-
Incoloy 800HT mechanism of the pipe was analyzed. 2013 Elsevier
Ltd. All rights reserved.
der hiered pon reer plof 8 g
versized pthe main c
the failure was stress corrosion. Khalaf et al. [11] reported
the failure of lube oil feed tube and revealed that the mainwas
oversized thermal stress at high temperature. Thermal stress is
caused by the thermal expansion and thermal ctivity. Since the
yield stress of the structure decreases with the increasing
temperature, the thermal stress will causedeformation if it exceeds
the yield stress and then leads to the cracking, warping, loosening
and other failure form of
1350-6307/$ - see front matter 2013 Elsevier Ltd. All rights
reserved.
Corresponding author at: School of Materials Science and
Engineering, Tianjin University, Tianjin 300072, China. Tel.: +86
22 27402439.E-mail address: [email protected] (Y.D. Han).
Engineering Failure Analysis 31 (2013) 375386
Contents lists available at SciVerse ScienceDirect
Engineering Failure
Analysishttp://dx.doi.org/10.1016/j.engfailanal.2013.01.013rosion
and erosion. Sun et al. [10] investigated the failure of urban pipe
and found that the causes were oand corrosion. Attia et al. [12]
studied the buried piping made of 316L and cold drain vessel, and
found thathe cor-ressureause ofcause
onduc-plasticsion of 800H containing extra Ni, Si, Mn, Ti, etc.
The heat and corrosion resistant of the Incoloy 800HT are
excellent. Hence,Incoloy 800HT has been widely used to resist high
temperature corrosion. The workability is very good and it has
outstand-ing resistance to stress corrosion, especially at the
temperature above 500 C. These advantages make it possible to
reducethe wall thicknesses and consequently improve the behavior of
structure submitted to thermal transients [2,3,5]. Becausethis kind
of alloy operates at elevated temperature, the thermal stress will
arise in the component. Thus, thermal stresscracking is an
important issue [68]. Besides, the problem caused by the stress
corrosion will also arise. To date, some similarinvestigations on
failure analysis have been reported. Gong et al. [9] analyzed the
failure of bursting on the inner pipe of ajacketed pipe which was
made of SA516Gr55 in a tubular heat exchanger, and found that the
failure was caused by tCoordination deformation
1. Introduction
The failure of the components unerally, nickel based alloys are
considloys have high strength and corrosicomponents, rockets and
nuclear powapplications [14]. With the densitygh temperature was
commonly caused by thermal stress and heat corrosion. Gen-romising
candidate materials to fulll these challenges. Since the nickel
based al-sistance at elevated temperature. Therefore, they are
widely used in jet-engineants, as well as in food-handling and
chemical-processing equipments, andmarine/cm3, Incoloy 800HT is a
kind of alloy of Incoloy 800 series and is a modied ver-Failure
analysis of Incoloy 800HT pipe at high temperature
L.Y. Xu a,b, P. Zhu a,b, H.Y. Jing a,b, K. Guo c, S.X. Zhong c,
Y.D. Han a,b,a School of Materials Science and Engineering, Tianjin
University, Tianjin 300072, Chinab Tianjin Key Laboratory of
Advanced Joining Technology, Tianjin 300072, Chinac SINOPEC SABIC
Tianjin Petrochemical Co. Ltd., Tianjin 300271, China
a r t i c l e i n f o
Article history:Received 25 October 2012Received in revised form
19 January 2013Accepted 20 January 2013Available online 5 March
2013
Keywords:Numerical simulationHeat stress
a b s t r a c t
In this study, the failure behavior of Incoloy 800HT pipe was
investigated. Metallographicanalysis and X-ray Diffraction were
conducted to nd the causes of the failure. It wasobserved that the
microstructure of the failed area was same as that in the
undamagedarea, which indicated that the failure of the pipe was not
caused by the change of micro-structure. Then, the nite element
method (FEM) was conducted. The results revealed thatonly axial
stress during service was larger than the yield stress at the
temperature of 890 Cin the inner wall of the pipe. While for the
case of 1032 C, both the axial and radial stresseswere larger than
the yield stress. Based on the stress and deformation analysis, the
failure
journal homepage: www.elsevier .com/locate /engfai lanal
-
Fig. 1. (a) Assembly drawing of the pipe; and (b) the pipes in
service.
