- Email: [email protected]
Failure analysis of a steam pipe weld used in power generation plant
Accepted Manuscript Failure analysis of a steam pipe weld used in power generation plant Qiaoling Chu, Min Zhang, Jihong Li, Yinni Chen, Hailong Luo, Qiang Wang PII: DOI: Reference:
S1350-6307(14)00166-6 http://dx.doi.org/10.1016/j.engfailanal.2014.05.019 EFA 2331
To appear in:
Engineering Failure Analysis
Received Date: Revised Date: Accepted Date:
3 April 2014 14 May 2014 20 May 2014
Please cite this article as: Chu, Q., Zhang, M., Li, J., Chen, Y., Luo, H., Wang, Q., Failure analysis of a steam pipe weld used in power generation plant, Engineering Failure Analysis (2014), doi: http://dx.doi.org/10.1016/ j.engfailanal.2014.05.019
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Failure analysis of a steam pipe weld used in power generation plant Qiaoling Chu, Min Zhang, Jihong Li, Yinni Chen, Hailong Luo, Qiang Wang College of Material Science and Engineering, Xi’an University of Technology, Xi’an, 710048
ABSTRACT In this paper a failure of steam pipe weld after 140000 hours of service in power generation plant has been analysed. Efforts were made to analyze the cause of failure in both experiment and finite element analysis (FEA) methods. Intergranular cavities and cracks were detected around the primary cracks. Study shows that the main failure mechanism is creep cracking. Cracks developped in this failed pipe as a result of cavity nucleation on grain boundaries, followed by cavity growth and linkage to form micro-and eventually macro-intergranular cracks. This paper brings out the details of investigation and suggests remedial measures to improve performance of this welded pipe under high temperature and pressure condition. Keywords: Creep cracking; Cavity ; Failure analysis; Welded joint 1. Introduction Components in power generating stations usually operate at high temperatures for long times. A large proportion of pressure components are manufactured from Cr-Mo-V steels which are selected for their high temperature creep resistance at optimum cost. When structural components are joined together by fusion welding, the thermal cycle introduces changes in the microstructure[1]. Welded pressure components are prone to creep deformation and fracture at high temperature, and the majority of the creep failures of pressurised components are associated with the welds[2]. During the maintenance halt of a power generation plant, a steam pipe connected with tee-junction structures was detected transverse cracks in welded joint (Fig. 1a). This failed pipe is manufactured from 1%Cr-0.3%Mo-0.2%V steel. Fig 1b displays the assemble diagram of the investigated steam pipe. The failed pipe is combined by shielded metal arc welding (SMAW) with matching electrodes (ASW E5515). The subsequent post-weld heat treatment(PWHT) (700℃×5h) is employed to alleviate the welding residual stresses. According to the record of operation for the steam pipe unit, the failed pipe has operated at around 140000 h. The pipe has an out-diameter of 450mm and a wall thickness of 50mm. The conveying medium is steam with average pressure of 17.4MPa and temperature of 510℃.
Cracks
a
b Fig. 1 General view of the cracks analyzed. (a) Cracks in welded joint; (b) Assemble diagram of the investigated steam pipe.
The aim of the present investigation is to identify the root cause of the failure and propose corrective measures. Finite
element simulations of the failed pipe were used as supplementary investigation techniques in order to illustrate the possible phenomena causing failure. 2. Experimental set-up The circumferential welded joint in Fig. 1a was cut for investigation (Fig. 2a). Chemical analysis by inductively coupled plasma (ICP) and carbon sulfur analyzer was applied to identify the matrix material of this failed pipe. The ex-service welded joint containing cracks was sectioned for detailed metallographic studies (Fig. 2b). Metallographic sections were prepared to examine the microstructural characteristic and difference between adjacent to and apart from the fracture regions. The metallographic samples were prepared by grinding devices and successive wet grinding emery papers up to #2000 grid and fine polishing in diamond paste. 4% nital was applied as the etch to reveal the microstructure. Cleaning was carrying out using ethanol followed by hot air stream drying. The fracture surfaces were observed in scanning electron microscope (SEM) equipped with an X ray energy dispersion spectrometer (EDX). Both the tensile and Charpy impact specimens were machined along the weld direction (Fig. 2c). The tests were performed at ambient temperature. Vicker microhardness test was also performed applying 200g of force in the welded joint. Finite element analysis procedures for predicting the magnitude and distribution of residual stresses were also developed in welding and as-service process, respectively. Cracks
Tensile specimen
Specimen C Specimen A
(SEM, microstructure,
(SEM, microstructure)
Charpy impact specimen
microhardness)
Weld metal
Specimen B Outer surface
a
inner surface
b
(SEM, microstructure)
c
Fig. 2 Failed circumferential welded joint cut for investigation. (a) Circumferential welded joint; (b) and (c) Sampling schematic.
3. Failure analysis 3.1. Chemical analysis The specified elemental composition of 1%Cr-0.3%Mo-0.2%V and measured composition of failed pipe are present in Table 1. The results indicated that the failed pipe material 1%Cr-0.3%Mo-0.2%V is consistent with the DL 5007 standards. Table 1 Chemical compositions of the failed pipe material (wt%). C
Mn
Si
Cr
Mo
V
Failed pipe material
0.09
0.52
0.34
1.01
0.28
0.30
DL 5007 standards
0.08-0.15
0.40-0.70
0.17-0.37
0.90-1.20
0.25-0.35
0.15-0.30
3.2. Microstructural analysis Non-metallic inclusion test was performed before metallographic detection in specimen C. The sample was prepared with fine polished in diamond paste. Fig. 3a displays the distribution of non-metallic inclusions. In accordance with the ISO 4967 standards, the degree of non-metallic inclusions conforms to D3, which is closed to the maximum limit and could be
considered high. SEM images of the non-metallic inclusions are shown in Fig. 3b and c. The inclusions presented circular morphology and chain distribution. EDX results indicated that the inclusions are rich in O(~24.0%) and Cr(~3.8%) elements. Some large pores are observed in Fig. 3c. It was obvious that cracks initiated from the pores. Residual inclusions were also detected in some pores. It was deduced that some large inclusions exposed were grinded off during the sample preparation. Previous research [3] investigated the failure of carbon steel pipes and results proved that the non-metallic inclusions could promote the initiation and propagation of cracks.
