Engineering Failure Analysis 47(2015)162-177 Contents lists available at Science Direct VGINEFNING Engineering Failure Analysis ELSEVIER journalhomepagewww.elsevier.com/locate/engfailanal Failure analysis on circulating water pump of duplex stainless steel in 1000 MW ultra-supercritical thermal power unit Yue-Yue Ma, Shi Yan, Zhen-Guo Yang Guo-Shui Qi, Xin-You He b Zhejiang Zhengneng Jiaxing Electric Power Co, LTD, Jiaxing, Zhejiang Province 314201, PR China ARTICLE INFO A BSTRACT With a large number of properties suc mechanical properties and excellent corrosion resistance, 2205 duplex stainle SS)has been extensively used in many eived in revised form 29 September 201 industries for the last decades. However welding procedures will induce embrit Available online 23 October 2014 tlement of the weld joint, seriously decreasing the safety reliability of the weld joint. In this study, lots of unexpected fractures occurred on 2205 DSS which was used as the materia for making circulating water pump(CWP)in an ultra-supercritical thermal power plant of stainless steel (Dss) China for the first time. By means of diverse characterization methods, comprehensive Fracture investigation was carried out on the failed CWP. Analysis results reveal that many lack Circulating water pump penetrations(LOPs)in the weld joint and unbalance ferrite/austenite ratio induced by r welding procedures should be responsible for the fracture of the CWP. And effective countermeasures and suggestions were also proposed. So the analysis results have instructive significance for the fracture prevention of the CwP, even for ensuring safety operation of other equipment under similar seawater environment. e 2014 Elsevier Ltd. All rights reserved. 1 Introduction As one of the largest thermal power plants in the eastern part of China, Jiaxing power plant phase lll has two 1000 MW ltra-supercritical thermal power generating units, which are named by 7# and 8#f, respectively. these two units were put into commercial operation on June 23, 2011 and October 20, 2011, respectively, contributing a lot to the development local economic development. irculating water system is an important facility in a thermal power generating unit, which is used to pump seawater and then to feed seawater into a condenser to cool exhaust steam by heat exchange. Hereby, each of the two units of Jiaxing power plant phase lll was equipped with three CWPs, named 7A, 7B, 7C and 8A, 8B, 8C, respectively, all of which have same structure, designed and manufactured by Hitachi Pump Manufacture(Wuxi)Co., Ltd Fig. 1(a)and Fig. 1(b) show the external appearance and the structure of 8A CWP respectively Since the surrounding of the CWP's shell is natural seawater which usually contains high contents of salts, chloride ions and sediment particles, it requires a high performances such as excellent corrosion resistance and good mechanical proper ties for the material used in the cwps so dss is one of the best choices in this condition. In China. 2205 DSS was the first time used as the material of CWPs in thermal power plants. Corresponding author. Tel: +86 21 65642523: fax: +86 21 65103056. E-mail address: zgyangefudanedu cn(Z -G Yang). http://dxdoiorg/10.1016/jengfailanal2014.09.014 1350-6307/0 2014 Elsevier Ltd. All rights reserved
Failure analysis on circulating water pump of duplex stainless steel in 1000 MW ultra-supercritical thermal power unit Yue-Yue Ma a , Shi Yan a , Zhen-Guo Yang a,⇑ , Guo-Shui Qi b , Xin-You He b aDepartment of Materials Science, Fudan University, Shanghai 200433, PR China b Zhejiang Zhengneng Jiaxing Electric Power Co., LTD, Jiaxing, Zhejiang Province 314201, PR China article info Article history: Received 4 April 2013 Received in revised form 29 September 2014 Accepted 30 September 2014 Available online 23 October 2014 Keywords: Duplex stainless steel (DSS) Fracture Circulating water pump Failure analysis abstract With a large number of properties such as good mechanical properties and excellent corrosion resistance, 2205 duplex stainless steel (DSS) has been extensively used in many industries for the last decades. However, improper welding procedures will induce embrittlement of the weld joint, seriously decreasing the safety reliability of the weld joint. In this study, lots of unexpected fractures occurred on 2205 DSS which was used as the material for making circulating water pump (CWP) in an ultra-supercritical thermal power plant of China for the first time. By means of diverse characterization methods, comprehensive investigation was carried out on the failed CWP. Analysis results reveal that many lack of penetrations (LOPs) in the weld joint and unbalance ferrite/austenite ratio induced by improper welding procedures should be responsible for the fracture of the CWP. And effective countermeasures and suggestions were also proposed. So the analysis results have instructive significance for the fracture prevention of the CWP, even for ensuring safety operation of other equipment under similar seawater environment. 2014 Elsevier Ltd. All rights reserved. 