REFERENCES
Center for Smart Energy (2005). The Emerging Smart Grid: Investment and Entrepreneurial Potential
in the Electric Power Grid of the Future. (October 2005), p. 9.
Edison Electric Institute (2009). Supply and Demand: Supplies are Shrinking, Systems are Straining.
Website, accessed 5 February 2009, from http://www.getenergyactive.org/supply/problem.htm.
Federal Power Commission (1970). Guidelines for the Protection of Natural, Historic, Scenic, and
Recreational Values in the Design and Location of Rights-of-Way and Transmission Facilities.
Adopted by the Federal Power Commission in Order No. 414 issued November 27, 1970, and now
applied by Federal Energy Regulatory Commission. Copies may be obtained from the Office of Public
Information, Federal Energy Regulatory Commission, Washington, D.C. 20426.
IEA & LI (Institute of Environmental Assessment and the Landscape Institute) (1995). Guidelines for
Landscape and Visual Impact Assessment. London: E&FN Spon.
Morris, M. G. (2008). AEP To Pursue Development Of Transmission Superhighway To Transport
Renewable Energy Across The Upper Midwest. Website, news release, accesses 5 February 2009,
from http://www.aep.com/newsroom/newsreleases/?id=1515 (Dec. 2, 2008).
Nieminen, K. (1996). Poles & Structures Transmission Structures as Landscape Art Artistic and
structural design are blended to create a visually appealing transmission line. Transmission and
Distribution World. Website, accessed 8 January 2009, from http://www.tdworld.com (May 1, 1996).
Nowak, C. A., Ballard, B.D., Greninger, L. K., and Venner, M. C. (2008). Performance Standards for
Assessing Vegetation Management on Rights-of-Way: Case Study of New York DOTs Roadside
Rights-of-Way Vegetation Management Program, in Goodrich-Mahoney, J. W., Abrahamson, L.,
Ballard, J. and Takalsky S. (eds.), 8th International Symposium on Environmental Concern in Rightsof-Way Management, September 12-16, 2004, pp. 113-126. Elsevier Science.
Schmid, W. (2001). The emerging role of visual resource assessment and visualization in landscape
planning in Switzerland. Landscape and Urban Planning, 54, pp. 213-221.
Tierney, S. F. (2008). Vision for a 21st Century Interstate Electric Highway System, White paper
commissioned by AEP from Analysis Group, Boston, MA.
USDA (U.S. Department of Agriculture) (1974). National Landscape Management, Volume 2 (Visual
Management System). U.S. Government Printing Office, Washington, D.C.
USDA (U.S. Department of Agriculture) (1995). Landscape Aesthetics: a handbook for scenery
management. Agriculture Handbook No. 462. U.S. Government Printing Office, Washington, D.C.
USEPA (U.S. Environmental Protection Agency). 2009. Internet website:
www.epa.gov/ems/info/index.htm, accessed on 5/29/2009.
USFS (U.S. Forest Service) (2002). AEP 765kV Transmission Line (Jacksons Ferry, Virginia to
Oceana, West Virginia). Supplemental Draft Environmental Impact Statement, Volume II, pp. 2-20.
The authoring team would like to thank Robert Brandner of American Electric Power for his efforts reviewing and
editing this paper.
278
Southern Company Transmission / Alabama Power Company, 600 North 18th Street,
Birmingham, AL 35203; PH (205) 257-1959; FAX (205) 257-4364; Email:
[email protected]
ABSTRACT
Alabama Power has successfully installed the new 3M Aluminum Conductor
Composite Reinforced (ACCR) high temperature, low sag conductor on a 2.5 km (1.5
miles) section of a 230 kV transmission line, utilizing the existing lattice steel
structures located in a congested corridor with 230 kV and 500 kV line crossings. The
upgrade was necessary to provide additional capacity to meet future (short term) load
demand and line loading under contingency conditions. This paper provides a case
study of the evaluation, selection and installation of the 3M conductor and highlights
the challenges during the design and construction stages of the project.
INTRODUCTION
The Gorgas to Miller 230 kV transmission line was originally designed and
constructed in 1956 as the Gorgas to Boyles 110 kV transmission line (North leg).
The line consisted of multi-pole wood structures with 795 kcmil 26/7 ACSR
conductor. In 1972, with additional insulation and structure modification where
necessary, the line was up-rated to operate at the 230 kV voltage level.
With the construction of the Miller Steam Plant in 1975, the Gorgas to Boyles line
was sectionalized and looped through the newly constructed Miller SP Substation
creating the new line segments; Gorgas - Miller 230 kV transmission line (GMTL)
and Miller Boyles 230 kV transmission line. The two new line sections from the
new Miller substation to the existing transmission line were constructed on lattice
steel towers with 1351 ACSR conductor.
In 2006, Alabama Power Company (APCo) laid out a multimillion dollar plan to
install Flue Gas Desulfurization (Scrubber) facilities at its fossil fuel plants as part of
its environmental commitment to reduce sulfur and mercury based emissions. The
Miller Steam Plant (SP) was one of the four locations where the Scrubbers would be
installed and had a deliverable date of 2010 2011. Load flow and system reliability
studies by APCos Transmission Planning identified the existing GMTL as the line to
best serve the loading from the new substation required for the Scrubber facilities.
The objective was to have a line capacity to satisfy a loading of 1500 Amps (peak and
279
contingency loadings) by the end of 2010 and a maximum 1900 Amps for future load
increase.
The solution, as proposed in the report from Transmission Planning, was to replace
the existing conductor on the GMTL with new conductor and have a loop in loop out
configuration to the new Miller Scrubber substation as shown in Figure 1. The
transmission line segment between the Miller SP substation and the new Miller
Scrubber substation would therefore consist of a 0.97 km (0.6 miles) line section of
steel lattice towers and a 1.45 km (0.9 miles) line section of single and multi-pole
concrete structures including a 0.48 km (0.3 miles) relocated new line section. The
steel tower line section is located in a highly congested area due to the presence of
nine other transmission lines emanating out of the Miller SP substation (Figure 2).
The primary line design objective therefore, was to use the five existing towers to
support the new conductor with minimum or no structure modification.
CONDUCTOR SELECTION
The line parameters for the tower line section are;
Line voltage 230 kV;
Line Section Length 0.97 km (0.6 miles);
Ruling Spans 77 m, 204 m, 363 m, 263 m (254 ft, 670 ft, 1190 ft, 864 ft);
Existing Conductor 1351 kcmil 54/19 strands ACSR;
Allowable Maximum Conductor Tension (tower design) 77,872 N (17,500 lbs);
Design Maximum Line Loading 1900 A
Utilizing aerial survey data, the line sections were modeled in PLS-CADD. Method 4
models were developed for the concrete pole structures using PLS-POLE and method
1 models were used for the lattice steel towers. The initial recommendation from
Transmission Planning was to use 1351 54/19 ACSS for the conductor replacement.
However, initial line design analysis identified conflict in clearances to ground and
obstacle crossings including a railroad and a 230 kV transmission line. Consequently,
an alternative solution was explored with the new 1033 54/19 ACCR conductor.
The 1033 ACCR is from a relatively new family of high temperature, low sag
conductor manufactured by the 3M Company. The stranded core of metal matrix
aluminum composite wires are surrounded by aluminum-zirconium alloy strands
resulting in a composite conductor with superior mechanical and thermal properties
compared to the more traditional ACSR conductor.
The following details the result of the evaluation of both the 1033 ACCR and 1351
ACSS conductor as a replacement to the existing 1351 ACSR conductor.
Conductor Ampacity. Both proposed conductors have the capacity to satisfy the
design maximum line loading of 1900 A. At 40 C ambient temperature, the 1033
ACCR conductor operating at 200 C has ampacity of 1925 A and the 1351 ACSS
conductor operating at 160 C has ampacity of 1989 A.
280
Figure 1: Schematic of GMTL before and after the Miller Scrubber Substation
281
56
Span A
ad
lro
ai
R
V
0K
0
5
V
0K
3
2
C
an
Sp
Sp
an
56D
Spa
n
56C
B
an
Sp
56A
282
Structure Loading. The conductor loading criteria was the NESC Medium loading.
With the conductors auto sagged in PLS-CADD, the 25% bare final was the limiting
condition for both the 1033 ACCR and the 1351 ACSS conductor tension. The
maximum tension for the conductors as shown in Table 1, were both below the
allowable 77,872 N (17,500 lbs). Table 1 also shows the diameter and unit weight of
the 1033 ACCR and 1351 ACSS conductors compared to the existing 1351 ACSR.
With the lower tensions and equal or lower unit weight and diameter of the new
conductors, the steel towers would not experience an increase in loading compared to
the 1351 ACSR conductor; hence the ability to utilize the five existing towers in the
main line without any structural capacity improvement.
Table 1: Comparison of 1351 ACSR with 1033 ACCR and 1351 ACSS
Conductor Type
Description
1351
1033
1351
ACSR
ACCR
ACSS
Stranding
54/19
54/19
54/19
Cable Diameter (mm)
36.1
31.6
36.1
Weight (N/m)
25.3
16.5
25.3
Ultimate Strength (N)
206,000
157,833
161,080
Maximum Tension (N)
75,645
61,410
55,623
-6
-6
Coeff. of Expansion of Core (/ C)
6.4 x 10
3.5 x 10
6.4 x 10-6
Ampacity (Amps)
1456
1925
1989
& Conductor Temperature
100 C
200 C
160 C
-(@ 40 C Ambient)
Conductor Sag. With the conductors displayed at maximum operating temperature
of 160 C (320 F) for the 1351 ACSS and 200 C (392 F) for the ACCR, the final
sags were compared with the 1351 ACSR conductor at 100 C (212 F). The length
and final sags for each span are shown in Table 2.
Table 2: Comparison of Final Sags
1351 ACSR
Span
Span
Ruling
Final Sag
ID.
Length Span
@ 100 C
(m)
(m)
(m)
A
263
263
8.0
B
91
363
0.8
C
405
363
15.9
D
204
204
5.7
E
77
77
2.3
1033 ACCR
Final Sag
@ 200 C
(m)
8.0
0.8
15.7
5.7
2.4
1351 ACSS
Final Sag
@ 160 C
(m)
10.2
1.0
20.7
7.4
2.8
283
The 1351 ACSS conductor had higher final sags compared to the existing 1351
ACSR. Span E conductors terminating on the substation structure were sagged in at
reduced tension for the conductor models, thus accounting for the smaller variance in
final sags.
For the 1351 ACSS conductor, the PLS-CADD report identified clearance violations
in 3 of the 5 spans in the tower line section. Ground clearance was not achieved in
Span A and obstacle clearances were violated in Spans C and D (the shield wire of a
230 kV line and a rail road crossing respectively). Table 3 shows the extent to which
the clearances were exceeded in each span.
Table 3: 1351 ACSS Clearance Violation in Affected Spans
Spans
Span
Extent of Clearance
Object Under Line
ID.
Length (m)
Violation (m)
A
263
1.2
Ground / Road
B
91
C
405
2.4
Shield wire (230 kV TL)
D
204
0.6
Railroad
E
77
The higher sags of the ACSS conductor reduced clearances to ground and other
obstacles crossing under the transmission line. Consequently, the challenge was to
achieve the required clearances over ground and the other crossings.
Improving vertical clearances by installing two taller towers at structures 56 (+4.6 m)
and 56C (+5.5 m) in the GMTL created new conflicts with the 500 kV transmission
lines crossing above in spans A and C (See Figures 3 & 4). Consequently, three new
500 kV towers would be required as intermediate or replacement structures for the
crossing transmission lines. The installation of five new towers would present
significant challenges with transmission line availability, system exposure during line
outages, location of new 230 kV towers and structure installation in a congested and
possibly an energized environment.
The lower sags of the 1033 ACCR conductor provided an ideal solution given the
various constraints. The lower unit weight and lower coefficient of thermal expansion
resulted in final sags of the 1033 ACCR conductor at 200 C that were comparable
(within a 0.5 m variance) to the final sags of the 1351 ACSR at 100 C; hence the
clearances to ground and crossing obstacles were not violated.
Construction Cost and Time. The construction cost of the 1033 ACCR option was
estimated at $2.32 M which is 12% less compared to the cost for the 1351 ACSS
option at $2.64 M, representing a savings of over $300,000. Unaccounted cost that
could also be associated with the 1351 ACSS option includes (a) loss of revenue
during outages (minimum 6 days total) on the three 500 kV lines during tower
installation; and / or (b) incremental increase in cost for alternative but more
expensive power generation or load transfer during line interruptions.
284
Tower 56B
Tower
Tower56C
56C
285
Tower 56A
Figure 4: Conductor in Span A crossing (2) 500 kV transmission lines and a roadway
286
Another significant benefit of selecting the 1033 ACCR option was the time savings
that resulted from eliminating the design, manufacture and installation of new towers.
The 1351 ACSS option with the five new towers was expected to take a longer time
for the design, tower supply and construction; and would also require a higher
number of interruptions necessary for installation, including outages on the three 500
kV lines.
The selection of the 1033 ACCR conductor for the Gorgas to Miller Scrubber
transmission line job was therefore based on; (a) the adequacy of the conductor in
meeting the technical requirement of the project; (b) the overall lower installed cost
and shorter construction time; (c) the ability to utilize the existing towers; and (d) that
no interruption or modification to the 500 kV transmission lines would be required.
INSTALLATION OF THE 1033 ACCR
The construction contract for the job was awarded through the normal bidding
process. Presentations were made by APCos construction supervisors and the 3Ms
installation representative to familiarize the prospective bidders with the particular
requirements of the new ACCR conductor. After the bid selection and contract award
processes were completed, follow up meetings were held between the successful
bidder, 3M and APCo in working out the construction schedule and implementation
strategies in more details. 3Ms installation representative would be on the job site for
the entire construction period.
In consultation with the Alabama Control Center (ACC), the GMTL had to be
available for the summer period from June 1 to August 31, 2008. Therefore, the
implementation plan was to install the new ACCR conductor on the existing
structures up to the point where the relocation started, as the first phase of
construction. This phase of the job was done during the period January to May 2008.
The line would be returned to service June 1 to August 31 and then the relocation
section was done September 1 to December 31, 2008 as the second phase of
construction.
The installation of the 1033 ACCR required a few specialized tools and equipment
including a 100 ton press, a 10,000 psi pump and high temperature compound for the
compression dead end clamps. In order to minimize the bending radius of the
conductor, a bull wheel tensioner of 180 cm (72) diameter or larger was
recommended. Roller array blocks were used at vertical and or horizontal angles
exceeding 30, as well as break-over angles. Distribution Grips were exclusively used
to grip the ACCR conductor and could be re-used for pulling up to 3 times before
discarding.
The construction crew experienced no major problems with the installation of the
ACCR conductor. With the appropriate on site training and proper coordination at all
levels, the job was safely executed and both phases were completed ahead of the
planned schedule.
287
CONCLUSION
Given the challenges of working in a highly congested area involving some critical
230 kV and 500 kV transmission lines, the 1033 ACCR solution provided an ideal
application of the new 3M high temperature, low sag conductor on existing towers in
a cost effective manner.
The successful installation of the ACCR conductor was primarily due to the
professional and team work approach of the 3M representatives, the contractor,
APCos inspector, construction supervisors and transmission line designers and the
ACC.
Our ability to meet the system requirement of increasing the line capacity and
satisfying the load demand of the Miller Scrubber facility will serve to ensure the
future reduction of emissions and their impact on the environment.
REFERENCES
PLS-CADD, (1992), Power Line Systems Computer Aided Design and Drafting,
Power Line Systems, Inc., Madison, Wisconsin.
PLS-POLE, (2000), Analysis and Design of Structures with Wood, Laminated Wood,
Steel, Concrete and FRP Poles or Modular Aluminum Masts, Power Line
Systems, Inc., Madison, Wisconsin.
288
Staff Engineer, Entergy Services, Inc., 120 W. Mayes St., Jackson, MS, 39213, 601-368-2011,
[email protected]
2
Manager, Transmission Line Design, 2121 38th St., Kenner, LA, 70065, 504-463-2974,
[email protected]
289
290
Necessity, which includes a routing study and Environmental Impact Statement. The
northern third of the proposed route passed through a number of industrial properties
located southeast of Port Arthur, and continued south along Texas State Highway 87
for the remainder of the distance. A portion of the route had to be reassessed with the
proposed addition of a second LNG facility to be operated by Sempra Energy. The
second LNG will require construction of future Round Lake Substation located
between Port Arthur and Sabine Lake. The location of the new substation required
that the line be moved approximately seven tenths of a mile westward into the marsh.
Highway 87 is expected to follow the same route around the proposed facility, but
relocation of the highway has not yet occurred. A map of the route and surrounding
area is presented as Figure 1.
Initially there was considerable
resistance to the proposed route
from a number of sources. As
part of the permitting process, a
number of design constraints
were introduced to the project,
including limiting the height of
the proposed line and installing
devices on the wires that would
make them more visible to the
birds in flight.
With the
assistance of Thomas & Betts,
numerous line and structure
configurations were evaluated to
address feasibility considering
the project constraints. PUCT
approval of the selected pole and
line configurations came in
August of 2006.
o
o
o
Port
Arthur, TX
o
o
o
o
o
J. D.
Murphree
WMA
o
o
o
Future
Round
Lake
Substation
o
o
o
o
Denotes
Approx.
Soil Boring
Location
o
Additionally the owners of
industrial facilities on the north
o
o o o
end of the route voiced numerous
concerns
regarding
pole Sea Rim
o
placement, structure type, soil State
Keith Lake o o o
sampling, means of access, etc. Park
o
Substation
These concerns resulted in
negotiations that were not
Figure 1 Project Map
satisfactorily resolved until mid
2007. Soil data along industrialized portions of the route became available in
September, allowing for final designs and procurement of poles and steel caisson
foundations by December. Thomas & Betts received the order to supply the
galvanized steel poles and caissons. Project construction was completed between
April and July of 2008.
291
Soil Testing
A total of 30 soil borings were
performed in two phases to support this
project.
While numerous natural
variations in stratification and soil
properties were observed, soil borings
generally conformed to one of four
simplified soil profiles as shown in
Figures 2.
Initial borings were performed in June
of 2005 during preparation of project
estimates, and consisted of nine borings
along the southern two-thirds of the
proposed route.
Several existing
borings at the Port Acres Bulk
Substation on the north end of the
project were also available. At the time
the initial project estimates were
prepared it was assumed that the line
would be located just off the highway
right-of-way, although the Port Arthur
LNG (Sempra) was known to be in the
planning stage.
The second phase of collecting data
occurred concurrently with final design.
At this point the route was altered to
follow the future path of Highway 87,
and serve a second (future) substation
added for Sempra's proposed LNG
facility. Agreements had also been
reached with the owners of most of the
industrial properties allowing soil
borings and soil testing along most of
the northern third of the route. Six
additional borings were planned, but
never executed due to a lack of a signed
agreement with the property owner.
The approximate locations of soil boring
locations are also shown in Figure 1.
Near-surface soils showed the most
variability. On the north end of the
project and on the extreme southern end
0 7 water
Soft
Organic
Clay
c 0.3 ksf
d 70 pcf
M. Stiff
Clay
c 1.0 ksf
d 95 pcf
30
Stiff Clay
c 1.5 ksf
d 99 pcf
60
Figure 2(a)
North End of
Project
35
M. Stiff
Clay
c 1.0 ksf
d 95 pcf
60
Figure 2(b)
Refinery Ponds
& Grassy Marsh
0 2 water
Soft
Organic
Clay
c 0.3 ksf
d 70 pcf
10
Soft Clay
c 0.4 ksf
d 85 pcf
50
M. Stiff
Clay
c 1.0 ksf
d 95 pcf
90
25
60
Figure 2(c)
Marsh near
Waterways
M. Stiff
Clay
c 0.8 ksf
d 85 pcf
M. Stiff
Clay
c 1.0 ksf
d 95 pcf
Figure 2(d)
South End of
Project
of the project, the near surface soils tended to be medium stiff consolidated clay. In
between they tended to be soft unconsolidated clay with considerable organic content.
The water table was generally within a few feet of the surface, and in many instances
the ground surface is submerged for much of the year. Structures crossing the
refinery freshwater storage ponds were typically in three to seven feet of water.
Except at the north end of the project, unconsolidated clays with a low cohesive
strength and a high organic content underlay the near-surface soils.
Properties of unconsolidated clay were reasonably uniform with the exception that the
thickness of the layer varied considerably in depth. A layer of consolidated medium
stiff, to stiff clay was encountered at the bottom of most of the borings.
Shells and thin seams of loose sand were encountered at varying depths in a few
locations. Blow counts were generally low, and in only one location did they
approach values that might indicate some difficulty in driving a caisson foundation
through the layer.
Limitations on Design and Construction
Except for a few locations on the north and south end of the project, initial soil data
demonstrated that deep foundations would be required for almost all structures,
including the tangent structures. Wetlands permitting did not allow excavation or fill.
Similarly, negotiated access through industrial properties on the northern section of
the route did not allow for excavation or removal of soil (apart from soil testing
samples where this was eventually approved). Additionally, the agreements with the
industrial concerns required foundation footprints to be small thereby preventing
guying, large reinforced concrete pile caps, etc. Consequently, driven large diameter
steel caisson foundations were identified as the most viable solution.
Additionally, access for much of the route was severely restricted, and would have
required extensive use of matting, construction of temporary roads, and other means
of access. Given the difficulty in moving concrete and other materials along the
right-of-way, it was determined that helicopter construction would be cost effect
along many portions of the route. Consequently, base-plated steel caisson
foundations were selected to facilitate installation by helicopter. Erickson Air Crane
(EAC) was consulted, and based on expected availability of their S-64E aircraft; a
target lift weight of 16,000 lbs was established. The S-64E has a maximum hook
weight of 20,000 lbs, but 4000 lbs was reserved for fuel, rigging and additional forces
exerted from the rotor-wash acting downward on the component being lifted. With
the assistance of Thomas and Betts (T&B), numerous transmission structure and
foundation configurations were evaluated against the target lift weight.
The ability to erect the foundations and structures using helicopters required use of
some special details to assist in alignment, and placed some restriction on the
structural details being considered. These aspects of the design required a great deal
292
of input from EAC and T&B. One of the more typical alignment guide details is
shown as Figures 3.
One of the concerns initially expressed by
operators of the Valero refinery was the
potential for excessive disturbance of
sediment in the freshwater intake ponds
used to supply process water to their
plant. That and the relatively shallow
pond depths eliminated the potential to
work that section of the line from heavy
barges. Access to that area with heavy
equipment was limited due to numerous
underground pipelines and other utilities
that would have to be crossed, and by
concerns for the stability of the pond
levees.
Finally, apart from the navigable waterway crossing structures, the Texas Parks and
Wildlife Department limited structure height to 85 feet, for protection of birds, on the
portion of the line adjacent to the Sea Rim State Park, and the adjacent J. D.
Murphree Wildlife Management Area. Although this did not directly affect the
foundation design, it was recognized early on that foundations would have to be
driven to their full depth regardless of the resistance encountered. Consequently
every effort was made to obtain adequate geotechnical data so as to minimize the
potential for difficulty in driving caisson foundations.
Foundation Design
This line and its structures and foundations were designed considering a 140 mph
wind speed. This design wind speed is determined by the load requirements in the
National Electrical Safety Code. Two and three pole structures were used where
possible to divide the foundation loads and allow foundations to be as small as
possible.
The foundations were designed using MFAD.
Steel caisson foundations were
modeled as piers, with stiffness of the caisson foundations determined and entered
manually. Where feasible foundations were designed: to maintain a maximum
ground line deflection of 3; a maximum foundation rotation of 1.5 degrees; a nonrecoverable ground line deflection of 1; and a non-recoverable foundation rotation of
1.0 degrees. In marshy areas with surface soils contributing little to the foundation
strength, the targeted limits were 6, 3 degrees, 2 and 2 degrees, respectively. To
accomplish these limits, and maximize foundation stiffness, diameters were selected
that were generally 6-7 inches larger than the anchor bolt circle. The typical pole to
foundation connection is illustrated in Figure 4.
293
Many of the foundations on the northern end of the project, including five large angle
dead-end structures, had to be designed using data extrapolated from available
geotechnical data. Six additional borings were planned between Port Acres Bulk
Substation and the northernmost crossing of Taylors Bayou, but the borings could
not be obtained. Foundations were designed considering the weak soils encountered
at Taylors Bayou rather than the comparatively strong soils encountered at Port
Acres Bulk. This was done to with the intent of producing the most conservative
design possible, but ultimately did complicate installation of the foundations.
Naturally, caisson lengths varied due to changes in soils and loading, but total lengths
for tangent structures generally ranged between 30 and 45 feet. Total caisson lengths
for dead-end structures typically ranged between 50 and 60 feet, but approached 70
feet in some cases.
The target weight of steel caisson foundations was not possible to achieve in all cases.
Many of the foundations for dead-end and navigable waterway structures had to be
installed conventionally. Fortunately, access was generally available in these areas.
Several other caissons had final weights in excess of 17,000 lbs. There was a
helicopter staging area relatively close to the structure locations, so these caissons
were able to be flow to their intended location and installed by helicopter. But the
additional weight came at the expense of fuel weight and affected flying time and
overall installation efficiency.
After initial pole and steel caisson foundation designs were complete, it was
determined that two of the 2-pole dead end-end structures would in all likelihood
need to be installed by helicopter due to access limitations imposed by the industrial
facility crossings involved. With no way of reducing the pole or caisson sizes at this
point, it was decided to fabricate these caissons in two pieces and weld them on site.
294
Foundation Construction
The complexity of permitting, Right Of Way
acquisition and design, and a rigid in-service
date, resulted in a very aggressive construction
schedule.
Conventional construction and
helicopter construction were conducted in
parallel to the extent possible to make best use
of time. In general, construction by helicopter
was done through the marsh and through the
refinery ponds. Conventional construction was
planned on the north and south ends of the
project, and included construction of the
navigable waterway crossings from barges. As
stated previously, not all foundations could be
designed to meet the lift limits of the
helicopters. Consequently, these structures had
to be added to the scope of the conventional
construction crews.
Figure 6a Matting
295
controllers.
Ground crews were typically
transported via marsh buggy to the pole location
prior to the flight. In marshy areas, the initial
lifts at each location consisted of timber matting
used to support the hydraulic unit and hammer
(figure 6a). At the refinery ponds crews were
allowed to bring in small modular shallow draft
barges that were secured together and served as
work platforms and support the hammer and
hydraulic unit (figure 6b). Following this, the
hammer and hydraulic units were flown to the
site. The hydraulic unit consisted of two
sections (figure 7a), a base section containing
the hydraulic components and an upper clam
shell section used to plumb and secure the
caisson foundation during driving. The two
sections had to be flown to pole locations
separately due to weight limitations. The
hammer was also selected considering the
helicopters capabilities, and was an American
Piledriving Equipment (APE) model 150.