Fig. 2. Photos of the failed area: (a) overall view; and (b)
local view.
376 L.Y. Xu et al. / Engineering Failure Analysis 31 (2013)
375386
-
component, which limits the reliability and lifetime of the
components [13]. Thus, it is valuable to understand the
failuremechanism of heat stress in engineering practice.
2. Background
In the present study, the system was composed of Incoloy 800HT
and ceramic. The ceramic was used to insulate heat inthe cavity of
the Incoloy 800HT (shown in Fig. 1a). There was a V type weld along
the outside wall of the pipe. The pipe wasused to transport pintsch
gas (mainly containing hydrocarbon) at the temperature of 844 C.
And the operating pressure wasabout 0.0798 MPa during the operating
process. The photo of operating process was shown in Fig. 1b.
During the operatingperiod, the maximum temperature may reach 890 C
(some pipes may reach 1032 C) for a short time.
After about 1000 h, an obvious necking was observed using the
endoscope in the inner wall of many pipes (shown inFig. 2a and b)
and thus led to the stoppage of operation.
According to the drawing shown in Fig. 1a and the deformed pipe
shown in Fig. 2, it was surprised to nd that the de-formed place
occurred just under the welded joints in the inner wall. However,
the damage caused by hot corrosion wasnot apparent. Consequently,
the objective of this paper is to nd the causes of failure of the
pipe during operation.
3. Experiments
The chemical composition of Incoloy 800HT was listed in Table 1.
The samples were extracted from the failed area andundamaged area
in the inner wall of Incoloy 800HT to conduct metallographic
analysis test and XRD. Before test, the sampleswere washed by the
ultrasonic agitation in acetone.
Table 1Chemical composition of Incoloy 800HT (wt%).
Alloy C Al Si S Ti Cr Mn Fe Ni Co CuIncoloy 800HT 0.077 0.51
0.34
-
Fig. 4. X-ray Diffraction of the failed area.
Table 2Result of tensile test for Incoloy 800HT.
25 C 844 C 890 C
Yield stress (MPa) 355 107 96Tensile stress (MPa) 624 141
130
Table 3Result of expansion coefcient test for ceramic and
Incoloy 800HT.
Name 844 oC (ppm/K) 890 oC (ppm/K)Incoloy 800HT 15.0209
15.940Ceramic 4.2113 4.4210
Table 4Parameters of welding.
Voltage (V) 10Current (A) 100Welding speed (mm/min) 15Interpass
temperature (C) 100Preheating temperature (C) 150Cooling condition
air coolingheat transfer coefcient w/(m2 K) 20Height of the weldout
(mm) 6
Fig. 5. Material properties of Incoloy 800HT.
378 L.Y. Xu et al. / Engineering Failure Analysis 31 (2013)
375386
-
L.Y. Xu et al. / Engineering Failure Analysis 31 (2013) 375386
3793.1. Macro-analysis
From Fig. 2, it can be clearly observed that there was large
plastic deformation in the inner wall which was necked. How-ever,
the features of necked and undamaged surfaces were almost the same.
And the failure caused by hot corrosion was notobvious.
3.2. Metallographic analysis
In order to further investigate the failure causes, a
metallographic experiment was also conducted on the failed
andundamaged area. The photos were shown in Fig. 3a and b.
Fig. 6. Temperature and stress distribution after welding: (a)
temperature; (b) plots of the temperature prole of points 1 and 2
over time; (c) radial stress;(d) circumferential stress; and (e)
axial stress.
-
380 L.Y. Xu et al. / Engineering Failure Analysis 31 (2013)
375386As known to all, the normal microstructure of the Incoloy
800HT was austenite. From Fig. 3a, the metallographic struc-tures
were homogenous and there was not any impurities or particles
precipitated from the grain spreading along the grainboundary. By
calculating the size of the grain, the result was about 24 lm2.
While from Fig. 3b, the grain size was about22 lm2, which meant
that it was smaller than that in undamaged area and the
metallographic structures were homogenoustoo. Besides, the
cavities, which may grow and gather together under stress and then
lead to the crack, were not seen in thegraphic. By analyzing, the
metallographic structures of the two areas were also austenite. In
other words, the metallographicstructures were not changed, which
indicated that the failure was not caused by the change of the
metallographic structure.Besides, the creep cavities were not
found, which meant that the failure was not caused by the creep
deformation.