1. Introduction As one of the largest thermal power plants in the eastern part of China, Jiaxing power plant phase III has two 1000 MW ultra-supercritical thermal power generating units, which are named by 7# and 8#, respectively. These two units were put into commercial operation on June 23, 2011 and October 20, 2011, respectively, contributing a lot to the development of local economic development. Circulating water system is an important facility in a thermal power generating unit, which is used to pump seawater and then to feed seawater into a condenser to cool exhaust steam by heat exchange. Hereby, each of the two units of Jiaxing power plant phase III was equipped with three CWPs, named 7A, 7B, 7C and 8A, 8B, 8C, respectively, all of which have same structure, designed and manufactured by Hitachi Pump Manufacture (Wuxi) Co., Ltd. Fig. 1(a) and Fig. 1(b) show the external appearance and the structure of 8A CWP respectively. Since the surrounding of the CWP’s shell is natural seawater which usually contains high contents of salts, chloride ions and sediment particles, it requires a high performances such as excellent corrosion resistance and good mechanical properties for the material used in the CWPs, so DSS is one of the best choices in this condition. In China, 2205 DSS was the first time used as the material of CWPs in thermal power plants. http://dx.doi.org/10.1016/j.engfailanal.2014.09.014 1350-6307/ 2014 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. Tel.: +86 21 65642523; fax: +86 21 65103056. E-mail address: zgyang@fudan.edu.cn (Z.-G. Yang). Engineering Failure Analysis 47 (2015) 162–177 Contents lists available at ScienceDirect Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Y-Y Ma et aL/ Engineering Failure Analysis 47(2015)162-177 163 However, in this event, after only ten months'operation, substantially less than the design lifetime of 30 years, a number of severe fractures occurred on the CwP of thermal power unit 8, as shown in Fig. 2, causing substantial economic losses as well as potential safety problems. By means of visual inspection, it was easy to find that most of the fractures occurred on the weld joints rather than the base materials, just as Fig. 2(a)-(c)show. Material property, manufacturing technology, equip- nt operation, service environment, routine maintenance or other factors, which were the main causes for inducing these abnormal fractures, were urgently investigated. Consequently, a comprehensive failure analysis including a variety of characterization methods was conducted to identify the root cause based on our previous failure analysis experiences 1-9]. 田 1. Appearance and structure of CWP: (a)appearance of SA CWP and(b)structure of the CWP
However, in this event, after only ten months’ operation, substantially less than the design lifetime of 30 years, a number of severe fractures occurred on the CWP of thermal power unit 8, as shown in Fig. 2, causing substantial economic losses as well as potential safety problems. By means of visual inspection, it was easy to find that most of the fractures occurred on the weld joints rather than the base materials, just as Fig. 2(a)–(c) show. Material property, manufacturing technology, equipment operation, service environment, routine maintenance or other factors, which were the main causes for inducing these abnormal fractures, were urgently investigated. Consequently, a comprehensive failure analysis including a variety of characterization methods was conducted to identify the root cause based on our previous failure analysis experiences [1–9]. Fig. 1. Appearance and structure of CWP: (a) appearance of 8A CWP and (b) structure of the CWP. Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177 163
164 Y-Y Ma et aL/Engineering Failure Analysis 47(2015)162-177 The mechanism of these fractures on the dss used in CWP was carefully discussed. Finally, effective countermeasures and suggestions were proposed as well Actually, many researchers have focused on the properties of DSS, such as its fatigue behavior, welding property, corro- sion resistance [10-17. but such an engineering practical study of mechanical degradation on DSs applied in CWP of 1000 MW ultra-supercritical thermal power unit has been rarely reported. What's more, the phenomenon that large num- bers of fractures occurred on the flanges of the Cwp is even less reported. Therefore, the analyses and results given in this study have not only important engineering values in failure prevention of the CWPs used under seawater environment, but also practical significance in ensuring safety operation of other equipment under similar condition. 2. Experimental The 8A CWP is located in the Number 3 CWP house of Jiaxing power plant phase Ill, with a vertical structure and th ngth of the underground part is 17.1 m As shown in Fig. 1(b), 8A CWP is mainly composed with two parts, i.e. the pump shell that weighs 42 tons and the shaft that weights 26 tons. The pump shell mainly consists of an inlet bellmouth, a bell pipe four connecting pipes and a bent outlet from bottom to top. Each connecting pipe is constituted of two round flanges and a cylindrical body by means of welding. Hereby the two flanges are located on both ends of the connecting pipe, and each flange is made up with four same flange arcs with a thickness of 46 mm by means of welding technology The cylin- drical body was manufactured by a process of rolling and welding, with an outer diameter of 2200 mm and a thickness of 14 In this event, more than twenty severe cracks were discovered on the surface of 8A CWP. Among the damaged pipes, the everest one is the second connecting pipe counted from bottom to top, whose macroscopic appearance is showed in Fig. 3(a). Two target pairs of cracking samples analyzed in this study were both from the flange of this damaged pipe. One pairs crack occurred on the weld joint of the flange, noted by cracking a and the other occurred on base material of the flange, noted by cracking B. The location of the two samples and the appearances are displayed in Fig 3(a-c) 2. 