With the hydraulic unit in place, the caisson
was flown to the site and secured using the
upper section of the hydraulic unit (figure 7b).
The driving hammer was then lifted in place
and the caisson driven (figure 7c). Generally
speaking,
caissons supporting tangent
structures were driven without difficulty due to
their comparatively small diameters and depths.
However, in a few tangent locations and in most
angle and dead-end locations, driving resistance
was considerably greater than expected,
primarily due to a lack of familiarity with the
light hammer being used. While most caissons
were eventually driven with the model 150, in
some cases driving times were considerable. In
several of the more accessible locations, a crane
and a model 200 hammer had to be brought to
the site to complete caisson driving operations.
Fortunately, dry weather preceded construction
allowing better access that might otherwise
have been possible.
As was expected, the most problematic of the
foundations during construction were the two-
296
297
(including stringing dollies) were flown to the site in one lift where possible (figure
8b).
Slip joints of two part poles were secured together using steel lift straps. Where poles
were too heavy to fly in one lift, alignment guides like those shown in figure three
were used to align the sections.
Light helicopters flown by Air2, LLC assisted with installing the pull cables through
the dollies; and were also used to assist in clamping in the conductors (figure 8c).
Following installation of conductors, Air2 also installed bird diverters to shield wires
in locations where they were required to meet Texas Parks and Wildlife Department
requirements.
Acknowledgements
Assistance provided by the following companies was instrumental in the successful
completion of this project: Air2 (aerial conductor installation); Erickson Air Crane
(aerial foundation and pole installation); Fitz & Shipman (surveying); New Park
(matting); North Houston Pole Line (conventional pole and caisson installation); S.
Bomack (barges and cranes); Stork Southwestern Laboratories (geotechnical testing);
T&D (conventional pole and caisson installation) Thomas & Betts (pole and steel
caisson foundation design and manufacturing); Turner Industries (on-site welding);
and QC Laboratories(geotechnical testing).
The cooperation received from the VALERO refinery was particularly crucial to this
project, and the assistance provided by Randy Lowe and Steve Clegg is greatly
appreciated.
Conclusions
1. Emphasis on caisson driving during foundation design is important, but it is
extremely important for caissons driven using helicopters. Prior to this
project, Entergy had very little experience using the comparatively smaller
APE Model 150 driving hammer. Consequently, much of the difficulty
encountered in driving was unanticipated.
2. For helicopter installation greater standardization of vibe-hammer driving ears
is warranted. Bolt circle variations resulted in variation in driving ear spacing.
Frequent adjustments had to be made, and often the hammer had to be flown
back to a staging area for adjustment. This affected Helicopter fly-time, and
overall efficiency. Additional material costs associated with holding the
driving ear spacing constant would have been money well spent.
3. The two-piece steel caisson foundation was an acceptable solution, and
necessary in this instance. But the solution came with numerous challenges
and will only be repeated as a last resort. The large caissons could not be
driven with the helicopter, and it was very fortunate that conventional
298
299
Wood County Electric Cooperative, PO Box 1827, 501 S. Main St., Quitman,
75783; PH (903) 763-2203, email: [email protected]
2
DiGioia, Gray & Associates, LLC, 570 Beatty Road, Monroeville, PA 15146: PH
(412) 372-4500, email: [email protected]
3
PBS&J, 18383 Preston Rd, Suite 110, Dallas, TX 75252; PH (972) 588-3158, email:
[email protected]
ABSTRACT
Wood County Electric Cooperative is serving a new pumping station for the Dallas
Water Utilities on Lake Fork in eastern Texas. The optimum route required a lake
crossing of over 1,738m (5,700 ft) for a double circuit 138 kV line. Lake Fork is an
internationally recognized bass fishing environment, holding the Texas record for the
largest bass, weighing over 8kg (18 lbs). The placement of the foundations for the
crossing structures presented several unique challenges including limiting the number
of structures set in the lake, providing clear channels for boating traffic, minimizing
environmental impacts and working in lake depths over 15m (50 ft). This paper
discusses the types of foundations considered and the design and installation of the
most effective design.
Project Description
Surface reservoirs for water have been built in east Texas for the past 30 years as
preparation for future potable water needs in the expanding Dallas, Texas metroplex.
In 1975 the Sabine River Authority dammed a branch of the Sabine River and created
a water storage area known as Lake Fork. Until recently the lake has served as a
recreational area, probably best known internationally as a prime tournament class
bass fishing habitat. The top 6 bass and 35 of Texas' top 50 bass caught have been
taken from Lake Fork. The lake is poised to enter its primary function as a source of
potable water for Dallas via construction of a 2.75m (108 inch) diameter pipe-viaduct
transferring water to Lake Tawakoni and eventually to the Dallas municipal water
system. The 63,360 kL/D (240 MGD) pump station will be powered by a substation
connected to a dedicated 138 kV line. Part of the transmission line supporting the
substation will cross the lake, a distance of over 0.6 km (1 mile) in length, see Figure
1. Three transmission structures were installed in the lake, supported by foundations
placed at depths of 15-21m (50-70 ft) of water.
The transmission assets are owned and operated by Wood County Electric
Cooperative (WCEC). The proposed line also provides an integral part of a loop
connection between two substations that were radially fed. The construction of the
lake crossing provided the final segment of the loop to afford greater reliability to the
WCEC service area and the pumping station. The land-based transmission line
section was constructed with double circuit 138 kV structures on concrete poles. The
300
design of the lake crossing posed several unique geotechnical and construction
challenges.
Figure 1 - Project Area Map Showing Lake Fork and Transmission Lines
301
Design Considerations
After the selection of three structures was established, a construction
process/foundation design practicality study was initiated. The type of foundation, its
size and costs would be driven by the structure loading requirements and the
subsurface conditions at the structure sites. The line route was mapped using a
hydrograph to determine the lake bottoms contours. The lake elevation was
established by maximum pool elevation and a desired ten foot reveal to the base of
the poles. See Figure 2 which shows the hydrograph data as a profile for design.
For a preliminary design of the foundations, several sources of existing boring data
were found that provided beneficial insight into the geology of the area. The Sabine
River Authority (SRA) shared with the project team the borings they had used in
designing the earth retention dam that formed the lake. SRA also had monitoring
wells installed along the toe of the dam to detect water leaching and undermining of
the structure. In exchange for their sharing of the dam boring data, the project shared
its lake bottom hydrograph cross-section with the SRA. They had several cross
sections but the crossing area offered them an additional data point for approximating
the true retention volume of the lake. This was valuable information to their river
management responsibilities.
Additional boring information was obtained from the Texas Department of
Transportation (TXDOT. Less than a kilometer upstream of the transmission line
crossing location, TXDOT had recently upgraded a bridge across the lake and they
shared their boring information with the design team.
Following several preliminary designs of the foundation size, the diameter and
embedment depth were designed to support the conductor and structure loads at the
base of the steel pole and the pedestal length of the structures length from pole
base to lake bottom. The most practical installation was a tubular steel caisson that
would be driven into the lake bottom a sufficient length and then extends to 3m (10
ft) above the maximum pool elevation. The actual pole structure would be designed to
be socketed into the top caisson section, aligned and grouted in place. Installing the
caissons into the lake bottom could have been accomplished with one of three
techniques. The boring logs showed that the lake bottom consisted of a dense sand
and clay mixture for a depth of over in most locations. There was no bedrock or
302
boulders that would interfere with the placement. The potential placement techniques
are addressed below with their perceived pros and cons for the Lake Fork project:
Vibrated Caisson the caisson would be lowered into position on the lake bottom and
then vibrated with a vibratory hammer to the desired depth. The lack of rock made
this an attractive alternative. The amount of clay and density of the sands were
concerns relative to the ability to vibrate the caisson to the desired depth. The skin
friction that would detract from this methodology could be mitigated by using an
auger to remove some of the spoils inside the caisson and relieve the friction forces.
Advantages of the vibrated caisson were the simplicity of the operation (relative in
comparison to the others; not in reference to working in 15-21m (50-70 ft) of water
and the minimal disturbance to the lake bottom.
Jetting using water pressure and jets at the bottom of the caisson, water injection
disturbs the soils sufficiently to allow the caisson to be advanced to its embedment
depth. This option presented a major effect that was an immediate detractor the
water injection would create a sizable plume of lake bottom material being disbursed
into lake water. This was deemed a major environmental concern and a significant
potential impact on the fishing habitat. We mentioned the notoriety of Lake Forks
bass fishing, did we mention that the construction window was overlapping or in
close proximity to an international invitation bass classic tournament. Both WCEC
and the SRA were ever mindful of the fishing habitat impacts.
As you can guess, the vibratory installation method seemed most practical and
environmentally acceptable.
Lake Features Shaped Project
Two additional lake factors played into the project execution: lake elevation and
winds. North and East Texas had been experiencing several years of drought and lake
elevations were low when the project was being designed. The lower lake elevation
increased the issues concerning the still standing timber in the lake; would the barges
and working boats be able to maneuver to the structure locations and launch areas.
During the season that the line crossing was actually built, rainfall had raised lake
elevations above normal and near full pool elevations. This condition increased the
water depth that the crews had to work in placing the foundation caissons.
Paradoxically, the water elevation could be a problem if it was high or low as noted
above plus the need to reach the exposed caisson top above the lake elevation. If the
lake was high, the barge work was in greater depth. If the lake elevation was down,
the water depths were better for working but access to the caisson top and the impact
of insitu trees were both more of a concern.
In Figure 2, the lake bottom profile shows the west shore on the left. Note that the
lakes deep channel is on the east and that the west shore is shallower. The barge
ramp and access was from the west shoreline near the crossing location. The issue of
the standing timber in the lake was a concern in the shallower area where the top of
the trees that were 7-11m (25-30 ft) tall when growing, would now be close to the
surface.
303
The open exposure and length of the lakes open water created windy conditions that
again made barge working difficult during the hydrograph, soil boring and
construction period. Several barges were tried until one sufficiently large and stable
enough provided a platform for the soil boring efforts. The construction crews were
familiar with working on lakes, setting transmission poles and they had the proper
equipment for executing the foundation installation.
Crossing Design
Each end of the transmission line crossing was anchored with a self supporting twopole deadend structure. The east shore structure was an angle structure. The west
shore structure had one circuit continue on the tangent line towards another WCEC
substation. The pump house feed circuit deadended on the west structure and dropped
into a dedicated substation. Figure 3 is a photograph of the west shore termination
with the two-pole structure and the pump house under construction adjacent to the
line.
The three structures located in the lake required a minimum of 24m (80 ft) long
tubular steel caissons that were embedded into the lake bottom a depth of 9m (30 ft)
as determined by the foundation design program MFAD. A 50m (165 ft) tall double
circuit steel pole structure was socketed into each caisson. The caissons were
galvanized, and then coated with corrosion preventive coating along the length of the
caisson that could be exposed to air as the elevation of the lake changed seasonally.
Construction Summary
The construction process was uneventful. The caisson vibration process went
smoothly. None of the caissons required internal auguring; the caissons were vibrated
to their prescribed depth. The caisson locations had been marked prior to construction
304
with buoys. When the barge with the caissons were brought to the buoy and jacked
into position, see Figure 4. The specific pole location was spotted from the shoreline
by survey. Then the steel caisson was stood vertically and lowered to the lake bottom
and held in position by a template frame on the edge of the barge, Figure 5. The
vibratory hammer was attached to the top of the caisson and was driven to depth, see
Figure 5.
Once the caissons were driven to depth, they were filled with low strength concrete to
the elevation of the socket plate that was dropped into each caisson to serve as a base
to support the pole base section while it was grouted in place, (Figure 6). The original
construction specification called for filling the caissons with crushed aggregate to the
plate elevation. During construction, however, the concrete backfill was found to be
more efficient and less costly than dealing with aggregate handling from the shore
onto the barge and then from the barge into the standing caisson. The barge then
brought the pole to the location and the base was socketed inside the pile with grout.
The three pole base sections were socketed in place before the crew came back and
set the top pole sections with attached arms and insulator strings attached. Helicopters
were used to pull in a lead line and the conductor was installed. To minimize phase
spacing and avoid galloping conditions, the conductor was damped. The dampers are
visible in Figure 7, where the line is displayed looking west to east.
305
306
Conclusions
The development of the Lake Fork crossing was built through a delicate coordinated
effort between the project owner, Wood County Electric Cooperative, the Sabine
River Authority and engineering staffs at PBS&J and DiGioia, Gray & Associates,
LLC to provide a lake crossing that met all code and engineering requirements plus
minimalized the project's impact on the environment, which in this case was a worldclass bass fishing habitat. Through the cooperative effort, the preliminary
optimization of the design strategy was accomplished by sharing geotechnical
information, environmental and recreational constraints presented by the lake and
construction experience of several transmission line contractors who frequently erect
lake crossings. The collaborative information quickly identified that vibrated tubular
steel caissons would be the optimum foundation option to support the three lake
crossing structures. Pole site geotechnical information obtained from borings
provided the detailed information along with pole loadings to design the caisson
foundations. Finally, construction of the crossing proceeded without events, as the
preparation in design and collaborative efforts with the Sabine River Authority
provided optimum lake access and efficient installation.
307
308
flow
Figure 1. Skagit River and transmission right of way. Note transmission right of way
crossing the left bend in the river and the proximity of two towers to the river.
Seattle City Light aerial photograph
The Shannon and Wilson study, after reviewing prior measurements and historical
aerial photographs, found that the rate of erosion had increased more rapidly in the
recent past. From 1997 to 2001 the average rate of erosion was 2.6 ft per year and
from 2001 to 2004 increased to 13.3 ft per year. The Shannon and Wilson report
concluded that tower stability could be reduced in about five years.and could
conceivably occur within one to two years Reducing risk to the transmission
structures was elevated to a high priority in the utility.
309
The four transmission lines supported by the two towers supply about one-third of the
power to Seattle. Loss of these lines would have serious lifeline consequences and
cause widespread outages critical to public safety, government response and law
enforcement, hospitals, sewer and water operation and last for an undetermined
period of time. The loss of these lines would also cause regional constraints on the
power system beyond the Seattle City Light service territory. The probability of loss
of these towers due to erosion was greater than the occurrence of a design seismic
event typically assumed in building codes.
The original lattice towers were 152 feet tall. The first alignment was constructed in
the 1930s and the second in the 1950s. The 1930 foundation is a cast in place
battered column on a 5ft x 5ft pad buried about 10 feet. The 1950s foundation was a
steel grillage foundation buried approximately 8 feet.
310
Wild and Scenic River watershed. Also the length of the rip-rap would extend well
beyond the 200 foot width of the transmission right of way onto private property and
would also require periodic inspection and maintenance. A sheetpile wall around the
lattice towers was also considered. This would create an island supporting the lattice
towers if the riverbank were to erode significantly in the future. The sheetpiling
would have to be driven well below the bottom of a future channel. Driving the
sheet through dense gravel and cobbles could potentially make installation difficult.
And the cost of the sheetpiling was estimated to be similar or greater than a drilled
pier solution, another option. Moving the towers further from the river bank dictates
that the towers be taller, following the rising shape of the catenary curve. The
maximum working tension already was 14,500 lbs so increasing the tension to offset
increased structure height was not considered. Moving the lattice towers further from
the river and raising them or putting the legs on pedestals was considered too. But
even so it would be difficult to guarantee what would be a safe distance from the
river. The selected solution was to use steel monopoles on drilled pier foundations
located nominally further from the river to minimize the increased height. The new
spans would increase by 60 feet to 1478 feet. The drilled pier foundations would act
as bridge piers in the future condition that the river were to erode the bank and
surround the piers. This would leave the concrete pier approximately 40 ft above the
bottom of the river and require a very deep foundation.
DESIGN LOADING
The National Electrical Safety Code (NESC) weather conditions are generally used as
the design criteria at Seattle City Light and they have performed well over the years,
but as is clearly stated in the first part of Section 1 they address a limited set of load
conditions.
The purpose of these rules is the practical safeguarding of persons during the
installation, operation, or maintenance of electric supply and communication
lines and associated equipment.
These rules contain the basic provisions that are considered necessary for the
safety of employees and the public under the specified conditions. This Code
is not intended as a design specification or as an instruction manual.
As loading of the structures was considered it became clear that this was a unique
situation and not all loading was covered in the NESC or ASCE Manual 74. NESC
Heavy, NESC Ice and Wind, NESC Extreme wind, and broken conductor criteria are
typical SCL transmission design loadings. Other loadings considered were a 100 year
flood event, earth pressure on one side of the foundation and a seismic event.
Flood
A 100 year flood event was selected as the design flood event because of the critical
nature of the transmission line and the effort to establish a durable solution to the
erosion problem. The 100 year flood event estimated by The Corp of Engineers is
103,000 cfs. To estimate the average velocity of the Skagit River at the transmission
311
crossing a survey of the channel bottom profile was conducted and the flood stage
elevation was determined. Both the American Association of State Highway and
Transportation Officials (AASHTO) Standard Specifications for Highway Bridges
and ASCE 7-02 were consulted for guidance for flood design loading. AASHTO
guidance for converting velocity to a vertical pressure distribution against a bridge
pier was followed.
2
Pavg = K (Vavg )
Pavg = average stream pressure
K = 0.7 for circular piers
Vavg =average velocity of water
in ft/sec.
Q
Vavg = 100 , calculated for the
A100
100 year event.
Q100 =cfs flow for 100 year event
A100 =cross sectional area based
on survey
AASHTO directs that the pressure against the pier shall be taken as a triangular
distribution with the pressure at the top of water equal to twice the average pressure.
Pmax = 2 Pavg
AASHTO also requires that the effects of drift buildup shall be considered. Two
scenarios for drift buildup were considered. The potential exists for a debris pile or
log jam to develop against the foundation pier and if the river were to erode a
relatively small distance beyond the pier a debris pile or log jam potentially could
form in this space. The flood loading used for design was a 100 year flood event
against such a debris pile pushing on the foundation.
The design also considered a large log carried in a flood event striking the pier.
ASCE 7-02 contains a section in the appendix that discusses the impact loading from
water borne debris and suggests for the Pacific Northwest that a 4 kip log can be
considered typical. The following expression for impact force is presented.
F=
312
CB = blockage coefficient
Rmax = maximum response ratio
for impulsive load
g = acceleration due to gravity =
32.2 ft/s2
t = the impact duration
Loads were calculated for a log of this size striking the pier in the 100 year flood
event.
Scour
The geo-technical investigation was asked to provide an estimate of scour depth
around the pier. Scour depth determination was based on the empirical formulas of
the Colorado State University equations for piers found in the Integrated Streambank
Protection Guidelines 2003. This additional depth was considered when calculating
the foundation design forces
Soil pressure
Another design case considered was the situation where the erosion might progress
only to the point of removing soil on one side of the foundation leaving the
foundation pier subject to earth pressure.
Seismic
Transmission structures are generally not designed for earthquake forces. For most
steel monopoles on drilled pier foundations the top of the foundation is at or near
ground level and seismic forces do not control the design of the pole or foundation.
ASCE Manual 74 (Appendix F) states that:
Typically, transmission structures are not designed for ground induced
vibrations caused by earthquake motion. The standard transmission structure
loadings caused by wind/ice combinations and broken wire generally exceed
design earthquake loads.
ASCE Manual 96 states that:
Transmission lines have been very resistant to earthquake damage: their
main vulnerabilities are foundation failure of transmission towers or the loss
of a tower due to landslides. Both occurrences are relatively rare. It would
appear that the low natural frequencies of lines decouple their mass from the
high energy content of earthquakes, and the design for extreme wind, ice, and
longitudinal load combinations is adequate for earthquakes.
There is minimal guidance in the design literature for seismic design of transmission
line structures and foundations. The Appendix to Chapter 14 of the National
Earthquake Hazard Reduction Program (NEHRP) 2000 Edition of the Recommended
Provisions for Seismic Regulations of New Buildings and Other Structures includes
transmission structures as non building structures and provides the most explicit
313
314
E
0.670 g
0.223 g
1.36
3.12
Cs
W
R
I
The equivalent static force approach was followed assuming two concentrated
lumped masses representing the steel pole and the concrete pier above the river bed.
In addition a modal analysis was done using the RISA3D software. In the
Commentary to Chapter 14 of the 2000 NEHRP/FEMA 369 provisions a damping
value of 2% is recommended. The results were scaled up to 85% of the equivalent
static force per the 2000 NEHRP provisions and distributed to the steel pole and to
the exposed foundation.
Summary of loading:
60 degree F, no wind
NESC Heavy
NESC Extreme wind
NESC Extreme Wind and Ice
2 broken conductors
Normal river flow
100 year flood event
100 year flood event with debris pile/log jam
Earth pressure on one side
Seismic event
LOAD COMBINATIONS
There are a number of codes and design guides to turn to for load combinations but
none address the range of loadings that include wind and wire tensions, a significant
flood event and seismic activity. Building codes and highway codes in a reciprocal
fashion dont address transmission wire tension and wind design loads. Transmission
line design guides typically dont include provisions for seismic and flood loads. To
proceed to design however reasonable load combinations had to be determined and
will be briefly discussed here. Wind direction was assumed to align with the
hydrostatic pressure of river flow. Flood events in the Pacific Northwest usually
occur with relatively warm wet weather often in conjunction with melting snow. For
this reason flood events were not combined with the NESC ice conditions. A high
wind event occurring simultaneously with a flood event or seismic event is not likely.
The probability of a design seismic event and a flood event occurring simultaneously
is small. All river flows, typical and extreme were assumed to act against a debris
pile recognizing that if a debris pile were to form it could remain in place for some
time. If erosion were to occur on only one side of the foundation pier earth pressure
against the pier could be sustained for an extended period and coexist with the NESC
weather conditions. Following is a list of load combinations that were considered for
design.
Load combinations used for design:
60 degree F wire tension, no wind and normal river flow
NESC Heavy and normal river flow
NESC Extreme wind and normal river flow
315
The seismic load combination was the governing combination for foundation design.
The foundation and pole designs were considered without the interaction of the
conductor stiffness because of the lack of readily available information and the
additional sophistication required of such an analysis.
Soil borings were taken in January of 2006. Alluvium, or soil deposited by rivers,
was encountered over the full 98 foot investigation depth. The upper 25 ft consisted
of very soft to medium stiff silt. Gravel, gravel with silt and sand, and sand with silt
and gravel were encountered to 38 feet. Sand and sand with silt was encountered to
the bottom of the borings. Ground water was encountered at 20 feet. Scour depth
was estimated at 8 foot for a 10 foot diameter pier. Liquefaction analysis was
conducted for a 475 return period earthquake. The risk was found low so no
mitigation was recommended for liquefaction. Foundation design software, L-Pile,
parameters were provided for foundation design. The size of the pier foundation is 90
feet deep from existing ground elevation and 10 feet in diameter.
Site class E soils identified in the geotechnical report require that the pier design meet
the detailing requirements required in higher seismic regions. Originally the
confinement reinforcing was called out as continuous spiral reinforcing but the timing
316
317
318
319
320
2 of 10
321
Route Description
The Middletown Norwalk 345-kV Transmission Project traverses the state of Connecticut
from near the center (Middletown area) to the Southwest corner (Norwalk area).
Between Middletown and Norwalk, the project passes though 18 separate municipalities,
with the overhead portion of the transmission line extending from Middletown to the
town of Milford. The entire line winds its way though rural, suburban, and urban areas.
With the exception of small portions of the line, the new route follows existing corridors
of 115-kV and 345-kV lines. By making use of several different pole configurations and
by replacing and upgrading large portions of the existing lines, the final route minimized
the amount of disturbance and additional right-of-way required. In some instances small
bypasses were constructed to accommodate residential developments that have been
constructed along the corridors. With the expanding development in the area, right-ofway access and expansion became a large coordinated effort. Typical of transmission
line projects, construction access also became an issue.
Terrain and Soil Conditions
The first item to consider in foundation design is the condition of the soil in which the
foundation is to be placed. With a project of this size traversing 45 miles of ROW, there
are numerous variations in the soil conditions. The second issue is the size of the
foundations that would be needed. Burns and McDonnell subcontracted Haley and
Aldrich, a geotechnical engineering firm with offices in CT to perform soil borings at
each proposed structure location. These locations were based on the initial line designs.
As we will discuss, changes to these initial locations provided some of our greatest
challenges. Various all-terrain drilling rigs were used to perform the necessary borings
and collect soil and rock samples. Where required, small track-mounted rigs were used
to gain access to areas of steep terrain and heavy vegetation. Laboratory tests including,
grain size analysis, moisture content, unconfined compressive strength of soil and rock,
split tensile strength of rock, and soil chemistry analysis were performed to provide
pertinent design information. Using the information gathered from these soil borings,
Burns & McDonnell was then able to begin conceptualizing the types of foundation
designs that might be feasible. There was now significant evidence that most of the
foundations used would have to be placed in rock. However, given the rolling and
sometimes mountainous terrain the depth of the different types of soil and rock
sometimes changed between adjacent soil borings let alone the change that occurred
between miles of line.
As a generalization, testing results showed bedrock at slightly higher subsurface
elevations in Segment 1 of the project. Segment 1 consisted of everything North and East
of the Beseck Substation (seen in Figure 1). Bedrock in this area consisted mainly of
schist, gneiss, and amphibolite. Topsoil in this area was typically less than 1 ft thick,
with thin layers of Subsoil and Colluvium beneath it at maximum thicknesses of 3.5 ft
and 3 ft respectfully. Soil in this area consisted of sandy silt to silty sand with increasing
amounts of gravel in the Subsoil and beneath. Glacial till consisting of much denser silty
sand and gravel containing large cobbles and boulders was found below this. Bedrock in
3 of 10
322
the remainder of the project was slightly less shallow, and was composed of Sedimentary
and Volcanic rock. In these areas various amounts of fill material was encountered at a
maximum depth of 23 ft. Topsoil in the area was generally less than 5 ft in thickness
with Subsoil up to 5 ft in thickness as well. In some areas, wind blown deposits of loose
to medium dense silty sand and sandy silt up to 2 ft thick was found beneath the Subsoil.
Beneath the Subsoil and wind deposits more of the same Glacial till was encountered.
Foundation Options
There are numerous structure configurations supporting single and double circuit
conductors on structures of moderate to extremely tall heights. Additionally, various
required conductor tensions produced structures with a variety of loading conditions.
