Applied Engineering Letters Vol.3, No.

1, 13-19 (2018) e-ISSN: 2466-4847

LIFETIME ESTIMATE OF PIPE SYSTEMS

UDC: 624.042.62
Original scientific paper https://doi.org/10.18485/aeletters.2018.3.1.3

Samir Dizdar1, Radoslav Tomović2, Adisa Vučina3

1
BerDiz Consulting AB, Gothenburg, Sweden
2
University of Montenegro, Faculty of Mechanical Engineering, Podgorica, Montenegro
3
University of Mostar, Faculty of Mechanical Engineering and Computing, Bosnia and Herzegovina

Abstract: ARTICLE HISTORY

Micro cracks have been found near some of the welds on the pipe unit. This Received 04.01.2018
paper investigates the root cause of this damage and suggests Accepted 22.02.2018
improvements to prevent such damage from occurring. Available 15.03.2018
The investigation includes several simulations using Pipestress and Ansys
Mechanical, a welding and materials investigation, and a more theoretical
creep investigation using the Larson-Miller parameter. A sensitivity study of KEYWORDS
the outlet pigtail installation and support configuration is performed as well Cracks,
creep,
as a cyclic plastic analysis. A complementary analysis of the catalyst tubes,
fatigue,
inlet and outlet system shows that the only areas where the stress and strain plastic analysis,
indicate an elevated risk of damage are the areas around the ends of the lifetime.
outlet pigtails.
Suggested improvements include reconfiguration and load reduction of
constant hanger supports, improving welding procedures in order to lower
residual stresses and undertake more material and on site investigations to
verify causes of failure.

1. INTRODUCTION 2. OBJECTIVE AND SCOPE

In a refinery in Norway where a new pipe The scope of this paper contains an
system was installed, they had serious problems investigation into different damage mechanisms
with cracks already after less than a year in use. It such as creep, fatigue and welding procedure to
is very rare to have similar problems after such a find a root cause for the micro cracks. An
short period of time. investigation into the previously made calculations
is made to establish weather this damage could
Micro cracks have been found on several of the
have been predicted. Using the results of the
pipes in the pipe unit [1]. To continue to operate, different analyses, suggestions on improvements
the root cause need to be determined and are made.
potential actions need to be taken. This paper treats the most highly stressed areas
The purpose of this paper is to analyze the micro and where the damages have been observed.
cracks in the outlet system and try to establish a Hence, not all parts of the system are evaluated in
root cause for this damage, as well as suggesting detail. The focus lies on the parts containing the
how to avoid further damage or prevent such welds where micro crack damages have been
damage in future installations. In addition, a found, in particular three locations where the most
complimentary evaluation is presented in Section severe damages are observed. Following the
notation of [1], these three welds are called S3, S4
4.
and S7. Their locations are shown in Fig.1.

CONTACT: Radoslav Tomovic, [email protected]

 S. Dizdar et al. / Applied Engineering Letters Vol.3, No.1, 13-19 (2018)

Several different methods are used to try to 2.1 Pipestress analysis

determine the cause of the observed damages. The
system is investigated with numerical finite The software Pipestress [4] is used to make a
element (FE) analyses as well as with analytical and model of the piping system and to obtain loads and
theoretical methods. Welding and inspection movements at different points throughout the
system.
books are studied to analyze the choice of
Pipestress is a one dimensional software that is
materials and welding procedure. The different used to model large piping systems including
methods are meant to provide a substantial width supports of different kinds. The software is very
in the investigation and hopefully increase the efficient at calculating the response of the system
understanding and to prevent damages like these to different loads and thereby identifying weak
from occurring in the future. points in its design. It is also very useful when
The system is evaluated according to the ASME wanting to investigate different design solutions.
B31.3 standards [2]. API papers are also
considered. ASME B31.3 is used for process piping 2.2 ANSYS finite element analysis
and will be used to evaluate stresses in the system
The software ANSYS Mechanical [5] is used as a
for design purposes. A Swedish code BSV97 [3] is
verification of the Pipestress results as well as to
used for calculation of wind and snow loads on provide detailed results regarding stress
different structures. concentrations and stress directions at points of
interest. Furthermore, thermal transients and
effects due to creep properties of the material are
investigated using ANSYS.