3.3. XRD analysis of the corrosion products
A preliminary analysis of an unexposed 800HT was studied by Yin
[2]. It was observed that the Incoloy 800HT consistedentirely of a
single-phase Cr0.19Fe0.7Ni0.11 without grain boundary precipitates.
In order to investigate whether the phase ofthe corrosion products
changed, a sample was extracted from the failed area to conduct the
XRD analysis. The result wasshown in Fig. 4.
It could be seen that the phase of the necked area in the inner
wall and the base metal was austenite. Consequently, itcould be
concluded that the damage was not caused by hot corrosion, which
may lead to the change of the metallographicstructure.
4. Numerical simulation analysis
Since the failure of the pipe was not caused by the hot
corrosion or the phase change, it is necessary to simulate the
pipefrom welding process to operating process to see whether it was
caused by thermal stress during service. Before the test,some tests
were conducted to nd out the 0.2% yield strength (YS), the ultimate
tensile strength (UTS) and the expansion
Fig. 6. (continued)
-
L.Y. Xu et al. / Engineering Failure Analysis 31 (2013) 375386
381coefcients of the Incoloy 800HT at temperature of 25 C, 844 C
and 890 C, respectively. The expansion coefcients of theceramic
were also tested.
4.1. High temperature tensile test
The test specimen was prepared in accordance with the ASTM test
method E8M. The mechanical properties of the spec-imens were
investigated using an Instron Microtester, at a cross-head speed of
0.5 mm/min at the temperature of 25 C,
Fig. 7. Distribution of temperature and stress after operating
at 890 C for 1 h: (a) scaleplate; (b) radial stress; (c)
circumferential stress; (d) axial stress; (e)axial stress in 3D
model; (f) value of radial stress along inner wall; and (g) value
of axial stress along inner wall.
-
382 L.Y. Xu et al. / Engineering Failure Analysis 31 (2013)
375386844 C and 890 C. Through the test, the stressstrain curves
were obtained. Consequently, the 0.2% yield strength (YS) andthe
ultimate tensile strength (UTS) were reaped from the curve. The
result was shown in Table 2.
4.2. Expansion coefcient test
The expansion coefcients (CTEs) were tested by the Automatic
thermo-mechanical analyzer TMA 2940, and were ob-tained by
measuring the displacement of the specimen as a function of
temperature at 844 C and 890 C, respectively.
Fig. 7. (continued)
-
The lengths of the specimens were 10 mm with the diameter of 4
mm. The rate of heating was 5 C/min. The height of eachspecimen was
about 30 mm. For each kind of material, at least three groups of
data should be obtained. And their valueswere averaged. The CTE
results were shown in Table 3.
4.3. Finite element modeling
Thermal stress simulations by the nite element method (FEM) were
widely used to analyze the damages caused bymechanical factors
[14]. The pipe was manufactured by welding before service. Hence,
it was necessary to simulate the weld-ing process. Since the pipe
was axisymmetric, it was reasonable to simplify the model as 2D
axial symmetry model. Hence, anite element (FE) model of the
problem was built using ABAQUS to study the stress during welding
and operating process.The method of sequentially coupled
thermal-stress analysis was used to simulate both welding and
operating process. Theheat transfer element type was used to
simulate the heat transferring process. Then the element type was
converted to stresselement type for calculation of the stress eld.
During the welding and operating process, both ends of the pipe
were fullconstrained and the ceramic was contacted to the Incoloy
800HT. The parameters used were shown in Table 4. After welding,it
was cooled down in the air for 10 h and the temperature was below
200 C
During operation after welding, operating pressure was about
0.0798 MPa. And the temperature in the inner wall was
L.Y. Xu et al. / Engineering Failure Analysis 31 (2013) 375386
383844 C for about 1000 h. However, the pipe will be worked at
scorched state for about 1 h at the temperature of 890 C,and some
pipes may reach 1032 C. Then the pipe was returned to operate at
844 C for about 100 h. During the processesmentioned above, the
heat would be exchanged between the outer wall of the pipe and the
air. The material properties ofIncoloy 800HT were shown in Fig.
5.
