Kansas Department of Transportation Design Manual

3.4 Analysis
Table of Contents
3.4.1 General .................................................................................................................. 1
3.4.2 Methods of Analysis.............................................................................................. 1
3.4.3 Discretization ........................................................................................................ 2
3.4.4 Distribution of Load .............................................................................................. 3
3.4.4.1 Beam Bridges ............................................................................................................ 3
3.4.4.2 Slab Bridges .............................................................................................................. 3
3.4.5 Longitudinal / Transverse Model .......................................................................... 3
3.4.5.1 Longitudinal Model................................................................................................... 4
3.4.5.2 Transverse Model...................................................................................................... 4
3.4.5.2.1 Slab Bridge .......................................................................................................... 4
3.4.5.2.2 Beam Bridge ........................................................................................................ 4
3.4.6 Superstructure........................................................................................................ 5
3.4.7 Substructure........................................................................................................... 5
3.4.7.1 Pier Beam .................................................................................................................. 5
3.4.7.2 Column ...................................................................................................................... 5
3.4.8 Fixity ..................................................................................................................... 5
3.4.9 Construction Load Analysis .................................................................................. 5

Volume III US (LRFD) Bridge Section

Version 4/08 3 - 2 - i
Kansas Department of Transportation Design Manual

Disclaimer:
Disclaimer:ThisdocumentisprovidedforusebypersonsoutsideoftheKansasDepartmentof
Transportationasinformationonly.TheKansasDepartmentofTransportation,theStateofKansas,nor
itsofficersoremployees,bymakingthisdocumentavailableforusebypersonsoutsideofKDOT,does
notundertakeanydutiesorresponsibilitiesofanysuchpersonorentitywhochoosestousethis
document.ThisdocumentshouldnotbesubstitutedfortheexerciseofapersonsownProfessional
EngineeringJudgement.Itistheusersobligationtomakesurethathe/sheusestheappropriate
practices.AnypersonusingthisdocumentagreesthatKDOTwillnotbeliableforanycommercialloss;
inconvenience;lossofuse,time,data,goodwill,revenues,profits,orsaving;oranyotherspecial,
incidental,indirect,orconsequentialdamagesinanywayrelatedtoorarisingfromuseofthis
document.

TypographicConventions:
Thetypographicalconventionforthismanualisasfollows:

NonitalicreferencesrefertolocationswithintheKDOTBridgeDesignManuals(eithertheLRFDorLFD),
orHyperlinksshowninred,asexamples:

x Section3.2.9.12Transportation
x Table3.9.21DeckProtection


ItalicreferencesandtextrefertolocationswithintheAASHTOLRFDDesignManual,forexample:

x Article5.7.3.4

ItalicreferenceswithaLFDlabelandtextrefertolocationswithintheAASHTOLFDStandard
Specifications,forexample:

x LFDArticle3.5.1

Volume III US (LRFD) Bridge Section

Version 4/08 3 - 2 - ii
Kansas Department of Transportation Design Manual

3.4.1 General
In general all analysis will fall into two categories, elastic and inelastic, that is, behaving in a lin-
ear or non-linear fashion. Of these, there are two parameter types to consider, geometric and mate-
rial behaviors. With the exception of P-delta analysis for long columns in substructure design, all
KDOT designs will be designed as linear elastic using first order analysis. Allowing inelastic
material resistance without doing a second order analysis is considered a lack of compatibility.

3.4.2 Methods of Analysis

Below is a list of some of the commonly used analysis methods for analyzing bridge structures.
The level of complexity in the analysis is governed by complexity of the structure and the results
needed. Commonly, more than one method is used on a structure or member with nodal displace-
ment being the common tie between methods.

3.4.2.1 Hand Analysis Methods

Any analysis/design method that can be performed completely by hand (even if it is sometimes or
often programmed into a spreadsheet or computer program). Examples include the Moment Dis-
tribution (Hardy Cross), V-Load Method, Conjugate Beam Method, Castiglianos Second Theo-
rem, Energy Methods (Bernoulli), etc.

3.4.2.2 Line Girder Analysis Methods

Any analysis/design method that isolates a single tangent girder from the rest of the superstruc-
ture system and evaluates that girder individually, with the rest of the superstructure system con-
sidered only by means of boundary conditions, live load distribution factors, etc.