The loads that resulted from some of these structures necessitated extremely large
foundations. Drilled shafts are a commonly used foundation for tubular steel
transmission structures. Due to there design, drilled shafts tend to have the most capacity
for resisting large overturning moments from transmission line structures. In many
designs the drilled shafts make use of rock sockets near the base of the foundation to
provide the required capacity. In the case of Middletown Norwalk there was initially
significant concern expressed with the extremely hard and very shallow rock that existed
in some areas. With compressive strengths of over 30,000 psi, there was a concern that
the rock could not be drilled with conventional drill rigs. The designers and the
construction contractors both believed there could be problems drilling to the required
depths for drilled shaft foundations through considerable solid rock. Blasting was
considered but was ruled out due to the proximity to businesses and residences in some
areas as well as the potential for damage to the remaining rock that was needed to support
the foundation. In an attempt to decrease the depth of foundations, two additional
foundations types were considered. The idea being that in instances where extremely
hard rock was deep or nonexistent, conventional drilled shafts would be used; however,
when extremely hard rock was relatively shallow one of the other two designs would be
considered, depending on the specific depth of rock. Where more moderate strength rock
was encountered, even at shallow depth, conventional rock socketed drilled shafts were
still considered most economical.
The first of the alternate foundations was an anchored mat design. This foundation is
similar to a spread footing or mat. Instead of installing full length anchor bolts into a
thick mat or pedestal, a steel plate placed in the mat is used to transfer the loads from the
pole to grouted rock anchors attached to the foundations. These anchors extend from the
bottom of the mat and anchor the foundation to solid rock below. This keeps the size of
the mat to a minimum while resisting the overturning moments. These types of
foundations were intended to be used when very hard rock was extremely shallow, or
within a few feet of the ground surface. See Anchored Mat Foundation figure 2.
The second of the alternate foundations was an anchored drilled shaft design. In
instances where hard rock was found to be shallow enough to require a rock socket, but
not shallow enough to necessitate an anchored mat foundation, an anchored shaft could
be used. The concept of these foundations was that a standard drilled shaft foundation
4 of 10
323
could be designed to the depth of rock. In order to prevent excessive drilling into rock,
the foundation would be anchored to the rock with grouted rock anchors that protruded
from the base of the foundation. The rock anchors would span the height of the drilled
shaft portion of the foundation before extending though the bottom and into the rock.
The anchors were assumed to be prestressed to limit shaft deflection and rotation. See
Anchored Shaft Foundation figure 2.
324
Final Decisions
At this point a framework had been established regarding design concepts. With this
framework in place, preliminary design was performed to establish bid quantities. As
usual a large majority of the foundation designs incorporated the standard drilled shaft
approach. This left around 20% of the foundations, which fell within the parameters
initially lending themselves to one of the two alternate designs. Areas where rock was
present within a few feet of the groundline utilized mat foundation designs. The
remaining areas of concern used the anchored shaft design.
Even though the alternate foundations would ideally save time and money required for
the drilling of large quantities of very hard rock, there were several problems with using
these foundations discovered through the bid process. The first of which was testing. It
became obvious that large scale testing of the rock anchors would be required at each
structure site. The problem with on site testing is that the consequences of a test failing
would typically render the foundation useless. If one of the rock anchors did not hold its
required capacity and pulled out of the rock below, that foundation would no longer be
sitting in a viable location. In turn, this would require moving the structures location
and creating another foundation. Another issue dealt with materials. The materials
needed to construct the alternative foundations tend to have long lead times. This was
especially true of the large metal plates needed for the mat anchor foundations. Each
plate was several inches thick to resist the bending moments with unique hole locations
placed in each mat. Because final designs were being completed as the installations of
previously designed foundations were in progress, this caused serious concerns for
procurement. Discussions with the bidders also revealed that the alternate foundations
required considerably more labor due the significantly more complex construction. The
plates were very heavy and would be difficult to handle in the field and the grouted rock
anchors turned out to be a more expensive option than initially considered.
Once construction began, it became apparent that the contractor was not having as much
trouble drilling through the hard rock as had initially been expected. With this in mind,
the decision was made to redesign all of the alternate foundations so that, if at all
possible, drilled shafts could be used. If the rock were to pose serious problems, the
original alternate foundation design could still be recalled and put into place. Even
though it sometimes required an extended timeframe to drill the newly designed rock
sockets in hard rock, it proved to be a more efficient method of foundation placement.
Ultimately, there were no structures that have required alternative foundations.
Geotechnical Considerations
Burns & McDonnells structural department was faced with determining the reactions at
each of the roughly 900 different foundation locations for drilled shaft design. Through
discussion with NU, parameters for foundation deflection and rotation were determined
to be 3 inches and 1.5 degrees respectfully for the entire project. These parameters were
used as the basis for the foundation designs. The inconsistent rock depths and varying
soil conditions made this a complicated task.
6 of 10
325
While the presence of shallow rock resulted in shallower drilled shafts for some
structures, the internal shear stresses built up in the rock sockets required very tight shear
ring spacing for the more heavily loaded drilled shafts.
There were a small number of structures that were placed in areas with very loose soils
with shallow groundwater and no rock to provide stability. Providing the required
foundation capacities and maintaining the required deflection and rotation criteria in
these locations required quite deep foundations. The resulting foundation designs
required some of the largest volumes of concrete in the entire project.
Due to the quantity of structures, organizing the different boring logs and keeping them
coordinated with the different foundation designs became extremely important. Troubles
along the way included changes in the structure numbering during design that caused the
boring log numbers to be mismatched with the structure numbering scheme. This was
overcome by a coordinated effort in organization. As indicated some individual structure
locations changed following the completion of the soil borings. These changes made it
necessary to make engineering judgments regarding the actual soil profiles and depth to
rock in these new locations.
Environmental Considerations / Permitting
As with any project, particularly one of this magnitude, environmental considerations
always need to be kept in mind. Burns & McDonnell employs environmental scientists
to do exactly that. As usual it was necessary to establish relations with the different
permitting agencies. For Middletown Norwalk the span of the project included a span of
different permitting agencies. When it came to foundations there was an increasing
concern for the amount of earth that would be displaced during construction. However,
the most important environmental issue was the affect the construction would have on the
wetlands. In order to decrease wetland disturbances, a great amount of effort was put
forth to keep as many of the new structures out of existing wetlands as possible. Once
the structures final locations were known, the overall changes and specific structure
locations were compiled so that the overall displacement could be provided to the various
permitting agencies as needed. As the project progressed the final displacement
information was continuously updated to keep all interested parties up to date.
Overall more than 250 documents were submitted to different permitting agencies,
municipalities, and the siting council for review and approval. Permits ranged from items
such as road and railroad crossing for the Connecticut DOT to Tidal Wetlands Permits for
the Office of Long Island Sound Program. The U.S. Army Corps of Engineers was
involved with wetland protection, while the U.S. Fish and Wildlife Service, National
Marine Fisheries Service, and the Connecticut DEP were all involved to oversee various
other environmental and wildlife concerns. As stated, the Middletown Norwalk Project
passes through 18 separate municipalities. Each of these municipalities had various
concerned parties ranging from Tree Wardens to Planning and Zoning Commissions.
Additionally, approval from the Federal Energy Regulatory Commission and the New
England Regional ISO was essential for the technical and monetary authorization of the
7 of 10
326
project. Lastly, a large majority of the required submittals consisted of Development and
Management Plans requested by the Connecticut Siting Council. Any design changes,
extended work hour requests, and contractor location information were submitted through
these D&M plans.
Construction
Once the Middletown Norwalk Project moved into the construction phase a whole new
set of challenges presented themselves. Hard to reach locations caused mobilization
issues, the quantity of foundations caused headaches in procurement, and the variable
depth to rock over short distances forced many redesigns for structures that were moved
even slightly during construction. Actual site conditions caused significant field changes.
Timing and coordination between the owner, contractors, designers, and the field
representatives became critical. Time required to drill rock sockets determined the pace
of foundation placement. Other field changes had to be coordinated with drilling
schedules so that the affect on overall time line could be minimized.
As stated, rocky soil conditions were found to pose less of a problem to drilling than they
did to engineering. The upper limits of this decision have been tested, and have proved to
hold true. Some of our largest rock sockets extended up to 15 feet in depth. Using large,
crane mounted drilling equipment the contractor has been able to drill though the largest
rock sockets in less than a week. For most rock sockets rock augers and core barrels
were used to drill a series of smaller diameter holes to the full length of the rock socket.
By drilling a series of smaller holes, the full diameter of the foundation could be reached
by breaking up and removing the remaining rock to the appropriate diameter. The exact
amount of time it takes to drill these sockets of course depends on the size of the drill rig
in use and the actual strength and depth of the rock in question. In the worst cases, it was
not a large enough loss of time to make the material lead time and additional construction
cost required for the alternate foundation designs worth while.
One issue that caused some scheduling problems was the difference between what could
be considered competent rock for design purposes and what was considered rock for
payment to the contractor for drilling. The competent rock that is assumed to provide the
needed rock socket for support of the foundation designs often laid beneath layers of
weathered or less competent rock. Unfortunately, the drilling contractors assumed that
any material not specifically called out as rock would be much easier to drill though. In
reality the less competent rock still required significant effort to excavate. The
contractors drilling rates for rock were uniform. Any soil that drilled like rock was
charged as rock. Once this was realized, the required drilling could be estimated more
affectively. This allowed for a much more accurate construction schedule.
In several locations, loose, sandy soils below the water table were encountered during
drilling that required installation of temporary casing. In a few instances, the casing was
not properly sealed and the sands were found to flow into the shaft during excavation,
resulting in relatively large areas of subsidence surrounding the foundations. In areas of
subsidence, the soils were assumed to have become very loose. Pressure grouting was
8 of 10
327
needed to stabilize the soils surrounding foundations and re-establish the anticipated soil
strength required to support the foundation. Detailed grouting plans and specifications
were prepared to accomplish this work.
Like all projects, field changes had to be dealt with as construction progressed. For
foundations, the most significant changes occurred due to the actual depth of competent
rock. Initially soil borings were taken for every structure in the initial design of the
project. However, issues arose for foundations where borings were taken for a structure
that was later relocated, or where borings were used for structures that were in close
proximity but not at the exact location. As the foundation installation took place it was
monitored by the contractor as well as onsite representatives for the geotechnical
consultant and Burns & McDonnell to ensure that the conditions were as anticipated. In
the event that the conditions failed to conform to what was expected, a new foundation
design could be generated to account for the existing conditions. As noted, the most
common cause for a foundation redesign was a significant difference in the depth of
competent rock.
When a foundation redesign was required, Burns & McDonnells on site construction
superintendents relayed the information back to the design engineers located in Kansas
City, Missouri, where they had a very short amount of time to return a new foundation
design that would support the revised subsurface conditions. This type of communication
need brings up the most important challenge faced in the Middletown Norwalk Project.
Without good communication and timing between all parties involved, there could have
been a significant amount of added delay and cost to the project. Burns & McDonnells
on site construction superintendents proved to be the vital link between the field and the
design engineers. They have coordinated the transfer of all pertinent information from
the on site geotechnical engineers, who provide the required expert opinion of the
existing soil conditions. They also coordinated the needs of the construction contractors
so that they were able to efficiently use their time and equipment. In addition to
coordinating with the field, the design engineers also had to maintain coordination with
the pole supplier, Thomas and Betts. In the case of foundation design, all parameters
were based on the loads and dimensions that each specific pole originally required.
Through cooperation with pole manufactures, delivery of pole designs was coordinated to
meet the timing needs of the field for foundation construction. Last, but not least,
coordination with NU was critical. Not only so that they could stay informed of progress,
but so that their engineers could have a chance to review and make comments regarding
the multiple aspects the project.
Conclusion
Overall, NUs Middletown Norwalk 345-kV Transmission Project has been a great
success. The overhead line design was completed six months ahead of schedule and
under budget. It has shown again the benefits of teamwork within Burns & McDonnell
and also within the industry. The foundations in transmission and distribution projects
are sometimes overlooked due to the more impressive nature of the projects other
aspects, but as engineers we know that these foundations offer the hidden support the
9 of 10
328
project requires. Overall, the most important lesson learned from Middletown Norwalk
has been that an engineer can never be too organized. No matter whether the project in
question will require a hundred miles of line or just a few. Being able to offer answers
during construction, instead of asking additional questions, has proven to be worth while.
It provides a perfect example of the need to prepare for all foreseeable situations in
design, so that construction can proceed as scheduled.
10 of 10
329
330
331
At the 10-km point the line left the active coal mine area and proceeded into the mountains north
of Healy where guyed X-Towers with a wider footprint were used. (Foothills Section)
Subsurface conditions in this section consisted of uplifted lake beds and were found to be
densely packed alluvial gravels. With the gravels being non-frost susceptible, a shallow drilled
pipe pile grouted in place was used. Existing mining and exploration roads provided vehicle
access to about half of the next 38-km; however, the remaining 19-km (11-miles) was restricted
to helicopter access only. No disturbance of the root mat was allowed in these areas which made
for creative solutions to provide level work pads. (Figures 3a and 3b)
332
The transmission line exits the mountains at kilometer 48, extending 96-km (60-miles) across the
low-relief terrain of the Tanana River Valley (Tanana Flats Section). The guyed X-Tower was
used exclusively for all tangent and light angles across the flats, with the un-guyed Swingset
Tower at deadends and large angles. The tower foundations and anchors consisted of single
steel pipe piles. Short term loading from lateral and uplift forces caused by wind and ice
controlled the design as sustained loads were relatively small for the X-towers, but much higher
for the swing-set towers.
The design of the foundations and anchors had to consider both the frozen and unfrozen state of
the soils across the Tanana Flats. Geotechnical investigations for this portion of the project
examined an existing GVEA 138 kV right-of-way approximately 5-km (3-miles) west of the
Northern Intertie route. When constructed in 1966 the underlying soils consisted of an
approximate 9-m (30-ft.) layer of frozen silts and gravels. The original 33-m (100-ft.) wide
right-of-way clearing removed all trees and shrubs, leaving the root mat intact. In the 31 years
between 1966 and 1997 the permafrost degraded completely in the cleared area as can be seen in
Figure 4. This phenomenon has been well established and a well known study by the U.S. Army
Corps of Engineers between 1946 and 1972 showed the clear link between the vegetation
removal and the resulting degradation of permafrost. (Linell, 1973)
The data collected during November 1997 showed that the permafrost was very marginal and
isothermal with ground temperatures warmer than -0.3 C (31.5 F). Isothermal ground
temperature is a characteristic indicator of permafrost susceptible to changes in surface
conditions. (Musial, 2007)
333
The original foundations on the 138 kV line consisted of 9-m (30-ft.) wooden piling and upon
degradation of the permafrost the frost heave jacked the foundations out of the ground 1- to 2-m
(3- to 6-ft.). Resulting maintenance costs for this type of foundation failure can be substantial
and was a major concern when designing the Northern Intertie.
With the marginal permafrost across much of the area, it was not if it thawed, but when it
thawed that was of concern during the design. When frozen, the soils will provide a sustainable
adfreeze resistance of between 488 and 976 kg/m (1000 and 2000 psf) (Golder Associates, May
2 and July 18, 2000). When thaw occurs, that resistance is reduced to an average of about 244
kg/m (500 psf). These conditions regularly create high frost-heave forces.
Frost heave forces will lift a pile out of the ground each season in what is known as frost jacking.
This phenomenon was found to control the length of piles used along the route. Significant
heave forces were expected because the soils were frost susceptible, meaning that the soils would
wick the groundwater up to the freezing front. Frost heave occurs when the pore water in the
active layer expands and ice lens form at the freezing front in freezing weather.
The active layer in the Tanana Flats, or that layer subject to annual freeze-thaw cycles, varied
between 0.6 and 2.4 m (2 and 8 ft.). The active layer thickness and the adfreeze strength was the
most significant factor in determining the relatively high frost heave forces that were expected
across this area.
Adfreeze skin friction developed by the frozen soil in the active layer will cause the soil around
the pile to heave, resulting in an upward load applied to the pile. Figure 5 shows a conceptual
frost heave model for a single pile.
334
During the initial design we used a 1.2 to 1.5 m (4 to 5 ft.) active layer depth and 2.8 kg/m (40
psf) adfreeze within the active layer. The estimated embedment depths that were developed to
resist frost heave ranged from 6 to 24 m (20 to 80 ft.) depending on loads and soil conditions.
The 96 km (60 mile) section across the Tanana Flats consisted of the most unstable soil and
would require the deepest embedment depths. GVEA worked with the transmission line
designer, Dryden & LaRue, Inc. and the geotechnical consultant, Golder & Associates, Inc. to
use probability and statistics in the design process. To develop a business risk model, we
estimated the variability in frost heave potential along the alignment. The X-tower has long been
used in Alaska due in part to its ability to withstand relatively high differential movement
between leg foundations. As the owner, GVEA had to evaluate the question of whether it was
important to have no foundation movement through the installation of deep piles, or would ongoing maintenance mitigate the significant frost heave (more than 20 cm [8 in.]) if shallower
piling were used. Although a probalistic model was developed, it was ultimately thinking in
terms of the risk of change over the life of the project that provided the design guideline.
Whether it be in business operations or engineering design:
It is better to be roughly right than to be precisely wrong. (John Maynard Keynes)
The decision was made to install piles to a minimum depth of 11 m (36 ft.) in the flats. For those
structures with higher loads, such as large angle swingset towers the depth was 25 m (83 ft.).
Some areas in the flats encounter very loose soils and piles were driven to depths in excess of 27
m (90 ft.).
To
accommodate
the
differential
movements of the foundation and provide
adjustable connections, pile clamps were
developed for ease of lowering or raising
the tower leg as part of the on-going
maintenance. Pin connections to the tower
leg also allowed the tower to be assembled
flat on the ground and tipped upright into
place. (See Figure 6 at left.)
335
The X-towers were guyed fore and aft for longitudinal support and the anchors were also driven
piles. Guy strand shear release mechanisms were developed to protect the structures from large
compressive forces that develop when the towers experience foundation jacking and move
differentially to the anchor piling. On GVEAs existing 138kV transmission line in the same area
a shear release device was included in the guying system with poor results. Several structures
suffered compressive failures when the release mechanisms failed to operate. It appeared that the
strength of the shear bolts used in that system were inconsistent and failure occurred in the
structure prior to failure of the shear bolt.
The new shear release mechanism uses the same shear pins used in the AB Chance screw anchor
torque limiter (Chance catalog #C303-045). The new shear release mechanism was manufactured
with positions for seven shear pins in double shear and one in single shear. A shear pin retainer
plate keeps the pins in place and allows for replacement if necessary. By varying the number of
pins used, release loads between 3,740 Kg (8,250 lbs.) and 7,485 Kg (16,500 lbs.) were
developed for the project. Components of the shear device are shown in Figure 7 below.
336
The final 10-km (6-mile) segment of the line into the Fairbanks area (Fairbanks / Tanana River
crossing section) consisted of two 56 m (185 ft.) crossing towers and single leg Y-towers and
side post structures into the substation in Fairbanks. Soils in the area were either within the thaw
bulb around the Tanana River or in areas where the permafrost had degraded due to past
development.
CONSTRUCTION ISSUES
It was natural to divide the project into four separate contracts as they were made distinct by
such issues as terrain, logistics and seasonal restrictions. The two largest contracts were the
Foothills Section and the Tanana Flats Section which were also subject to the most
restrictions on construction. The contracts for the two 10-km (6-mile) sections at each end of the
transmission line were easily accessible and subject to few restrictions.
Communication on any project is key to a projects success, whether it be through the contract
documents, pre-construction meetings or during construction itself. Whether through the
Owners inability to communicate the importance of subsurface conditions or the contractors
inability to listen, the issue that created the most challenges was the contractors lack of
understanding of these subsurface conditions.
As part of the bid process, geotechnical information collected during the design phase was made
available to contractors at both the Owners offices and the Engineers offices; but were not
337
included as part of the bid package. Unfortunately, several of the contractors did not examine the
geotechnical data during the bid process. Add to this the absence of meaningful site visits (some
contractors did fly-bys) led to claims of change-in-conditions, etc.
It became apparent early on that pile driving contractors were accustomed to projects where
refusal was the standard for piling installation. Because of frost heave forces and uplift that
were expected and in some cases the erosion possibilities, all of the foundations were specified to
a minimum installation depth. The contract recommended pre-drilling to 80% of the diameter
and the use of hardened pile tips, but did not require them.
In the Usibelli Section the subsurface soils consisted of material ranging from sands to
boulders and no permafrost. Within this section the pile driving contractor did not request the
geotechnical data or visit the site during the bidding process. Without a clear understanding of
the reasons behind a minimum depth burial, the contractor refused to pre-drill or use pile tips,
thinking that refusal provided an adequate foundation. Figure 9 shows the results when held to
the design depths without pre-drilling.
338
339
Fig. 11a (above left) Truck with water tanks and snow guns (Used with Permission)
Fig. 11b (above right) Hydraulic vibratory driver on 30 in. pile (Used with Permission)
In developing the construction contracts it became apparent that with several uncontrollable
factors such as snowfall and temperature extremes it was necessary to be flexible in the
completion time for the contracts. This was done by giving a completion date of two years, but
removing liquidation damage and any penalty clauses from the contract. The contractor would
therefore be able to use a third winter season if needed. Even with the marginal snowfall the
contractor was able to complete the project on time.
It is imperative that contractors are qualified to construct projects in an Arctic environment. The
effect of permafrost on subsurface construction, decisions in the equipment used and the logistics
of working in extreme cold with little daylight are critical in a projects success.
340
REFERENCES
Musial, M. and Wyman, G.E., 2007, Marginal Permafrost a Foundation Material in Transition,
Permafrost Foundations: State of the Practice, Technical Council on Cold Regions
Engineering Monograph, pp 30-49, 2007.
Crowther, G. Scott, 1990, Analysis of Laterally Loaded Piles Embedded in Layered Frozen
Soil, Journal of Geotechnical Engineering, Vol. 116, No. 7, July 1990.
Linell, K.A., 1973, Long Term Effects of Vegetative Cover on Permafrost Stability in an Area
of Discontinuous Permafrost, Proceedings of the Second International Conference on
Permafrost, Yakutsk, USSR, pp 688-693, 1973.
Morgenstern, N.R., Roggensack, W.D. and Weaver, J.S., 1980, The Behavior of Friction Piles
in Ice and Ice-rich Soils, Canadian Geotechnical Journal, No. 17, pages 405-415, 1980.
Reese, L.C. and Wang, S.T., 1997, Documentation of Computer Program LPILE, Plus Version
3.0 for Windows, Ensoft, Inc., Austin, Texas, 1997.
Weaver, J.S. and Morgenstern, N.R., 1981, Pile Design in Permafrost, Canadian Geotechnical
Journal, Volume 18, pp 357-370, 1981.
341
342
James W. Graham
Marlon Vogt
Alliant Energy
Ulteig Engineers
343
344
Cedar Rapids. Unfortunately, this circuit switcher was submerged in flood waters
and not serviceable. An emergency purchase order was issued, and a new circuit
switcher arrived six days later.
This same substation was scheduled to receive a metal clad switchgear in mid-June.
The switchgear comprised a 1200A main breaker, a 1200A bus tie breaker, 1200A
bus, four 1200A distribution circuits, and associated relay panels. The switchgear
alone met neither the 8 circuit nor the 50 MVA requirements.
A small rural substation was also being built in rural north central Iowa. Reviewing
the material available from this project yielded a low tension dead end structure, a 69
kV gang operated switch, and a steel box structure comprising a transformer bay and
three distribution bays. The bus conductor could support 50 MVA of load but the
main disconnect switch was rated 2000A, slightly below the transformer's required
full load current of 2091A at 13.8 kV. We debated whether the switch we had would
be adequate, but this became a moot point when we salvaged two sets of 15 kV
3000A disconnect switches from yet another project.
On Tuesday, June 17th, two days after the water receded, the design team developed
the one-line shown in Figure 1, along with an equipment plan and section views for
the substation based on the available equipment. The design used the 69 kV switch
and a 34.5 kV circuit switcher for primary side transformer protection. We could not
locate a 15 kV 3000A circuit breaker in inventory, and lead times for new 15 kV
breakers were unacceptably long. This forced us to connect the transformer
secondary directly to the box structure bus through the 3000A disconnect switches
installed on the transformer main bay. We installed reclosers on the three distribution
circuits. Because all four bays of the box were used, we ran a flying tap conductor
connecting the cross bus on top of the box structure to the roof bushings of the metal
clad switchgear enclosure.
With three circuits coming off the box structure bus and four circuits coming off the
switchgear bus, we were still one circuit short. A quick review of the switchgear
drawings and a call to the manufacturer confirmed the bus tie position could be
converted to a distribution circuit position in the field. This gave us the required
eighth distribution circuit.
After consulting their load flow studies, planning was able to assign sufficient loads
to the three recloser circuits to keep the remaining load current below the switchgear
main breaker's 1200A continuous current rating. This arrangement allowed us to load
the transformer to its 50 MVA top rating. Figure 1b shows the final equipment oneline defining the scope of the substation project.
345
346
The foundation designs, by necessity, were low-tech. We did not have time to
remove the concrete parking surface, pour concrete pads, or drill and construct pier
type foundations so we needed other options. Both our general office and our area
maintenance facility were inaccessible due to the flood, which meant the tools
available to lay out the equipment foundations were a 100 ft. cloth tape, a can of
spray paint and our ingenuity.
We aligned the main bus with the transmission pole which would bring the source
34.5 kV line into the substation by eye and painted the bus center line on the concrete
surface. Working from the equipment plan, we measured the center lines for the
major structures and equipment, and painted the intersecting centers.
Transformer Foundation
At first glance, a parking lot surface looked level, but in fact the pavement sloped to
drain surface water to the storm sewer inlets. Our first structural challenge was to
devise a stable, level surface to set the power transformer on. We did not have the
time to remove the existing concrete and pour a traditional concrete pad.
We considered setting the transformer on the concrete and shimming the base, but the
amount of shims required resulted in an unacceptable stability risk. The contractor
suggested building a gravel pad for the power transformer (see Figure 2). The pad
was sized to extend about four ft. past the transformer base on all sides to provide a
working surface after the transformer was set in place. A line of 8" high railroad ties
was set next to the lower edge of the gravel pad. The ties served two purposes. First,
to establish the pad thickness, and second, to prevent the road stone from washing
away during heavy rains.