2.3 Loads

The system is evaluated for sustained loads and

thermal range loads. The sustained loads
considered are internal pressure and dead weight.
The loads are evaluated in accordance with
reference [2]. No wind loads have been applied in
the calculation, but recorded data suggest that the
Fig.1. The location of the three most severe
contribution is small. The temperatures (T) and
damages, (the damages at S3, S4 and S7 are all located
on the southern side of the mid-section of the
pressure for the different parts of the system,
manifold) according to specifications, are presented in Table
1.

Table 1. Temperature and Pressure conditions according to specifications

Design Operation
Part
Temperature [°C] Pressure [MPa] Temperature [°C] Pressure [MPa]
Transfer Line 300 2.8 250 2.48
Manifold (Sub Header) 890 2.8 860 2.48
Pig Tails 890 2.8 860 2.48

The load combinations are presented in Table 2. Table 2. Load Cases

Thermal range is the load resulting from Thermal Load case Loads Allowed stress
Expansion (TE). Thermal Range is defined as the Sustained Loads DW+DP SA (from table)
difference between the cold and hot states of the Thermal Range TE (+DW) SA=1.25 Sc+0.25 Sh
system. In these cases, Dead Weight (DW) is Sustained Loads DW+OP SA (from table)
included together with TE. Thermal Range TE (+DW) SA=1.25 Sc+0.25 Sh
Combined (Creep) DW+OP+TE -

14
 S. Dizdar et al. / Applied Engineering Letters Vol.3, No.1, 13-19 (2018)

2.4 Welding and materials investigation installation and inspection books together with
comparisons with literature.
An investigation into what impact welding The materials used for the different parts are
procedure and material quality may have had on shown in Table 3.
the occurrence of damages is made using
Table 3. Parts and Materials
Temperature Allowable Young’s Thermal Expansion
Part Material [°C] Stress [MPa] Modulus [GPa] Coefficient [mm/m]
20 172 204 0
Transfer Line A 387 GR. 11 CL. 2 250 130 190 2.6
300 125 186 3.8
20 116.7 196 0
Manifold
PARALLOY CR32W 860 15.18 143 15.54
(Sub Header)
890 12.64 141 16.1
20 115 196 0
Pig Tails ASTM B 407 N08811 860 9 143 15.35
890 6.9 141 15.59

3. RESULTS cases considered are the conditions for design and

operation. Temperature and pressure conditions
Many of the results are presented with the term according to specifications are given in Table 1 and
utilization. The utilization for a component is the utilizations for the different load cases for the
fraction between the actual and allowed stress, different models are given in Table 4. The model of
the current state of the system is also run with the
U= S_actual/S_allowed. (1)
sliding boundary condition at the top of the Pig
A utilization below 1 (𝑈 < 1) means the Tails.
component is qualified, a utilization above 1 (𝑈 > Table 5 shows the stresses at the welds S4, S7
1) means that a more detailed analysis needs to be described in the Technical report, reference [1], as
performed or requires a change in design. well as the stress at the node with highest
utilization in the model. The pressure causes a
3.1 Pipestress Results stress of 1.68 MPa in design and 1.48 MPa during
operation conditions. Adding the moment stress
Results are presented for Sustained Loads and gives the total sustained stress presented in the
Thermal Range Loads with utilization in accordance table.
with ASME B31.3 302.3.5 (c) and (d) [2]. The load
Table 4. Utilization
Design Operation
Part
Sustained Thermal Range Sustained Thermal Range
Current State 1.36 0.90 1.02 0.83
C.S. Sliding 0.82 0.77
Improved State 0.84 0.90 0.63 0.83
Table 5. Stresses at S4, S7 and the highest utilization stress MPa, (the node number where the stress occurs in
Pipestress is presented in parenthesis)
Design stress [MPa] Operation stress [MPa]
Part Stress
Sustained Thermal Range Sustained Thermal Range
Current Maximum 9.36 (PO3L) 130.4 (PJ27) 9.16 (PO3L) 120.4 (PJ27)
State S4 (S518-PD10) 7.45 110.2 7.25 104.4
S7 (S004-PE30) 7.39 117.1 7.19 110.7
Current Maximum 11.64 (PM1L) 119.1 (S002) 11.45 (PM1L) 112.5 (S002)
State S4 (S518-PD10) 7.46 108.3 7.26 102.5
(Sliding) S7 (S004-PE30) 7.4 118.7 7.2 112.1
Improved Maximum 5.76 (PQ82) 130.4 (PJ27) 9.49 (SS02) 120.4 (PJ27)
State S4 (S518-PD10) 4.17 110.2 3.97 104.4
S7 (S004-PE30) 3.88 117.1 3.68 110.7