3.4.2.3 Grid Analysis Methods

The characterization as a Grid analysis is meant to mostly address 2D Grid or Grillage analysis
methods. This includes any analysis/design method that includes a computerized structural analy-
sis model where:
the superstructure is typically modeled as a two-dimensional array of nodes and line ele-
ments, and where the girders and cross frames or diaphragms are typically modeled using
line elements,
the analysis displacements are solely vertical displacements and rotations about axes in a hor-
izontal plane, and where the loads considered are primarily out of plane vertical (gravity)
loads. Two variants are listed below (Plate and Eccentric Beam Grid and 3D Grid).

Volume III US (LRFD) Bridge Section

Version 4/08 3-4- 1
Kansas Department of Transportation Design Manual

3.4.2.4 Plate and Eccentric Beam Grid Analysis Methods

A variant on the 2D Grid/Grillage analysis model, where the deck is modeled using plate or shell
elements, while the girders and cross frames are still modeled using line elements offset from the
deck elements.

3.4.2.5 3D Grid Analysis Methods

This is a modification of a 2D Grid analysis, where more degrees of freedom are modeled. Some
typical additions that separate 3D Grid methods from 2D Grid methods include:
modeling of warping stiffness and warping response of I-shaped girders
modeling of the shear stiffness of cross frames or diaphragms,
modeling of girder supports, lateral bracing, and/or cross-frames or diaphragms at their phys-
ical elevation within the structure.

3.4.2.6 3D Analysis Methods

Any analysis/design method that includes a computerized structural analysis model where the
superstructure is modeled fully in three dimensions, including:
modeling of girder flanges using line/beam elements or plate/shell/solid type elements
modeling of girder webs using plate/shell/solid type elements
modeling of cross frames or diaphragms using line/beam, truss, or plate/shell/solid type ele-
ments (as appropriate)
modeling of the deck using plate/shell/solid elements.

3.4.3 Discretization
Prior to the advent of desktop computers, structures were mathematically disassembled to sim-
plify the analysis. Traditionally, the reaction of individual components were used in loading adja-
cent members, thus limiting the degree of external indeterminacy and simplifying the work of the
analyst. For planer (1-D) analysis, this method still has considerable merit in that conservative
boundary conditions can be applied to each member thus maximizing the force effects in each.
Also, these methods lend themselves to closed form solutions. Modern methods do not need this
level of simplification, however the principles used still apply. All modern structural analysis pro-
grams use stiffness methods to model the structures behavior. This involves mapping a members
degrees-of-freedom, through the member stiffness matrix, to the structure degrees-of- freedom
creating a global structure stiffness matrix with boundary conditions consistent with the actual
structure. Inverting the structure stiffness matrix and multiplying by the load vector yields the
structures displacement vector. From this, all member end forces are obtained. The general form
is:
1
{ }= [ K] {P }
Where:
{ } = Displacement
1
[ K ] = Inverted Structure Stiffness Matrix
{ P } = Load

Volume III US (LRFD) Bridge Section

Version 4/08 3 -4 - 2
Kansas Department of Transportation Design Manual

The structure being analyzed shall be reduced to members or elements connected with the same
number of degrees-of-freedom which represents the attributes of the physical structures. Members
which have only slight cross-sectional changes can be modeled as prismatic members with nodes
at convenient locations such as tenth points. Non-prismatic components may be modeled by dis-
cretization of the components into a number of frame elements with stiffness properties represen-
tative of the actual structure at the locations of each element. The use of beam elements with
nodes having six degrees-of-freedom will accomplish the desired goal. When using finite ele-
ments, plates or shells, limit the aspect ratio to 5:1 to reduce errors. Solids should only be used as
load transferring devices or to prove structural continuity. It is important to remember that rods,
beams, plate and shells derive their stiffness from constants multiplied outside of the stiffness
matrix, whereas solids derive their stiffness as geometry from nodal locations. Solids are therefore
more sensitive to geometrically induced errors.

The degree of continuity and the location within the cross-section where load transfer occurs
could conceivably effect the analysis results for members that are not considered axially rigid. If
the member stiffness matrix considers an axial degree-of-freedom, then the load path will follow
the centroid of the member and supports resisting these forces will form a couple, resulting in
internal moments. These considerations must be well understood by the analyst. Using a member
offset creates a rigid link which will pick up these internally coupled effects.

Boundary conditions for the model vary with assumed structural responses of the supports. They
can be idealized as pinned or fixed for both translation and rotation in either a yielding or non-
yielding condition. This can be refined where the rotational or translational stiffness is repre-
sented by springs with the same stiffness as the members framing into that location. One common
example of this, in bridge design, is the modeling of the substructure elements. For example, the
analyst may represent pile as short lengths assumed by fixed at a point below the ground corre-
sponding to where the maximum moment in the pile would occur. This can be determined by a P-
Y analysis performed with programs like L-Pile or COM624. Without such information it is rec-
ommended that this point of fixity be considered at 8 feet below the ground line for cohesive
materials and 10 feet for cohesionless materials. If the boundary conditions can change during the
life of the structure, as in the case of scour, then both limits will be investigated.