An 8 inch layer of gravel (crushed rock with fines typically used to build secondary
roads in rural areas) was then laid directly on the concrete surface and compacted,
starting at the ties and gradually tapering off up slope. This gave us the required level
surface for the transformer. Gravel was chosen because it compacts well, thus giving
us the desired stable surface.
Next, copperweld ground mats were laid on top of the base and bonded together to
provide a ground plane. The ground mat array was then bonded to the substation
ground grid. Finally, a 2" layer of top soil mixed with road stone was placed over the
ground mats. This layer provided a measure of self-leveling as the weight of the
transformer pressed down onto the pad.
The result of this design was more than satisfactory. After the transformer was set in
place, we put a level on the transformer base and found all four sides were almost
perfectly level.
347
348
under one corner of the structure were required to level the building. Out of curiosity,
a worker put a level on the buildings footing rail, and it was only a partial bubble off
true level.
349
350
351
Substation Completion
Substation physical construction was essentially completed on Friday, June 27th.
Fence installation was completed and the site secured the following day. On
Wednesday July 2nd, the power transformer was energized as system check out
continued. Commissioning checkout of the equipment and control panels continued
through July 2nd.
With the total destruction of Cedar Substation and the limited availability of other
distribution substations to phase to, very careful attention was paid to existing
phasing. All 13.8 kV distribution feeders left the substation underground, and all
paper documents on phasing were destroyed in the flood, so no clear indications of
pre-flood phasing existed. Phasing was traced from the 34.5 kV line through the
damaged substation to the riser poles, where the phases were taped and color-coded.
Starting from the same point, we repeated the process through the temporary
substation to the feeder exists. Phases for each distribution line were taped and colorcoded to match the riser pole coding. On 3 July, 2008, the temporary substation was
in service and ready to serve the available loads.
Typical construction time for a substation similar to Cedar Temporary Substation is
six weeks. We were able to go from conception to design to construction completion
(see Figure 5) in 18 days.
11
352
While the Cedar Temporary Substation was being built, the 13.8 kV underground
network circuits serving the downtown loads and fed from this substation were
repaired and tested by the underground construction team.
In parallel with these efforts, the transmission provider built a 161-34.5 kV temporary
substation and a 161 kV line tap to provide a source of power to the 34.5 kV
subtransmission system. Some reconfiguring of existing 34.5 kV lines was required
to route power to the substation.
Thanks to a combination of planning, innovation, and good fortune, the transmission,
substation, and underground distribution projects were all completed on the same day.
Epilogue
The 161kV transmission and 34.5 kV subtransmission systems were returned to near
normal conditions by early fall 2008. At that time the temporary 161-34.5kV
substation and the associated line tap were dismantled, but much work remains to
restore the previous reliability levels to the transmission and subtransmission systems.
The 13.8 kV distribution network is fully restored. Cedar Temporary Substation is
the primary source for the network and is expected to remain in service for a
minimum of two years from its in service date. Efforts are continuing to acquire a
suitable site and develop plans for a permanent substation to carry the downtown
distribution network.
12
353
354
355
PROJECT LOCATION
The project is located near Panama City, Florida, U.S.A. The transmission line
connects Tyndall Air Force Base and the Town of Parker, in St. Andrew Bay, as
shown in the map below (see Figure 1).
DuPont Bridge
Runway
Figure 1.
Map of Area
356
The overhead route was selected for purposes of more reliability, more readily
available maintenance, less environmental impact, and less cost. Alternate routes
were evaluated and with each route the new circuits would still have to cross the bay
and would have to traverse through residential neighborhoods where there was
considerable opposition from property owners. The direct route across the bay was
the least obtrusive and the shortest. The second most favorable route would have
added several miles to the project and would have been attached to the DuPont Bridge
with opposition from the FDOT. Other crossings would have been closer to the
runways at Tyndall AFB and would have been longer.
It was necessary to provide sufficient span at the middle of the bay for intracoastal
barge traffic. Environmental considerations to protect shoreline transmission and
substation structures and extension of stormwater concrete pipe had to also be taken
into account.
TRANSMISSION SOLUTION
This project consisted of two new circuits designed for 115kV operations, but
operating at 46kV, starting at Wewa Road Substation, crossing St. Andrew Bay and
going to Tyndall Field Substation on the Tyndall AFB property, the total combined
length of the two circuits is approximately 8 miles.
The middle span in the bay was made as wide as possible for barge traffic and the
centerline of the middle span coincides approximately with the centerline of the
Intracoastal Waterway, all in order to provide a low probability of a barge hitting the
structures.
The spans were also made longer than the span between piers of the DuPont Bridge,
which is located approximately 1-1/2 miles southeast of the bay crossing and where
barge traffic goes into the Tyndall AFB property.
Federal Aviation Regulation Part 77 describes the standards for determining
obstructions in navigable air space. A study was made to ensure that the new
overhead line and structures did not constitute an obstruction. The height of the
structures is well below the maximum height restrictions described by the FAA
regulations. A report of the new construction was provided to the FAA to provide
information to be added to navigation charts. The FAA did not require marker balls,
lights, or painting of the structures in the bay area for this design. Two very tall poles,
one on each side of the bay, would have caused an obstruction, would have required
lighting, and would have been extremely impractical to merge the design back into
the existing circuits on each side of the bay. The structures were provided with
provisions for climbing for maintenance purposes. Anti-climbing devices were
attached to the lower portion of the poles to discourage anyone in a boat from trying
to climb a pole. The vertical clearance above the waterway was dictated by the Corps
357
of Engineers Federal Register rules and regulations for clearances of aerial electric
power lines crossing navigable waters of the United States.
The following are the conductor and overhead groundwire used in the line:
Conductor
Overhead Groundwire
Design Tension in the
Bay Area for Conductor
Design Tension in the
Bay Area for Overhead
Groundwire
Ruling Span in Bay Area
STRUCTURAL SOLUTION
The crossing of St. Andrew Bay required four structures to be located in the water.
The following three types of steel structures were considered in the design:
1. Single steel pole with davit arms with one circuit on each side with steel
caisson foundations.
2. Steel H-Frame with bracings and davit arms with one circuit on each side with
steel caisson foundations.
3. Lattice tower with cross arms with one circuit on each side with caisson
foundations.
The single steel pole with davit arms with one circuit on each side with steel caisson
foundations was selected based on being aesthetically pleasing due to the small foot
print and due to the overall cost.
The following loads were considered in the design:
1. 2002 National Electrical Safety Code
a. Combined ice and wind loading: Light loading 0 radial thickness of
358
Gulf Power Company design loads are critical in comparison with NESC and ASCE
loads. Therefore, the structures are designed for Gulf Power loads. Tornado loading
was not considered at this location.
The four required structures are numbered 1 through 4 as shown in Figure 2. The
engineered poles were designed using flange joints in three sections. The utility
company wanted to have spare sections in order to replace any of the poles in water
that failed due to heavy hurricane/tornado loading higher than the design loadings.
The top sections for all four water structures were designed identically such that the
spare top pole sections could be used in any of the four locations. Sufficient spare
sections were fabricated to facilitate the assembly of one complete replacement pole
for any one of the water structures No. 1 through No. 4.
A comprehensive geotechnical investigation was performed with borings taken at all
four water structure locations using a barge mounted drill rig. A soil boring log for
one of the middle structures is shown in Figure 3. The soil generally consists of soft
to hard clayey sand. The standard penetration blow count number varies from 0 (Nvalue) at the bottom of the bay (at mud line) to 60 (N-value) at 54 ft. deep into the
soil.
The caisson foundations were designed using the engineering software LPILE Plus,
developed by Ensoft, Inc. The foundation caissons were designed to use a permanent
pipe pile casing that replaced the interior excavated soil with cast-in-place reinforced
concrete. The 6ft. diameter thick steel casing was to be the caisson form during
construction and then act as a protective covering for the concrete caisson over the
life of the project. Due to the salt water in the bay, the steel casings were shop
painted, after fabrication, with 20 mil dry film thickness of CorroCote II Classic on
the exterior surfaces for corrosive protection. The reinforced concrete placed within
the casings was the structural component of the foundation while the thick steel
casing was considered as form work and long-term protection.
359
360
See
Figure 3
Figure 2.
Drawing Showing Transmission Line, Steel Poles, and Caisson Foundations
Figure 3.
Caisson and Soil Boring Detail
361
Construction was performed from two barges. For the steel casings, 6 ft. diameter
thick smooth pipes were selected instead of 12-sided pole-type geometry. Steel
caissons were installed using the vibratory hammer method. After installation, the
material inside each caisson was removed completely down to the bottom level of the
caisson. The smooth steel pipe caisson interior helped in effective cleaning, better
than a 12-sided geometry pipe. 5500 psi concrete, with approximately 9 of slump
maintained by use of a plasticizer, was brought in from an adjacent paper mill loading
dock which was approximately one mile from the construction site. From the truck,
concrete was transferred by gravity to three 3-cubic yard buckets at the end of the
barge. It took approximately 20 minutes for the barge to take the concrete to the site
of each foundation. Concrete was placed using the tremie process (underwater
placement). With the proper mix design, concrete was plastic throughout the
placement time. See Figures 4 and 5.
Figure 4.
Smooth Pipe Caisson
Installed Using
Vibratory Hammer
362
Figure 5.
Reinforcement Bar Cage Placed Inside Caisson
363
ENVIRONMENTAL SOLUTION
Environmental issues and regulations for any activity in the bay had changed
considerably since the time the original 46kV underwater circuits had been installed
40+ years ago. Current environmental issues that had to be considered and actions
taken during the design and construction of the new 115kV overhead circuit lines in
the St. Andrew Bay crossing are discussed below.
Coordination was required with the Florida Department of Environmental Protection,
Corps of Engineers, and the Federal Aviation Administration for the St. Andrew Bay
crossing.
Sea grasses, oyster beds, and sea life, such as the manatee and small and large turtles,
required protection from construction. The manatee, a docile creature, is sometimes
called a sea cow since they eat sea grasses. In order to protect sea life and grasses,
water structures were located away from shore lines at a minimum distance of 60 ft.
from sea grass areas. If a manatee was to come within 50 ft. of the work area, the
work activity was to be shut down to avoid any harm to the animals. Motorized
watercraft had to avoid any sea grass areas where the water level was less than 1 ft.
from the sea grasses.
Turbidity curtains were required in the work area. These are floating barriers with
curtains designed to control the dispersion of silt and sediment in a body of water.
Turbidity curtains surrounded the work area around each structure. Construction
forces were to leave nothing in the water, such as drilling mud, dirty water sediment
or sludge, and storage tanks were provided on the work barges to contain the waste. A
water quality monitoring system was required. The Florida DEP along with Gulf
Power Environmental personnel monitored the water conditions for turbidity.
In general, the osprey nesting season begins in January and ends in August in this
area. The construction activities were scheduled in such a way that there was no
potential for interference with the osprey nesting.
Erosion protection was required on both shores to minimize the environmental impact
to the shoreline due to new construction and to protect the new structures on land
nearest to the shore from erosion concerns. Erosion protection consisted of
embankments on both shores with filter fabric, unopened Quickrete concrete mix
bags, and 6-10 riprap as shown in Figure 6. Unopened Quickrete bags, after
wetting in place, formed a good foundation for the embankments. An existing
drainage pipe on the Panama City side of the bay was modified and protected with
riprap.
364
Figure 6.
Embankment Drawing
365
CONCLUSION
The project was successfully completed in 2004 due to the excellent pre-planning,
dedication, and cooperation from various personnel from the following organizations:
1.
2.
3.
4.
5.
Figure 7.
Completed Caissons and Poles with Wires in Place
366
367
368
Conductor clearances along the length of the line was checked in 3-dimension in the
swinging condition and finalized the tower locations. Towers are designed to carry one
OPGW and six twin conductors (ASCR-SQUAB).
Space available for design and installation of the tower base was limited to 2.5 Sq. m (27 Sq.
ft). There was no space to provide conventional cross arms due to electrical clearance
requirements hence the vertical configuration tower was designed where conductors were
attached to body directly. Crossarms were provided to accommodate jumpers.
Block foundations of 4 Sq. m (43 Sq. ft) were designed within the space available near the
water cannel. Due to high water table 9 m (30 ft) deep fully reinforced block foundation were
designed to take care of uplift.
Since the transmission line had to follow the road route, all angle towers were provided. (See
exhibit Nos.: 4-9)
369
Stringing had to be completed section by section, since every tower was angle tower and
finally transmission line was charged and in operation since 2007. (See exhibit Nos.: 10-11)
370
Survey is carried out using high end GPS system and the route is finalized. Access roads are
created for movement of heavy duty piling rigs and erections cranes. Sand dunes and extreme
hot (+50C) temperatures made working conditions difficult. (See exhibit Nos.: 12-13)
Vertical as well as battered piles were designed with pile diameters varying from 3.5 ft to 8
ft. For heavier uplift and compressive loads multiple piles and tie beams were required. (See
exhibit No.: 14)
Towers were erected in two stages. A 150 ton crane was used for erection up to the waist
level and then a 250 ton crane is used for second lift. Civil aviation paint was applied to all
towers. (See Exhibit No.: 15)
Due to congestion of lines, 400 kV multi circuit transmission lines are designed to carry 4
circuits of quad conductors making the construction much more challenging. (See exhibit
No.:16)
The weight of the towers ranged from 101 MT to 242 MT since it has an uplift loads range
from 300 tons to 900 tons which lead to design a leg with a built up of 4 legs of section
250X250X35 mm (largest angle available in the world).
371
Exhibit No. 16: 400 kV quad multi circuit undergoing prototype testing
Due to non-availability of right of way through the city, line had to pass through Arabian Sea
to reach the substation. Barge were used for construction. (See exhibit Nos.: 17-18)
Exhibit No.17: Erection of tower in Arabian Exhibit No. 18: Double-deck gantry near
GIS substation
Sea
Hybrid tower was designed and installed to turn the line. Huge cantilevered auxiliary
crossarms lifting was quite difficult due to weight of structures. (See exhibit No.: 19)
372
373
Kazakhstan:
Other extreme condition was faced during construction of 600 miles of 500 kV single circuit
where the temperature is extremely cold (-40C) and the site is fully covered by snow and
inaccessible for four months each year.
The transmission line had to pass through seven earthquake and wind zones. Towers were
configured as guyed towers supported by 4 guys per tower. Guying allowed the towers to
became much lighter in weight and the resulting foundations were smaller as well.
With the engineering solution of using precast foundations and prefabricated towers, all
towers structures in eight parts were kept ready in the non-working season in the factory.
Similarly foundations were casted in the factory. An electric heating system was used for
curing of concrete. (See exhibit Nos.: 22-23)
Dully assembled towers and precast foundations together were transported to sites. Cranes
were used for installation of foundations and erection of towers. (See exhibit Nos.: 24-27)
374
375
Libya:
A 250 mile 400 kV single circuit line has been constructed in the Sahara Desert. Horizontal
configuration towers were designed for a triple conductor bundle. Since the line had few
obstructions or angles, long spans were used with tall towers by providing body extensions
ranging from 90 ft to 240 ft. (See exhibit No.: 31)
376
377
378
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
379
Network Mapping, Unit B1G, Fairoaks Airport, Chobham, UK, GU24 8HU; PH
+441276 857800; email: [email protected]
2
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
380
This paper will first outline the problems facing ROW vegetation management
personnel and the traditional techniques employed to identify vegetation threats. The
basic concepts of airborne LiDAR, also known as Aerial Laser Survey (ALS), will
then be discussed with a methodology proposed for its practical application to ROW
vegetation management. The technique is evaluated with a case study on
implementation with the National Grid UK vegetation management program.
TRANSMISSION RIGHTS-OF-WAY AND HAZARDOUS VEGETATION
A transmission line ROW is a cleared strip of land occupied by a transmission line or
lines that is maintained by the owning utility. In the USA, ROWs are normally held
in easement or fee. Typical easement language often includes the right to ...trim,
cut, remove or control by any other means at any and all times such trees, limbs and
underbrush within or adjacent to said right of way as may interfere with or endanger
said structures, wires or appurtenances, or their operations. (NERC, 2004)
Vegetation control is generally carried out in accordance with regionally or
nationally accepted clearance requirements, such as IEEE Standard 516-2003 and the
requirements of the National Electrical Safety Code (NESC). They are used to
develop vertical and horizontal clearance margins to ensure that no vegetation can
encroach close enough to conductors to cause a flashover or physical damage, even
at maximum rated operating temperature. However, even with adequately maintained
vertical clearance margins, a second risk is posed by falling trees that may contact
the line or infringe clearance [Fig.1.].
Figure.1: A schematic of the two principle hazards posed by vegetation. (a) A tree
that has grown in close proximity to the transmission line, and may cause a
flashover. This hazard would need to be removed or pruned by arborists on a short
return period to prevent infringement of statutory clearance margins. (b) A tree that
does not currently infringe statutory clearance, but may infringe clearance or impact
the line if it fell.
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
381
An effective TVMP is required that mitigates risk to a transmission line through the
management of both these hazards, as required by NERC Standard FAC-003-01.
Conventional management techniques based on visual estimates can be over-zealous
in their maintenance of grow-in infringements, and reactionary to falling tree events.
The trim schedules for these conventional management programs are often based on
time, and not on need, which can result in inefficiencies and increased risk (PSERC,
2007). To reduce the risk of these hazards, average return periods of 5 years (NERC,
2004) for time based trim schedules should therefore be reduced, or an entirely new
methodology is required with factual clearance information at its core.
SYSTEM AND ENVIRONMENTAL VARIABLES
In order to create a management system based on need, it is essential to acknowledge
the system and environmental variables that increase risk. The increased risk stems
from either those factors that reduce conductor clearance to vegetation, or the
increased likelihood of a hazardous event occurring.
Conductors sag in a catenary curve between attachment points on adjacent strain or
suspension structures. As demand levels increase, the higher load levels generate
greater demands on both active power (MW) and reactive power (MVAr). Resistance
within a conductor creates heat, and hence increased flows along the transmission
line will cause the conductor metal to expand, and consequently further sag. It is
therefore imperative to ensure that clearance to vegetation is determined at the
maximum operating temperature for a line, when the conductor sag is at its lowest
point. This is required under NERC FAC-003-01.
The weather is a critical component that must be considered when examining
clearances to vegetation. Wind can cause the conductor to swing perpendicularly,
and clearance should be maintained even at this extended position. Again, this
consideration is required when determining clearance requirements for vegetation
under NERC FAC-003-01. Ice loading also has a dramatic impact on conductor
position, through increased sag. Differential ice accretion in adjacent spans will
result in differential sag.
The risk posed to transmission lines varies subtly with vegetation type and climatic
zone across the globe, but generally this hazard exists for all lines. However, some
vegetation on or adjacent to ROWs should be earmarked for more intensive
inspection, due to a higher risk induced by natural hazards. Despite the limitations
within the source dataset (cf. PSERC, 2007), Figure 2 shows the possible level of
correlation between vegetation outages and storm conditions for a geographical
locality.
In geographical areas prone to earthquakes, high winds, permafrost activity, icing
and heavy wet snow, utilities must acknowledge the threat posed by vegetation to
transmission lines. They should take measures to mitigate this risk in order to ensure
the reliability of the power supply. Lines traversing woodland that becomes
desiccated in the summer months should also be checked frequently, as flashovers to
encroaching vegetation could start forest fires.
3
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
382
Figure.2: A graph showing the vegetation related outage summary for a 32 feeder
substation, of which 8 feeders were monitored. Adapted from PSERC, 2007. Despite
the limitations of the data (cf. PSERC, 2007), the graph provides an insight into the
significantly high percentage of vegetation outages that may be storm induced in
some areas.
TRADITIONAL VEGETATION MANAGEMENT TECHNIQUES
Traditionally, and to a certain extent currently, vegetation management needs along
transmission line ROWs were identified with terrestrial visual assessment
techniques. Experienced foresters would conduct visual inspections along ROWs to
identify vegetation that appeared to be encroaching pre-determined minimum
clearances based on line voltage. These minimum clearances would typically include
a high operating temperature sag factor to account for high electrical loading
periods. This approach often results in generally more vegetation removal than may
be required, to be on the safe side. This approach also underestimated conductor
sag in long spans, and led to inadequate pruning.
In specific instances where accurate vegetation height with respect to conductors was
required (orchards for example), land surveyors would take measurements to the top
of any vegetation under or adjacent to the transmission lines using traditional survey
equipment. Five points on the conductors and the conductor attachment points would
also be surveyed to allow the conductors to be accurately modeled. All clearances
from conductor to vegetation could then be calculated. This information was then
used to clear any vegetation infringing on the minimum clearance. An example of an
output diagram created through this method is provided below [Fig.3.].
This ground survey technique is time and labor intensive, and is consequently
impractical for application on more than a handful of spans. Application of this
technique on an entire transmission system would be time consuming, costly and
unnecessary given the availability of alternative techniques.
4
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
383
A description is now given of Airborne Laser Survey and associated output data [e.g.
Fig.4.], and how this addresses the inherent problems of traditional vegetation
management techniques, whilst ensuring the system and environmental variables are
adequately considered in order to comply fully with NERC FAC-003-01. Laser
scanning technology presents a shift from management techniques based on time and
judgment, to one based on quantitative data.
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
384
Figure.5: A depiction of laser interaction with the ground and an above ground
feature. The helicopter mounted laser scanning system fires a pulse of infrared
energy. The first contact this makes is with part of the tree structure, which produces
backscatter. This is recorded by the sensor head as the first return. The laser then
continues and contacts the ground. As the ground is non-permeable, this causes the
last backscatter from the original pulse, and is recorded by the scanner as the last
return. LiDAR technology can therefore be used to create ground models even when
vegetation coverage is present.
Airborne laser scanning systems are usually equipped with a high resolution
downward looking metric camera [Fig.6.], providing photographs for a complete
coverage of the survey swath. These images are later orthorectified, to ensure the
distance between two objects is representative of the actual real world distance.
6
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
385
Aircraft can be equipped with forward looking cameras, and additional downward
looking sensor types such as multispectral and thermal cameras.
Following the survey, the collected data is processed to produce a 3D point cloud and
photography that is spatially referenced. In order for the data to be meaningful to
utilities, it must be populated with attribute data. Feature types within the 3D point
cloud are classified by assigning a code to each feature. For example, roads,
vegetation and buildings, may be feature codes 108, 109 and 130 respectively. Once
imported into PLS-CADD, the computer can discriminate between these feature
types. Statutory clearance values are entered for each feature type according to a
countrys electrical code. For vegetation clearances in the USA, transmission
companies normally base this on published standards, such as ANSI Z133.1 and
IEEE Standard 516-2003, which specify minimum clearances for vegetation from
various voltage lines. The structures and conductors are modeled within PLSCADD to produce an accurate model of the transmission line at the time of data
capture. This model incorporates electrical load information provided by the line
operator and meteorological data from locally deployed weather stations and in-flight
measurements. PLS-CADD allows users to define the size and type of conductor,
shield wire or optical ground wire along with associated engineering properties, such
as yield strength. Simple frames of structures can be created from circuit diagrams of
a route, or full engineering structure models can be modeled or imported. PLSCADD models can be used for a range of applications across all areas of overhead
line management, including new line routing and design, thermal rating assessments,
re-conductoring projects, and the structural analysis of towers and poles.
Figure.6: Optech Incorporateds ALTM 2050 airborne laser scanner. The laser is
attached to the front of a Squirrel helicopter with a purpose built mount. The metric
camera is installed just behind the scanner, and both of these sensors feed data to the
recorder contained within the aircraft.
Field inspection teams previously estimated the conductor position at maximum
rated temperature, with inherent subjectivity. PLS-CADD allows the user to sag
the wire to this position, and then find the vegetation violations for various
combinations of conductor temperature, wind and clearance requirements. LiDAR/
7
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
386
PLS derived infringement locations therefore represent not only a great improvement
in survey speed, but a quantitative step change for analysis.
The raw datasets used to create the PLS-CADD models can also be used to
produce models in other power line software packages.
VEGETATION MANAGEMENT USING ALS
The seasonal nature of vegetation necessitates the rapid turnaround of data from inflight acquisition to the final delivery of vegetation intelligence that will be used for
ROW management. Rapid vegetation check deliverables can be offered within 30
days of data capture, and these depict all vegetation within a pre-defined radius from
a conductor, for example, 25 ft. This enables compliance with section R1.5. of
NERC FAC-003-1 which states action must be taken on vegetation conditions that
present an imminent threat of a transmission line outage. Network Mapping uses the
TerraSolid plug-in for Microstation to produce this deliverable, with the simple
vectorization of wires and the classification of above ground features as trees. An inbuilt function then determines any vegetation within the pre-defined distance from
these wires to give an idea of the extent of vegetation violations. Violations are
highlighted in a plan sheet, and displayed against stock imagery [Fig.7.].
Figure.7: Initial vegetation check plan sheet for compliance with NERC FAC-003-1
R1.5. The conductors are vectorized and vegetation points classified. An algorithm
highlights all hazardous vegetation within a predefined radius which is displayed
against stock imagery.
Pinpoint vegetation checks can be provided 30 40 days following data capture. To
produce the deliverable, classified Aerial Laser Survey data is imported into PLS8
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
387
CADD, and the conductors are graphically sagged. The Danger Tree Locator
tool is then used to determine any grow-in or falling tree violations that occur under
given parameters, for example maximum sag or blow-out conditions. Plan sheets are
produced from this report, and display visually the type and context of each violation
within a span [Fig.8.]. Anticipated growth buffers can be incorporated in the
analysis. A buffer of 20% may be added to a line that traverses through forests of fast
growing trees, where as only 2% may be added to lines traversing forests of
predominantly slow growing trees. The report provides intelligence on the three
outage categories detailed in FAC-003-1. These are grow-ins inside the ROW, fallins inside the ROW and fall-ins outside the ROW.
Figure.8: Pinpoint vegetation analysis plan sheet. This displays the nature of the
infringement, either grow-in or falling tree, against high resolution orthorectified
photography. The algorithm used to produce this example plan sheet considered
conductor position at maximum sag and a 10% growth buffer.
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
388
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
389
REFERENCES
Addison, B., Jacobs, K., Richardson, P., Ray, G., Rivkin, L., Merrett, S., (2003).
Wide-scale Application of Lidar Data for Power Line Asset Management in
National Grid Transco plc. International Council on Large Electrical Systems
(CIGRE) Committee SC22 Conference Proceedings.
11
Prepared for the 2009 Electrical Transmission and Substation Structures Conference
Nicholas Ferguson
390
Addison, B., Ryder, S., Redhead, S., Richardson, P., Ray, G., Jacobs, K., Rivkin, L.,
Browne, M. and Sullivan, T. (2003) The Application & Business Benefits of Aerial
Laser Survey Techniques used in the Management of Vegetation Encroachment
within Transmission Rights of Way International Council on Large Electrical
Systems (CIGRE).