2. Characterization methods In order to figure out the failure causes and mechanisms, a variety of characterization methods were successively con- ducted Oxygen nitrogen hydrogen(ONH)analyzer, carbon sulfur analyzer (CSA), and inductively coupled plasma atomic emission spectroscopy(ICP-AES) were used to inspect their chemical compositions. Optical microscopy(OM) was utilized to observe their metallographic structures and the austenite/ferrite ratio of the butt weld was obtained by electron back ter diffraction(EBSD)and dyeing calculation method under metalloscope, respectively. The impact toughness of the d used in the CWP was also measured by Charpy impact test and the constituents of the seawater were detected by ion chro- tography(IC)and ICP-AES. Meanwhile, besides further observation of the macroscopic morphologies of the ruptures on the two samples, three-dimensional stereomicroscopy(3D-SM) scanning electron microscopy (SEM)and energy dispersive spectrometry(eds) were adopted to analyze their microscopic morphologies along with micro-area compositio 3. Results and discussion 3.1. Matrix materials 3.1.1. Chemical compositions The chemical compositions of the cylindrical body, the flange and the weld joint of the CWP are listed in Table 1 respec- tively. It can be concluded that the materials used in the cylindrical body and the flange are the same, both of which are in accordance with the requirements of the UNS31803 grade dss [ 18(equals to the 2205 DSS in GB/T 21833-2008 19). Flux cored duplex stainless steel welding wire and gas shielded welding were used according to the manufactory. However, seen in the third row of Table 1, the carbon content at the weld is much higher than that at the cylindrical body and the flange. It meant that the quality of the weld joint was unqualified and it would induce the embrittlement of the weld joint. 3. 1.2. Metallographic structure the materials used in making the cylindrical body and flange are the same kind of DSS, which consists of ferrite and austenite distributing very evenly. The ferrite acts as the matrix, whose color is grey, while the austenite in white color distributes in the ferrite matrix. The grain of the two phases is quite clearly, so is the boundary. Fig 4(c) presents the metallograph ture of the weld joint, which is also consisted of ferrite and austenite but quite different with those of the cylindri and flange. It is obviously that the amount of ferrite is much more than that of the austenite with a dendritic grain shape. By means of EBSD, the ratio of the two phases in the microscopic field can be calculated. Just as the Fig. 5 shows, the amount of
The mechanism of these fractures on the DSS used in CWP was carefully discussed. Finally, effective countermeasures and suggestions were proposed as well. Actually, many researchers have focused on the properties of DSS, such as its fatigue behavior, welding property, corrosion resistance [10–17], but such an engineering practical study of mechanical degradation on DSS applied in CWP of 1000 MW ultra-supercritical thermal power unit has been rarely reported. What’s more, the phenomenon that large numbers of fractures occurred on the flanges of the CWP is even less reported. Therefore, the analyses and results given in this study have not only important engineering values in failure prevention of the CWPs used under seawater environment, but also practical significance in ensuring safety operation of other equipment under similar condition. 2. Experimental 2.1. Visual observation The 8A CWP is located in the Number 3 CWP house of Jiaxing power plant phase III, with a vertical structure and the length of the underground part is 17.1 m. As shown in Fig. 1(b), 8A CWP is mainly composed with two parts, i.e. the pump shell that weighs 42 tons and the shaft that weights 26 tons. The pump shell mainly consists of an inlet bellmouth, a bell pipe, four connecting pipes and a bent outlet from bottom to top. Each connecting pipe is constituted of two round flanges and a cylindrical body by means of welding. Hereby, the two flanges are located on both ends of the connecting pipe, and each flange is made up with four same flange arcs with a thickness of 46 mm by means of welding technology. The cylindrical body was manufactured by a process of rolling and welding, with an outer diameter of 2200 mm and a thickness of 14 mm. In this event, more than twenty severe cracks were discovered on the surface of 8A CWP. Among the damaged pipes, the severest one is the second connecting pipe counted from bottom to top, whose macroscopic appearance is showed in Fig. 3(a). Two target pairs of cracking samples analyzed in this study were both from the flange of this damaged pipe. One pair’s crack occurred on the weld joint of the flange, noted by