46 2 Vol.46 No.2 21 2 195 ACTA METALLURGICA SINICA Feb. 21 pp.195 Đ ³ Ì Ó Ö ßß Öß ¼»¹ ( À ÅÈ, 445) ½º¾ ( Þ, «½ 142 41, ¾ ) Р º À ½Ê ß Û ¹Ä Ñ», À Ðû Üß Û. ĐºÑÜÆ ßÜÖß Û Đ ÃÛ ÜÖßà ± Ü, Ð À Û ßÑ», ½ ÂÓ ßÀ, ÂÓÜÖ ß ß. ÔÝ Û, Ñ», ½Ê,   º Ò Ø TG47 «A µïø 412 1961(21)2 195 6 NUMERICAL SIMULATION OF WELDING RESIDUAL STRESSES IN A MULTI PASS BUTT WELDED JOINT OF AUSTENITIC STAINLESS STEEL USING VARIABLE LENGTH HEAT SOURCE DENG De an College of Materials Science & Engineering, Chongqing University, Chongqing 445 KIYOSHIMA Shoichi Department of Technical Development, Research Center of Computational Mechanics, Inc., Tokyo 142 41, Japan Correspondent: DENG De an, professor, Tel: (23)65112879, E-mail: deandeng@hotmail.com Manuscript received 9 7 3, in revised form 9 12 11 ABSTRACT Recent discoveries of stress corrosion cracking at welded joints in pressurized water reactors and boiling water reactors have raised wide concerns about the safety and integrity of plant components. It has been recognized that residual stress and applied stress on their surfaces largely increase the expanding risk of initial stress corrosion cracking. Therefore, it is very important to investigate the welding residual stress in welded joints. It is very expensive and time consuming to measure the residual stress, and sometime is impossible. As an alternative approach a computational procedure on the basis of finite element method is effective in solving non linear problems such as thermal and mechanical nonlinearity in a welding process. Accurately simulating welding residual stress not only needs generally a long computational time, but also strongly depends on the analyst s experience and know how which is a main hindrane for the welding process simulation. Therefore, it is an urgent task to develop a time effective numerical simulation procedure to calculate welding temperature field and residual stress distribution. In this