15
 S. Dizdar et al. / Applied Engineering Letters Vol.3, No.1, 13-19 (2018)

3.2 ANSYS results Table 6. Load Cases

Maximum stress
Overall, the ANSYS results are in good Current load [MPa]
agreement with the Pipestress results. However, cases
more details are modeled in ANSYS, which results Inside Outside
in stress concentrations and higher peak stresses. Sustained 36 13
(PD+DW)
Another fundamental difference from the Design Thermal 75 150
Pipestress results is that circumferential stresses expansion
due to inner pressure is included in the ANSYS (TE890 °C +DW)
results. The absence of these stresses in Pipestress Sustained 32 12
is a result of Pipestress only being concerned with (PO+DW)
Operation Thermal 70 137
evaluating longitudinal stresses.
expansion
Three different types of analyses are performed (TE820 °C+DW)
with ANSYS: Increasing 40 25
1. Static structural analyses of the sustained Thermal (860-20) oC/4h
(pressure + dead weight) loads and the thermal Transient Decreasing 130 60
(20-860) oC/4h
expansion load (including dead weight)
2. Creep analyses, to study the creep strain over
time at different temperatures.
3. Thermal transient analyses, to study the
temperature and local thermal stress
distributions at start-ups, shut-downs and
trips.

The maximum stresses observed in the

considered system parts for the static structural
and thermal transient analyses are summarized in
Table 6 for the respective load cases. The stresses
on the inside and outside of the pipes are
presented separately. The internal pressure has
little effect on the outside but may cause high
stresses on the inside while the dead weight
typically affects the outside most. The von Mises
stress on the inside and outside surface of a section
of the manifold is plotted in Fig.2. One may note in
Table 6 and Fig.2 that the stress under sustained
loads is significant on the inside. This occurs on the Fig.2. The von Mises stress on a section of the
manifold’s outer surface (top) and its inner surface
inside of the manifold at the pig tail connection.
(bottom) under PO+DW loads, (the central southern pig
One may also in Table 6 note the large stress tail group is on the bottom of the plots and thus S4 is
due to thermal expansion. This stress arises at the where the maximum is labelled)
end of the pig tails by the manifold and is
considered in detail in the context of the observed The main results concern the observed damages
damage at the S7 weld. at S3, S4 and S7. An overview of the maximum
principal stresses at these points for different load
cases is presented in Table 7, where also the
maximum utilization among the points is given for
each load case.

16
 S. Dizdar et al. / Applied Engineering Letters Vol.3, No.1, 13-19 (2018)

Table 7. Overview of the maximum principal stress, which roughly coincides with the longitudinal direction, at the
areas around the observed damages
Maximum Principal stress Utilization
Current load cases (Longitudinal) [MPa]
S3 S4 S7 Overall Stress limit [MPa]
Design Sustained (PD+DW) 7.5* 14.7 7 1.16 12.64
Thermal expansion (TE890 °C +DW) 20 63 175 - -
Thermal range (890-20 °C) 14 50 168 1.15 145.5
Operation Sustained (PO+DW) 8.5* 13.3 7 0.88 15.18**
Thermal expansion (TE820 °C+DW) 19 54 158 - -
Thermal Range (820-20 °C) 13 41 151 0.98 154