cracking A and the other occurred on base material of the flange, noted by cracking B. The location of the two samples and the appearances are displayed in Fig. 3(a)–(c) 2.2. Characterization methods In order to figure out the failure causes and mechanisms, a variety of characterization methods were successively conducted. Oxygen nitrogen hydrogen (ONH) analyzer, carbon sulfur analyzer (CSA), and inductively coupled plasma atomic emission spectroscopy (ICP-AES) were used to inspect their chemical compositions. Optical microscopy (OM) was utilized to observe their metallographic structures and the austenite/ferrite ratio of the butt weld was obtained by electron backscatter diffraction (EBSD) and dyeing calculation method under metalloscope, respectively. The impact toughness of the DSS used in the CWP was also measured by Charpy impact test. And the constituents of the seawater were detected by ion chromatography (IC) and ICP-AES. Meanwhile, besides further observation of the macroscopic morphologies of the ruptures on the two samples, three-dimensional stereomicroscopy (3D-SM), scanning electron microscopy (SEM) and energy dispersive spectrometry (EDS) were adopted to analyze their microscopic morphologies along with micro-area compositions. 3. Results and discussion 3.1. Matrix materials 3.1.1. Chemical compositions The chemical compositions of the cylindrical body, the flange and the weld joint of the CWP are listed in Table 1 respectively. It can be concluded that the materials used in the cylindrical body and the flange are the same, both of which are in accordance with the requirements of the UNS31803 grade DSS [18] (equals to the 2205 DSS in GB/T 21833-2008 [19]). Flux cored duplex stainless steel welding wire and gas shielded welding were used according to the manufactory. However, as seen in the third row of Table 1, the carbon content at the weld is much higher than that at the cylindrical body and the flange. It meant that the quality of the weld joint was unqualified and it would induce the embrittlement of the weld joint. 3.1.2. Metallographic structure The metallographic structures of the matrix are displayed in Fig. 4. Fig. 4(a) and (b) present the metallographic structures of the material used in the flange and cylindrical body, both of which show a typical 2205 DSS structure. It is obviously that the materials used in making the cylindrical body and flange are the same kind of DSS, which consists of ferrite and austenite, distributing very evenly. The ferrite acts as the matrix, whose color is grey, while the austenite in white color distributes in the ferrite matrix. The grain of the two phases is quite clearly, so is the boundary. Fig. 4(c) presents the metallographic structure of the weld joint, which is also consisted of ferrite and austenite, but quite different with those of the cylindrical body and flange. It is obviously that the amount of ferrite is much more than that of the austenite with a dendritic grain shape. By means of EBSD, the ratio of the two phases in the microscopic field can be calculated. Just as the Fig. 5 shows, the amount of 164 Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177
Y-Y Ma et aL/ Engineering Failure Analysis 47(2015)162-177 (c) (d) Fig. 2. Fractures on the CWP's flange: (a) welding joint 1# (b)welding joint 2#, (c) welding joint 3# and (d) base material
Fig. 2. Fractures on the CWP’s flange: (a) welding joint 1#, (b) welding joint 2#, (c) welding joint 3# and (d) base material. Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177 165
Y-Y Ma et aL/Engineering Failure Analysis 47(2015)162-177 (b) Fig 3. Macroscopic appearance of the failed CWP pipe and samples: (a)appearance of the failed connecting pipe, (b)crack A and (c) crack B Chemical composition of the base material and weld joint of the CwP (wte Element SS of the cylindrical body 8.18 290 1.0-23.0 2.5-35 GBT21833 21.0-23.0 45-6.5 2.5-3.5
Fig. 3. Macroscopic appearance of the failed CWP pipe and samples: (a) appearance of the failed connecting pipe, (b) crack A and (c) crack B. Table 1 Chemical composition of the base material and weld joint of the CWP (wt%). Element C Cr Ni Mo N DSS of the flange 0.015 22.26 5.20 3.22 0.17 DSS of the cylindrical body 0.019 22.52 5.47 3.02 0.17 Weld joint 0.35 22.22 8.18 2.90 0.11 ASTM-A790/A790M-09 <0.03 21.0–23.0 4.5–6.5 2.5–3.5 0.08–0.20 GB/T 21833 <0.03 21.0–23.0 4.5–6.5 2.5–3.5 0.08–0.20 166 Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177
Y-Y Ma et aL/ Engineering Failure Analysis 47(2015)162-177 167 Fig 4. Metallographic structures of the DSS:(a)flange(b) cylindrical body and(c) butt weld. austenite in the weld joint is only about 23.0%, far below the standard of qualified 2205 DSS, indicating an unbalanced aus- tenite/ferrite ratio in the weld joint. By means of determined metallographically point count method, the same conclusion was drew out that the amount of austenite in the weld joint is less than 25% By the same method the amount of the aus- tenite in the cylindrical body and flange are nearly 48. 13% and 49.38% respectively both of which are in accordance with the tandard of 2 205 dss