Cracks Residual inclusions
Chain-like distribution inclusions a
b
20μm
c
Pores
500μm
Fig. 3 The morphology of non-metallic inclusions. (a) Observed by optical microscopy; (b) and (c) Observed by SEM.
Metallographic inspection was performed in both cross and longitudinal section of the weld line. Specimen A and B were contained the primary cracks. Fig. 4a depicts the crack morphology in specimen A. As can been seen, some secondary cracks initiated from the outer surface of weld metal. Fig. 4b presents intergranular cracks observed near the primary cracks. Regions adjacent to the primary cracks characterized with dark scales. EDX test revealed that the dark regions were rich in O element. The cracks were most probably oxidized in high temperature operation. Fig. 4c-f display the microstructure of specimen C. The microstructure of base metal was consist of ferrite and pearlite(Fig. 5c). The pearlite was spheroidized with the grade of 2-3. Fig. 4d-f present the closed cracks observed. As can be seen, some cracks propagated along the chain-like distribution pores. Most of the non-metallic inclusions were grinded off during the sample preparation. Residual inclusions were also detected in the pores. The morphology of the non-metallic inclusions was different from the one observed in Fig 3a, which was probably due to the corrosion. Secondary cracks
Primary cracks Outer surface of weld metal a
b
c
Chain-like distribution pores
Chain-like distribution pores
d
e
f
Fig. 4 Microstructure of failed welded joint. (a) and (b) Microstructure observed in specimen A; (c) Microstructure of base metal observed in
specimen C; (d)-(f) Inner cracks observed in specimen C.
The primary cracks observed in welded joints (Fig. 1a) propagated along the thickness with the deepth of 45mm. The microstructure of specimen B is present in Fig. 5. Fig. 5a and b display the morphology observed in crack tip. Compared to the microstructure of base metal (Fig. 4c), the grains observed in this regions were much finer. It was deduced that this regions where crack arrested were probably fine-grained heat-affected zone (FG-HAZ). Intergranular cracks were observed in crack tip regions. Fig 5c presents the metallographic structure in fusion zone with coarse grain morphology. The microstructure of the weld metal was mainly consist of ferrite and bainite. Spheroidization morphology was also detected in bainite microstructure and the grade was the same as the base metal (range of 2-3). SEM images of specimen B are exhibited in Fig. 6. Fig. 6a-c and d-f display the morphology observed in Fig. 5b and c, respectively. As can be seen, some grain boundaries were completely decorated with cavities. In some regions the neighbouring cavities were coalesced, which developed into microcracks ultimately. In the regions far away from the primary cracks, the number and size of the cavities were much reduced. Since some non-metallic inclusions were grinded off during the sample preparation, the pores left probably acted as cracks initiation or cavities. In Fig. 6f the pores presented chain-like distribution. L. Falat[4] reported that the coalescence of cavities was an important factor cause the creep failure in heat resistant steel. Previous investigation[5] reported a reheater tube leakage after 185000h. Typical creep failure with abundant cavitation damage in the vicinity of the failure surface was detected. Heat affect zone Fusion line
b Cracks
c
a
Weld metal
c
b
Fig. 5 Microstructure of specimen B. (a) and (b) Crack tip with intergranular cracking morphology; (c) Cracks propagating in fusion zone. Pores
c
Pores
b 50μm
a
b
20μm
20μm
c
Pores
e
d
20μm
e
10μm
f
Coalescence 10μm
Fig. 6 Scanning electron microscopy images of specimen B. (a)-(c) Morphology observed in crack tip regions and (d)-(f) in weld metal.
3.3. Fracture analysis After metallographical inspection, specimen A was pulled open along the primary cracks for further study and fracture surfaces are shown in Fig. 7a. It was examined both macroscopically and with a high optical stereoscopic microscope. Visual examination of the fracture sufaces showed no sign of macroscopic plastic deformation. The fracture was featured with an adherent oxide film which was probably due to successive exposure of the opened crack to the hot steam. The oxide layer could not be removed by conservative methods, which impaired the microfracture examination. Fig. 7b and c present the SEM images of the fracture surfaces. Microfractography confirmed the brittle nature of the fracture surfaces. Many cracks were observed in the oxide film.
a
b
10μm
c Fig. 7 Fracture morphology of specimen A.
3.4. Mechanical analysis Three impact specimens were machined in the dimensions of 10×10×55mm from the failed pipe weld. Cracks were detected in the specimens indicated by arrows (Fig. 8a). The specimens ruptured along the cracks and the charpy toughness values were nearly zero at ambient temperaure. Fig. 8b depicts the SEM image of the impact fracture. High concentration of oxygen was detected. The oxide morphology observed in this paper was similar to the one presented in previous research[6]. Regions free from oxide film exhibit dimpled fracture morphology in Fig. 8c.
a
20μm b
20μm c
Fig. 8 Impact specimens and fracture. (a) Specimens with cracks; (b) Oxide and (c) Dimple observed in impact fracure.