study, a new method on the basis of the variable length heat source was developed to simulate the welding residual stress in a multi pass butt welded joint of austenitic stainless steel. Meanwhile, the experiment was carried out to obtain the welding residual stress in the butt welded joint. Comparing the simulated with experimental results, it was found that this method could not only save a large amount of computational time but also provide a highly accurate numerical result for the residual stress in multi pass butt welded joints. KEY WORDS welding residual stress, numerical simulation, austenitic stainless steel, variable length heat source * ÖÏ : 9 7 3, ÖÈ : 9 12 11 È : Û,, 1968 Å,,»Ë DOI: 1.3724/SP.J.137.9.521
196 µ 46»«Î, ±Ü Å Ü Â ±, Ø, Ü Â³ Æ» [1 4]. Ü Â 3 ÁØÚµ, ßÕÇ Â ÍÜ ( Ì Ü Æ Ü )., ²Ê Í µ É, Ø, Đ»¾Ã²Û Ú Ü. [3], ÍÜ ÍÕ ²Ê Ü Â ± Ø. Æ Î ²Ê µçá, Ü ÍÐÁ ØÚ. Ð ÎÖ, Ä ÃÔ Ö Ç ÖÇ Äµ ÎÜÝÊ Ü, б ĵ À Ó¹ Æ, Æ, ĵ ÎÜÝ Á Ü Ð¾Ã. Ì«7 «ÔÎ Î Ô»ÒÝÇ Æ [5] Ò¼Đ µ ÃÔ «Ä ÎË Ü ÅµÃ, ¾, Ò ³Ý º. ¾ Æ ³ Abaqus, Ansys Marc, ÔĐ Quick Welder Syswled, ÆÃÔĐ É Ü µã, Ì Å¹ [6,7]. É» Å Í À Ç, Ñ» ĐÁÎ Ù Ü, Đ» Ì º Æ, ± Á Ú Ã [8]. Õ, Ô Ð µ Î Á, Á : (1) Æ ºÆÐÆ, Ò Î Á [9] ; (2) µ, Û, Á [1] ; (3) Ó ²Ê Å Ç ²ÊµÕ Ç,»ĐÁÎ «³ÅĐÁ, Á [11]. Ð Î Ê ³» ĐÁ õû ĐÁ [12] Î Á. Ó 3D ²ÊÍ Ë¼, ± Á» ĐÁ, 25 mm, ÙÒ 17 Å¾Ë Ü ºÅ Ò¼Đ ; Quick Welder [13], Đ É», º ² IJ, ºÐ õû ĐÁ Ç. 1 Å Ð 1 Å¾Ë Ü Í ÊÕ. Å¾Ë SUS34, Õ Æ» (»Ò, %) : C.8, Si.45, Mn 1.96, P<.45, S<.3, Ni 9.5, Cr 18.5, Fe. Èà 1 3 mm, Ê 3 mm 25 mm ÅÆ, ÇÃÉ 2 Ê., Ð Ç Å Ò ( Å TIG ), ÕÎ 2 Ê, X Á ÇÉ 17 Æ Ù, Ý 8 Æ, ¼Ý 9 Æ, Æ 1 Ê. É, Å 1 Ô¼. Î Y38L, Õ Æ» (»Ò, %) : C.4, Si.34, Mn 1.92, P<.19, S<.2, Ni 9.6, Cr 19.5, Fe. º 1.2 mm, Ò Å Ar, Á ÅÑ 15 L/min. Ä Ý B Đ» A Å C Á Ý ¼ Ý Ü ( Ý 1 Ê), Ý B z Ê 5 mm, A C ÝÅ B Ý 1 mm. Æ, ¾, Äܵƻ¹ ¼ Ý, Æ Æ, ÔÜ³Ü Ï ÒÇܵ¼. Å ÒÇܵ, Ü Ü ¼Ã σ x = E(ε x + νε z )/(1 ν 2 ) σ y = E(ε z + νε x )/(1 ν 2 ) 1 Ì Fig.1 Butt welded joint and its dimensions in mm (section A is at z=49 mm, section B is at