Char, C., and Noda, R. (2006). Transmission Line Management Using LiDAR and
PLS-CADD. Electrical Transmission Line and Substation Structures Conference,
286-297.
Federal Energy Regulatory Commission [FERC] (2004) Utility Vegetation
Management And Bulk Electric Reliability Report From The Federal Energy
Regulatory Commission. Vegetation Management Report.
McGarrie, D. (2008) Vegetation Management and the Challenge of ESQCR.
http://www.sigmaseven.co.uk/files/Vegetation_Mgt_Industry_Opinion.pdf>(Jan.05,
2008) Sigma Seven.
North American Electric Reliability Corporation [NERC] (2004) Final Report on
the August 14, 2003 Blackout in the United States and Canada: Causes and
Recommendations. U.S.-Canada Power System Outage Task Force.
North American Electric Reliability Corporation [NERC] (2006) Standard FAC003-1
Transmission
Vegetation
Management
Program.
http://www.nerc.com/files/FAC-003-1.pdf>(Oct.04,2008).
North American Electric Reliability Corporation [NERC] (2008) VegetationRelated Transmission Outage Report: Third Quarter 2008.
Power Systems Engineering Research Centre [PSERC] (2007) Reliability Based
Vegetation Management Through Intelligent System Monitoring.
Shen, B., Koval, D. and Shen, S. (1999) Modelling extreme-weather-related
transmission line outages. IEEE Canadian Conference on Electrical and Computer
Engineering, 3, 1271-1276.
12
391
M.IEEE
P. Eng., M.IEEE
Craig Pon 4
Zsolt Peter 5
Principal Engineer, American Electric Power (AEP), 700 Morrison Road, Gahanna,
OH 43230, Phone: (614) 552-1944, email: [email protected]
Senior Engineer-Utility Products, Alcan Cable, Suite 1600, Three Ravinia Drive,
Atlanta, GA 30346, Phone: (770) 677-2613, email: [email protected]
Senior Engineer, Utility Markets Canada, Alcan Cable, Suite 602 East Tower,
2700 Matheson Blvd. East, Mississauga, Ontario L4W 5H7, Canada,
Phone: (905) 206-6874, email: [email protected]
Manager, T&D Mechanical Laboratory, Kinectrics, Inc., 800 Kipling Avenue,
Toronto,
Ontario
M8Z-6C4,
Canada,
Phone:
(416)
207-6741,
email: [email protected]
Senior Engineer, Kinectrics, Inc., 800 Kipling Avenue, Toronto, Ontario M8Z-6C4,
Canada, Phone: (416) 207-6000 Extn 5501, email: [email protected]
Abstract
Like many utilities, American Electric Power (AEP) has concerns regarding the
longevity of new high temperature low sag conductors using non-traditional
materials. For example, conductor designs such as ACCC (Aluminum Conductor
Composite Core) and ACCR (Aluminum Conductor Composite Reinforced) that have
entered the market in recent years as well as other new conductor designs that may be
coming in the future. While vendors of the new conductors generally run their
product through individual tests such as sheave, galloping, aeolian vibration, etc.,
they typically do not run a single sample through a sequence of tests to represent the
mechanical loadings that generally occur during installation and operation of an
actual line. For example, conductor installation (sheave testing), exposure to
galloping conditions, aeolian vibrations and tension cycles due to varying weather
loads.
AEP has developed a testing protocol with the goal of evaluating new conductor
designs in a few months (approx. 90 days) rather than having to wait years. In order
to verify that the testing protocol is not excessively burdensome, AEP has partnered
with Alcan Cable and Kinectrics to run samples of commonly used Drake ACSR
(Aluminum Conductor Steel Reinforced) and Drake ACSS (Aluminum Conductor
Steel Supported) through the same protocol. The only difference between ACSR and
ACSS is that ACSR is stranded using hard-drawn aluminum wires, while in ACSS
the aluminum wires have been annealed to remove any hardness and strength due to
the wire drawing process. This paper discusses the findings of the test program. See
Table 1 for the conductor properties.
Page 1 of 12
795 kcmil
26 x 7 ACSR
1.108 in.
31,500
1.093 lb./ft.
795 kcmil
26 x 7 ACSS
1.108 in.
25,900
1.093 lb./ft.
Background
Utilities have had extensive experience using ACSR (Aluminum Conductor Steel
Reinforced) for about 100 years and ACSS (Aluminum Conductor Steel Supported,
which is similar to ACSR but with the aluminum strands fully annealed) for over
30 years. When installed with appropriate consideration for weather loadings,
vibration concerns, etc., lines built using these conductors have operated successfully
for many decades.
In recent years, new types of bare overhead conductor have entered the market,
primarily designed for reconductoring existing overhead facilities. Two products
currently on the market are:
ACCC Aluminum Conductor Composite Core, manufactured by Composite
Technology Corporation (CTC)
ACCR Aluminum Conductor Composite Reinforced, manufactured by
Minnesota Mining and Manufacturing (3M)
These new conductor types make attractive claims of improving the electrical
capacity of the transmission grid when compared to standard ACSR. For
example [1]:
392
An obvious question for the utility considering installing a new type of conductor is:
Will the new conductor type last 50 years? 10 years? 5 Years?
Typical Testing by Manufacturers
Conductor manufacturers have typically run their new products through various
individual, stand alone, tests such as sheave, aeolian vibration and galloping. Often
these conductor tests are based on IEEE Std 1138 [2], which contains various tests for
optical overhead ground wires (OPGW). However, once the conductor sample has
passed any of these tests, the test specimen is typically discarded and a new sample
is prepared for the next test. But do these individual tests represent the real world?
Sequential Test Protocol
When a conductor is installed, it is first pulled over a series of sheaves to place the
conductor in position on the structures. After the conductor has been sagged and
clipped in, over the next few years it will be exposed to aeolian vibrations and it may
also be in an area subject to conductor galloping. Additionally, over the years and
seasons the conductor tensions will increase and decrease through numerous cycles.
Thus, AEP came up with the following Sequential Test Protocol in an attempt to gain
more knowledge regarding the long term performance of new conductors.
Except for the sheave test, all tests are to be run on samples of conductor, using
suitable compression-type fittings or other termination intended for deadending the
conductors on an actual utility project. If excessive elongation of the aluminum
strands due to compressing the fittings in short test span is a concerns, then poured
epoxy terminations may be used.
Sheave Test
A suitable length of new conductor (the longer the better) is subjected to multiple
back-and-forth passes over a sheave of a diameter and other dimensions as
recommended by the conductor manufacturer, for an installation tension of 20% of
the conductors rated strength for the Total Conductor Angle Over Sheave values
listed in Table 1, below. In the absence of a manufacturers recommendation, a lined
sheave with a minimum diameter to the bottom of the groove of [20 x DC 4] inches,
where DC is the diameter of the conductor would be used.
The conductor should be at the maximum stringing tension(s) recommended by the
conductor manufacturer for each of the Total Conductor Angle Over Sheave values
listed in Table 2.
Table 2 Sheave Test Angles and No. of Passes
Page 3 of 12
393
394
The conductor shall have a vibration loop antinode displacement (peak-topeak) of one-half () of the conductor diameter, DC.
Page 5 of 12
395
A suitable shaker is used to excite the cable in the vertical plane with its
armature securely fastened to the cable in the vertical plane.
Page 6 of 12
396
397
90
Cycle 2
Cycle 3
Cycle 4
80
70
60
50
40
30
20
10
0
0
60
120
180
240
300
The conductor sample is carefully disassembled (including the deadend fitting) and
all layers and components of the conductor are examined for cracks, etc., that could
lead to moisture coming into contact with critical materials. Carefully examine the
outer surface of the core for any defects:
Cut cross sections from the core, and
Examine the cross sections using a microscope, etc.
Perform dye penetration tests using Zyglo or other appropriate media.
Tension Cycle Test Part C Test to Failure
A new termination is installed on a suitable length of the remaining tortured specimen
and tested to failure, which should meet or exceed 95% of the rated strength of the
conductor.
Test Results for ACSR and ACSS
In order to verify that the testing protocol is not excessively burdensome, AEP, Alcan
Cable and Kinectrics partnered to prove the validity of the sequential testing protocol.
Samples of Drake ACSR and Drake ACSS (both 795 KCM 26/7) were tested at
Kinectrics using the criteria described above.
Sheave Test
Due to the setup in the lab (see Figure 1) the actual sheave portion of the testing was
somewhat more conservative than specified, as shown in Table 3.
Table 3 Actual Sheave Test Angles and No. of Passes
Specified Conductor Specified Minimum
Actual Number
Angle Over Sheave Number of Passes
of Passes
10 Degrees
20
-20 Degrees
7
27
30 Degrees
3
3
Aeolian Vibration and Galloping Tests
The ACSR conductor was supported in suspension by a standard suspension clamp
installed directly on the conductor strands for the galloping and aeolian vibration
tests.
For the galloping and aeolian vibration tests, the ACSS conductor was (initially)
supported by a standard suspension clamp installed over a set of armor rods. This
arrangement met the minimum recommendations of Alcan Cable and also was AEPs
standard practice for this product. When the armor rods were removed in preparation
Page 8 of 12
398
for the tension tests, it was found that a number of the annealed aluminum strands had
failed in fatigue in the vicinity of the suspension clamp (see Figure 4).
Page 9 of 12
399
After removal of the suspension assembly, we saw no damage to the outer aluminum
strands on the ACSS. We carefully tried to examine the inner strands (without
damaging the soft aluminum) and did not see any damage.
We then proceeded to the tension tests using poured epoxy terminations.
Tension Tests
The ACSR was then subjected to the tension cycling tests using poured epoxy
terminations (see Figure 7) without any undue occurrence. The tension in the ACSR
sample was then increased to failure, which occurred at 103.6% of the rated breaking
strength.
The second ACSS sample was also subjected to the tension cycling tests using poured
epoxy terminations without any undue occurrence. The tension in the ACSS sample
was then increased to failure, which occurred at 109% of the rated breaking strength.
Page 10 of 12
400
401
(in tension) but had partially cracked during the aeolian vibration and/or galloping
test and would have eventually severed at some point in the future. At this point this
is not known. Additionally, later examination of this ACSR sample at Alcan's
Technical Center revealed extensive fatigue damage to the sample, with fatigue
breaks in the inner layer under the suspension assembly.
Typical Tensile Failures
Under Metal
Clamp (first test)
ACSS
Under Shaker
Clamp
Under Cushioned
Clamp
= Steel Strands
Page 11 of 12
We feel that these results prove the rigor of the testing protocol. Although it will not
provide any direct measure of the potential longevity of a new conductor type, the use
of the Sequential Mechanical Testing of Conductor protocols described above may be
a valid method of collecting information and may assist in identifying potential
weaknesses in conductor types. At the moment there are few data points in this
collection and conclusions cannot be drawn more precisely.
At AEPs request, Composite Technology Corp. (CTC) ran a sample of their ACCC
conductor through the above test protocol. As a result of this, they are considering
including new design criteria and recommendations to users regarding design
limitations for their conductor as presently specified in their data sheets.
References:
[1] From CTC and 3M websites:
(http://www.compositetechcorp.com/cabadv.htm);(http://solutions.3m.com/wps/porta
l/3M/en_US/EMD_ACCR/ACCR_Home/Proven_Benefits/More_Amps/)
[2] IEEE Std 1138-1994 (R2002) IEEE Standard Construction of Composite Fiber
Optic Overhead Ground Wire (OPGW) for Use on Electric Utility Power Lines,
Copyright by the Institute of Electrical and Electronics Engineers, Inc.
Page 12 of 12
402
403
President, SANPEC, Inc, 11819 Skydale Drive, Tomball, TX 77375; PH: (832) 3924230, FAX: (832) 550-2634; email: [email protected]
2
INTRODUCTION
404
A. Background
Transmission lines T436 (now T522) and T437 were constructed in 2002 by utilizing
direct embedded steel poles. These lines are located in Corpus Christi area. AEP first
noticed a rust spot on pole # 106 (T522) in 2004. The assumption was made that the
pole had been struck by lightning. The rusty area was painted with cold galvanize
paint. Four years later, Jan 2008, AEP reported a visible exterior corrosion on another
pole # 71 (T437). Structures #106 and #71 are adjacent to each other and are
separated by approximately 60 feet.
LCRA Transmission Lines:
T437 #71
T522 # 106
LCRA SAMPLING
On June 2008, LCRA sampled suspected toxic atmospheres within six (6)
transmission poles on both lines (T437 and T522 - Pole # 70, 71, and 72 on T437 and
Pole # 105, 106, and 107 on T522). The sampling target compound was H2S gas.
The sampling team confirmed the presence of H2S gas within two (2) poles. Pole #
106 contained harmful concentrations of H2S gas and elevated levels of flammable
gas. Pole # 71 contained deadly concentrations of H2S gas and explosive
2
405
concentrations of flammable gas. Both the poles had severe external corrosion on the
top of the poles.
SAMPLE RESULTS:
The following results represent data gathered on site on June 2nd and 3rd, 2008 with
two (2) calibrated Industrial Scientific M40 4-gas monitors, and four (4) calibrated
ToxiRAE 3 H2S personal monitors.
Structure
#
106
106
106
106
106
106
71
71
~
Height
(ft.)
*0
**25
**50
**60
**70
***75
*0
**25
H2S
(ppm)
0.0
0.0
0.0
0.0
0.0
47
0.04
0.0
% Lower
Explosive
Limit
0
0
0
0
0
8
0
0
Oxygen
21.0
21.0
21.0
20.9
20.9
< 19.5
21.0
21.0
Personal
H2S
Monitoring
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
71
71
71
71
**30
**50
**60
**70
0.0
0.0
0.0
0.0
0
0
0
0
21.0
21.0
20.9
20.9
0.0
0.0
0.0
0.0
71
***80
> 500
> 10
< 19.5
0.0
Notes
5 minute
sample
peak
Seam at
upper/
lower
structure
connection
2 second
sample
III.
406
Pole # 71
Pole # 106
Pole # 106
407
408
C. H2S Physical and Chemical Properties: Two atoms of Hydrogen and one
atom of sulfur
Description: Colorless Gas
Gas density: 1.4 g/L @ 25 C
STEL: 20 ppm
409
E. Corrosion Theory
Corrosion is defined as the destruction of a metal by a chemical or electrochemical
reaction. It occurs when a metal in contact with water forms a corrosion cell. The
corrosion cell has four components, the aqueous phase (water) which acts as an
electrolyte (through which ions migrate), an anode on the metal surface (where the
metal is oxidized and goes into solution as metal ions), a cathode (where excess
electrons are consumed) and a metallic path connecting the cathode to the anode.
410
Bisulfide ion may react with the dissolved metal ions and precipitate as metal
sulfides. Hydrogen sulfide is a weak acid when dissolved in water, and can act as a
catalyst in the absorption of atomic hydrogen in steel, promoting sulfide stress
cracking (SSC) in high strength steels
Fe + H2S -> FeS + H2
Hydrogen is evolved as a corrosion by-product. Main corrosion by-product that
indicates H2S corrosion is Pyrite (FeS2 ), Pyrrhotite (Fe7 S8 ), and troilite (FeS), which
are iron sulfides. In environments with hydrogen sulfide (H2S) corrosion, the most
common types include uniform corrosion, pitting corrosion, corrosion fatigue, sulfide
stress cracking, hydrogen blistering, hydrogen embrittlement, and stepwise cracking
IV.
CONCLUSION
For public safety, LCRA decided to remove the embedded steel poles, which had the toxic
and potentially explosive H2S gas entrapped inside. Two replacement base-plated steel poles
were ordered. Traditional concrete pier foundations were used with Type V cement (high
sulfate resistance) and Class F fly ash (beneficial to sulfate resistance). The conductors and
shield wires were transferred from existing poles to new poles
A. New base plated galvanized steel poles Installation:
411
B. Unique design and method to remove H2S gas and embedded steel poles:
412
10
V.
413
REFERENCES:
Currah, Mitch, and Cooper, Andy (2008), Transmission & Substation Design &
Operation Symposium (TSDOS) -138kV Transmission Pole H2S
Corrosion Remediation
Gundiler, Dr. Ibrahim (2005), New Mexico Water and Infrastructure Data System
http://octane.nmt.edu/waterquality/corrosion/H2S.htm
CUDD Energy Services (2009), Power Transmission Pole Purge
REA BULLETIN 1751F-670 (1993), Outside Plant Corrosion Considerations
11
414
II.
415
this study was to find the limits of potential error, which required in over 37
billion calculations to reach the following observations. The variations
identified are the extreme limit of fluctuation, the actual amount of error in
each case is expected to be less. The variables include:
a. Date & Time The date and time variables are easy input variables to
be very accurate. The 2006 update to the IEEE Standard 738 added
sensitivity to the equations to account for small incremental
adjustments to the suns angle to the wire and the resulting impact on
the conductor. There is a period of time required to survey a
transmission line, this time would introduce some error as many of the
input variables could vary. The amount of this error is very difficult to
predict and for the purposes of this paper will be assumed as minimal.
b.
c. Latitude & Elevation The latitude and elevation input variables are
also easy to be very accurate as they relate to geospatial location of the
transmission line under evaluation. Each transmission line is different
and that would cause some variation in the input variables, but if either
latitude or elevation caused significant variation in the results of the
IEEE calculations, each section of line could use the average latitude
and elevation of the corresponding section. The amount of this error is
assumed minimal for the purposes of this paper.
d. Line Azimuth The azimuth of a transmission line is also an input
variable that can be determined per section of line, although any type
of running angles present in a section of line makes it difficult to
determine since the azimuth varies. Once again minimal error is
expected by small variations in an azimuth angle of a line; although, if
the angle varies significantly, 15 to 20 degrees F. of variation occurs
and it would be expected that some error would be included in the
IEEE results for heavy and light loaded transmission lines.
Wind Pressure The wind input variable is also a difficult variable to
collect. Weather data is usually collected during the survey, but it is
typically near a survey control point which could be some distance
from the transmission line and the weather station is typically at an
elevation near the ground and possible shielded by surrounding
vegetation. The transmission conductor on the other hand is typically
25 50 feet above the ground, where the expected wind would be
416
greater. The variation results between a 0 foot per second wind and an
8 foot per second wind vary from 50 degrees F. on lightly loaded lines
to 380 degrees F. on heavily loaded lines, it is expected that significant
error would be included in the IEEE results. The amount of error
expected would also increase as the load on the transmission line
increases.
Figure 1 - This graph shows the thermal gain comparison between a "No Wind" condition
and an "8 ft/s Wind" condition.
Figure 2 - The difference between the curves in Figure 1 is represented this graph to show
the potential error that could exist between the two wind pressure values of "No Wind"
and "8 ft/s".
417
Figure 3 - Similar to the effect of Wind Pressure is the Wind Angle, this graph shows
difference in thermal gain between a parallel wind and a perpendicular wind on the wire.
418
Figure 4 - This graph shows difference in thermal gain between the extreme values of
Emissivity.
The variation results for solar absorption was much less and varied
from 25 degrees F. on a lightly loaded line to 20 degrees F. on a
heavily loaded line. The amount of error expected would be between 5
10 degrees F.
419
great application for finding hot spots in substations, transmission lines are
typically too high to measure and the sky washes out the sensors ability to
measure. The other difficulties would be to calibrate the solar absorption and
emissivity coefficients which as discussed earlier can dramatically impact the
predicted temperature.
Direct Measurement method
There are a few instruments that have been produced that can be mounted on a
hotstick and safely measure the surface temperature of the conductor while energized
or the resistance of the conductor, which then can produce the conductor temperature.
I. Temperature Probe
These instruments have proven to be accurate to less than one degree for the
surface temperature of a conductor. There are three drawbacks for this
method; first, instrument availability; second, utility operations practices for
hotline clearances and third, there is an increase in cost to collect these
measurements. The availability of these instruments has diminished due to
lack of demand from the utility sector and therefore that are currently no
known manufacture of a hotline temperature probe. Many utilities have
adopted safety policies which prohibit any contact with a transmission line
while circuit reclosures are in place therefore making it more difficult to get
hotline clearances, especially on heavier loaded lines. Some utilities have
reviewed the procedure of collecting a temperature reading and were able to
get an exception to that type of policy and still collect temperature readings.
The last concern of cost for the additional personnel, which is typically a very
small percent of LiDar line rating project and easily offset by the value of the
information collected, where the cost will be offset by the savings of being
more accurate.
420
Figure 6 - Sensorlink's Ohmstik modified to make resistance readings over five feet instead
of one foot to increase the accuracy of the instrument.
421
422
423
equivalent wind loading and direct dynamic analysis of TL structures under dynamic
wind loading for two TL structures.
The first study concerns a guyed tower, which is more flexible than typical
self-supporting towers and have a dynamic behaviour that cannot be captured easily
by a static approach. The dynamic interaction between the cables (conductors,
ground-wire and guyed cables) and the tower mast could greatly contribute to
possible dynamic amplification under dynamic wind loading.
The second study deals with a river crossing, which have towers and cable
spans with much greater dimensions than typical TL structures. Spatial variability
under turbulent wind loading has more influence on the overall magnitude of wind
loading applied to the structure.
ANALYSIS PRINCIPALS
For the dynamic wind analysis of these TL, it is necessary to: (i) generate a whole
wind field with the right spatial and temporal correlation to be applied on the TL; and
(ii) model adequately the tower and conductors so the structural behaviour is
accurately represented. The methods used for the wind generation and structural
modelling are presented in this section.
Transmission line finite elements model. The FE models of the TL structures
include geometrically non-linear static and transient-dynamic (TD) analyses. The
cables (conductors, ground-wires and guy-cables) are modelled using tension only
cable element with large displacement-small strain kinematics. For the damping
consideration, aerodynamic damping is modelled in accordance with current practice
(Gani & Lgeron, 2008). The structural damping is assumed as 0.5%. When
truss/lattice tower is modelled, in order to reduce the computational burden, the lattice
tower mast was modelled as equivalent beam-column elements using equivalent
plate-thickness theory (Meshmesha et al., 2006).
Wind loading. The wind forces were applied on a set of chosen points along the span
and the tower. The complete application of these wind points on the FE model will be
shown in the next sections. The static-equivalent wind loading is calculated in
accordance with the current industry documents (e.g. IEC-60826, 2003; CENELEC
EN-50341-1, 2000; ASCE Manual No.74 Draft, 2005).
In a real structure, under a real wind, at a given time, the dynamic wind force
on the conductor is varying along the span. The spatial correlation is important as it
has considerable influence on the propagation of the conductor force along the
conductor span. The dynamic wind forces are also varying along the tower height at a
given time. Wind force on structures depends on the total wind speed U(z, t), which is
random in nature. The wind speed U(z, t) is divided into two parts, the mean wind
part U(z) and the turbulent wind part u(z, t).
For TD analysis, the turbulent wind part u(z,t) was generated numerically
using the in-house program WindGen (Hang et al. 2005), based on spectral
representation method as detailed in Deodatis (1996). Using this numerical approach,
424
the spatial and temporal correlations of the wind loading are taken into account.
Verifications of the generated wind speed and the target spectra showed good
agreement. An example of the auto-spectra verification is given in Figure 1. The
generated wind speed was then incorporated into the FE model as a time-varying
loading. Further details on the dynamic wind loading analysis were given in Gani &
Lgeron (2008). It is to be noted that if real measured wind data are available for the
structure to be analysed, this real time wind record could also be incorporated to the
FE model accordingly. However, due to the size of a real structure it is unpractical to
record real wind data for the analysis.
425
Description of the structure and wind loading. A single mast guyed lattice tower
(suspension type) with four guy cables and a pinned base connection was chosen
(Figure 2 &Figure 3). The guy cable configuration is as such that torsion effect is
minimized and possibility of large moment due to unbalanced longitudinal load is
reduced. The tower geometry is adapted from the guyed tower for 450kV DC used in
study by Oakes (1971). To reduce the computation burden, the truss tower masts were
simplified to equivalent beam-column pole model. Details of the tower modelling
used in this study are given in Gani & Lgeron (2008). Figure 2 shows the
calibration between the full truss model and the simplified model.
To evaluate different dynamic guyed tower response, two guyed towers were
analysed (see Figure 3): a two-phase tower that would be used for a direct-current
(DC) line with both conductor arms located above the guy-cable attachment point;
and a three-phase tower that would be used for an alternating current (AC) line with
two conductor arms located below and one conductor arm located above the guycable attachment point. The tower height is 52.18 m, the height of guy cable
attachment point on the tower mast is 38.04 m and the conductor and ground wire
span length is 480 m.
A total of four loading cases, with and without ice, were studied. For brevity,
only two cases will be detailed here. The two phase guyed tower with mean wind
U(10) = 35 m/s normal to the line (Case 2P90w) and the three phase guyed tower
with the same wind condition (Case 3P90w) will be discussed here. Wind points
applications were as shown in Figure 4. The terrain type is B in which the roughness
characteristics are categorized as open country with very few obstacles (IEC-60826
2003) with roughness length z0 = 0.05 m. For comparison with static-equivalent of
IEC-60826 (2003), the TD analysis used 10 wind record samples of ten minutes each.
The TD results detailed herein are the mean of maximum of the available samples.
426
Figure 3. FE model outline: (a) two-phase guyed tower; (b) three-phase guyed
tower
427
428
tend to result in lower response than the dynamic method on locations that could be
considered critical for the structural design on the tower mast.
Therefore, it can be concluded that dynamic effects due to wind loading for
guyed tower type supports is not fully taken into account in the static-equivalent
method provided in IEC-60826 (2003). Considering the complexity of the transient
dynamic analyses performed in this study, it would be of interest to find a simpler
way that bridges the gap between the two methods and includes the dynamic effects
of guyed towers.
Table 1. Summary of results for guyed tower wind study: Static-equivalent vs.
TD analyses
Parameters
Tower
transversal
reaction (kN)
Average
conductor
tension (kN)
Case 2P90w
IEC-60826
TD
Case 3P90w
IEC-60826
TD
3.96
11.75
14.49
18.29
48.92
45.95
34.39
33.59
SRSS could predict relatively well the TD response, except for the area near the guycable attachment point in Case 2P90w, Case 3P90w where the static-equivalent
approach provides good results.