Table 2 displays the tensile test data. The results revealed that the failed weld metal experienced an excellent ultimate strength. However, the average ductility value was 5.8%, which was much less than the specified values in standards. The material was probably embrittled during 140000 h operation in high temperature and pressure condition. With respect to the microstructure heterogeneity of the welded joint, hardness measurement was carried out in specimen C. The base metal and weld metal exhibited an average microhardness of 196HV0.2 and 292HV0.2, respectively. In accordance
with the technological requirements of this material, the upper limitation of base metal is 180HV and weld metal 280HV. Consequently, the failed pipe material suffered a slightly higher hardness. Table 2 Tensile test results of failed pipe weld. σb/MPa
δ5/%
Failure pipe material
627
5.8
DL 5007 standards
470-640
≥21
3.5. Finite element analysis (FEA) In fact, the real stress state in this failed pipe is thermo-mechanical stress combined with the welding residual stress. In order to facilitate the discussion of stress state, two analysis steps were established. The first step was established for the analysis of welding residual stresses, and the second was for the thermo-mechanical stresses in as-servive condition. The second step analysis was based on the first one. Although the real structural stress analysis was a 3-D procedure, it was often considered sufficient to represent a circumferential weld with an axi-symmetric FE model. The investigated pipe has six welds in all(Fig. 1b) and the welds near the elbow were detected cracks. Therefore two welds were modeled for the purpose of simplification. The FEA model is exhibited in Fig. 9a, which was consist of 1980 elements. For the first step, birth and dead element method was applied to simulated the filling process of weld metal. The preheat temperature of the failed pipe was 300℃. The interpass temperature should be controlled between 200~300℃. PWHT(700℃×5h) was proposed to alleviate the residual stresses and improving the mechanical properties of the welded pipe. For the second step, the mode was subjected to a combination load consisted of temperature (510℃) and pressure (17.4MPa). The nonlinear, Newton-Raphson method was employed in both computation. For all results, MegaPascal (MPa) unit was used to describe the Von Mises stress. Fig. 9b depicts the Von Mises stress distribution during welding process. It was derived that the maximum Von Mises stress (452 MPa) was developed in heat-affected zone and elbow regions. The residual stresses experienced a slightly drop after PWHT with the maximum Von Mises value of 402MPa. Residual stresses still concentrated in heat-affected zone and elbow regions. Fig. 9c exhibits the stress distribution in as-service condition. The magnitude and distribution of Von Mises stresses differed from the welding condition. Weld metal, heat-affected zone and elbow regions exhibited the stress concentration. A full-scale pipe mode was also developed as supplementary investigation techniques. In this step, only pressure and temperature were taken into consideration. The maximum Von Mises stress (109MPa) was developed in elbow regions(Fig. 10). The results revealed that the stress concentration still occurred in spite of the welding residual stresses.
Weld metal
Beld metal
a
b
c
Fig. 9 FEA results. (a) FEA model; (b) Welding residual stress distribution; (c) Thermo-mechanical stress distribution.
Fig. 10 FEA results. Thermo-mechanical stress distribution without welding residual stresses.
4. Discussion In this investigation, the steam pipe was detected cracks after 140000 h service. Intergranular cracks and microstructural spheroidization were observed in welded joint. A lot of intergranular cavities were detected near the primary cracks. Creep cracking is identified as the main root causing the failure. Intergranular creep usually occurs by either of the two fracture process[7,8]: (1) triple point cracking; (2) grain boundary cavitations. Shunsheng Zhou[9] investigated the mechanism of creep cracking in low-alloyed heat-resistant steels. The research results revealed that grain boundary cavitations were the most acceptance one in heat-resistant steel creep failure. Cracks developped as a result of cavity nucleation on grain boundaries, followed by cavity growth and linkage to form micro-and eventually macro-intergranular cracks. Fig. 11a displays the formation of intergranular cracks. Fig. 11b and c exhibit the intergranular cavities and cracks observed in the failed weld. The coalescence of creep cavities was an important factor cause the creep failure of heat resistant steel[4]. Hopkin [10] investigated the failure mechanism of 1%Cr-0.5%Mo steel at the temperature of 550℃. The results indicated that the intergranular cavities observed in the failed material gradually turned into intergranular cracks. Creep cavitation is an important degradation mechanism in steels, where tensile stresses are experienced for prolonged periods at elevated temperature (＞450℃)[11]. Nucleation of cavities generally occurs continuously over the whole creep life. There are a number of mechanisms by which cavities can grow during creep life, for example plastic deformation, diffusion controlled and constrained cavity growth, but the one which is operative is dependent on the prevailing temperature
and stress. The cavitation damage is promoted by a critical combination of residual stress, temperature and microstructure. Constrained weld geometries where the welding process can create severe triaxial stress fields are particularly susceptible to crack initiation. Lei Zhao et al[12] studied the creep rupture time of ASME P92. It revealed that the creep rupture time was associated with the applied stress. Coleman et al[13] investigated the reheat cracking in as-welded austenitic stainless steel components. The results indicated that tensile residual stresses in the weld relaxed during high temperature operation, resulted in the development of creep cavitation damage. The welding residual stresses probably released during the operation. However, stress concentration still occurred in spite of the welding residual stresses. The investigated pipe had six welds. Only the welds near the elbow exhibited the cracks. FEA results revealed that stresses concentrated in elbow and HAZ regions in welding and as-service conditions. Xirui Hou[14] investigated the safety technology of power plant boiler pressure pipeline. It reported that stress concentration would be occurred in elbows and tee-junctions. In this study transverse cracks were detected in weld metal. However, simulations indicated the highest stress concentration within the HAZ and elbow regions The discrepancy between experiment and simulation probably due to the influence of non-metallic inclusions which was not included in the simulations. The non-metallic inclusions could act like initiation sites for cavity nucleation and subsequent cracking. Stress
Stress
(a) Nucleation
(b) Growth
(c) Formation of transverse crack
b
Stress
a
20μm
Stress
(d) Formation of tortuous crack
(e) Connection of tortuous cracks
c
10μm
Fig. 11 The formation of intergranular cracks. (a) Schematic diagram; (b) Creep cavitation; (c) Intergranular cracks.