z=5 mm and section C is at z=51 mm) 2 ß ÆÂÈ Fig.2 Groove and welding sequence of butt welded joint as (1) listed in Table 1 1 ß Table 1 Welding conditions for the butt welded joint in Fig.2 Weld Current Voltage Welding speed Heat input pass A V mm/min kj/mm 1 and 2 17 1 8 1.275 3 8 17 11 1 1.122 9 17 9 1.918 1 and 11 17 1 8 1.275 12 17 17 1 1 1.2
2 ĐÚ Ð : Á³Á ¹ Ð ¼É Þ Ú 197 É, ε x ε z»¹ Ä Ý ( ºµ ) Á ( ŵ ) ÒÇܵ; E ν ÍÒÇ Đ Possion, Ü Á, E ν»¹³ 198.5 MPa.3 [1]. 2 ² Ì Á,»±«Æ ÀÔ» º, ± ÀÔ» Ð, 3 Ê. Æ ÅÖ ÅÄ Ü Ù ÕÎ, 4 Ê. ĐÁ, µ ³ÁĐÁÔ»Æ 24 ¼, Æ ÏÁ 8 ÒÝÅ 6 µýå Æ, Ò 1.2 1 4, Æ Ò 1.3248 1 4. Ë µé, Đ É», ½Ã Æ Â, ² Ô» µé Ü ĐÁ. Đ Ó» Ç Ç µõ, Á Å Å» Ç [1].» Ç Ç [14]. 2.1 ÜÆ É, Ñ» Æ Õà ǻµÉ [2] Þ ( 2 T ) ( 2 T ) ( 2 T ) λ x 2 + λ y 2 + λ z 2 + q v = ρc T t (2) É, T, λ, q v, ρ c»¹í» Õ¹Ò»»Õ Ú» ; x, y z µ 1 Ê. Ñ µé, Æ Å (Ä ) ºÅ» Ù. ĐÁÅ Ä ÑµÉ» Ù Ã Newton µé [2] Þ q a = h a (T s T a ) (3) É, q a ÍÆ Å Ä» Ù, h a Í Ñ» Ù¹Ò, T a ÍÄ (25 ), T s ÍÆ Ý., Å h a ; µõ ż [13], Å 6 1 6 W/(mm 2 ).» Đ» q r à q r = εσ[(t s + 273) 4 (T a + 273) 4 ] (4) 3 ß À Fig.3 Finite element model of butt welded joint (number of nodes: 1.2 1 4, number of element: 1.3248 1 4 ) 4 À ß Ø ÔÍ Fig.4 Weld pass and their sequences defined in the finite element model É, ε Í» ¹Ò, ε ³¼.8 [2,12] ; σ Í Stefan Boltzman ÀÒ. ĐÁ, Ú Å» ÎĐ». À ˼, ٠ŵ ĐÁ»Æ à [2,12] N»,» ÎĐ Ñ». Á, ĐÁ õû ÎĐ Ü. õà ĐÁ» Þ Ô¼ µý Ó: (1)  ½¼ ³» Ù» Ç Í ; (2)  ٠Á, Á ²Ê Ý ; (3) ¾ Í µý Í µ Ü Ý Ë¼, Ü ³ Ç ½¼» ÐÆ» Ò; (4) Ù, Ù ½ Ù Æ Ü Ý Ä [12]. õû ԼР²: (1)  ٻ ÍÐÁ Ú» Å», Ý Å Ü Ù ÝÝ. (2) Ù,» à µõ². Æ Ã ÅÄ Ë Í Ó, Æ Ã ÃÔÍ³Æ Ù Ã. Ä, ÍÑ», ÍÓ». (3) ½ Ù, ÃÔ Ã», Ù ÃÔ» Ý º Ñ». ĐÁ, X Á Ç ¼Ý ÕÎ, ½ Ù, º, Æ Ý
198 µ 46 Ù. Ù Ò 17, Æ»ÒÝÇ Á Ã, Õ, Þ Ü» à µ Đ», µõæ, µ Ù À» µ à µã» ĐÁλØĐ É, ± à ¾ Đ Ë¼, Á. õû ĐÁ,» Ù ( 4 1, 2, 9 1 Ù), û,» à ĐÁà 1/2, ÙÆ Ù» Æ. û ÅÓ» [15]»ÏØ; Ù ( 4 3 6 12 15 Ù),» à ĐÁà 1/4; Ý Ù ( 4 7, 8, 16 17 Ù),» à ŠÆà Ð, Đ Áà 1/24, ¼Ï Ä ËÃ, ÐÅ Ä Ñ» Ç», à λØĐ Ñ». Đ É, û ¾ µ ˼, ³ Ò, Ñ Ñ Û, Æ Ë Æ Å ÐÁ ², 1.5 1 3 2. 