Figure 5. Transversal bending moment: (a) Case 2P90w; (b) Case 3P90w
RIVER CROSSING WIND STUDY
River crossings generally consist of long-span conductors that are usually outside the
range of span and height considered in transmission line industry documents. As
defined in CIGR SCB2.08 (2009), river crossings or large overhead line crossings, are TL
structures with spans larger than 1000 m and/or with tower heights larger than 100 m. In IEC
60826, the limitation for normal span range is between 200 and 800 m and height of
tower is up to 60 m. In this standard, for span over 800 m and height above 60 m, the
criteria given there, could still be used, but the validity of the approach could be
questioned. This is due to the fact that the spatial correlation for very long-span
conductor (limitation of gust widths) is generally not taken completely into account in
the current industry documents.
Comparison between the wind loading defined in the industry documents and
the one actually encountered in the field has been a continuous effort for transmission
line designers (Houle et al., 1991). Their study were based on normal range of
conductor span (max span tested was 457 m), with wind speed up to 30 m/s in terms
of gust velocity at the conductor level, for terrain with very few obstacles, typical for
normal transmission lines. Over the years, with the continuous improvement, it has
429
been well accepted that the requirements given in the industry documents are
providing relatively good agreement for wind loading of conductor within the normal
span range. For overhead line crossing, recent study by Paluch et al. (2007), which
was based on field observations, did not include comparison with the requirements
from the code utilized in that study.
Description of the river crossing. This study was focused on a typical overhead
transmission line crossing. The maximum span of this river crossing is 1100 m and
the maximum height of the tower is 145 m (Ghannoum et al., 1990). It is composed
of 3 suspension towers and 2 dead-end towers. The two-phase transmission line is
spaced out by 26 m and it carries two conductors and one ground wire per phase. The
conductors are attached to the towers by 4 insulator chains. For an overview of the
overhead-line crossing, see Figure 6. For this study, only the insulator strings and the
main spans of the conductors were modelled.
Comparison with the current industry documents. Before pursuing the wind
loading analysis on the river crossing, a benchmark study was done using normal
three-span conductors (380 m - 450 m - 380 m). While for the river crossing, the
three-span conductors consists of 930 m -1100 m - 917.6 m, respectively. The
summary of results is given in Table 2. From the results obtained, it could be seen
that for the benchmark, the transmission line industry documents IEC-60826 and
CENELEC EN-50341 are comparable to dynamic analysis, with less than 8%
difference. The code ASCE Manual 74 underestimated the wind loading by as much
as 20% if compared to the other documents and the dynamic analysis result. For the
river crossing three-span conductors, it could be observed that this time the difference
for transversal force between the dynamic results and the ones from the documents is
more pronounced, by up to 18% if compared to CENELEC. Hence, it has been
highlighted that for the river crossing, the percentage of difference between the
industry documents and the dynamic analysis is relatively higher than for the normal
span range from the benchmark study. To illustrate that the industry documents
430
provide relatively higher wind loading, Figure 7 shows a sample of 10 minutes time
history of the transversal forces obtained from dynamic analysis versus the values
obtained from these documents.
Table 2. Forces transmitted to the tower
Model
Force (kN) IEC ASCE CENELEC NL Dynamic
Benchmark
Transversal 43.47 34.41
44.77
41.35
River-crossing Transversal 228.9 223.00
249.00
210.71
431
432
Total
transversal
(kN)
204.67
196.61
192.00
Longitudinal
5.01
6.86
9.73
16.52
23.26
35.16
Vertical
2.80
4.19
5.74
Torsion
(kNm)
82.60
290.75
879.00
CONCLUSION
The present study has shown the versatilities of dynamic analyses for TL structures.
Each structure presented here pointed out different aspects of the necessity for more
complete dynamic analyses.
For the guyed tower study, depending on the tower configuration and the type
of climatic loading, the dynamic effect could have a large influence in the overall
response of the tower mast under wind loading. Therefore, the application of staticequivalent method, which is an algebraic summation of mean wind effect and
turbulent wind effect, does not always provide a conservative estimate of the possible
dynamic response.
While in the study of the river crossing, a more complete consideration of the
spatial correlation resulted in lower wind loading than the ones obtained from the
industry documents. Also, the necessity to reconsider the current practice for large
parallel spacing for the tower cross-arms was highlighted.
Overall, the present study is a first step towards a more reliable and
economical design for overhead line crossings where the conductor spans are longer
than what are covered in the current transmission line industry documents. Further
study that involves comprehensive field measurement of the wind loading on site
specific terrain or river crossing would be interesting in order to validate the
numerical approach proposed here.
ACKNOWLEDGEMENT
This study was carried out as part of the research projects of the Natural Sciences and
Engineering Research Council of Canada (NSERC)/Hydro-Qubec Trans-nergie
(HQT) Industrial Research Chair on Overhead Transmission Line Structures at
Universit de Sherbrooke, Quebec. The financial support provided by NSERC and
HQT is well appreciated.
REFERENCES
ADINA R&D Inc. (2004). Automatic Dynamic Incremental Nonlinear Analysis (ADINA).
ADINA R&D Inc.
ASCE Manual No.74 Draft (2005). Guidelines for electrical transmission line structural
loading, American Society of Civil Engineers, USA.
Battista, R.C., Rodrigues, R.S. and Pfeil, M.S. 2003. Dynamic behavior and stability of
transmission line towers under wind forces. Journal of Wind Engineering and
Industrial Aerodynamics, 91, 1051-1067.
CAN-CSA-S37.01 2001. Antennas, towers and antenna supporting structures. Canadian
Standard Association.
CENELEC EN-50341-1 (2000). Overhead electrical lines exceeding AC 45 kV, European
Committee for Electrotechnical Standardization.
CIGR SCB2.08, (2009). Large overhead line crossings. CIGRE. Technical brochure to be
published by CIGR.
Deodatis, G. (1996). Simulation of ergodic multivariate stochastic processes, Journal of
Engineering Mechanics, 122(8), 778-787.
Gani, F. and Lgeron, F. (2008), Dynamic wind loading analysis of transmission lines with
guyed towers, submitted to Canadian Journal of Civil Engineering.
Ghannoum, E., Bell, N., Binette, L., Ouellet, P. , Manoukian, B. (1990). Use of reliability
techniques in the design of Hydro-Quebecs 1100 km Radisson-Nicolet-des-Cantons
450kV HVDC line, Proceedings of the 1994 35th International Conference on
Large High Voltage Electric Systems: Part 1, Paris, France, no. 22-201 (7p).
Hang, S., Gani, F. and Lgeron, F. (2005). Gnrateur de vent appliqu au gnie civil:
WindGen, Manuel du logiciel, Universit de Sherbrooke.
Houle, S., Hardy, C. and Ghannoum, E. (1991),Static and dynamic testing of transmission
lines subjected to real wind conditions, CIGRE Symposium: Compacting Overhead
Transmission Lines, Leningrad, USSR, Paris, France, no. 200-02 (6p).
IEC-60826 (2003). International Standard: Design criteria of overhead transmission lines,
International Electrotechnical Commission.
Mathur, R.K., Shah, A.H., Trainor, P.G.S. and Popplewell, N. (1987). Dynamics of a guyed
transmission tower system. IEEE Transactions on Power Delivery, 2(3), 908-916.
Meshmesha, H., Sennah, K. and Kennedy, J.B. (2006). Simple method for static and
dynamic analyses of guyed towers. Structural Engineering and Mechanics, 23(6),
635-649.
Oakes, D.L.T. (1971). Nelson River HVDC transmission line towers. In Proceedings of the
Manitoba Power Conference EHV-DC, Winnipeg, Manitoba, Canada, 7-10 June
1971, 441-458.
Paluch, M.J., Cappellari, T.T.O. and Riera, J.D. (2007). Experimental and numerical
assessment of EPS wind action on long span transmission line conductors, Journal
of Wind Engineering and Industrial Aerodynamics, 95, 473-492.
433
Simiu, E. and Scanlan, R. (1996) Wind Effects on Structures, 3rd Edition. John Wiley & Sons
Inc.
Sparling, B.F. and Davenport, A.G. (1998). Three dimensional dynamic response of guyed
towers to wind turbulence. Canadian Journal of Civil Engineering, 25, 512-525.
Tsui, Y.T. (1978). Dynamic behavior of a pylne chanette line: Part I Theoretical
studies. Electric Power Systems Research, 1, 305-314.
Yasui, H., Marukawa, H., Momomura, Y. and Ohkuma, T. (1999). Analytical study on
wind-induced vibration of power transmission towers. Journal of Wind Engineering
and Industrial Aerodynamics, 83, 431-441.
434
435
436
437
Tower members are designed to carry tension and compression loads only. The detailing
of tower member joints takes special care to avoid inducing bending in these members.
The removal of any tower member may create bending stress in the adjacent remaining
members. Often times the remaining members can safely carry the additional forces
caused by removing a member. The PLS-TOWER model is very useful for removing
members and checking the structural stability of the tower. Removal of multiple members
should be avoided if possible. This type of removal may induce bending stress in the
adjacent members, which the members are not designed to carry. When multiple
members are to be replaced, the best practice is to replace one at a time. Replacement of
tension members such as the diagonal at the bottom of the tower can be achieved with the
addition of a come-a-long where the member is located before it is removed. Member
removal for this type of repair work requires checking by a qualified individual
experienced with this type of repair. Remember, tower members are under stress and
construction personnel will need to use extreme caution while making repairs.
438
becomes inefficient. It is important to note that sometimes clearances may dictate the
radius of the crane, as minimum clearances to high voltage lines must be maintained.
ASME 30.5 shows minimum required clearances to energized lines of various normal
voltages for cranes.
439
The next step after determining the correct crane radius is to determine the crane size.
This is not as simple as it may seem. Let us investigate what OSHA regulations say about
sizing the crane.
1926.550(a)(1)
The employer shall comply with the manufacturer's specifications and limitations applicable to the
operation of any and all cranes and derricks. Where manufacturer's specifications are not
available, the limitations assigned to the equipment shall be based on the determinations of a
qualified engineer competent in this field and such determinations will be appropriately
documented and recorded. Attachments used with cranes shall not exceed the capacity, rating, or
scope recommended by the manufacturer.
This specification only allows 85 % (for outriggers fully extended and set) of the ultimate
capacity of the crane to be published in the manufacturers load charts. The specification
requires a 25 % reduction for cranes without outriggers fully extended and set. This
reduction is already taken into account in crane manufacturers load tables.
Assuming outriggers to be fully extended and set for most utility cranes, this translates
into a safety factor
F.S. = 1/.85 = 1.18
However, the reaction load calculated on the tower is a dynamic load (due to variable
wind loading). A minimum factor of safety of 2.0 is recommended. Ultimate capacity =
F.S. x S.W.L.
This is in line with most utilities safety standards. Therefore, a factor must be applied to
the safe working loads of the tower reaction to have a factor of safety of 2.0 on the crane
capacity.
Correction Factor = 1 / (1.18 / 2.0) = 1.7 (use instead of 2.0 for sizing the crane because
a 15% strength reduction is already used in the crane chart). It is the purpose of this factor
to translate a static load factor into a dynamic load factor. This factor can be used by field
personnel to multiply the working loads in order to use the manufacturers load tables
directly.
Note that this results in a crane that has 44 % more capacity than if the minimum OSHA /
ASME factors are used.
440
ASME 5-1.11 States that for cranes used for non-lifting purposes the load ratings shall be
established by a qualified person. This qualified person, either from the utility or
contractor, must understand the load factors or apply them correctly. Many times, there
are only a limited number of crane sizes available. A crane can always be chosen that far
exceeds the capacities specified here. In many cases, availability will dictate if this is the
case.
Many times, a factored load with an overload factor already applied is given by the
responsible engineer or utility. It is important to know what overload factor was used so
as not to apply more than one load factor to the safe working load (SWL).
Welded Support
It has been found to repair localized damage to towers, a temporary support can be
welded to the tower member spanning the damaged area, the damaged area removed and
butt spliced, and then removed.
Several
drawbacks
to
this
construction method make it less
desirable. These include but are not
limited to: wariness of field welds
by utilities, many variations in
support member configurations /
sizes, and the need for certified
welders in the field. The temporary
welds that attach the temporary
member to the tower are usually
ground flush and recoated to avoid
future corrosion. The temporary
support member should be
engineered and sized for load and
the welds or bolted connections
should be specified.
Figure 5 Example of Welded Temporary Support
Member
This is quite difficult, as field personnel tend to use what is readily available for this
member.
441
442
443
The streets of Muscat were almost deserted, thousands of residents were evacuated
from Eastern Coastal areas. The Army, Police and Civil defense troops were
mobilized and about 18,000 people were evacuated across the Country.
444
445
446
447
Exhibit No. 11: Unique solution for immediate Restoration of power by using
wooden frame
448
449
KEC, organized materials, men and machines at site on 12th June 2007 and
completed the entire proposed work with in 8 days and commissioned one
circuit on 20th June 2007.
450
451
452
ABSTRACT
Existing structures were aluminum double-circuit lattice towers with a single circuit
installed and had to be evaluated for the proposed addition of a second circuit.
Original design data drawings were not available for these towers. Therefore, field
measurements of towers with an energized single circuit were performed to determine
overall tower geometry, individual member sizes and connection configuration. Field
hardness testing and sample coupons were taken for laboratory analysis to determine
material properties.
Loadings for the towers were determined in accordance with the 2007 edition of the
National Electrical Safety Code. Analysis and qualification of the towers were
performed using the software program Tower by Power Line Systems (PLS
Tower), and the Guide for the Design of Aluminum Transmission Towers as
published in the December 1972 Journal of the Structural Division, Proceedings of
the American Society of Civil Engineers. The existing foundations were checked
against the support reactions from the PLS Tower Analysis.
This paper describes the various analysis challenges encountered in this project.
453
PROJECT REQUIREMENTS
Four existing structures were aluminum double circuit lattice towers with a single
circuit installed and had to be evaluated for the proposed addition of a second circuit.
Three towers are dead-end towers and one tower is a tangent tower supported by a
portal frame as shown in Figure 1.
FIELD INSPECTION FOR TOWER DIMENSIONS
Original design data drawings were not available for these towers and, therefore,
tower dimensions and member sizes were not known.
Mesa personnel along with linemen from Shaw Energy Group field measured the four
existing double-circuit aluminum transmission towers to determine overall tower
geometry, individual member sizes, and connection configuration. Dimensions of legs
and web members were carefully taken by maintaining electrical clearance from the
energized conductors. Cross arm dimensions were taken from the un-energized side
of the towers.
Field measured dimensions of one of the towers are shown in Figure 2. Cross section
area and moment of inertia of the unusual shape of the members with bulb or stem at
the end were calculated using AutoCAD program.
454
Figure 1.
Tangent Tower Supported By A Portal Frame
(Linemen taking dimensions, individual member sizes, and connection
configuration on this energized tower)
455
Figure 2.
Field Measured Dimensions of One of The Towers
456
457
Alloy
6070-T6
6061-T6
Fy
45 ksi
35 ksi
Fu
48 ksi
38 ksi
Modulus of
Elasticity
10,000 ksi
10,000 ksi
5/8 diameter tower bolts are assumed as aluminum alloy 2024-T4 with the following
properties: Fs = 37,000 psi, Fu = 62,000 psi
Test results for the above materials are attached at the end of this paper, see Figures 3
and 4.
DESIGN LOADS
Line loads were determined in accordance with the 2007 edition of the National
Electric Safety Code.
NESC
Section
250B
250C
250D
Description
Medium wind and ice
Extreme wind
Light wind on extreme ice
458
Values
40 mph wind on radial ice at 15
90 mph wind, no ice at 60
30 mph wind on radial ice at 15
ANALYSIS
Using the field measurements, the towers were modeled in the software program
Tower by Power Line Systems (PSL Tower). The members were checked against
two independent design criteria: ASCE 10-97, Design of Latticed Steel
Transmission Structures, and the article Guide for the Design of Aluminum
Transmission Towers as published in the Journal of the Structural Division,
Proceedings of the American Society of Civil Engineers, December 1972. In general,
the ultimate stresses allowed by the ASCE Guide are more conservative than those
allowed by ASCE 10-97.
The maximum stress level in each tower from the PLS Tower program is given
below:
Tower
No.
1A
2A
3A
4A
250B
250C
250D
Diagonals
73.8%
97.5%
67.3%
46.4%
46.1%
35.8%
51.7%
54.6%
38.2%
Diagonals
58.0%
91.7%
69.6%
56.4%
81.8%
38.6%
31.5%
53.2%
15.8%
Diagonals
95.7%
94.8%
85.5%
57.7%
49.4%
44.8%
64.6%
57.5%
49.0%
Diagonals
87.2%
98.0%
57.1%
22.9%
25.3%
18.5%
25.7%
29.7%
20.7%
The steel portal frame supporting one of the four towers and spanning over an access
road was analyzed in STAAD Pro using the tower support reactions from PLS Tower.
All existing foundation footings consist of four steel piles each with a concrete pile
cap and concrete pedestal.
The existing foundations, including steel piles, were checked against the support
loads from the analysis in assumed soft clay and found to meet the appropriate factor
of safety requirements.
CONCLUSION
All four steel towers, the steel portal frame, and the foundations are capable of safely
supporting the addition of a second circuit with no structural modifications.
The project was successfully completed in the year 2007 due to excellent preplanning, dedication, and cooperation from various personnel from the following
organizations:
1.
2.
3.
4.
459
Figure 3.
Test Results, Page 1
460
Figure 4.
Test Results, Page 2
461
462
3
4
Ziqin Li 2
PhD P.Eng., SM.IEEE
Bruce Freimark 3
P.E., F.ASCE
Teja Rao 4
AVNG
Chief Engineer, Kinectrics, Inc., 800 Kipling Avenue, Toronto, Ontario M8Z-6C4,
Canada, Phone: (416) 207-6047, email: [email protected]
Principal Engineer, Kinectrics, Inc., 800 Kipling Avenue, Toronto, Ontario
M8Z-6C4,
Canada,
Phone:
(416)
207-6000
Extension
6489,
email: [email protected]
Principal Engineer, American Electric Power (AEP), 700 Morrison Road, Gahanna,
OH 43230, Phone: (614) 552-1944, email: [email protected]
Engineer, American Electric Power (AEP), 700 Morrison Road, Gahanna, OH
43230, Phone: (614) 552-1865, email: [email protected]
Abstract
Starting in 1969, American Electric Power (AEP) built a 765kV transmission network
that amounted to over 2,000 circuit miles by 1984. The system utilized four
conductor bundles of Rail and Dipper conductors. Operational experience with the
network revealed issues around audible noise due to corona from the lines.
In 1990, when the need arose to expand the network, it was decided that audible noise
issues would be addressed by utilizing a new bundle design that consisted of a
6-conductor bundle using Tern conductors. The intent of the design change was to
eliminate corona and thus resolve the audible noise issues associated with previous
designs. The new design would also allow the use of non-ceramic insulators, which
are susceptible to damage from long term exposure to corona. The move to the
6 x Tern phase bundle posed significant mechanical and electrical challenges to the
engineers involved in hardware development.
This paper describes the extensive electrical testing programs carried out to prove the
performance of the new 6-conductor bundle design.
The Traditional Method of Corona Testing
Traditionally corona testing has been performed in laboratories by mounting a
single-phase full size mock-up of the conductor-hardware-insulator assembly at a
height above the ground representing the minimum design clearance of the line and
applying 110% of the rated line-to-ground operating voltage. If the test setup is
Page 1 of 11
shown to be free of corona by this test, then it is considered that the assembly will be
free of corona under operating conditions. This method does not appear in any
standards, but is used as a generally accepted test method. In spite of its general
acceptance, this test method can give erroneous results. This is due to the fact that the
inception of corona occurs at a given electric field gradient rather than a given
absolute voltage. Under actual operating conditions, the electric field gradient at the
conductor-hardware-insulator assembly is a function of phase spacing, the local
geometry, the applied 3-phase voltage, and the lines elevation (air density). In order
to correctly perform such a test in a laboratory, it is essential that the maximum
gradients occurring on the conductors in the field be reproduced in the laboratory test.
Introduction to the Improved Method
To reflect the true operating environment, corona testing of equipment should be
carried out at a specified gradient related to the actual system-operating gradient. To
establish this gradient in the test setup, three alternatives are possible. These are:
1. Build an accurate mock up of the three-phase system utilizing representative
energized conductors, phase spacing and ground planes (including the
structure), etc.
2. Test the hardware on a conductor of the same size upon which will be used in
service and set the test gradient to that representative of the maximum 3-phase
gradient present at the location where the test object will be installed. The test
gradient can be obtained by calibrating the test setup through the use of
corona calibrating spheres.
3. Test the hardware as above in 2, but calculate the applied voltage necessary to
achieve the required gradient for the test setup used.
The first of these alternatives requires the use of a 3 phase supply, and an extremely
large laboratory in which real setups can be constructed. In most High Voltage labs
this is not possible. Therefore, the latter two alternatives are of most interest. These
methods are based on the use of calculated or calibrated field gradients which are
used during the testing.
The third alternative is based upon the calculation of the electric field surrounding the
insulator assembly or hardware under test. Currently, there are a number of electric
field plotting software packages available. Some of these are commercially available,
while others have been developed in-house by research and development institutes
engaged in testing, which requires characterization of the electric fields surrounding
high voltage equipment. While such software has undergone marked improvement
over the past decade, there is still significant concern on the effects of modeling
details on calculated results. The second method is based on the use of gradient
calibrating spheres. While this method is slightly more time consuming to
implement, it has the advantage of actually measuring the electric field in areas of
Page 2 of 11
463
interest in the test setup while avoiding the need to create an accurate mathematical
model.
A calibrating sphere is a small sphere, which is mounted on a circular clip and can be
attached to a conductor [1, 2]. Figure 1 shows the design details of the calibrating
sphere, and how it is attached to the conductor under test.
464
corona-free was particularly important in this case since polymer insulators were used
extensively on the line. The tests were intended to demonstrate that the new design
would be corona free under normal operation, eliminating the possibility of in-service
corona discharges and the potential life reduction of the polymer insulators and also
improving the audible noise performance. Kinectrics, in Toronto, Canada worked
with AEP in the design and performance of the corona and RIV test program.
Test Program
The first step in the test program development was the calculation of the values of the
maximum midspan operating gradients characterizing the line design. Based on
these, the voltage gradient at which the testing was to be performed was established,
which included factors to account for aging of the hardware and for the lower air
density at the elevations that the lines would be built. Following the establishment of
the target voltage gradient, full scale suspension and dead-end assemblies were
erected in the high voltage laboratory and subjected to visual corona testing.
Calculation of the mid-span conductor surface gradients
The mid-span conductor surface gradients were calculated for the line deign provided
by AEP. The maximum operating voltage of the AEP 765-kV system is 805 kV.
Based on this value, the maximum mid span sub-conductor surface voltage gradient
on the operating line with a ground clearance of 13.7 m (45 ft) was calculated as
being 18.27 kV/cm. In order to account for the reduced air density in the highest area
of the AEP service territory in which this design could be utilized, and for the surface
aging that occurs on line hardware exposed to nature, the value of the subconductor
surface voltage gradient set for acceptance was set at 1.3 times the maximum
operating gradient. Therefore, the assemblies were tested at a subconductor surface
gradient of 23.7 kV/cm.
Calibration of the single phase test setup
The first step in the calibration of the full single phase test setup comprised the
calibration of the corona calibrating sphere which would be utilized in establishing
the relationship between the voltage applied to the full single phase test setup and the
electric field at the outside surface of the individual subconductors [3].
The diameter of the calibrating sphere chosen for use was 3mm. Calibration of the
calibrating sphere on a conductor of the same diameter as that of the bundle
subconductors was performed so that the sphere could be used to set the test voltage
to the value which would produce the required electric field at the bundle
subconductor surface.
The calibration procedure is based on the use of a geometry characterized by a known
and easily calculable electric field distribution. Typically the geometries used are
either a single conductor positioned above a ground plane or a coaxial geometry with
the conductor placed in the central axis of a metal cylinder. Both of these geometries
Page 4 of 11
465
allow for accurate and simple calculation of the electric field at the conductor surface.
The calibrating sphere is mounted on a conductor at its midpoint. The length of the
conductor used is sufficient to ensure that the electric field at the center of the
conductor is not subject to end effects. In this case the calibration of the corona
calibrating sphere was done utilizing a single conductor positioned a given distance
above the laboratory ground plane (i.e. the floor).
The calibration was performed by raising the voltage applied to the conductor and
determining the applied voltage at which the corona calibration sphere went into
positive corona. This was repeated 5 times for two different heights of conductor
above ground. The positive corona inception voltage was established as the average
of the 5 measurements taken at each of the two heights above ground. This value was
then used to establish the electric field occurring on the conductor surface at the
inception of positive corona on the calibrating sphere. For the single conductor above
ground geometry utilized, the relation between the applied voltage and the electric
field at the conductor surface is given by:
E=
V
r ln
where: E
V
r
h
2h
r
This value of electric field was then assigned to the particular corona calibration
sphere as the electric field at which the calibrating sphere goes into positive corona
when mounted on this particular diameter of conductor.
Figures 2 and 3 show the setup used for calibration of the corona calibrating sphere
and the details of how the calibrating sphere is attached to the conductor. The results
of the calibration are shown in Table 1.
466
Page 6 of 11
467
468
Vr =
Es
Vc
Ec
where:
Vr = The required applied test voltage
Es = The voltage gradient at which the test assembly must be free of positive
corona
Ec = The positive corona inception voltage gradient for the sub-conductor
mounted calibrating sphere (in this case 15.49 kV/cm)
Vc = The applied voltage corresponding determined as the positive corona
inception voltage obtained during the bundle calibration.
Table 2 shows the results. Based on these data, it was agreed that 530 kV was to be
used as the pass/fail voltage criterion of the transmission suspension assemblies.
Page 7 of 11
469
337.0
516
347.5
531
353.2
540
340.0
520
333.0
510
If corona extinction did not occur until the test voltage was reduced to below 530-kV,
then the assembly was considered to have failed the test.
Sample photographs of the double V, single V, and the quadruple tension assemblies
showing the setup, and the corona extinction and inception details are shown in
Figures 5 to 7.
In many cases the assemblies passed the test successfully, while in others they failed.
On assemblies which initially failed to pass the test, modifications were made during
the testing. In some cases, the assemblies could be made to pass the test through
simple modifications such as moving, adjusting, or re-orientating the corona rings.