5. Conclusion In this paper, a welded pipe was detected failed after 140000h operation under internal pressure loading of 17.4MPa at 510℃. Efforts were made to analyze the root cause of the failure in both experiment and finite element analysis methods. Non-metallic inclusions with the degree of D3 were observed in the failed weld metal. Metallographic study revealed that the material of the failed pipe was spheroidized The main failure mechanism is intergranular creep cracking. Creep cavitation is an important degradation mechanism in steels, where tensile stresses are experienced for prolonged periods at elevated temperature. FEA results revealed the stress concentration in welded joint. Non-metallic inclusions could act like initiation sites for cavity nucleation and subsequent cracking. Both the number and distribution of the non-metallic inclusions should be strictly controled. And steels with higher creep-resistant are recommended for such conditions. For future studies the creep damage will be taken into account in simulations.
Acknowledgements This work is funded by the National Science Fundation of China, Grant No.51274162, National High Technology Research and Development Program of China (863 program), Grant No.2013AA031303 and Shaanxi Province Education Department Fund, Grant No.2012JC16. References [1] Warwick Payten, Damien Charman, Andrew Chapman et al. Life assessment and inspection of a hot reheat turbine bifurcation. Engineering Failure Analysis, 18(2011): 1445-1457. [2] D.J. Smith, N.S. Walker, S.T. Kimmins. Type IV creep cavity accumulation and failure in steel welds. Science Direct, 80(2003): 617-627. [3] Jongmin Lee, Shinho Han et al. Failure analysis of carbon steel pipes used for underground condensate pipeline in the power station. Engineering Failure Analysis, 34(2013): 300-307. [4]
L. Falat, A. Vyrostkova et al. Creep deformation and failure of E911/E911 and P92/P92 similar weld-joints. Engineering Failure Analysis, 16(2009): 2114-2120.
[5] Jorma Salonen, Peritti Auerkari, Olli Lehtinen et al. Experience on in-service damage in power plant components. Science Direct, 14(2007): 970-977. [6] G. Atxaga, A.M. Irisarri. Study of the failure of a duplex stainless steel value. Engineering Failure Analysis, 16(2009): 1412-1419. [7] Stedfeld RL, Desterani JD, Dieterich DA, editors. Fractography. Metal Handbook, 1987, 12(9): 14-41. [8] N. Ejaz, I.N. Qureshi, S.A. Rizvi. Creep failure of low pressure turbine blade of an aircraft engine. Engineering Failure Analysis, 18(2011): 1407-1414. [9] Shunsheng Zhou. Low Alloy Heat-resistant steel, 1796. [10] L.M.T.Hopkin. The Iron and Steel Institute, 1964, 202(11): 929. [11] P.J. Bouchard, P.J. Withers, S.A. McDonald et al. Quantification of creep cavitation damage around a crack in a stainless steel pressure vessel. Science Direct, 52(2004): 23-34. [12] Lei Zhao, Hongyang Jing et al. Numerical investigation of factors affecting creep damage accumulation in ASME P92 steel welded joint. Material and Design, 34(2012): 566-575. [13] Coleman MC, Miller DA, Stevens RA. Reheat cracking and stragegies to assure integrity of Type 316 weld components. Proceedings of the International Conference on Integrity of High Temperature Welds, 1998: 79-169. [14] Xirui Hou. Safety technology of power plant boiler pressure vessel pressure pipeline. Electic power press, 2005: 107-109.
Cracks
b
a
Fig. 1 General view of the cracks analyzed. (a) Cracks in welded joint; (b) Assemble diagram of the investigated steam pipe.
Cracks
Tensile specimen
Specimen C Specimen A
(SEM, microstructure,
(SEM, microstructure)
Charpy impact specimen
microhardness)
Weld metal
Specimen B a
Outer surface
inner surface
b
(SEM, microstructure)
c
Fig. 2 Failed circumferential welded joint cut for investigation. (a) Circumferential welded joint; (b) and (c) Sampling schematic.
Cracks Residual inclusions
Chain-like distribution inclusions a
b
20μm
c
500μm
Pores
Fig. 3 The morphology of non-metallic inclusions. (a) Observed by optical microscopy; (b) and (c) Observed by SEM.
Secondary cracks
Primary cracks Outer surface of weld metal b
a
c
Chain-like distribution pores
Chain-like distribution pores
d
e
f
Fig. 4 Microstructure of failed welded joint. (a) and (b) Microstructure observed in specimen A; (c) Microstructure of base metal observed in specimen C; (d)-(f) Inner cracks observed in specimen C.
Heat affect zone Fusion line
b Cracks
a
c
b
c
Weld metal
Fig. 5 Microstructure of specimen B. (a) and (b) Crack tip with intergranular cracking morphology; (c) Cracks propagating in fusion zone.
Pores
c
Pores
b 50μm
a
b
20μm
20μm
c
Pores
e
20μm
d
e
10μm
f
Coalescence 10μm
Fig. 6 Scanning electron microscopy images of specimen B. (a)-(c) Morphology observed in crack tip regions and (d)-(f) in weld metal.
a
b
10μm
c Fig. 7 Fracture morphology of specimen A.
a
20μm b
20μm c
Fig. 8 Impact specimens and fracture. (a) Specimens with cracks; (b) Oxide and (c) Dimple observed in impact fracure.
Weld metal
Beld metal
a
c
b
Fig. 9 FEA results. (a) FEA model; (b) Welding residual stress distribution; (c) Thermo-mechanical stress distribution.
Fig. 10 FEA results. Thermo-mechanical stress distribution without welding residual stresses.
Stress
Stress
(a) Nucleation
(b) Growth
(c) Formation of transverse crack
b
Stress
a
20μm
Stress
(d) Formation of tortuous crack
(e) Connection of tortuous cracks
c
10μm
Fig. 11 The formation of intergranular cracks. (a) Schematic diagram; (b) Creep cavitation; (c) Intergranular cracks.