3.4.4 Distribution of Load

This section addresses load distribution consideration for different structure types and modeling
methods. The LRFD Specifications allow the use of either a refined or an approximate method of
analysis. An approximate method of analysis can be utilized to determine the lateral live-load dis-
tribution to individual girders for typical highway bridges. Lateral live-load distribution factors
are dependent on multiple characteristics of each bridge. There are specific ranges of applicability
for the use of approximate methods of analysis. Extending the application of such approximate
methods beyond the limits requires sound and reasonable judgement. Otherwise, used a refined
analytical method 2.

Volume III US (LRFD) Bridge Section

Version 4/08 3-4- 3
Kansas Department of Transportation Design Manual

3.4.4.1 Beam Bridges

Equations and tables for live load distribution factors are provided in the LRFD Specifications.
For typical beam bridges distribution factors are provided in Tables 4.6.2.2.2b-1 thru 4.6.2.2.3c-1
for interior beam flexure (single lane, multiple lanes, and fatigue), and interior beam shear (single
lane, multiple lanes, and fatigue). When applying the skew correction factor, use the controlling
factor for all supports when considering shear force effects.

Use lever rule and distribution formulas to determine the amount of live load carried by the exte-
rior beam. LRFD C4.6.2.2.2d provides a formula for computation of an additional distribution
factor for bridges that have diaphragms or cross-frames. Use of the rigid cross-section or pile
equation distribution factor is not required for design of exterior beams. Recent studies have
shown this to be overly as it is thought to be overly conservative1.

3.4.4.2 Slab Bridges

The distribution factor (E) will be investigated for changes in span length, single lane loading, and
for multiple lanes loaded according to Article 4.6.2.3. Average the distribution factor for adjacent
spans over the piers.

When using the empirical methods, the designer should be aware that the multiple presence
factor has been included in each respective formula.
For analysis of fatigue, load one lane loaded and use only one truck loading as the loading
condition with dynamic allowance specified in Article 3.6.2.

3.4.4.3 Structures Under Fill

This section is to be completed later.

3.4.5 Longitudinal / Transverse Model

Typically a bridge structure is modeled as two independent structures with the longitudinal model
determining the superstructure force effects, and the transverse model determining the substruc-
ture force effects. For both model types, where transient load application would produce a large
amount of load cases, it is recommended that influence diagrams be used to simplify determining
maximum force effects.

3.4.5.1 Longitudinal Model

The beam/girder or slab is usually analyzed as a one dimensional structure with the resulting dead
load being the self-weight and the tributary area. The live load is moved to the location which
maximizes the particular force effect (shear or moment) being designed. Because of the number
of load cases the use of influence diagrams is recommended. The controlling live load results are

Volume III US (LRFD) Bridge Section

Version 4/08 3 -4 - 4
Kansas Department of Transportation Design Manual

multiplied by the distribution factor, the multiple presence factor, and the dynamic allowance.
This results in the unfactored service live load.

Stability of Steel Bridges- Using a one dimensional analysis assumes that the structure is stable
both laterally and torsionally. Lateral stability means the structure can resist horizontal loads with-
out excessive deformations. Torsional stability means that the structure can resist twisting without
excessive deformations. The LRFD Specification does not limit the spacing of diaphragms as it
has in the past, and with the use of higher strength materials, stability must be provided by design.
This is most critical during erection and at the time of concrete placement. Because of this, it is
KDOTs policy to limit the maximum spacing of the cross-frames and diaphragms to 25-0 for
the interior bays on tangent steel bridges. The exterior bays which support the screen require a
detailed torsional analysis.

Horizontally curved bridge structures have geometric ecentricities that require special consider-
ations both during construction and in service. The definition of what constitutes a curved bridge
is not presented in the LRFD Specification. This can be found in the scope section of the 2003
AASHTO Document Guide Specifications for Horizontally Curved Highway Bridges. Ignore cur-
vature effects when computing primary bending moments for structures with the characteristics
detailed in Article 4.2 of the Guide Specification.

3.4.5.2 Transverse Model

The pier or abutment bents are usually modeled as 2-D structures because the inherent nature of
loading conditions. Dead load reactions, self weight, wind, water, braking, seismic restraint, drift,
and ice are all loads of known location and magnitude. Live load reactions can be moved across
the pier and therefore must be reduced to the individual components. The maximum live load
reaction is separated into truck and lane on a per girder line or per foot basis.