In these cases, the new positioning and orientation of the rings was included as a part
of an immediate design revision. In other instances, more extensive modifications
were required. Here, solutions such as changing the shape or increasing the diameter
or the size of tubing used in the ring construction were identified and proven to work
in mock-up form. Following the identification of the potential solutions, the design
was revised, new rings were manufactured, and the assemblies were later re-tested to
ensure that the revised designs met the test requirements.
b) Corona inception
a) Test Set-up
c) Corona extinction
Page 9 of 11
470
b) Corona inception
a) Test Set-up
c) Corona extinction
b) Corona inception
a) Test Set-up
c) Corona extinction
Page 10 of 11
471
Conclusions
Voltage gradient based visual corona tests [1,2] were carried out in accordance with
accepted standards and guidelines for visual corona and RIV testing for transmission
line insulator assemblies and line hardware. The tests were performed on several
assemblies for use in AEPs new 765-kV transmission construction. The tests
confirmed that several of the assemblies would operate free of positive corona in fair
weather conditions. In some instances, the initial designs did not meet the test
requirements and in these cases, improvements to the design were formulated and
their effectiveness was checked during the testing. Assemblies requiring significant
modification were re-tested after the identified modifications had been made.
During the tests, several instruments were used to detect and observe visual corona on
the insulator assemblies and hardware. These included a standard image intensifier,
a Daycor camera and a Coronascope. It was also determined that a tripod mounted
(non-commercial) camcorder with low-light capabilities was quite effective in
recording the tests. All of the instruments proved capable of detecting the presence of
positive corona, and resulted in the same assessment of corona inception and
extinction levels.
References:
[1] CAN/CSA-C411.4-98, "Composite Suspension Insulators for Transmission
Applications"
[2] IEC Publication 61284, 1997, "Overhead lines Requirements and tests for
fittings"
[3] O. Nigol, Development and Testing of Corona-Free High Voltage Line and
Station Hardware, 3rd International Symposium on High Voltage Engineering,
Milan, Italy, 1979.
Page 11 of 11
472
473
APPLICATION
Broken conductors are one of the extreme event loads used to design high voltage
transmission lines. This load case can be a controlling load for the design of a tower or
components, such as cross-arms. The broken conductor load is a short duration load
compared to the traditional static loads used to design for wind and ice. It has been
demonstrated in other structural engineering applications that members have higher
compression capacities when subjected to short duration loads.
Traditional high voltage transmission line towers are designed using industry
codes that base member capacity on nominal strength values. Tension members and
compression members with low slenderness ratios are designed using the nominal yield
strength, whereas compression members with high slenderness ratios are designed using
Euler buckling, which is independent of yield strength.
New transmission line tower designs currently use the nominal member capacities
to resist all design loads. In special cases where an existing high voltage transmission line
is being assessed for upgrading and reliability, capacities higher than the nominal value
may be used in structural assessment of the transmission line. Investigation of
transmission line failures involving short duration loads may benefit by considering the
increase in compression capacity of individual members.
Tension members are designed to the nominal yield strength. A tension member
can have an increase in capacity on the order of 1020 percent, when the actual yield
strength is considered. This would also be true for low slenderness ratios compression
members. The compression capacity of members with high slenderness ratios are not
controlled by yield strength. In this case the increased as a result of RL compression
capacity could be beneficial.
TESTS
Both static and RL Tests were performed as part of this research. Angles tested
were of two (2) different cross sectional areas and four (4) different slenderness ratios.
The angle sizes were 76 X 76 X 6.4 mm (3 X 3 X 1/4 in.) and 89 X 89 X 9.5 mm (3-1/2
X 3-1/2 X 3/8 in.), and the slenderness ratios, L/rs were 70, 100, 140 and 180. The
angles were cut from 12.2 m (40 ft.) sections and piece marked accordingly in order to
account for material property variations. To determine the yield strength, a coupon was
cut from each 12.2 m (40 ft.) section and tested in accordance with ASTM E8 (2001).
Members were first tested under normal static loading conditions to determine the
static capacity of the member. These tests are referred to as the Pre-Static Tests. The
capacity of the member is defined as the point where the member could no longer support
additional applied load. Three (3) Pre-Static Tests were performed for each angle size and
L/r. The RL Tests began once the Pre-Static Tests were complete. At least five (5) RL
Tests were performed for each angle group. A group consists of angles of the same
slenderness ratio, cross sectional area and end support.
The RL Test was performed in a series of load steps that were each a percentage
of the ultimate load estimated using the Pre-Static Test capacity and calculated values.
The percentage varied slightly from test to test. A typical RL Test started at 50% of
474
ultimate static load and continued in 20% steps until failure. Failure is defined as the load
at which a permanent set in lateral deflection occurred or when the load resistance of the
member did not increase. Each RL step started and ended at a load equal to zero.
Whenever possible, a static load test to failure was conducted for each member after it
failed in the RL Test.
TEST SETUP
Three types of end supports were tested with the two angle sizes and all four
L/rs. They included: 1) Ball Joints at each end, 2) Single Leg Bolted at both ends and 3)
One end with Single Leg Bolted and the other with Partial End Restraint.
End Support Type 1, ball joint at each end, is shown in Figure 1. The angle was
bolted to a channel that was welded to a base plate on each leg at both ends. The angle
had slotted bolt holes so that when loaded, the angle would slip into place and bear
directly onto the end plate. This was to ensure that the load applied was transferred
through the plate and into the angle, as opposed to being transferred through the bolt. The
load was applied through the Center of Gravity (CG) of the angle.
End Support Type 2, single leg bolted at both ends is shown in Figure 2. The
angle was bolted on one leg at both ends of the member. This end support was also
hinged on both ends. This is different from End Support Type 1 not only because of the
hinge, but also because the angle did not bear against the plate. The load was instead
transferred through the base plate and then through the angle via the bolts. The end
support consisted of a staggered five bolt-hole pattern. The hinge allowed rotation about
an axis along the outside surface of the bolted leg. The load was not centered on the
centroid of the angle, but at the outside of the bolted leg at the centroid of the bolt pattern
representing normal frame eccentricity.
475
Figure 3, Single-Leg Bolted on one end and Partial End Restraint on the other end,
Type 3
476
SUMMARY OF RESULTS
Short Members L/r of 70 and 100
The Average RL Factors (Avg. RL Factor) for both size angles with L/r of 70 and
100 are very close to 1.0, indicating that there is no increase in the load resistance of the
members as the rate of loading increases. A RL Factor close to 1.0 for Short Members is
consistent with information provided by others (Ari-Gur, J., Weller, T., and Singer J.,
1982, Galambos, T. V., 1998, and Hoff, N. J., 1951).
Long Members L/r of 140 and 180
The Average RL Factors (Avg. RL Factor) for both size angles with L/r of 140
and 180 are larger than 1.0, indicating that there is an increase in the load resistance of
the members as the rate of loading increases. The values range from 1.25 to 2.8. RL
Factors greater than 1.0 for Long Members is consistent with information provided by
others, per the listed literature search references in Appendix A (Ari-Gur, J., Weller, T.,
and Singer J., 1982, Galambos, T. V., 1998, and Hoff, N. J., 1951).
For members with L/r of 140 and 180 the Avg. RL Factors consistently trend
lower as the member cross section increases. The mass of the member increases as the
member size increases. It is the authors opinion that the trend to smaller RL Factors as
the member size increases is caused by the increase in inertia of the larger cross section
angle. The increase in lateral deflection as a result of lateral inertial effects could cause
yielding at a lower test load.
CONCLUSION
The results of this research show that an increase in compression load capacity
can be expected for long members (L/r of 140 and 180) when subjected to rapid loads.
Based on the test results of this study a Rapid Load Factor of 1.5 may be a reasonable
value to represent this type of behaviour. This increased compression capacity was not
seen in short members (L/r of 70 and 100). This is encouraging because long members
fail by compression buckling while short members fail by yielding. A buckling failure
load of long members is not dependent on the yield strength of the member, thus yield
strength higher than that specified is of no benefit. In short members, the capacity is
dependent on the materials yield strength, which is normally higher than that specified
by the material specification (nominal yield strength). In addition, strain hardening of the
steel occurs before a yielding failure. Short members have the potential for additional
compression capacity due to the actual yield strength exceeding the nominal yield
strength.
These results of this study are consistent with published papers (Ari-Gur, J.,
Weller, T., and Singer J., 1982, Galambos, T. V., 1998, and Hoff, N. J., 1951) which
address rapid loading of structural members for applications other than transmission line
towers. Additional research should be performed on individual members with other
slenderness ratios. Also substructure and full-scale tests should be performed to
determine if the individual members of these assemblies attain the increased compression
capacity reported here, when subjected to a more complex structural load path.
477
REFERENCES
Ari-Gur, J., Weller, T., and Singer J., (1982) Experimental and Theoretical Studies of
Columns under Axial Impact, International Journal of Solids and Structures, Vol.
18, No. 7.
CIGRE B2-308 (2004) Testing and Numerical Simulation of Overhead Transmission
Line Dynamics under Component Failure Conditions, Vincent, P., Huet C.,
Charbonneau M., Guibalt P., Lapointe M., and Banville D., et. al., Paris, France.
Galambos, T. V., (1998) Selected Topics in Dynamic Stability, Guide to Stability
Design Criteria for Metal Structures, Fifth Edition, John Wiley & Sons, Inc.
Hoff, N. J., (1951) The Dynamics of the Buckling of Elastic Columns, Journal of
Applied Mechanics, American Society of Mechanical Engineers.
CEATI INTERNATIONAL, INC. CONTACT INFORMATION
CEAT International, Inc. (CEATI), 1010 Sherbrooke Street West, Suite 2500, Montreal,
Quebec, Canada H3A 2R7, Website: www.ceati.com
478
ABSTRACT
Over time, steel structures are affected by corrosion and their capacity is therefore
affected. To estimate the capacity of structures with corrosion, it is necessary to
estimate the residual capacity of corroded members. The objective of this article is to
evaluate the structural behaviour of corroded steel angle members under compressive
load in order to provide practical data to engineers. An accelerated corrosion
procedure was used on 16 angle members. The angle members were then tested in
compression. The influence of corrosion on compressive capacity was measured and
compared to analytical methods accounting for weight loss. Recommendations are
drawn from this research to provide guidance to engineers on how to evaluate
compressive capacity of corroded members. Needs for future work are also
highlighted.
INTRODUCTION
Steel structures exposed to environmental condition are subject to corrosion. Even
galvanized steel can experience corrosion after some time. Corrosion can be
particularly important in the foundation, for example, where steel is in direct contact
with the soil. For companies managing large structure networks, such as transmission
line utilities, it is extremely difficult to assess the residual capacity of an existing
structure and determine when it is no longer safe. One method based on qualitative
parameters was developed by Hathout (2004) to evaluate the reliability of
transmission lines. To examine the problem in more details, it would be useful to
develop a model predicting the residual capacity of corroded members.
The phenomenon of corrosion is well-studied. However, very little research has been
done on how the capacity of steel members is affected once corrosion has developed.
The objective of this study is to provide data on compressive capacity of corroded
angle steel members that could be used by practicing engineers. A model to predict
residual capacity of members is evaluated.
479
EXPERIMENTAL PROGRAM
In order to evaluate the residual capacity of angle members, an experimental program
was performed. The main parameters of this study are: (i) the overall slenderness ratio
(KL/r), (ii) the width-to-thickness ratio (w/t), and (iii) the corrosion level.
The slenderness ratio (KL/r) is one of the main factors affecting the capacity of steel
members under compression load. This is the ratio of the unsupported length
(effective length KL) to the radius of gyration. The length of the test specimens were
selected such that two levels of this variable are investigated: (i) slenderness ratio of
approximately 25, and (ii) slenderness ratio of approximately 70. Originally, K is
assumed to be equal to 1.
The width-to-thickness ratio (w/t) is defined as the ratio of the leg flat width to the leg
thickness. This ratio affects the value of Fcr of the member based on the definitions of
the ASCE 10-97 standard (1997), and therefore the calculated compression capacity
of the member. The two levels of width-to-thickness ratio studied are 5.72 and
12.4.The corrosion level is measured as a mass loss percentage. Three levels of this
variable were investigated: 0% (not corroded), 25% (moderate corrosion), and 40%
(severe corrosion).
Table 1 shows all specimens and their level for each variable mentioned above, as
well as the section, length, and gross area.
Accelerated corrosion procedure. The steel members were corroded by the galvanic
corrosion process. The members were immersed in a conductive solution saturated
with copper sulphate, and connected to the positive node of an electric circuit. The
galvanized steel piece then plays the role of the anode. A copper plate cathode was
also immersed in the solution and connected to the negative node of the circuit.
Current was applied with a constant intensity generator to oxidize the steel member
(from the anode to the cathode). The steel members were submitted to the process for
durations between 253 and 1819 hours. Measurements of the leg width (b) and leg
thickness (t) were taken during and after the corrosion process at four locations along
the member. Current intensity and voltage were monitored. Standard assemblies for
the corrosion process are shown in Figure 1.
Experimental set-up. The experimental set-up developed by Morissette (2008) to
study the compression capacity of single diagonal bracing members was used here to
assess the strength in compression of the corroded members. A sketch of the test setup is shown in Figure 2.
480
481
Specimen
Section
L (mm)
% of
corrosion
w/t
KL/r
A (mm2)
48S-00-1
L64X64X4.8
500
12.4
39.7
582
48S-00-2
L64X64X4.8
500
12.4
39.7
582
48L-00-1
L64X64X4.8
1358
12.4
108
582
48L-00-2
L64X64X4.8
1358
12.4
108
582
95S-00-1
L64X64X9.5
500
5.72
40.3
1120
95S-00-2
L64X64X9.5
500
5.72
40.3
1120
95L-00-1
L64X64X9.5
1358
5.72
110
1120
95L-00-2
L64X64X9.5
1358
5.72
110
1120
48S-25-1
L64X64X4.8
500
25
12.4
39.7
582
48S-25-2
L64X64X4.8
500
25
12.4
39.7
582
95S-25-1
L64X64X9.5
500
25
5.72
40.3
1120
95S-25-2
L64X64X9.5
500
25
5.72
40.3
1120
48L-25-1
L64X64X4.8
1358
25
12.4
108
582
48L-25-2
L64X64X4.8
1358
25
12.4
108
582
48S-40-1
L64X64X4.8
500
40
12.4
39.7
582
48S-40-2
L64X64X4.8
500
40
12.4
39.7
582
95S-40-1
L64X64X9.5
500
40
5.72
40.3
1120
95S-40-2
L64X64X9.5
500
40
5.72
40.3
1120
95L-25-1
L64X64X9.5
1358
25
5.72
110
1120
95L-25-2
L64X64X9.5
1358
25
5.72
110
1120
95L-40-1
L64X64X9.5
1358
40
5.72
110
1120
95L-40-2
L64X64X9.5
1358
40
5.72
110
1120
48L-40-1
L64X64X4.8
1358
40
12.4
108
582
48L-40-2
L64X64X4.8
1358
40
12.4
108
582
482
The length of the member taken in the calculation is the theoretical value from
centroid of connection to centroid of connection. In accordance with ASCE 10-97,
the radius of gyration of the member is the radius about the principal minor axis.
Experimental evidence has however shown that the effective length may be different
from theoretical length and that buckling may not occur around the minor axis. This
phenomenon is due to the effect of the connections and is currently studied at the
University of Sherbrooke. Nevertheless, because the gusset plates in the tests had
flexibility very close to the flexibility of gusset plates used in practice, it was
considered that the code should take this effect into account, at least in a limited way.
Indeed, ASCE 10-97 allows modifying KL/r for members with normal framing
eccentricities at both ends of the unsupported panel and for members partially
restrained against rotation at both ends of the unsupported panel (ASCE 10-97, 1997).
EXPERIMENTAL RESULTS
Accelerated corrosion procedure. The steel members reached corrosion levels
between 20% and 50% when submitted to the accelerated corrosion process. The
main properties after corrosion are shown in Table 2. The leg width (b) for corroded
members was approximately constant at 63 mm.
To provide more information on the uniformity of the corrosion, the thickness of
some members was measured at each point of the grid shown in Figure 4. A sample
of the results is shown in Figure 5. These measurements show that the uniformity of
the corrosion seems to decrease with the level of corrosion increasing. It is also
5
483
observed that the corrosion is more uniform for the 9.5 mm thick members than for
the 4.8 mm thick members, suggesting that corrosion uniformity is less for smaller
thicknesses.
Compression tests. Figure 6 shows the results of the tests in terms of ratio between
residual strength and uncorroded strength. The value of the uncorroded strength is
taken as the average of the experimental strength of the two members of the same
series that are not corroded. Table 2 also shows the capacity obtained in laboratory
and the observed failure modes. It should be noted that the type of failure was either
local or global and the buckling occurred at various locations along the members. In
two cases, complete perforation of the member was observed. The type of failure is
an important parameter and should be considered in the analysis of the results. It
should be noted that the classification of failure modes is not straightforward and
therefore these observations should be analysed with care.
6
484
485
Specimen
% of
corrosion
48S-00-1
Exp.
Failure
A
capacity mode*
2
(mm )
(kN)
582
n/a
3
(mm)
w/t
(mm)
KL
KL/r
(ASCE)
0.00
4.76
12.4
39.7
79.9
48S-00-2
0.00
4.76
12.4
39.7
79.9
582
114
48L-00-1
0.00
4.76
12.4
108
114
582
91
48L-00-2
0.00
4.76
12.4
108
114
582
88
95S-00-1
0.00
9.53
5.72
40.3
80.2
1120
323
95S-00-2
0.00
9.53
5.72
40.3
80.2
1120
335
95L-00-1
0.00
9.53
5.72
110
115
1120
168
95L-00-2
0.00
9.53
5.72
110
115
1120
167
48S-25-1
26.4
4.32
13.8
40.0
80.0
526
57
48S-25-2
26.6
4.18
14.3
40.0
80.0
509
n/a
95S-25-1
26.9
7.28
7.79
40.5
80.3
864
253
95S-25-2
25.3
7.18
7.91
40.5
80.2
853
251
48L-25-1
27.4
3.86
15.6
108
114
471
80
48L-25-2
27.3
3.84
15.7
108
114
469
79
48S-40-1
37.2
3.51
17.2
39.8
79.9
430
66
48S-40-2
37.3
3.37
18.0
39.8
79.9
413
63
95S-40-1
37.1
6.43
8.95
40.4
80.2
769
216
95S-40-2
36.1
6.25
9.24
40.4
80.2
748
202
95L-25-1
22.4
7.80
7.21
110
115
922
151
95L-25-2
21.5
7.98
7.02
110
115
942
150
95L-40-1
30.7
6.79
8.43
110
115
809
99
95L-40-2
27.2
7.08
8.04
110
115
842
99
48L-40-1
40.7
3.28
18.5
108
114
403
65
48L-40-2
47.1
2.63
23.3
108
114
324
25
* Legend for failure modes: 1- global buckling, 2- local buckling near connections, 3- local buckling
near center, 4- mode not clearly identified in the test.
Perforated members.
Specimen 48S-00-1 and 48S-25-2 did not produce any compressive capacity results
due to difficulties in the experimental procedure.
It was observed that the effective length remained constant throughout the corrosion
process. The radius of gyration varied moderately due to the moderate change of the
leg width which affects the moment of inertia of the member. The slenderness ratio is
7
directly dependent on the variation of the radius of gyration and therefore was subject
to minor modifications only. However, the section area of the members varied
significantly, causing major changes to the compressive capacity of the angle steel
members.
ANALYSIS OF RESULTS
Comparison of experimental and theoretical results. The experimental results for
the capacity in compression of the angle members are compared in Table 3 to the
theoretical results from the ASCE standard. The calculation of the theoretical
strength was based on an average thickness measured in laboratory, assuming that the
thickness of corroded members is constant. It is however obvious from the failure
modes observed in Table 2 that the corrosion was not uniformly distributed. The
compression capacity is largely dependent on the extent and the location of the
corrosion. This non-uniformity of the phenomenon explains in large part the
difference between the experimental and theoretical results. In fact, complete
perforation was found on two members (see Figure 7). In particular, the capacity of
member 48L-40-2 was greatly reduced due to perforation because most of one leg
was lost near the center. Another example is member 48S-25-1 which is particularly
weaker than its predicted capacity. The failure mode of this member was very
localized on one leg near the center. The failure in this case was due to the large w/t
8
486
and is not influenced by its KL/r value. The thickness near its failure point is
approximately 3.3 mm, nearly 1 mm smaller than its average measured thickness.
The experimental capacity of 57 kN would be obtained theoretically for a thickness of
approximately 3.1 mm with the ASCE 10-97. As seen in these two examples, the
compression capacity is largely dependent on the extent and the location of the
corrosion. Concentration of corrosion reduces in an acute way the capacity of the
member, specifically when such concentration occurs near the middle of the member.
Corrosion is critical at the middle of the member because this is where the moment
due to the eccentricities of the connections is a maximum. This non-uniformity of the
phenomenon explains in large part the differences between the experimental and
theoretical results.
Figure 8 shows the ratios of the experimental to the theoretical compression
capacities versus the level of corrosion. This figure clearly shows the weakened
capacity of the 48S-25-1 member.
To further analyse the results obtained and to focus on capacity reduction due to
corrosion rather than the theoretical prediction of compression strength, the
experimental and theoretical results are presented in Figure 9. The vertical axis is the
ratio between the experimental and predicted percentage of residual capacity as
compared to the uncorroded strength. This value compares the predicted loss of
capacity and the loss of capacity measured during the test. From this figure, it is seen
that there is a wide scatter in the results. Up to about 25% of corrosion, predictions
are in line with experimental results. For corrosion level above 25%, experimental
results and predictions do not match. However, there is not enough data between 0
and 25% to conclude that the analytical approach is satisfactory in this range of
corrosion level.
It is interesting to note that predictions of the capacity of members with 9.5mm legs
are better than the predictions on 4.8mm leg members. The average of the predictions
is actually close to experimental results for 9.5mm leg members and variability is
acceptable when we consider the rough method of estimating the capacity as
compared to the complex development of corrosion. The predictions for members
with 4.8mm leg thickness present more variability. The scatter in the ratio
experimental/prediction observed on 4.8-mm leg thickness members might be
attributed to the irregularity of the corrosion process when corrosion level increases.
487
Specimen
% of
Exp. capacity
corrosion
(kN)
Prediction
Exp. /
ASCE
48S-00-1
0.00
n/a
ASCE (kN)
142
48S-00-2
0.00
114
142
0.80
48L-00-1
0.00
91
88.6
1.03
48L-00-2
0.00
88
88.6
0.99
95S-00-1
0.00
323
278
1.16
95S-00-2
0.00
335
278
1.21
95L-00-1
0.00
168
168
1.00
95L-00-2
0.00
167
168
0.99
48S-25-1
26.4
57
121
0.47
48S-25-2
26.6
n/a
114
n/a
95S-25-1
26.9
253
214
1.18
95S-25-2
25.3
251
212
1.19
48L-25-1
27.4
80
70.4
1.13
48L-25-2
27.3
79
69.9
1.13
48S-40-1
37.2
66
80.6
0.82
48S-40-2
37.3
63
73.0
0.86
95S-40-1
37.1
216
191
1.13
95S-40-2
36.1
202
186
1.09
95L-25-1
22.4
151
137
1.10
95L-25-2
21.5
150
140
1.07
95L-40-1
30.7
99
121
0.82
95L-40-2
27.2
99
126
0.79
48L-40-1
40.7
65
54.3
1.20
48L-40-2
47.1
25
32.1
0.78
10
n/a
488
Figure 8 Ratio of the experimental to the theoretical compression capacity (ASCE 10-97) versus
the level of corrosion.
11
489
490
particular for members with level of corrosion between 0 and 20%. Additional data
for other leg thickness and KL/r would also be interesting. For utilities, residual
strength curves as a function of time are also necessary to plan asset repairs and
replacements. For this purpose, additional research is required to transform corrosion
level into time, depending on environment aggressiveness. Finally, it is also
necessary to verify that the actual corrosion pattern found on steel structures is similar
to the corrosion pattern obtained in this experiment through an accelerated corrosion
method.
REFERENCES
ASCE 10-97. (1997). Design of Latticed Steel Transmission Structures. American
Society of Civil Engineers, USA.
ASTM E 8-61T. (1961). Tentative Methods of Tension Testing of Metallic Materials.
American Society for Testing and Materials, USA.
Hathout, I. (2004). Damage assessment and soft reliability evaluation of existing
transmission lines. In Proceedings of the International Conference on
Probabilistic Methods Applied to Power Systems, Piscataway, NJ, 12-16 Sept.
2004, IEEE, pp. 980-986.
Morissette, E. (2008). valuation des normes de calcul et du comportement des
cornires simples en compression utilises comme contreventements dans les
pylnes treillis en acier. Thse de matrise, Dpartement de gnie civil,
Universit de Sherbrooke, Sherbrooke, QC.
13
491
492
Introduction
Below Grade Coatings is a means of protecting steel structures at the ground line &
below grade. Figure 1 is an example of galvanized steel structures with a below
grade coating applied.
Protective coatings provide a barrier coating that protects the steel from corrosion and
should be applied whenever there is a possibility of corrosive soil conditions. The
dielectric properties of urethane coatings insulate poles from stray currents that may
be present in the immediate area. Figure 2 illustrates a typical spray application
process for a Below Grade Coating
493
Background
Over the past several years Thomas & Betts has performed numerous and extensive
tests through the efforts of our Research & Development Department on below grade
coating products in order to identify the best method for surface preparation and
subsequent coating applications.
We considered a range of coatings formulas, as well as surface preparation methods.
Some of the tests were successful and some did not perform as originally
hypothesized. The testing program played a significant role by helping us increase
our knowledge base and improve our processes and product performance.
Since we are only aware of a minimal number of field performance problems when
coating weathering steel, we used A871 panels to benchmark our results. History
indicates that most performance challenges were associated with coating over
galvanized structures made of A572 steel.
Testing
We identified and tested multiple combinations of blasting media, surface profiles,
coatings formulations and adhesion promoters.
Test panels of each combination were prepared. Substrate temperature, dew point,
temperature, humidity, time and any special comments were recorded immediately
prior to coating. Galvanize thickness and coating thickness measurements were taken
in three places for each panel and surface profiles were recorded.
After the panel preparation was complete, we selected samples that best represented
each individual test group. A fully accredited ISO registered third party test facility
performed all testing.
Testing consisted of:
X-Cut Adhesion per ASTM D6677
Pull-Off Adhesion per ASTM D4541
Cathodic Disbondment per ASTM G95 modified to 60 days
QUV Accelerated Weathering testing per ASTM G154-06 at 1000 hour intervals.
All adhesion tests were performed 5 days after application. Additional adhesion tests
were conducted after 1000 hours of QUV for a degradation check/comparison. No
significant changes in performance were observed when compared to the initial tests.