Highlights 1 A failure of ex-service steam pipe weld from power generation plant has been analysed. 2 Both experimental and computational methods were used to study the failure. 3 The material contained a high level of non-metallic inclusions and was spheroidized. 4 The main failure mechanism is intergranular creep cracking.
S1350-6307(14)00166-6 http://dx.doi.org/10.1016/j.engfailanal.2014.05.019 EFA 2331
To appear in:
Engineering Failure Analysis
Received Date: Revised Date: Accepted Date:
3 April 2014 14 May 2014 20 May 2014
Please cite this article as: Chu, Q., Zhang, M., Li, J., Chen, Y., Luo, H., Wang, Q., Failure analysis of a steam pipe weld used in power generation plant, Engineering Failure Analysis (2014), doi: http://dx.doi.org/10.1016/ j.engfailanal.2014.05.019
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Failure analysis of a steam pipe weld used in power generation plant Qiaoling Chu, Min Zhang, Jihong Li, Yinni Chen, Hailong Luo, Qiang Wang College of Material Science and Engineering, Xi’an University of Technology, Xi’an, 710048
ABSTRACT In this paper a failure of steam pipe weld after 140000 hours of service in power generation plant has been analysed. Efforts were made to analyze the cause of failure in both experiment and finite element analysis (FEA) methods. Intergranular cavities and cracks were detected around the primary cracks. Study shows that the main failure mechanism is creep cracking. Cracks developped in this failed pipe as a result of cavity nucleation on grain boundaries, followed by cavity growth and linkage to form micro-and eventually macro-intergranular cracks. This paper brings out the details of investigation and suggests remedial measures to improve performance of this welded pipe under high temperature and pressure condition. Keywords: Creep cracking; Cavity ; Failure analysis; Welded joint 1. Introduction Components in power generating stations usually operate at high temperatures for long times. A large proportion of pressure components are manufactured from Cr-Mo-V steels which are selected for their high temperature creep resistance at optimum cost. When structural components are joined together by fusion welding, the thermal cycle introduces changes in the microstructure[1]. Welded pressure components are prone to creep deformation and fracture at high temperature, and the majority of the creep failures of pressurised components are associated with the welds[2]. During the maintenance halt of a power generation plant, a steam pipe connected with tee-junction structures was detected transverse cracks in welded joint (Fig. 1a). This failed pipe is manufactured from 1%Cr-0.3%Mo-0.2%V steel. Fig 1b displays the assemble diagram of the investigated steam pipe. The failed pipe is combined by shielded metal arc welding (SMAW) with matching electrodes (ASW E5515). The subsequent post-weld heat treatment(PWHT) (700℃×5h) is employed to alleviate the welding residual stresses. According to the record of operation for the steam pipe unit, the failed pipe has operated at around 140000 h. The pipe has an out-diameter of 450mm and a wall thickness of 50mm. The conveying medium is steam with average pressure of 17.4MPa and temperature of 510℃.
Cracks
a
b Fig. 1 General view of the cracks analyzed. (a) Cracks in welded joint; (b) Assemble diagram of the investigated steam pipe.
The aim of the present investigation is to identify the root cause of the failure and propose corrective measures. Finite
element simulations of the failed pipe were used as supplementary investigation techniques in order to illustrate the possible phenomena causing failure. 2. Experimental set-up The circumferential welded joint in Fig. 1a was cut for investigation (Fig. 2a). Chemical analysis by inductively coupled plasma (ICP) and carbon sulfur analyzer was applied to identify the matrix material of this failed pipe. The ex-service welded joint containing cracks was sectioned for detailed metallographic studies (Fig. 2b). Metallographic sections were prepared to examine the microstructural characteristic and difference between adjacent to and apart from the fracture regions. The metallographic samples were prepared by grinding devices and successive wet grinding emery papers up to #2000 grid and fine polishing in diamond paste. 4% nital was applied as the etch to reveal the microstructure. Cleaning was carrying out using ethanol followed by hot air stream drying. The fracture surfaces were observed in scanning electron microscope (SEM) equipped with an X ray energy dispersion spectrometer (EDX). Both the tensile and Charpy impact specimens were machined along the weld direction (Fig. 2c). The tests were performed at ambient temperature. Vicker microhardness test was also performed applying 200g of force in the welded joint. Finite element analysis procedures for predicting the magnitude and distribution of residual stresses were also developed in welding and as-service process, respectively. Cracks
Tensile specimen
Specimen C Specimen A
(SEM, microstructure,
(SEM, microstructure)
Charpy impact specimen
microhardness)
Weld metal
Specimen B Outer surface
a
inner surface
b
(SEM, microstructure)
c
Fig. 2 Failed circumferential welded joint cut for investigation. (a) Circumferential welded joint; (b) and (c) Sampling schematic.
3. Failure analysis 3.1. Chemical analysis The specified elemental composition of 1%Cr-0.3%Mo-0.2%V and measured composition of failed pipe are present in Table 1. The results indicated that the failed pipe material 1%Cr-0.3%Mo-0.2%V is consistent with the DL 5007 standards. Table 1 Chemical compositions of the failed pipe material (wt%). C
Mn
Si
Cr
Mo
V
Failed pipe material
0.09
0.52
0.34
1.01
0.28
0.30
DL 5007 standards
0.08-0.15
0.40-0.70
0.17-0.37
0.90-1.20
0.25-0.35
0.15-0.30
3.2. Microstructural analysis Non-metallic inclusion test was performed before metallographic detection in specimen C. The sample was prepared with fine polished in diamond paste. Fig. 3a displays the distribution of non-metallic inclusions. In accordance with the ISO 4967 standards, the degree of non-metallic inclusions conforms to D3, which is closed to the maximum limit and could be
considered high. SEM images of the non-metallic inclusions are shown in Fig. 3b and c. The inclusions presented circular morphology and chain distribution. EDX results indicated that the inclusions are rich in O(~24.0%) and Cr(~3.8%) elements. Some large pores are observed in Fig. 3c. It was obvious that cracks initiated from the pores. Residual inclusions were also detected in some pores. It was deduced that some large inclusions exposed were grinded off during the sample preparation. Previous research [3] investigated the failure of carbon steel pipes and results proved that the non-metallic inclusions could promote the initiation and propagation of cracks.