1 3 ². Á» Û Î Ã» ¾µ, Æ Æ Ê Ð Á ². ± µ à H = αq/(α v) (5) É, H Í Ã», Q Í Ñ ÇÕ, v Í Û, α α»¹í Q v ¹Ò. ÑÇÕÅÖ [1] Q = ηui (6) É, η Í Ñ Õ, Å η.75 [16] ; U Í Ñ ; I Í Ñ. Ú Å» Đ Á, Ã Æ Ì ÅÖ Á Ò: ÐÁÍ»»Õ q v, ÐÁÍ»»Á t h. q v Ã Ê É, V h Í» Å. q v = αq/v h (7) û ¾µ ˼,»»Á t h à É, L h Í» Ã. t h = L h /(α v) (8) ¼Ý½ Í, ÃÔ Â Ù Î Ó ¹Ò, Ô ÃÝ Ë [12]. 2.2 ÜÆ Å Á µ ÆĐÁ, Á Ý Â É», ³³Á É Ñ Üµ Ü. Ü», Ó Ç µõ Ç [12]. Đ, ÒÇÜ ÜµÎ¹ Â Ç Hooke ÅÔ, Ý ÇÅ Von Mises ; ÆÞÕÇ Â Ç ÎÞ, Á Ó Û Ý, Ó ÝÄ Û Á, ÆÞÕ Å µ. ĐÁ, Ð Û, ÝÇܵ, Ô±ß Û Ü; Á, Û, ØÃÔ ÆÞÕ., ±«Đ», Ø, Ó Û Ü Ý Í»±. Û Å 8 [12]. Ä É Ø Ð, Ô, ÐÆ Î ¾ ÆĐ Á Å. Ò¼ Í Á ß (CPU 3.2 GHz, RAM 4. GB) Æ, Á 38.7 h. Ñ», Ó¹ Á Ì Ò Á, õû ÃÔÆ Á. 3 Û ÐձŠÐÕÎÞ 3.1 Ý Ý µ Ü (ԼŠÁ Ü )» ÅIJ 5a 6 (a) A-line B-line 4 C-line FEM Longitudinal stress, MPa Longitudinal stress, MPa - -15-1 -5 5 1 15 6 (b) 4 - -15-1 -5 5 1 15 x-coordinate, mm 5 Ü» ÜßÀ Û Fig.5 Calculated longitudinal residual stress distributions and the experimental data on top (a) and bottom (b) of section B in Fig.1 (A, B, C lines show the experimental data of sections A, B and C, respectively)
2 ĐÚ Ð : Á³Á ¹ Ð ¼É Þ Ú 199 Ê. ÃÔ, Á Ü» ÅÄ Ý Ü» Ã Þ Ð, ±«Ý Æ ÍÜ ÅIJ», ¾ ±Đ», ÍÜ ² ļ. ±Ã Í Đ,» Ã Ä Ñ» Ã,» ʵ» Ä ÄÉ, ± ºµ Ä Ð Æ (½¼ ) ±«Ú Ä ±«Ö, Ýǵñ Ð, Ô ÍÜ ±«Ä ±«ÖÊ. ÆĐÁ, Â Ù Ã Ø Å, ±Ã ÍÙ Ø Ð., ÆÞÕĐÁ Û ² Ï Ü» ² Ã Æ ÐÅ Ý. 5b Ý ¼ Ý Á Ü» Å Ü Ä². Ã, ±Đ» Đ ² ÅIJ Ó», Ð ± ±«¹. ØÍ, Đ Ø Ó É Ç Ü Ý. Đ» ½¼», É Ü. ±Ð Ü Í, ±«½¼, ¾ÂÔ Ü, Ô Ü Úµ Ý» Ð ¼Î [17]. Ý (B Ý) Đ» Á Ü» 6 Ê. ² ÅIJ ÃÔ, ± Æ Ü ¾ À ² Ð ( 27 MPa [12] ). Ü ØÍ, ±» ÝǵÃ, Å¾Ë ÆÞÕ À ¾, Ø, Ýǵñ«² Ü Ú,, ± Ýǵà ±«Ú Ä ² Ü. 3.2 Ù 7a b»¹ Ä Ý ¼ Ý Å ºµ Ü (Ô¼Å Ü )». ÃÔ, Ò¼Đ ² ÅÄ ² Þ Ð, Ð ² ÅIJ. Ç Ü, ±«² ÅIJ ¾, ļ ². ØÍ, ÐµÝ Ä Ý ÄܵÆ, Ý Ç ÎÓ ß Æ, ÆÝ Ü µ; еÝ, Đ É Ø ÅÂÁ ÙÃ Ý Ù. ØÐ Æ, Ð ² ŠIJ» à ½Þ ², Ø ÃÔ¼ : Ý Ü Í. Ý (B Ý) Đ» Ü» 8 Ê. Ã, ± Ü µ Ç Í. 3.3 ÍÚÑ B Ý µ Ü» Transverse stress, MPa Transverse stress, MPa 5 3 1 (a) A-line B-line C-line FEM - -15-1 -5 5 1 15 (b) - -15-1 -5 5 1 15 x-coordinate, mm 7 Ü» Üß Û Fig.7 Calculated