austenite in the weld joint is only about 23.0%, far below the standard of qualified 2205 DSS, indicating an unbalanced austenite/ferrite ratio in the weld joint. By means of determined metallographically point count method, the same conclusion was drew out that the amount of austenite in the weld joint is less than 25%. By the same method, the amount of the austenite in the cylindrical body and flange are nearly 48.13% and 49.38% respectively, both of which are in accordance with the standard of 2205 DSS. Fig. 4. Metallographic structures of the DSS: (a) flange (b) cylindrical body and (c) butt weld. Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177 167
Y-Y Ma et aL/Engineering Failure Analysis 47(2015)162-177 ■ Ferrite:70%■ Austenite23.0% Fig. 5. EBSD of the butt weld. 3.1.3. Mechanical test In order to identify whether the dSs used in the water circulating pump is qualified, the impact toughness of samples from the flange and the cylindrical body were tested respectively by the method of Charpy impact test. It revealed that the Charpy impact value of the flange and the cylindrical body is both greater than 300 J exhibiting a superior toughness quality. To further investigate into the toughness of the base material, SEM morphology analysis was applied. As seen in Fig 6, a number of dimples were found on the Chary impact fractography under magnification of 500, confirming the excel lent ductility of the dSs used in CWP. Based on the analysis above, it was concluded that the 2205 dSS used in the CwP was fully qualified and the root cause of the failure did not ascribe to the selection of materials. Thus, the scope of the root cause for the failure was narrowed down to the quality of the weld. 3. 2. Environmental media By means of ICP-AES and IC, the main element constituents of the seawater are revealed in Table 2, which conformed to the normal compositions of natural seawater- a high content of chloride ions. Judging from the appearance of the CWP. there is no severe corrosion, thus selecting 2205 DSS as the base material of the CwP is right and essential because of the strict demand for corrosion resistance 3.3.1. Macroscopic morphologies Fig 7 shows the macro morphologies of sample A. Just as Fig. 7(a)shows, besides the butt weld in the flange there is also a weld joint that joins the flange and the cylindrical body, and these two weld joints are converged in a point. It also shows hat the crack propagated along the weld joint of the flange, noted by the arrow in Fig. 7(a). Further studies were conducted on the cross-sections of this fracture. Fig. 7(b) shows sample A's two corresponding cross sections(marked by Sections 1 and 2), on each lie a long and deep ditch respectively. This kind of ditch throughout the weld joint of the flange is a serious defect in welding, which is so called lack of penetration(LOP). Besides the LoP inside the butt joint(marked by LOPI), an obvious crevice, which is also a LOP, can be found in the weld joint connecting the cylindrical body and the flange(marked by LOP2) ust as Fig. 7(b)represents. As known, there is more residual stress in the weld joint, especially for those without a well heat treatment after welding 20. As Fig. 7(b)shows, the two serious defects mentioned above are linked to an area, greatly increasing the stress concentration and residual stress, and the intersection becomes the weakest area of this weld joint. According to the characteristics of the fracture surface, it can be learnt that the crack initiated from the intersection of he two weld joints, and propagated along the weld joint of the flange, as showed in Fig. 7(b). With the help of 3D-SM hologies of the section surface can be observed more clearly. As Fig. 7(c)shows, the welding flux did not fill the gap between the cylindrical body and the flange, leaving a serious defect that the cylindrical body and the flange were jointed only by two small areas. Judging by the morphologies of cross section, crack origin site can be located, as illustrated in Fig. 7(d), which is the weakest area of the weld joint
3.1.3. Mechanical test In order to identify whether the DSS used in the water circulating pump is qualified, the impact toughness of samples from the flange and the cylindrical body were tested respectively by the method of Charpy impact test. It revealed that the Charpy impact value of the flange and the cylindrical body is both greater than 300 J, exhibiting a superior toughness quality. To further investigate into the toughness of the base material, SEM morphology analysis was applied. As seen in Fig. 6, a number of dimples were found on the Chary impact fractography under magnification of 500, confirming the excellent ductility of the DSS used in CWP. Based on the analysis above, it was concluded that the 2205 DSS used in the CWP was fully qualified and the root cause of the failure did not ascribe to the selection of materials. Thus, the scope of the root cause for the failure was narrowed down to the quality of the weld. 3.2. Environmental media By means of ICP-AES and IC, the main element constituents of the seawater are revealed in Table 2, which conformed to the normal compositions of natural seawater – a high content of chloride ions. Judging from the appearance of the CWP, there is no severe corrosion, thus selecting 2205 DSS as the base material of the CWP is right and essential because of the strict demand for corrosion resistance. 