Cracks Residual inclusions
Chain-like distribution inclusions a
b
20μm
c
Pores
500μm
Fig. 3 The morphology of non-metallic inclusions. (a) Observed by optical microscopy; (b) and (c) Observed by SEM.
Metallographic inspection was performed in both cross and longitudinal section of the weld line. Specimen A and B were contained the primary cracks. Fig. 4a depicts the crack morphology in specimen A. As can been seen, some secondary cracks initiated from the outer surface of weld metal. Fig. 4b presents intergranular cracks observed near the primary cracks. Regions adjacent to the primary cracks characterized with dark scales. EDX test revealed that the dark regions were rich in O element. The cracks were most probably oxidized in high temperature operation. Fig. 4c-f display the microstructure of specimen C. The microstructure of base metal was consist of ferrite and pearlite(Fig. 5c). The pearlite was spheroidized with the grade of 2-3. Fig. 4d-f present the closed cracks observed. As can be seen, some cracks propagated along the chain-like distribution pores. Most of the non-metallic inclusions were grinded off during the sample preparation. Residual inclusions were also detected in the pores. The morphology of the non-metallic inclusions was different from the one observed in Fig 3a, which was probably due to the corrosion. Secondary cracks
Primary cracks Outer surface of weld metal a
b
c
Chain-like distribution pores
Chain-like distribution pores
d
e
f
Fig. 4 Microstructure of failed welded joint. (a) and (b) Microstructure observed in specimen A; (c) Microstructure of base metal observed in
specimen C; (d)-(f) Inner cracks observed in specimen C.
The primary cracks observed in welded joints (Fig. 1a) propagated along the thickness with the deepth of 45mm. The microstructure of specimen B is present in Fig. 5. Fig. 5a and b display the morphology observed in crack tip. Compared to the microstructure of base metal (Fig. 4c), the grains observed in this regions were much finer. It was deduced that this regions where crack arrested were probably fine-grained heat-affected zone (FG-HAZ). Intergranular cracks were observed in crack tip regions. Fig 5c presents the metallographic structure in fusion zone with coarse grain morphology. The microstructure of the weld metal was mainly consist of ferrite and bainite. Spheroidization morphology was also detected in bainite microstructure and the grade was the same as the base metal (range of 2-3). SEM images of specimen B are exhibited in Fig. 6. Fig. 6a-c and d-f display the morphology observed in Fig. 5b and c, respectively. As can be seen, some grain boundaries were completely decorated with cavities. In some regions the neighbouring cavities were coalesced, which developed into microcracks ultimately. In the regions far away from the primary cracks, the number and size of the cavities were much reduced. Since some non-metallic inclusions were grinded off during the sample preparation, the pores left probably acted as cracks initiation or cavities. In Fig. 6f the pores presented chain-like distribution. L. Falat[4] reported that the coalescence of cavities was an important factor cause the creep failure in heat resistant steel. Previous investigation[5] reported a reheater tube leakage after 185000h. Typical creep failure with abundant cavitation damage in the vicinity of the failure surface was detected. Heat affect zone Fusion line
b Cracks
c
a
Weld metal
c
b
Fig. 5 Microstructure of specimen B. (a) and (b) Crack tip with intergranular cracking morphology; (c) Cracks propagating in fusion zone. Pores
c
Pores
b 50μm
a
b
20μm
20μm
c
Pores
e
d
20μm
e
10μm
f
Coalescence 10μm
Fig. 6 Scanning electron microscopy images of specimen B. (a)-(c) Morphology observed in crack tip regions and (d)-(f) in weld metal.
3.3. Fracture analysis After metallographical inspection, specimen A was pulled open along the primary cracks for further study and fracture surfaces are shown in Fig. 7a. It was examined both macroscopically and with a high optical stereoscopic microscope. Visual examination of the fracture sufaces showed no sign of macroscopic plastic deformation. The fracture was featured with an adherent oxide film which was probably due to successive exposure of the opened crack to the hot steam. The oxide layer could not be removed by conservative methods, which impaired the microfracture examination. Fig. 7b and c present the SEM images of the fracture surfaces. Microfractography confirmed the brittle nature of the fracture surfaces. Many cracks were observed in the oxide film.
a
b
10μm
c Fig. 7 Fracture morphology of specimen A.
3.4. Mechanical analysis Three impact specimens were machined in the dimensions of 10×10×55mm from the failed pipe weld. Cracks were detected in the specimens indicated by arrows (Fig. 8a). The specimens ruptured along the cracks and the charpy toughness values were nearly zero at ambient temperaure. Fig. 8b depicts the SEM image of the impact fracture. High concentration of oxygen was detected. The oxide morphology observed in this paper was similar to the one presented in previous research[6]. Regions free from oxide film exhibit dimpled fracture morphology in Fig. 8c.
a
20μm b
20μm c
Fig. 8 Impact specimens and fracture. (a) Specimens with cracks; (b) Oxide and (c) Dimple observed in impact fracure.