The non-linear behavior of creep makes the unlikely that micro cracks also could be found on
results from the creep analyses very sensitive to the inside of the manifold.
the particular values used for the temperature and The following list summarizes the conclusions
material properties. Thus, the creep results are which can be drawn from the ANSYS results
here chosen to be summarized with a plot rather regarding the damages at S3, S4 and S7.
than a table. In this way, one may judge the S3: The evaluated stresses around S3 for the
behavior of the creep in the model. One of the different load cases do not on their own suggest
locations found to be quite sensitive to creep is the that there would be any damages there at this
upper part of the weld between manifold and pig time nor in the near future. Perhaps together
tail weldolets, e.g. where the damage at S4 where with the residual stresses, discussed in the next
found. The strain at such a point is plotted against section, they could contribute to damages.
time for various temperatures in Fig.3. The strong
S4: The stresses due to sustained loads around S4
temperature dependence is evident from the plot
exceed the allowed limit in the design case and
where it can be seen that changing the
temperature from 820 to 860 °C reduces the time are highly utilized in the operational case.
to reach a particular strain almost by a factor of 10. However, the region is shown to be sensitive to
creep strain but with current data it is difficult
to make accurate predictions regarding its life
time. Furthermore, the thermal stress range is
moderate and fatigue will thus give negligible
contributions.
S7: The thermal stress range at this weld is large
but lie 2% below the allowed limit in the
operational case. However, this limit is rather
conservative, in particular in this case where the
actual number of cycles is so small.
Furthermore, a comparison of the strain range
Fig.3. The strain over time at the weld between
manifold and weldolet, e.g. S4’s location, (the 1% strain with the low-cycle fatigue data of the similar
limit at which creep damages might emerge is plotted INCOLOY 800H alloy [6] shows that the
as a dashed line) evaluated strain is roughly 50 times smaller than
that which would cause failure.
The outside of weldolet/manifold was one of Even though the magnitude of the evaluated
the most highly stressed points under sustained stresses at the damaged regions do not provide
loads, which is why it is susceptible to creep. The striking, conclusive evidence of the cause of micro
point on the inside, mentioned earlier, where the cracks, their directions correspond well with the
maximum stress occurs during sustained loads, is orientation of the cracks. The number of tests and
located on the inside of the manifold by the pig tail observed cracks do not provide a good statistical
connection. Due to its higher stress, this point is basis but there seems to be a correlation between
even more susceptible to creep. Therefore, creep the directions of stresses and cracks. This may
damages might be worse on the inside and it is not suggest that the stress have had an impact on the
crack formation, perhaps in combination with

17
 S. Dizdar et al. / Applied Engineering Letters Vol.3, No.1, 13-19 (2018)

factors not accounted for here, e.g. residual expected lifetime for these temperatures at the
stresses in the welds. stress of 14.7 MPa is presented in Table 9.
From the results, it is clear that the materials
3.3 Creep Analysis Results differ in rupture life, with the ASME material being
considerably weaker. The ASME material has a
The evaluated utilization at the current time safety factor applied to it to get the allowable
with regard to expected lifetime varies by use of stress given in the code [2]. This explains most of
data from references [2,6,7]. The results for the the difference but since the determining material
two stresses 10 MPa and 14.7 MPa at different value is unknown a compensation for the safety
operational temperatures are presented in Table 8 factors cannot be made. Furthermore, the ASME
for the different references. The corresponding material is specified for the weldolets, to which two
of the critical welds are connected.

Table 8. Utilization at present time of operation for different operational temperatures

Operational PARALLOY, reference [7] INCOLOY, reference [5] ASME B31.3, reference [2]
Temperature 10 MPa 14.7 MPa 10 MPa 10 MPa 14.7 MPa 10 MPa
820 °C 0.002 0.005 0.002 0.01 0.16 0.43
840 °C 0.006 0.01 0.006 0.03 0.38 0.99
860 °C 0.02 0.04 0.02 0.11 0.86 2.20
880 °C 0.05 0.1 0.07 0.35 1.89 4.75

Table 9. Expected life (rupture life) [h] at constant stress 14.7 MPa
Operational PARALLOY, reference [7] INCOLOY, reference [5] ASME B31.3, reference [2]
Temperature 14.7 MPa h 14.7 MPa 14.7 MPa
820 °C 9610000 h 4320000 h 100000 h
840 °C 3250000 h 1270000 h 43700 h
860 °C 1140000 h 391000 h 19700 h
880 °C 417000 h 125000 h 9100 h

The contribution from fatigue is very low. The loads considered are the same as for the
Considering that the utilization from fatigue is previous analyses, Dead Weight (DW), Thermal
combined with the utilization from creep as a sum Expansion (TE) and Operational Pressure (OP). The
of squares, when the creep-fatigue utilization is Pipestress analysis evaluates both the Design case
evaluated, the contribution from fatigue becomes as well as the operational conditions while Ansys
completely dismissible. only considers the operational case. A summary of
the temperatures and pressures in the system is
4. COMPLEMENTARY ANALYSIS OF THE INLET presented in Table 10.
SYSTEM AND CATALYST TUBES Table 10. Load Cases