After reviewing the results of the various tests that were performed, we concluded
that the most meaningful information was obtained by conducting the Pull-Off
Adhesion test per ASTM D4541.
494
495
Testing (continued)
There was a direct correlation between the adhesion strength values obtained and the
surface preparation method and media used. The higher performing test plates had
both a higher profile and a higher surface roughness, which was achieved by
following the rigorous process parameters that were set and controlled along with the
blast media that was specified.
For the purpose of this report we are only going to focus on the correlation between
an improved Pull-Off Adhesion test value and the benefits of a superior surface
preparation practice. Figure 3 shows one of the Pull-Off Adhesion test panels after
testing. The face of the test stub is shown on the left and the coated steel surface
being tested for adhesion strength is shown on the right. Note that the percent of
coating and/or glue remaining on the steel plate provides indication of the failure
mode as shown in Table 1.
5000
4500
4000
Improved
Performance
3500
3000
2500
Typical range of
performance
specified by
coatings
manufacturers
2000
1500
1000
500
0
Baseline, A871
Surface 1, A572
Surface 2, A572
Sample Preparation
Figure 4 summarizes the Pull Of adhesion test results. A871 material was utilized as
a baseline for the test. Surface 1 represents a previously acceptable profile on
galvanized steel which is typical for the industry. Surface 2 represents improved
surface roughness characteristics on galvanized steel which was obtained by altering
the blast process and media. Surface 2 provided adhesion test results that were much
higher than the typical range of performance specified by coatings manufacturers for
galvanized surfaces.
4
496
Testing (continued)
ASTM D4541, Pull-off Adhesion Test
Fail ure Mode
Te st Grou p Su rfa ce
Ba se line, A871
1
Su rfa ce 1
Su rfa ce 2
#
5
7
3
6
6
7
Po int 1
Thi ckne ss
( Mils)
14.4
22.5
19.4
27.6
21.1
20.9
Po int 2
Po int 3
Thi ckne ss Thickne ss
( Mils)
(Mils)
10.6
10.6
12.3
15.3
17.8
18.1
27.1
27.3
16.0
17.7
16.7
15.7
Averag e
Mils
12
17
18
27
18
18
% Gl ue
95
100
0
0
80
85
% Co ati ng
Coh esion
3
0
5
5
5
5
%
Substrate
Ad hesion
2
0
90
95
15
10
%
Sub stra te
Cohe sion
0
0
5
0
0
0
Pull Off
Stren gth
(Actual PSI)
4 149
3 823
2 112
1 216
4 068
3 253
%
Sub stra te
Cohe sion
0
85
0
60
0
0
Pull Off
Stren gth
(Actual PSI)
3 579
1 623
2 031
2 194
4 557
4 638
%
Sub stra te
Cohe sion
0
0
10
0
0
0
Pull Off
Stren gth
(Actual PSI)
2 601
3 742
1 623
2 764
3 171
4 801
%
Sub stra te
Cohe sion
0
0
0
0
0
0
Pull Off
Stren gth
(Actual PSI)
4 231
3 660
2 031
1 786
2 927
4 149
%
Sub stra te
Cohe sion
0
0
0
0
0
0
Pull Off
Stren gth
(Actual PSI)
4 394
3 253
2 275
2 764
3 742
3 253
Me an PSI
398 6
166 4
366 1
Te st Grou p Su rfa ce
Ba se line, A871
2
Su rfa ce 1
Su rfa ce 2
#
7
8
5
7
2
3
Po int 1
Thi ckne ss
( Mils)
17.7
39.7
24.3
22.2
28.3
23.4
Po int 2
Po int 3
Thi ckne ss Thickne ss
( Mils)
(Mils)
16.9
18.4
35.2
35.2
25.5
25.6
19.6
21.2
26.3
26.3
25.6
25.0
#
5
8
3
7
6
7
Po int 1
Thi ckne ss
( Mils)
17.4
16.2
23.4
22.9
26.7
26.5
Po int 3
Po int 2
Thi ckne ss Thickne ss
(Mils)
( Mils)
17.1
18.2
16.7
17.9
25.9
24.5
21.6
21.9
26.6
25.9
28.0
25.8
Averag e
Mils
18
37
25
21
27
25
% Gl ue
35
3
2
0
100
98
% Co ati ng
Coh esion
5
12
8
5
0
2
%
Substrate
Ad hesion
60
0
90
35
0
0
Me an PSI
357 9
211 3
459 8
Te st Grou p Sa mp le
Ba se line, A871
3
Su rfa ce 1
Su rfa ce 2
Averag e
Mils
18
17
25
22
26
27
% Gl ue
100
100
20
90
100
93
% Co ati ng
Coh esion
0
0
5
5
0
5
%
Substrate
Ad hesion
0
0
65
5
0
2
Me an PSI
317 2
219 4
398 6
Te st Grou p Su rfa ce
Ba se line, A871
4
Su rfa ce 1
Su rfa ce 2
#
6
8
1
2
3
5
Po int 1
Thi ckne ss
( Mils)
17.1
18.2
26.7
21.8
18.9
23.8
Po int 2
Po int 3
Thi ckne ss Thickne ss
( Mils)
(Mils)
20.3
17.9
25.4
23.8
20.3
17.4
26.5
26.5
20.7
21.3
22.6
22.4
Averag e
Mils
18
22
21
25
20
23
% Gl ue
99
93
2
0
45
100
% Co ati ng
Coh esion
1
7
3
5
5
0
%
Substrate
Ad hesion
0
0
95
95
50
0
Me an PSI
394 6
190 9
353 8
Te st Grou p Su rfa ce
Ba se line, A871
5
Su rfa ce 1
Su rfa ce 2
#
2
3
4
7
4
6
Po int 1
Thi ckne ss
( Mils)
14.4
10.3
17.3
24.3
14.9
18.7
Po int 2
Po int 3
Thi ckne ss Thickne ss
( Mils)
(Mils)
15.0
14.0
12.1
9 .9
17.1
17.8
22.4
23.5
15.9
17.8
19.9
16.1
Averag e
Mils
14
11
17
23
16
18
% Gl ue
10
2
0
0
0
0
% Co ati ng
Coh esion
5
3
5
5
5
5
%
Substrate
Ad hesion
85
95
95
95
95
95
Me an PSI
382 4
252 0
349 8
Surface Preparation
A common practice for determining a surface profile which is accepted throughout
many industries entails measuring the height of the highest profile peak within a
sample area.
While the use of this method serves its intended purpose for surface measurements
involving many applications, we found that there are other variables, which when
defined and controlled will provided a more robust surface enabling better adhesion
for Below Grade Coatings.
Roughness is a texture measurement of a surface. It is quantified by the frequency
and short wavelength of surface deviations from its ideal form. If these deviations are
large, the surface is rough; if they are small the surface is smooth.
We found that identification and controlling of the optimum vertical deviation or
amplitude along with the slope and spacing of peaks and valleys (peak density) is
critical to obtaining a good anchor profile.
Having an optimum profile will increase the total surface area and provide superior
adhesion performance. The optimum profile must be determined for the specific
application being considered.
It is imperative that very strict process parameters are in place as well as the blast
media used.
Manufacturing Practices
Controlling the blasting media and process along with selection of a robust coating
product provide the best adhesion performance available.
These Items are Critical to Performance:
Leading-Edge Coating Formulations
Qualified Blast Media
Qualified and Certified Application Practices
Trained, Experienced and Certified Applicators
Monitoring of Process Control Parameters
Well Maintained Equipment PM Schedule
Performance Verification by Third Party Testing
497
Conclusion
Superior adhesion is the key for overall performance of a Below Grade Coating.
Aromatic coatings typically provide superior adhesion when applied using best
manufacturing practices and best surface preparation methods. Coatings of this type
have less gloss retention over an extended period of time, but out perform aliphatic
coatings when used for below grade applications. Performance vs. aesthetics must be
considered when choosing a Below Grade Coating.
It is critical that blast practices and media used be optimized in order to achieve the
best surface roughness possible in order to provide superior adhesion.
Robust Surface Roughness = Superior Adhesion = Superior Performance
Thomas & Betts has performed over five years of Research and Development which
involved four manufacturing plants, consultants, independent test laboratories and a
significant financial investment in order to provide our customers with the best below
ground coating protection available.
We have invested in new technology and equipment, developed process control
guidelines for surface preparation at our own plants and at galvanizing
subcontractors, and partnered with our coatings supplier to define the ultimate
coatings solution.
This study enabled us to surpass what was originally thought to be good adhesion by
a significant amount. Feedback from the field allowed us to confirm a marked
improvement in performance since we implemented these changes and refined our
processes.
Thomas & Betts is confident that what we are currently doing is the best combination
of surface preparation and coatings in our industry and we are committed to
continuing research to maintain that position.
498
Sargent and Lundy LLC, Power Delivery Services, 55 E. Monroe, Chicago, IL 60603; PH (312) 269-2000; email:
[email protected]
2
Sargent and Lundy LLC, Power Delivery Services, 55 E. Monroe, Chicago, IL 60603; PH (312) 269-2000; email:
[email protected]
Page 1 of 12
499
Electric Power Research Institute, Transmission Line Reference Book: Wind-Induced Conductor Motion, Palo
Alto, California, 1979.
4
Electric Power Research Institute, Transmission Line Reference Book: Wind-Induced Conductor Motion, Palo
Alto, California, 1979.
Page 2 of 12
500
8f
A4 = 80 * d * ln
50d
A1 = 0.7 A1
A
A2 = 4
4
A5 = 0.4 * A4
Where,
d = Wire diameter (meters)
f = Sag of wire at 32F; in. ice; no wind (meters)
5
Electric Power Research Institute, Transmission Line Reference Book: Wind-Induced Conductor Motion, Palo
Alto, California, 1979.
6
International Council on Large Electric Systems (CIGRE), State of the Art conductor Galloping, Task Force
B2.11.06, 2006
Page 3 of 12
501
To select wires for a sample set, several items and design parameters were considered. Since
galloping is a concern for longer span lengths, which are typical for higher voltages, conductors
for 115kV to 345kV were considered. A shield wire was also selected since depending on the
structure configuration, shield wire-conductor galloping interference can be a design
consideration. Only round wire was selected for this paper since galloping design conditions are
most prevalent when round wire is used. For the purpose of this paper, round wire is defined
as non wind-induced, motion-resistant (T2) wire.
With these parameters and considerations, the following three conductors and one shield wire
were selected as typical industry examples for regions that exhibit galloping.
795kcmil Drake ACSR
954kcmil Cardinal ACSR
2156kcmil Bluebird ACSR
7/16in. x 7 strand steel EHS
To compute the sag of the wires, the governing tension condition was selected as 20% rated
tensile strength (RTS) at 0 deg F final condition and 16% RTS at 0 deg F final condition for
conductors and shield wire respectively.
To display the conductor galloping ellipse information, ruling spans from 500 to 1200ft were
chosen. This range was selected so that both the ends of the galloping design spectrum were
accounted for. For simplicity of this paper, the ruling span was used to determine galloping
ellipse magnitude.
CIGRE Galloping Amplitudes
To gain a high level overview of the CIGRE method impacts, Figure 2 was created to display the
galloping ellipse amplitudes. Figure 1 displays the sample wire set for various ruling spans.
For the ruling span length of 900ft, the conductor amplitudes range from 27.2ft to 38.9ft. These
are substantial amplitudes when compared to the Toye method which, in S&Ls experience, is
the most common method for such span lengths. This difference is due to the assumption in the
Toye method that the wire will undergo double loop galloping and thus reduces the ellipse
amplitude significantly.
Page 4 of 12
502
7/16" EHS
45
40
35
30
25
20
15
10
500
600
700
800
900
1000
1100
1200
Davidson
Toye
50
45
Transition Zone
40
35
30
25
20
15
10
0
500
600
700
800
900
1000
1100
1200
Page 5 of 12
503
CIGRE
Davidson
Toye
50
45
Transition Zone
40
35
30
25
20
15
10
0
500
600
700
800
900
1000
1100
1200
Davidson
Toye
50
45
Transition Zone
40
35
30
25
20
15
10
0
500
600
700
800
900
1000
1100
1200
Page 6 of 12
504
CIGRE
Davidson
Toye
40
35
Transition Zone
30
25
20
15
10
0
500
600
700
800
900
1000
1100
1200
With the examination of figures 3 to 6, three key observations are noted. For all the sample
conductors chosen, the CIGRE method has larger amplitudes when compared to the Davidson
method for shorter spans (500ft to 700ft). The maximum difference observed is between 8.8 to
15.3ft in vertical amplitude which corresponds to percent increases of 78.8% to 129%.
The CIGRE method amplitudes are always substantially larger than the Toyes method by 300400% for the sample conductors. This is due to the methods fundamental assumptions regarding
double loop galloping. A change of this magnitude for longer spans could greatly affect the
configuration of the structure.
The 7/16x7 EHS displays the CIGRE method as the moderate method by having a smaller
amplitude than the Davidson method but a larger amplitude than the Toye method. This is a stark
difference compared to the sample conductors.
STRUCTURE CONFIGURATION IMPACT
To further highlight the differences between the different methods, three typical structure
configurations were selected to display the galloping ellipses. These examples were chosen in an
effort to best highlight structure configurations used in the utility industry.
138kV H-Frame
One of the most common structures used in the utility industry is the H-frame. For this example a
138kV H-frame with 15.5ft horizontal phase spacing is shown in figure 7. From the sample set
of conductors, the 795kcmil Drake ACSR was selected with a 600ft ruling span. Both shield
wires were selected as 7/16x7 EHS steel strands.
Page 7 of 12
505
Due to the moderate span length of 600ft, coupled with design practices observed by S&L, only
the CIGRE and Davidson galloping methods were applied. The CIGRE method in figure 7 is
displayed using a solid line while the Davidson method is displayed using a hidden (dashed) line.
Even though the CIGRE method has substantially larger ellipses, it is evident in figure 7 that
none of the galloping ellipses overlap. This is to be expected since over the many years and
numerous applications, H-frame structures have had minimal flashover occurrences due to
galloping events.
Page 8 of 12
506
requirement is used that ellipses shall maintain an additional 1.5ft7 as the EPRI Orange Book
recommends to avoid phase-phase flashover, then the arm would need to be a minimum of
19.75ft which is an increase of 36%.)
Not only does the increase in arm length change the phase spacing but it would also have an
adverse effect on structure cost due to the increased arm length and weight, increase pole shaft
size for extra capacity, and larger foundation to account for the larger overturning moment.
While some of these changes may seem relatively minor, for large projects this can add
substantial cost due to total quantities.
Electric Power Research Institute, Transmission Line Reference Book: Wind-Induced Conductor Motion, Palo
Alto, California, 1979.
Page 9 of 12
507
As was the case in the 138kV double circuit example above (figure 8), the 345kV double circuit
configuration has extensive galloping ellipse overlap for the CIGRE method and no overlap for
the Toye double loop method. If the current vertical spacing was held constant, the middle phase
arm would require an increase in length by 8ft (40%) to ensure no overlapping of the CIGRE
galloping ellipses. The shield wire arm would also require a 2ft extension (10% increase).
The extension of the middle phase arm 8ft would increase the required pole shaft and foundation
capacity but could also cause right-of-way (ROW) issues.
Page 10 of 12
508
With the implementation of the CIGRE method, a new set of questions arise. Since the amplitude
has a degree of conservatism built in the empirical equations additional ellipse spacing for
flashover may be overly conservative.
8
RUS Bulletin 1724E-200, DESIGN MANUAL FOR HIGH VOLTAGE TRANSMISSION LINES, Section 6.5.1,
May 2005
9
International Council on Large Electric Systems (CIGRE), State of the Art conductor Galloping, Task Force
B2.11.06, 2006, Figure 8.44
Page 11 of 12
509
510
One potential issue with the CIGRE method is maximum amplitude of the ellipses. The CIGRE
method does not have a maximum ellipse amplitude so for longer spans the amplitude keeps
increasing. This can be viewed as having an excessive design conservatism which may not be
economically justified. Similar to the Davidson & Toye method utilities could set a maximum
amplitude.
CONCLUSION
While the design criteria considerations mentioned above influence the galloping analysis, the
real impact of the galloping method chosen is most evident in the structure configuration. One
of the main differences when first examining the CIGRE calculations is the amplitude of the
galloping ellipses for longer spans. The CIGRE methodology when compared to A.E. Davidson
methodology has much larger amplitudes that can significantly affect structure configuration,
size, and cost. The effects on the structure configuration were shown above for a select set of
wires and structure types. The increase in structure size and therefore weight also affect the
foundation design and potential ROW width. If the ROW width is fixed which is typical for a
rebuild project, the options are limited to increasing phase spacing and thus structure height,
decreasing span length with additional structures, or changing conductor.
Design for galloping occurrences can be a challenging problem with many influencing factors.
The solutions to address the galloping concerns may prove to be economically inefficient
depending on the project. With these economic factors, a balancing act occurs between project
cost and project design when addressing galloping concerns.
REFERENCES
Wind-Induced
International Council on Large Electric Systems (CIGRE), State of the Art conductor Galloping,
Task Force B2.11.06, 2006
United States Department of Agriculture Rural Utilities Service, Design Manual for High
Voltage Transmission Lines, Bulletin 1724E-200, 2004
Page 12 of 12
A|B|C|D|E|F|G|H|I|J|K|L|M|N|O|P|Q|R|S|T|U|V|W|X|Y|Z
Author Index
A
Amundsen, T. A.
Analytical Techniques to Reduce Magnetic Force from High Fault Current on Rigid
Bus
Andrews, Trevor
Estimation of Containment Loads on a 230kV Steel Transmission Line Using Finite
Element Model
Ashby, Mathieu
Dynamic Wind Analyses of Transmission Line Structures
Asselin, Jean-Marie
Sequential Mechanical Testing of Conductor Designs
B
Baker, Brent
Are Wood Poles Getting Weaker?
Bazn-Zurita, E.
Seismic Design of Substation Structures
Beaulieu, L. -V.
Residual Capacity in Compression of Corroded Steel Angle Members
Beehler, Michael E.
Lessons Learned on Mega Projects
Bgel, C.
3D Automated Design of Substations
Behncke, Roberto H.
Review of Span and Gust Factors for Transmission Line Design
Benavides, C.
C
Calvert, Rodger
Temporary Support of Lattice Steel Transmission Towers
Analysis Challenges in the Evaluation of Existing Aluminum Towers
Carrington, Ron
The Effects of Ice Shedding on a 500 kV Line
Catchpole, P. G.
Large Catenary Structures for High Voltage Transmission Lines
Chau, Max
Aesthetic MitigationThe Challenge Confronting Future Expansion of Transmission
Lines
Cooper, Andy
H2S Entrapment and Corrosion on Direct Embedded Galvanized Steel
Transmission Poles
Crissey, Dana R.
345 kV High Ampacity OH/UG Project
DiGioia, A. M. Jr.
Seismic Design of Substation Structures
Deepwater Transmission Line Foundations Meet Trophy Bass Lake Environment
Drolet, Y.
3D Automated Design of Substations
F
Fan, Quan He
Lessons Learned from Design and Test of Latticed Steel Transmission Towers
Ferguson, N.
Vegetation Management Through LiDAR Derived CADD Models: Compliance with
NERC Reliability Standard FAC-003-1
Foster, Michael
Integration of Optimum, High Voltage Transmission Line Foundations
Freimark, Bruce
Laboratory and Field Testing of Steel Davit Arm Fatigue Failures on Concrete Poles
Sequential Mechanical Testing of Conductor Designs
Electrical and Mechanical Considerations in 765kV (Insulator) Hardware Assembly
Design
G
Gajan, Sivapalan
Transmission Pole Foundations: Alternate Design Methods for Direct-Embedded
Round, Wood Transmission Poles
Gani, Ferawati
Dynamic Wind Analyses of Transmission Line Structures
Goodwin, E. J.
Deepwater Transmission Line Foundations Meet Trophy Bass Lake Environment
Graham, James W.
The 2008 Iowa Floods: Structural Challenges and Solutions
H
Haldar, Asim
Estimation of Containment Loads on a 230kV Steel Transmission Line Using Finite
Element Model
Hastings, Aaron
Integration of Optimum, High Voltage Transmission Line Foundations
Ho, T. C. Eric
Review of Span and Gust Factors for Transmission Line Design
Hodges, Norman P.
230kV Lattice Tower Replacement: An Example of Design Loading Not Addressed
in the NESC
Hogan, James M.
Design and Construction Challenges of Overhead Transmission Line Foundations
(NUs Middletown Norwalk Project)
J
Jacobs, Edward M.
Steel Structure Surface Preparation for Below Grade Coatings
Jarenprasert, S.
Seismic Design of Substation Structures
Johnson, Kristofer
Geotechnical Investigations for a Transmission Line are More Than Drilled Borings
K
Kempner, Leon Jr.
Wind Load Methodologies for Transmission Line Towers and Conductors
Strength of Steel Angles Subjected to Short Duration Axial Loads
Kennedy, Scott
Aesthetic MitigationThe Challenge Confronting Future Expansion of Transmission
Lines
Khattak, Anwar
Integration of Optimum, High Voltage Transmission Line Foundations
Kile, John D.
Temporary Support of Lattice Steel Transmission Towers
Kilgore, Jason W.
Analysis Challenges in the Evaluation of Existing Aluminum Towers
Kuffel, John
Electrical and Mechanical Considerations in 765kV (Insulator) Hardware Assembly
Design
L
Lacroix, C.
3D Automated Design of Substations
Lakhapati, Deepak
Construction Challenges of Extra High Voltage Transmission Lines: Building in the
Most Difficult Terrain in the World
Power Restoration Solution after Major Cyclone: Gonu Hit Oman in June 2007
Langlois, S.
Residual Capacity in Compression of Corroded Steel Angle Members
Lgeron, F.
Residual Capacity in Compression of Corroded Steel Angle Members
Legeron, Frederic
Dynamic Wind Analyses of Transmission Line Structures
Lemelin, D. R.
3D Automated Design of Substations
Li, Ziqin
Electrical and Mechanical Considerations in 765kV (Insulator) Hardware Assembly
Design
Lynch, Otto J.
Common Sag and Tension Errors: Its Time for Template Technology to be Put in
the Drawer Forever
M
Mallik, Ajay
H2S Entrapment and Corrosion on Direct Embedded Galvanized Steel
Transmission Poles
Malmedal, Keith
Substation Bus Design: Current Methods Compared with Field Results
Malten, K. C.
Analytical Techniques to Reduce Magnetic Force from High Fault Current on Rigid
Bus
McCall, Christopher L.
Design and Construction Challenges of Overhead Transmission Line Foundations
(NUs Middletown Norwalk Project)
McNames, Cassie
Transmission Pole Foundations: Alternate Design Methods for Direct-Embedded
Round, Wood Transmission Poles
Monroe, Scott
Temporary Support of Lattice Steel Transmission Towers
Mueller, Wendelin H. III
Strength of Steel Angles Subjected to Short Duration Axial Loads
N
Niles, Kevin D.
Assessment and Repair of Steel Tower and Steel Pole Foundations
Norman, R.
Deepwater Transmission Line Foundations Meet Trophy Bass Lake Environment
O
Oliphant, Wesley J.
The Chaotic Confusion Surrounding Wood Equivalent Non-Wood Poles
Oster, J. L.
Analytical Techniques to Reduce Magnetic Force from High Fault Current on Rigid
Bus
P
Peabody, Alan B.
The Effects of Ice Shedding on a 500 kV Line
Perez, Jared J.
Strength of Steel Angles Subjected to Short Duration Axial Loads
Peter, Zsolt
Sequential Mechanical Testing of Conductor Designs
Peyrot, Alain H.
Wind Loading: Uncertainties and Honesty Suggest Simplification
Pon, Craig
Sequential Mechanical Testing of Conductor Designs
Prudhomme, S.
Residual Capacity in Compression of Corroded Steel Angle Members
Pugh, Archie
Aesthetic MitigationThe Challenge Confronting Future Expansion of Transmission
Lines
R
Rao, Teja
Electrical and Mechanical Considerations in 765kV (Insulator) Hardware Assembly
Design
Ray, R.
Vegetation Management Through LiDAR Derived CADD Models: Compliance with
NERC Reliability Standard FAC-003-1
Retz, David
Design and Construction Challenges of Overhead Transmission Line Foundations
(NUs Middletown Norwalk Project)
Rice, J.
Impact of Alternative Galloping Criteria on Transmission Line Design
Richardson, P.
Vegetation Management Through LiDAR Derived CADD Models: Compliance with
NERC Reliability Standard FAC-003-1
Rowland, Ronald L.
Golden Pass LNG 230kV Double Circuit: Foundations
Ruggeri, E. A.
Large Catenary Structures for High Voltage Transmission Lines
Ryder, S.
Vegetation Management Through LiDAR Derived CADD Models: Compliance with
NERC Reliability Standard FAC-003-1
S
Salisbury, Nick
Integration of Optimum, High Voltage Transmission Line Foundations
Samms, Howard A.
Alabama Power Increases Line Capacity Using 3M ACCR Conductor (On Existing
Towers)
Shanmugasundaram, Rengaswamy
Transmission Line at St. Andrew Bay
Analysis Challenges in the Evaluation of Existing Aluminum Towers
T
Thompson, Freeman
Integration of Optimum, High Voltage Transmission Line Foundations
V
Vaughn, Bruce
Sequential Mechanical Testing of Conductor Designs
Veitch, Maria
Estimation of Containment Loads on a 230kV Steel Transmission Line Using Finite
Element Model
Verth, S.
Vegetation Management Through LiDAR Derived CADD Models: Compliance with
NERC Reliability Standard FAC-003-1
Vogt, Marlon
The 2008 Iowa Floods: Structural Challenges and Solutions
W
Weelden, Tim Van
The 2008 Iowa Floods: Structural Challenges and Solutions
Wendelburg, John
Substation Bus Design: Current Methods Compared with Field Results
Whapham, Robert
Laboratory and Field Testing of Steel Davit Arm Fatigue Failures on Concrete Poles
Williamson, Ernest C. Jr.
Golden Pass LNG 230kV Double Circuit: Foundations
Wong, C. Jerry
Fact or FictionWeathering Steel Can Provide Effective Corrosion Protection to
Steel Structures
Wyman, Gregory E.
Transmission Line Construction in Sub-Arctic Alaska Case Study: Golden Valley
Electric Associations 230kV Northern Intertie
Z
Zhu, Meihuan (Nancy)
Laboratory and Field Testing of Steel Davit Arm Fatigue Failures on Concrete Poles