3.4.5.2.1 Slab Bridge

For a slab bridge, the lane load is reduced to a load per foot for lane load by multiplying the lane
reaction by (E) and dividing by 10 for a 10 foot load lane, and a load per wheel line for the truck
load. The maximum truck reaction is multiplied by (E) and divided by 2 to get the per wheel line
force. Knowing that a design lane is 12-0 in width and a load lane occupies 10-0 of that design
lane, between the barriers, the designer will divide the roadway by the maximum amount of full
12-0 design lanes. Within the design lane the 10-0 load lane will be placed to maximize the
force effects on the bent to the limits specified in Article 3.6.1.3.

3.4.5.2.2 Beam Bridge

The method used for a slab bridge applies to beam bridges except that instead of using (E) the
distribution factor (g) is used and the deck forces are transferred to the substructure at a discrete
girder locations. Other wise, the methods are similar.

3.4.6 Superstructure
This section looks at how the force effects for each design element are obtained.

Volume III US (LRFD) Bridge Section

Version 4/08 3-4- 5
Kansas Department of Transportation Design Manual

3.4.6.1 Deck Slab Analysis

The primary direction for analysis and design of the deck slab is perpendicular to the center line
of traffic. The deck slab is considered to be a continuous and a simply supported beam in the
transverse direction. The maximum design live load moments have been calculated in Table A4-1.

3.4.6.2 Slab Type Bridge

For slab type bridge structure analysis the primary direction for analysis and design of a slab type
bridge is parallel to traffic. Historically, design and analysis are based on a one foot strip that is
representative of a section taken near the center of the structure. The Slab distribution (E) is
assumed to be the width required to support one lane of traffic. There is, a reciprocal relationship
between the distribution factor and the controlling force effects, that is as (E) is decreased the
force effects increase.

3.4.7 Substructure
3.4.7.1 Bearing
To maximize the force effects use the controlling (obtuse corner) skew correction with the maxi-
mum load factor for all supports. Check uplift by using the lower skew correction factor (acute
corner) with the minimum load factor.

3.4.7.2 Pier Beam

Analysis consists of maximizing the force effects for live loads as described earlier. Web walls,
although typically ignored for stiffness, may be considered either structural or not, based on the
amount of connectivity provided. Seismic restraint must be provided according to Article
3.10.9.2. Friction will not be used to resist seismic forces and a positive connection must be pro-
vided.

3.4.7.3 Column
All loads must pass through the columns into the foundation. For vehicular collision design,
frame action can be considered if a crash worthy wall is cast between the columns. Columns will
be analyzed for moments caused by biaxial loads at the top and bottom of the column.

3.4.7.4 Footing/Foundations
This section is to be completed later.

3.4.8 Fixity
The overall fixity of the bridge should be examined in detail for bridges on steep grades, moderate
to severe curvature, or when the columns are tall and/or slender. Use the following guidelines for
providing fixity at bearings.

For short bridges on steep grades, the down hill abutment should be fixed. For longer bridges
the flexibility of each pier and its bearings need to be considered to determine the appropriate

Volume III US (LRFD) Bridge Section

Version 4/08 3 -4 - 6
Kansas Department of Transportation Design Manual

substructure units to fix. If pier flexibility and geometry permit, a minimum of two fixed
piers per expansion unit should be used.
For very flexible piers, such as pile bents or slender columns, the expansion bearings may be
redundant in that the pier may move before the bearings begin to slide.
For typical prestressed I-beam bridges with two sets of bearings surfaces on each pier (per
beam line), sufficient anchorage to the pier is provided by using strands that to extend into the
cast-in-past diaphragm that is anchored to the pier beam though doweled reinforcement.
For river piers and for spans over 145, feet designers will provide bearings that are guided
externally.

3.4.9 Construction Load Analysis

By Specification, all KDOT construction managed structures require erection plans. These are to
be developed, submitted and reviewed based on the geometry of the incomplete structure during
each phase or condition of construction in which the structure is considered to be self-supporting.
Structures which will have railroad or highway traffic passing under the partially completed
structure require a Professional Engineer to Seal the Erection plans. See KDOT Specification for
a complete list of requirements.

(1) NCHRP Report #592 (Project 12-62) Simplified Live Load Distribution Factor Equations
(2) Non-Standard Gage Finite Element Analysis using Virtis/Opis

Volume III US (LRFD) Bridge Section

Version 4/08 3-4- 7