transverse residual stress distributions and the experimental data on top (a) and bottom (b) of section B in Fig.1 6 ºÑÜÆ ßÜÖß ÜßÀ Û º Fig.6 Calculated longitudinal residual stress distribution on section B in Fig.1 8 Üß Û º Fig.8 Calculated transverse residual stress distribution of section B in Fig.1
µ 46 Residual stress, MPa 8 6 4 - Long. stress Trans. stress -4 5 1 15 2 25 y-coordinate, mm 9 Ü B ß Û º Fig.9 Longitudinal and transverse residual stress distributions along the centerline (y axis in Fig.1) of section B 9 Ê. Ã, µ Á Ü» à Ż Ø. µ (Á ) Ð Ð, Ø, Á Ü ¾ Ü. ³Á µ, Á Ü µ ÍÜ, Æ Ü ¼ Ý 8 mm ; Ü ÍÜ, Ü., ÙÒ, Ô Ü Êµ (x µ ), µ Ú, Ô µ (y µ ) Ç Í º». 4 Ð Þ»ÒÝÇ Æµ, õû ĐÁ Å¾Ë Ü Ò¼Đ, ¾ ÃÔ Á,. Ò¼Đ ² Þ ±ß Å¾Ë Ü», Ç È. [1] Deng D, Murakawa H, Liang W. Comput Mater Sci, 8; 42: 234 [2] Ogawa K, Deng D, Kiyoshima S, Yanagida N, Saito K. Comput Mater Sci, 9; 45: 131 [3] Hamada I, Yamauchi K. Nucl Eng Des, 2; 214: 25 [4] Koshiishi M, Fujimori H, Okada M, Hirano A. Hitachi Rev, 9; 58: 88 [5] Ueda Y, Yamakawa T. JWRI Trans, 1971; 2: 9 [6] Deng D, Murakawa H. Comput Mater Sci, 8; 43: 681 [7] Deng D. Mater Des, 9; 3: 359 [8] Deng D. PhD Thesis, Osaka University, 2 [9] Hong J K, Tsai C L, Dong P. Weld J, 1998; 77: 372s [1] Nishigawa H, Serizawa H, Murakawa H. Sci Technol Weld Join, 7; 12: 147 [11] Ogawa K, Yanagida N, Saito K. Proc PVP8, 8 ASME Pressure Vessels and Piping Division Conference, July 27 31, 8, Chicago, Illinois, USA, PVP8 61321, CD ROM [12] Kiyoshima S, Deng D, Ogawa K, Yanagida N, Satio K. Comput Mater Sci, 9; 46: 987 [13] Kiyoshima S. Quick Welder User s Manual. Research Center of Computational Mechanics Inc., Tokyo, 5: 23 ( Ô. Quick Welder «Å Æ. ß, ¾, 5: 23) [14] Deng D, Murakawa H. Comput Mater Sci, 6; 37: 269 [15] Terasaki T, Kitamura T, Akiyama T, Nakatani M. Sci Tech Weld Join, 5; 1: 71 [16] Deng D, Murakwa H. Comput Mater Sci, 6; 37: 29 [17] Leggatt R H. Int J Press Vessels Piping, 8; 85: 144