Table 2 displays the tensile test data. The results revealed that the failed weld metal experienced an excellent ultimate strength. However, the average ductility value was 5.8%, which was much less than the specified values in standards. The material was probably embrittled during 140000 h operation in high temperature and pressure condition. With respect to the microstructure heterogeneity of the welded joint, hardness measurement was carried out in specimen C. The base metal and weld metal exhibited an average microhardness of 196HV0.2 and 292HV0.2, respectively. In accordance
with the technological requirements of this material, the upper limitation of base metal is 180HV and weld metal 280HV. Consequently, the failed pipe material suffered a slightly higher hardness. Table 2 Tensile test results of failed pipe weld. σb/MPa
δ5/%
Failure pipe material
627
5.8
DL 5007 standards
470-640
≥21
3.5. Finite element analysis (FEA) In fact, the real stress state in this failed pipe is thermo-mechanical stress combined with the welding residual stress. In order to facilitate the discussion of stress state, two analysis steps were established. The first step was established for the analysis of welding residual stresses, and the second was for the thermo-mechanical stresses in as-servive condition. The second step analysis was based on the first one. Although the real structural stress analysis was a 3-D procedure, it was often considered sufficient to represent a circumferential weld with an axi-symmetric FE model. The investigated pipe has six welds in all(Fig. 1b) and the welds near the elbow were detected cracks. Therefore two welds were modeled for the purpose of simplification. The FEA model is exhibited in Fig. 9a, which was consist of 1980 elements. For the first step, birth and dead element method was applied to simulated the filling process of weld metal. The preheat temperature of the failed pipe was 300℃. The interpass temperature should be controlled between 200~300℃. PWHT(700℃×5h) was proposed to alleviate the residual stresses and improving the mechanical properties of the welded pipe. For the second step, the mode was subjected to a combination load consisted of temperature (510℃) and pressure (17.4MPa). The nonlinear, Newton-Raphson method was employed in both computation. For all results, MegaPascal (MPa) unit was used to describe the Von Mises stress. Fig. 9b depicts the Von Mises stress distribution during welding process. It was derived that the maximum Von Mises stress (452 MPa) was developed in heat-affected zone and elbow regions. The residual stresses experienced a slightly drop after PWHT with the maximum Von Mises value of 402MPa. Residual stresses still concentrated in heat-affected zone and elbow regions. Fig. 9c exhibits the stress distribution in as-service condition. The magnitude and distribution of Von Mises stresses differed from the welding condition. Weld metal, heat-affected zone and elbow regions exhibited the stress concentration. A full-scale pipe mode was also developed as supplementary investigation techniques. In this step, only pressure and temperature were taken into consideration. The maximum Von Mises stress (109MPa) was developed in elbow regions(Fig. 10). The results revealed that the stress concentration still occurred in spite of the welding residual stresses.
Weld metal
Beld metal
a
b
c
Fig. 9 FEA results. (a) FEA model; (b) Welding residual stress distribution; (c) Thermo-mechanical stress distribution.
Fig. 10 FEA results. Thermo-mechanical stress distribution without welding residual stresses.
4. Discussion In this investigation, the steam pipe was detected cracks after 140000 h service. Intergranular cracks and microstructural spheroidization were observed in welded joint. A lot of intergranular cavities were detected near the primary cracks. Creep cracking is identified as the main root causing the failure. Intergranular creep usually occurs by either of the two fracture process[7,8]: (1) triple point cracking; (2) grain boundary cavitations. Shunsheng Zhou[9] investigated the mechanism of creep cracking in low-alloyed heat-resistant steels. The research results revealed that grain boundary cavitations were the most acceptance one in heat-resistant steel creep failure. Cracks developped as a result of cavity nucleation on grain boundaries, followed by cavity growth and linkage to form micro-and eventually macro-intergranular cracks. Fig. 11a displays the formation of intergranular cracks. Fig. 11b and c exhibit the intergranular cavities and cracks observed in the failed weld. The coalescence of creep cavities was an important factor cause the creep failure of heat resistant steel[4]. Hopkin [10] investigated the failure mechanism of 1%Cr-0.5%Mo steel at the temperature of 550℃. The results indicated that the intergranular cavities observed in the failed material gradually turned into intergranular cracks. Creep cavitation is an important degradation mechanism in steels, where tensile stresses are experienced for prolonged periods at elevated temperature (＞450℃)[11]. Nucleation of cavities generally occurs continuously over the whole creep life. There are a number of mechanisms by which cavities can grow during creep life, for example plastic deformation, diffusion controlled and constrained cavity growth, but the one which is operative is dependent on the prevailing temperature
and stress. The cavitation damage is promoted by a critical combination of residual stress, temperature and microstructure. Constrained weld geometries where the welding process can create severe triaxial stress fields are particularly susceptible to crack initiation. Lei Zhao et al[12] studied the creep rupture time of ASME P92. It revealed that the creep rupture time was associated with the applied stress. Coleman et al[13] investigated the reheat cracking in as-welded austenitic stainless steel components. The results indicated that tensile residual stresses in the weld relaxed during high temperature operation, resulted in the development of creep cavitation damage. The welding residual stresses probably released during the operation. However, stress concentration still occurred in spite of the welding residual stresses. The investigated pipe had six welds. Only the welds near the elbow exhibited the cracks. FEA results revealed that stresses concentrated in elbow and HAZ regions in welding and as-service conditions. Xirui Hou[14] investigated the safety technology of power plant boiler pressure pipeline. It reported that stress concentration would be occurred in elbows and tee-junctions. In this study transverse cracks were detected in weld metal. However, simulations indicated the highest stress concentration within the HAZ and elbow regions The discrepancy between experiment and simulation probably due to the influence of non-metallic inclusions which was not included in the simulations. The non-metallic inclusions could act like initiation sites for cavity nucleation and subsequent cracking. Stress
Stress
(a) Nucleation
(b) Growth
(c) Formation of transverse crack
b
Stress
a
20μm
Stress
(d) Formation of tortuous crack
(e) Connection of tortuous cracks
c
10μm
Fig. 11 The formation of intergranular cracks. (a) Schematic diagram; (b) Creep cavitation; (c) Intergranular cracks.