3.3. Rupture of the sample A 3.3.1. Macroscopic morphologies Fig. 7 shows the macro morphologies of sample A. Just as Fig. 7(a) shows, besides the butt weld in the flange, there is also a weld joint that joins the flange and the cylindrical body, and these two weld joints are converged in a point. It also shows that the crack propagated along the weld joint of the flange, noted by the arrow in Fig. 7(a). Further studies were conducted on the cross-sections of this fracture. Fig. 7(b) shows sample A’s two corresponding cross sections (marked by Sections 1 and 2), on each lie a long and deep ditch respectively. This kind of ditch throughout the weld joint of the flange is a serious defect in welding, which is so called lack of penetration (LOP). Besides the LOP inside the butt joint (marked by LOP1), an obvious crevice, which is also a LOP, can be found in the weld joint connecting the cylindrical body and the flange (marked by LOP2), just as Fig. 7(b) represents. As known, there is more residual stress in the weld joint, especially for those without a well heat treatment after welding [20]. As Fig. 7(b) shows, the two serious defects mentioned above are linked to an area, greatly increasing the stress concentration and residual stress, and the intersection becomes the weakest area of this weld joint. According to the characteristics of the fracture surface, it can be learnt that the crack initiated from the intersection of the two weld joints, and propagated along the weld joint of the flange, as showed in Fig. 7(b). With the help of 3D-SM, morphologies of the section surface can be observed more clearly. As Fig. 7(c) shows, the welding flux did not fill the gap between the cylindrical body and the flange, leaving a serious defect that the cylindrical body and the flange were jointed only by two small areas. Judging by the morphologies of cross section, crack origin site can be located, as illustrated in Fig. 7(d), which is the weakest area of the weld joint. Fig. 5. EBSD of the butt weld. 168 Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177
Y-Y Ma et aL/ Engineering Failure Analysis 47(2015)162-177 169 Fig. 6. SEM morphologies of a sample's cross-section after Charpy impact test:(a)total morphology under low magnification and( b)dimples under high Table 2 Element constituents of the seawater(ppm, equals to mg/L). Fe Seawater 565×103 1.22x102 0.14 <0.002 <0.002 0025 33. 2. SEM and EDS In order to find the root cause of the fracture, Section 2 of sample a that was cleaned thoroughly by ultrasonic cleaning was studied further by SEM and EDS. The total morphology of the origin site showed in Fig 8(a) was selected to be observed thoroughly. Fig 8(b) shows the overall morphology of the origin site, where an uneven surface can be observed More def inite cracking origin site can be located in a further magnified scheme in Fig 8(c), besides, some impurities near the cracking rigin site can also be found, displayed by Fig 8(d Additionally, chemical compositions near the cracking origin site(sites 001, 002, 003 in Fig. 9)were detected by EDs. Based on the results in Table 3, very high carbon elements were detected in all the three sites, demonstrating the fracture urface of the weld joint was contaminated by organic substances under the seawater environment. Further experiments were conducted on the fillet weld Fig 10 shows the overall appearance of the fillet weld connecting the flange and the cylindrical body. With a high magnification, multiple microcracks were observed by SEM, which strongly enhanced the conclusion about the embrittlement of the weld joint 3.4. Rupture analysis of sample b Fig 11 shows the appearances of sample B, which does not locate on the weld joint. Fig. 11(a) shows the overall morphol- ogy of this crack on the sealing surface and Fig. 11(b)exhibits the two corresponding cross sections. As Fig. 11(c)and(d
3.3.2. SEM and EDS In order to find the root cause of the fracture, Section 2 of sample A that was cleaned thoroughly by ultrasonic cleaning, was studied further by SEM and EDS. The total morphology of the origin site showed in Fig. 8(a) was selected to be observed thoroughly. Fig. 8(b) shows the overall morphology of the origin site, where an uneven surface can be observed. More definite cracking origin site can be located in a further magnified scheme in Fig. 8(c), besides, some impurities near the cracking origin site can also be found, displayed by Fig. 8(d). Additionally, chemical compositions near the cracking origin site (sites 001, 002, 003 in Fig. 9) were detected by EDS. Based on the results in Table 3, very high carbon elements were detected in all the three sites, demonstrating the fracture surface of the weld joint was contaminated by organic substances under the seawater environment. Further experiments were conducted on the fillet weld. Fig. 10 shows the overall appearance of the fillet weld connecting the flange and the cylindrical body. With a high magnification, multiple microcracks were observed by SEM, which strongly enhanced the conclusion about the embrittlement of the weld joint. 3.4. Rupture analysis of sample B Fig. 11 shows the appearances of sample B, which does not locate on the weld joint. Fig. 11(a) shows the overall morphology of this crack on the sealing surface and Fig. 11(b) exhibits the two corresponding cross sections. As Fig. 11(c) and (d) Fig. 6. SEM morphologies of a sample’s cross-section after Charpy impact test: (a) total morphology under low magnification and (b) dimples under high magnification. Table 2 Element constituents of the seawater (ppm, equals to mg/L). Element Cl Mg Al Cu Fe Ti Mn Seawater 5.65 103 1.22 102 0.14 <0.002 0.92 <0.002 0.025 Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177 169