4.4. Results and discussion
The processes of welding and operating were simulated. The
distribution of temperature and stress in the inner wall
wasobtained. Points 1 and 2 near the V groove (showed in Fig. 6a)
were created to show the temperature prole over time andthe result
was shown in Fig. 6b.
From the results (showed in Fig. 6ce) in inner wall during the
welding process, it can be obviously observed that theresidual
stresses in the radial direction, axial direction and
circumferential direction in the inner wall were too low to
causethe failure. In other words, the failure was not caused during
the welding process. After welding, the inner wall of the
pipeworked at 844 C for 1000 h and then the temperature increased
to 890 C for 1 h and the results were shown in Fig. 7bdand the
value of radial stress and axial stress in the inner wall were
shown in Fig. 7f and g. After that, the pipe was returnedto work at
844 C for about 100 h. However, the axial stress near the free end
(see endpoint B in Fig. 8) in the inner wall was115 MPa, which was
higher than its yield stress at 890 C (96 MPa). The radial stress
was about 85 MPa.
However, for the pipe whose temperature increased to 1032 C
(results were shown in Fig. 9ac and the value of radialstress and
axial stress in the inner wall were shown in Figs. 9e to 7f), the
axial stress near the free end in the inner wall in-creased to 88.4
MPa, which was higher than the yield stress at 1032 C (75 MPa,
which was calculated according to theextrapolation method conducted
by using the yield stress of 844 C and 890 C [15]). Besides, the
radial stress was about80 MPa, which indicated that two directions
of stress exceeded the yield stress.
Each point position in the inner wall could be referred in Fig.
7a. In order to observe the axial stress apparently, the
three-dimensional models were created to show the axial stress in
Figs. 7e and 9d.
From the point of mechanics, only the axial stress near the
endpoint B was higher than the yield stress for the pipe oper-ated
at 890 C. Although the radial stress was very large near the
endpoint A, it was constrained and was difcult to deform.However,
for the case of 1032 C, the axial stress near the endpoint B was
larger than its yield stress. Besides, the radial stressalso
exceeded the yield stress. Under the effect of stresses, the Mises
stress was about 107 MPa, which led to the failure of the
Fig. 8. Scheme of specimen deformation.
-
384 L.Y. Xu et al. / Engineering Failure Analysis 31 (2013)
375386pipe according to the strength theory. And this was why most
pipes working at 890 C were not failed, while some pipeswere
failed.
Fig. 9. Distribution of stress after operating at 1032 C for 1
h: (a) radial stress; (b) circumferential stress; (c) axial stress;
(d) axial stress in 3D model; (e)value of radial stress along inner
wall and (f) value of axial stress along inner wall.
-
L.Y. Xu et al. / Engineering Failure Analysis 31 (2013) 375386
385From the point of deformation, the temperature of endpoint A of
the pipe along the thickness direction would be almostthe same
during the operating process. The area near endpoint A would expand
along the radial direction. Since the thicknessnear endpoint A was
very large, the area near the endpoint A in the inner wall would
deform to the left when the endpoint Aexpanded, so that the whole
system could remain stable. The farther the distance from endpoint
A in the inner wall, thesmaller the effect of this coordination
deformation would be. While the endpoint B of the inner wall was
free in the axialdirection, it would expand freely along the axial
direction. Meanwhile, since the endpoint B was free in the axial
direction,it was more difcult for heat to transfer in the inner
wall. Therefore, the temperature would be higher than other areas
in thepipe. Consequently, the heat stress in the endpoint B would
be higher than other places. Besides, the endpoint B could
Fig. 9. (continued)
Fig. 10. Node change between deformed state and undeformed state
(green line represents the deformed state). (For interpretation of
the references tocolor in this gure legend, the reader is referred
to the web version of this article.)
-
deform freely and the area near the endpoint A would deform to
the left. The area marked by green circle would be the crit-ical
point of coordinative deformation and free deformation (shown in
Fig. 8). Hence, the area would be in dangerous. Thegridding of the
pipe between deformed and undamaged state veried the explanation
above (shown in Fig. 10).