5. Conclusion In this paper, a welded pipe was detected failed after 140000h operation under internal pressure loading of 17.4MPa at 510℃. Efforts were made to analyze the root cause of the failure in both experiment and finite element analysis methods. Non-metallic inclusions with the degree of D3 were observed in the failed weld metal. Metallographic study revealed that the material of the failed pipe was spheroidized The main failure mechanism is intergranular creep cracking. Creep cavitation is an important degradation mechanism in steels, where tensile stresses are experienced for prolonged periods at elevated temperature. FEA results revealed the stress concentration in welded joint. Non-metallic inclusions could act like initiation sites for cavity nucleation and subsequent cracking. Both the number and distribution of the non-metallic inclusions should be strictly controled. And steels with higher creep-resistant are recommended for such conditions. For future studies the creep damage will be taken into account in simulations.
Acknowledgements This work is funded by the National Science Fundation of China, Grant No.51274162, National High Technology Research and Development Program of China (863 program), Grant No.2013AA031303 and Shaanxi Province Education Department Fund, Grant No.2012JC16. References [1] Warwick Payten, Damien Charman, Andrew Chapman et al. Life assessment and inspection of a hot reheat turbine bifurcation. Engineering Failure Analysis, 18(2011): 1445-1457. [2] D.J. Smith, N.S. Walker, S.T. Kimmins. Type IV creep cavity accumulation and failure in steel welds. Science Direct, 80(2003): 617-627. [3] Jongmin Lee, Shinho Han et al. Failure analysis of carbon steel pipes used for underground condensate pipeline in the power station. Engineering Failure Analysis, 34(2013): 300-307. [4]
L. Falat, A. Vyrostkova et al. Creep deformation and failure of E911/E911 and P92/P92 similar weld-joints. Engineering Failure Analysis, 16(2009): 2114-2120.
[5] Jorma Salonen, Peritti Auerkari, Olli Lehtinen et al. Experience on in-service damage in power plant components. Science Direct, 14(2007): 970-977. [6] G. Atxaga, A.M. Irisarri. Study of the failure of a duplex stainless steel value. Engineering Failure Analysis, 16(2009): 1412-1419. [7] Stedfeld RL, Desterani JD, Dieterich DA, editors. Fractography. Metal Handbook, 1987, 12(9): 14-41. [8] N. Ejaz, I.N. Qureshi, S.A. Rizvi. Creep failure of low pressure turbine blade of an aircraft engine. Engineering Failure Analysis, 18(2011): 1407-1414. [9] Shunsheng Zhou. Low Alloy Heat-resistant steel, 1796. [10] L.M.T.Hopkin. The Iron and Steel Institute, 1964, 202(11): 929. [11] P.J. Bouchard, P.J. Withers, S.A. McDonald et al. Quantification of creep cavitation damage around a crack in a stainless steel pressure vessel. Science Direct, 52(2004): 23-34. [12] Lei Zhao, Hongyang Jing et al. Numerical investigation of factors affecting creep damage accumulation in ASME P92 steel welded joint. Material and Design, 34(2012): 566-575. [13] Coleman MC, Miller DA, Stevens RA. Reheat cracking and stragegies to assure integrity of Type 316 weld components. Proceedings of the International Conference on Integrity of High Temperature Welds, 1998: 79-169. [14] Xirui Hou. Safety technology of power plant boiler pressure vessel pressure pipeline. Electic power press, 2005: 107-109.
Cracks
b
a
Fig. 1 General view of the cracks analyzed. (a) Cracks in welded joint; (b) Assemble diagram of the investigated steam pipe.
Cracks
Tensile specimen
Specimen C Specimen A
(SEM, microstructure,
(SEM, microstructure)
Charpy impact specimen
microhardness)
Weld metal
Specimen B a
Outer surface
inner surface
b
(SEM, microstructure)
c
Fig. 2 Failed circumferential welded joint cut for investigation. (a) Circumferential welded joint; (b) and (c) Sampling schematic.
Cracks Residual inclusions
Chain-like distribution inclusions a
b
20μm
c
500μm
Pores
Fig. 3 The morphology of non-metallic inclusions. (a) Observed by optical microscopy; (b) and (c) Observed by SEM.
Secondary cracks
Primary cracks Outer surface of weld metal b
a
c
Chain-like distribution pores
Chain-like distribution pores
d
e
f
Fig. 4 Microstructure of failed welded joint. (a) and (b) Microstructure observed in specimen A; (c) Microstructure of base metal observed in specimen C; (d)-(f) Inner cracks observed in specimen C.
Heat affect zone Fusion line
b Cracks
a
c
b
c
Weld metal
Fig. 5 Microstructure of specimen B. (a) and (b) Crack tip with intergranular cracking morphology; (c) Cracks propagating in fusion zone.
Pores
c
Pores
b 50μm
a
b
20μm
20μm
c
Pores
e
20μm
d
e
10μm
f
Coalescence 10μm
Fig. 6 Scanning electron microscopy images of specimen B. (a)-(c) Morphology observed in crack tip regions and (d)-(f) in weld metal.
a
b
10μm
c Fig. 7 Fracture morphology of specimen A.
a
20μm b
20μm c
Fig. 8 Impact specimens and fracture. (a) Specimens with cracks; (b) Oxide and (c) Dimple observed in impact fracure.
Weld metal
Beld metal
a
c
b
Fig. 9 FEA results. (a) FEA model; (b) Welding residual stress distribution; (c) Thermo-mechanical stress distribution.
Fig. 10 FEA results. Thermo-mechanical stress distribution without welding residual stresses.
Stress
Stress
(a) Nucleation
(b) Growth
(c) Formation of transverse crack
b
Stress
a
20μm
Stress
(d) Formation of tortuous crack
(e) Connection of tortuous cracks
c
10μm
Fig. 11 The formation of intergranular cracks. (a) Schematic diagram; (b) Creep cavitation; (c) Intergranular cracks.
Highlights 1 A failure of ex-service steam pipe weld from power generation plant has been analysed. 2 Both experimental and computational methods were used to study the failure. 3 The material contained a high level of non-metallic inclusions and was spheroidized. 4 The main failure mechanism is intergranular creep cracking.