Y-Y Ma et aL/Engineering Failure Analysis 47(2015)162-177 ection rigin of the crack body Fig. 7. Macro morphologies of sample A: (a)the crack, (b) two corresponding cross sections, (c)magnified morphology of LoP2 and(d)origin site in Section show, fatigue cracks could be clearly observed, propagating along the arrows marked in the figures. According to the surface morphologies, the crack seemed to be torn by external force, exhibiting a kind of secondary fracture rather than the first fracture So it can be sure that the fracture appeared in this location happened after the fracture in the weld joint of sample A. The fracture in the weld joint of the flange led to an unbalanced of the CwP and large force was applied suddenly on the of sample B, which cause it to be finally fractured. As a result, further studies are not needed to conduct for the 4. Failure analysis 4.1. Lack of penetration Based on the analysis results presented above, it is clear that lack of penetrations(LOPs)existing in the weld joints duced by improper welding procedures should be blamed for the serious fracture of the circulating water pump(CWP). Just as Fig. 7(b)illustrates, LOP not only lies in the weld joint that connects the flanges, but also in the weld joint that connects the flange and the cylindrical body decreasing the strength of the weld joint LOPI locating in the weld joint of the flange joint. These LOPs strength and fatigue life of the joints significantly It has been proved by many researchers that LOP, as a kind of common defect in the weld joint, would markedly decrease the strength of the weld joint and reduced greatly the fatigue life, leading to a severe fracture. Kim [21 studied the effect of LOP on the fatigue strength of butt weld by a number of fatigue tests of high steel containing partial and full penetration butt elds, revealing that fatigue strength of partial penetration butt weld was lower remarkably than that with full penetration, and the fatigue cracks initiated at the LOP section. Wahab 22 pointed out that the weld imperfections such as LOP, porosity, lack of fusion, undercut effectively reduced the fatigue crack propagation life and fatigue strength of welded joints b
show, fatigue cracks could be clearly observed, propagating along the arrows marked in the figures. According to the surface morphologies, the crack seemed to be torn by external force, exhibiting a kind of secondary fracture rather than the first fracture. So it can be sure that the fracture appeared in this location happened after the fracture in the weld joint of sample A. The fracture in the weld joint of the flange led to an unbalanced of the CWP and large force was applied suddenly on the location of sample B, which cause it to be finally fractured. As a result, further studies are not needed to conduct for the sample B. 4. Failure analysis 4.1. Lack of penetration Based on the analysis results presented above, it is clear that lack of penetrations (LOPs) existing in the weld joints induced by improper welding procedures should be blamed for the serious fracture of the circulating water pump(CWP). Just as Fig. 7(b) illustrates, LOP not only lies in the weld joint that connects the flanges, but also in the weld joint that connects the flange and the cylindrical body, decreasing the strength of the weld joint. LOP1 locating in the weld joint of the flange even runs through the weld joint. These LOPs would result in high stress concentration on the crack tip, and reduce the strength and fatigue life of the joints significantly. It has been proved by many researchers that LOP, as a kind of common defect in the weld joint, would markedly decrease the strength of the weld joint and reduced greatly the fatigue life, leading to a severe fracture. Kim [21] studied the effect of LOP on the fatigue strength of butt weld by a number of fatigue tests of high steel containing partial and full penetration butt welds, revealing that fatigue strength of partial penetration butt weld was lower remarkably than that with full penetration, and the fatigue cracks initiated at the LOP section. Wahab [22] pointed out that the weld imperfections such as LOP, porosity, lack of fusion, undercut effectively reduced the fatigue crack propagation life and fatigue strength of welded joints by Fig. 7. Macro morphologies of sample A: (a) the crack, (b) two corresponding cross sections, (c) magnified morphology of LOP2 and (d) origin site in Section 2. 170 Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177