5. Conclusions
The analyses of the failure of the pipe were investigated. The
obtained results can be summarized as follows:
(1) The corrosion and the welding process have insignicant
effect on the failure of the pipe.(2) The microstructure of the
Incoloy 800HT kept constant during the welding and operating
process. And the creep cav-
ities were not seen in the failed area, which meant that the
failure was not caused by the creep.(3) The heat stresses of radial
and axial direction of the pipes working at 1032 C were too large
and the Mises stress
exceeded the yield stress, which caused the failure of the
pipe.
Acknowledgements
386 L.Y. Xu et al. / Engineering Failure Analysis 31 (2013)
375386The authors acknowledge the research funding by National
Natural Science Foundation of China (Grant No. 51275341),Key
Project in the Science & Technology Pillar Program of Tianjin
(Grant No. 11ZCKFGX03000), Program for New CenturyExcellent Talents
in University (NCET-11-0375) and Specialized Research Fund for the
Doctoral Program of Higher Educationof China (20110032130002).
References
[1] Nganbe M, Heilmaier M. High temperature strength and failure
of the Ni-base superalloy PM 3030. Int J Plast.
2009;25(5):82237.[2] Yin RC. Carburization performance of Incoloy
800HT in CH4/H2 gas mixtures. Mater Sci Eng, A
2004;380(12):2819.[3] Fulger M, Mihalache M, Ohai D, Fulger S,
Valeca SC. Analyses of oxide lms grown on AISI 304L stainless steel
and Incoloy 800HT exposed to
supercritical water environment. J Nucl Mater
2011;415(2):14757.[4] Kalpakjian Serope, Steven R. Manufacturing
processes for engineering materials. Schmid: Prentice Hall;
2007.[5] Inman IA, Datta PS. Studies of high temperature sliding
wear of metallic dissimilar interfaces IV: Nimonic 80A versus
Incoloy 800HT. Tribol Int
2011;44(12):190219.[6] Sharan AM, Dhar D, Seshadri R. Heat
transfer/thermal stress analysis of semiconductors under transient
conditions. Model Simul Mater Sci
1997;5(3):24558.[7] Wingert PC. Testing of the
thermal-stress-cracking characteristics of silver-refractory
contacts. Electrical contacts. In: Proceedings of the annual
holm
conference on electrical contacts. Montreal, Can; 1995. p.
33845.[8] Hattali ML, Valette S, Ropital F, Mesrati N, Treheux D.
Effect of thermal residual stresses on the strength for both
alumina/Ni/alumina and alumina/Ni/
nickel alloy bimaterials. J Mater Sci 2009;44(12):3198210.[9]
Gong Y, Zhong J, Yang ZG. Failure analysis of bursting on the inner
pipe of a jacketed pipe in a tubular heat exchanger. Mater Des
2010;31(9):425868.[10] Sun P, Zhu W, Zheng J. Failure risk analysis
and control of urban ow pipe. In: Proc. international conference on
pipelines and trenchless technology
2009, ICPTT 2009: advances and experiences with pipelines and
trenchless technology for water, sewer, gas, and oil applications.
Shanghai, China;2009. p. 158995.
[11] Khalaf AM, Seibi AC. Failure analysis of lube oil feed tube
of a gas turbine operating in oil elds. Eng Fail Anal
2011;18(5):134150.[12] Attia MS, Ragab AR, El-Raghy S. Failure
analysis of buried piping and cold drain vessel. Eng Fail Anal
2011;18(3):93343.[13] Shin S, Kim KM, Song J, Kim HK, Choi DJ, Cho
HH. Thermal stress analysis of Ge1Sb4Te7-based phase-change memory
devices. IEEE Trans Electron Dev
(USA) 2011;58(3):78291.[14] Hino Y, Kiyota Y, Hattori Y. Thermal
stress analysis of BOF accounting for friction force experimentally
measured. ISIJ Int 2010;50(8):112533.[15] Sun CY, Liu JR, Li R,
Zhang QD. Constitutive modeling for elevated temperature ow
behavior of Incoloy 800H superalloy. Acta Metall Sin
2011;47:1916.
Failure analysis of Incoloy 800HT pipe at high temperature1
Introduction2 Background3 Experiments3.1 Macro-analysis3.2
Metallographic analysis3.3 XRD analysis of the corrosion
products
4 Numerical simulation analysis4.1 High temperature tensile
test4.2 Expansion coefficient test4.3 Finite element modeling4.4
Results and discussion
5 ConclusionsAcknowledgementsReferences