The additional analysis presented in this section Design Operation

serves as a current status check and verification of Part Temp. Press. Temp. Press.
the inlet system and catalyst tubes of system. To [°C] [MPa] [°C] [MPa]
Inlet 540 3.15 450 2.8
take into account interactions between the
Catalyst 922 2.8 880 2.48
different system parts, the outlet parts are
Outlet 890 2.8 860 2.48
considered in this analysis as well.
The Pipestress software is used to determine
The results of the Ansys analysis, show that no
the stress levels in the inlet system. Furthermore,
severe stresses or strains are found in the catalyst
Pipestress provides displacement results, which
are used as input in the more detailed ANSYS tubes or the inlet pigtails. The maximum creep
analysis of the Pig Tails and Catalyst Tubes, which strain in the inlet system is 0.07 % and is not in the
also models the development of creep strains in vicinity of any weld which might affect crack
the materials. formation and is thus not considered a risk area.
The only significant strain in connection to the

18
 S. Dizdar et al. / Applied Engineering Letters Vol.3, No.1, 13-19 (2018)

catalyst tubes is at the bottom of the reducer to the the system is very sensitive to small changes during
outlet pigtail where comparatively large stresses installation, e.g. mounting of insulation and
and creep strains occur. It is at the upper end of the cladding as well as the relative position on the
outlet pigtail where the maximum creep strains pipes in relation to the constant hanger support.
occur. These are approximately 0.2 % after 100 000 The Ansys analysis indicates that creep could be
h, and are at the same levels as at lower end of the the root cause of the damage at the S4 weld.
However, the point most sensitive to creep in the
pigtail towards the manifold, where cracks have
model lies on the inside of the manifold, which
been observed. These creep strain levels are not on
suggest that creep damages could be worse there.
their own worrying but taking to account the fact In general, the observed crack orientations are
that cracks have been observed at locations consistent with the direction of the principal
showing the same levels of stress and strain in stresses. However, the magnitude of the principal
combination with close proximity welds, these are stresses are typically not large enough to, by
considered risk areas. The areas with elevated risk themselves, be the cause of the observed damages,
for damage are encircled in Fig.4. indicating the presence unknown stresses, for
example residual stresses.
The complementary analysis of the inlet system
and catalyst tubes show that there are no regions
of significant stress or creep strain in these parts.
The analysis show that the ends of the outlet
pigtails (and adjacent reducer and manifold
respectively) acquire the largest creep strains.
Because of their similarity (in terms of creep strain,
material and close proximity to welds) to the
locations where cracks are observed these
locations are considered as areas with risk for
damage.

REFERENCES

[1] Teknisk rapport, Replikprovning AG, BerDiz

Consulting AB, Rev 01A, 2015.
Fig.4. Regions at the upper end of outlet pigtails
[2] ASME B31.3, 2012.
considered to have an elevated risk of damage [3] Boverket, Boverkets handbook om snö och
vindlast, 1997, Boverket 2nd edition. (In
Swedish).
5. CONCLUSION [4] DST Computer Services S.A., PIPESTRESS,
version 5.0.
The system seems to have been manufactured [5] ANSYS Mechanical, Release 16.0, ANSYS, Inc.
and evaluated in accordance with applicable [6] Catalog: INCOLOY alloy 800T & 800HT, Special
standards and codes, save the questionable Metals Corporation, 2004.
existence of an applied wind load, which would not [7] Catalog: Heat resistant alloy technology
influence the result significantly. centrispun tubes and statically cast fittings,
Moving the constant hanger supports according fabricated coils, catalyst tubes and manifolds.
to the improved Pipestress model would greatly DONCASTER PARALLOY, UK.
reduce the creep driving forces, e.g. sustained
loads. The analyses over all shows that the system The shorter version of this research was
is very sensitive to the loads from the constant presented at the "8nd International Scientific
hangers. The movements of the pipes from Conference IRMES 2017", 7 - 9 September 2017,
sustained loads are very small; this indicates that Trebinje, Bosnia and Hercegovina.

19