Y-Y Ma et aL/ Engineering Failure Analysis 47(2015)162-177 (b) (d) origin of the crack Fig 8. SEM morphologies of the fracture: (a)total morphology, (b)magnification of the cracks origin. (c)further magnification of the cracks origin and (d) further magnification of the impurity near the cracks origin. tudying a variety of weld imperfections. Kim[ 23 pointed out that the LOPs significantly reduced the fatigue lives of 9 mm thick transverse butt welded specimens without weld reinforcements, showing shorter fatigue life than JSSc-B Sanders and Lawrence [24 studied the effects of lack of penetration(LOP)and lack of fusion( LOF)on the fatigue behavior of butt welds. He concluded that LoP defects can seriously reduce the fatigue life, while inclined loF defects were generally less serious than loP defects Former studies25-27 also proved that LOP had much to do with welding procedure, including welding method, groove, welding speed, welding heat input and so on. In this event, the welding method used here was multi-layer and multi-pas The weld groove is one of the most important facts to affect the quality of welding. Here, the Double-V preparation with proper parameters is recommended, listed in Table 4. As for LOP2 shows in Fig. 7(c), the weld groove is improper because the depth of preparation is too short and the groove angle is quite small according to the standard of IS09692-1: 2003[28]. So, it is difficult for the weld flux to fill the gar ea though other operations are correct. Here, the k preparation with proper parameters is recommended also listed in Table 4.2. Unbalanced microstructure with excessive ferrite As the EBSD analysis result shows, there was a much higher ferrite to austenite ratio in the weld joint, also proved by means of determined metallographically point count method. As known, the toughness of dSS weld joint with higher con- tent of ferrite phase decreases remarkably According former researches [29-31 this phenomenon had much to do with the welding procedure especially the cooling rate or heat input. As Table 1 shows, the content of nitrogen in weld metal is in quirement with the standard, which indicates that the low austenite contend of the weld metal is rather due to the low heat input or high cooling rate employed in the welding than to nitrogen loss. Fig. 12 shows the vertical section of Fe- Cr-Ni ternary diagram based on Chromium-Equivalent presenting the phase transformation of 2205 DSS during the welding procedure Based on the Fig. 12, austenite wi to precipitate at about 1330C, above which all the phas in the weld joint is ferrite. However, during continuous cooling there is not enough time for austenite to precipitate until the temperature has decreased to about 1200C. when a sufficiently amount of nuclei has been formed 32. thus a high cooling
studying a variety of weld imperfections. Kim [23] pointed out that the LOPs significantly reduced the fatigue lives of 9 mm thick transverse butt welded specimens without weld reinforcements, showing shorter fatigue life than JSSC-B. Sanders and Lawrence [24] studied the effects of lack of penetration (LOP) and lack of fusion (LOF) on the fatigue behavior of butt welds. He concluded that LOP defects can seriously reduce the fatigue life, while inclined LOF defects were generally less serious than LOP defects. Former studies [25–27] also proved that LOP had much to do with welding procedure, including welding method, groove, welding speed, welding heat input and so on. In this event, the welding method used here was multi-layer and multi-pass welding and no sufficient heat treatment was conducted after the welding according to the manufacturer. The weld groove is one of the most important facts to affect the quality of welding. Here, the Double-V preparation with proper parameters is recommended, listed in Table 4. As for LOP2 shows in Fig. 7(c), the weld groove is improper because the depth of preparation is too short and the groove angle is quite small according to the standard of ISO9692-1:2003 [28]. So, it is difficult for the weld flux to fill the gap even though other operations are correct. Here, the K preparation with proper parameters is recommended, also listed in Table 4. 4.2. Unbalanced microstructure with excessive ferrite As the EBSD analysis result shows, there was a much higher ferrite to austenite ratio in the weld joint, also proved by means of determined metallographically point count method. As known, the toughness of DSS weld joint with higher content of ferrite phase decreases remarkably. According former researches [29–31], this phenomenon had much to do with the welding procedure, especially the cooling rate or heat input. As Table 1 shows, the content of nitrogen in weld metal is in requirement with the standard, which indicates that the low austenite contend of the weld metal is rather due to the low heat input or high cooling rate employed in the welding than to nitrogen loss. Fig. 12 shows the vertical section of Fe– Cr–Ni ternary diagram based on Chromium-Equivalent, clearly presenting the phase transformation of 2205 DSS during the welding procedure. Based on the Fig. 12, austenite will start to precipitate at about 1330 C, above which all the phase in the weld joint is ferrite. However, during continuous cooling there is not enough time for austenite to precipitate until the temperature has decreased to about 1200 C, when a sufficiently amount of nuclei has been formed [32], thus a high cooling Fig. 8. SEM morphologies of the fracture: (a) total morphology, (b) magnification of the crack’s origin, (c) further magnification of the crack’s origin and (d) further magnification of the impurity near the crack’s origin. Y.-Y. Ma et al. / Engineering Failure Analysis 47 (2015) 162–177 171
