Prediction of welding distortion during assembly process of thin plate structures Prediction of welding distortion during assembly process of thin plate structures

Prediction of welding distortion during assembly process of thin plate structures

  • 期刊名字:中国焊接
  • 文件大小:563kb
  • 论文作者:Luo Yu,Deng De'an,Jiang X
  • 作者单位:Shanghai Jiaotong University,Chongqing University
  • 更新时间:2020-11-22
  • 下载次数:
论文简介

Prediction of welding distortion during assembly process of thin plate structures153Prediction of welding distortion during assembly processof thin plate structuresLuo Yu, Deng De' an and Jjiang Xiaoling罗宇,邓德安,江晓玲*Abstract Ships and automobiles are fabricated from thin plates To assemble parts, welding is commonly employed.Houerer, welding distortion in large thin-plate panel structure is usually cased by buckling due to the residual sress In thisstudy, an elastic finite element method for predicting the welding distortion of three -dimensional thin-plate structures withconsidering welding sequence was proposed. In this method, the inherent strain was employed t0 model the local shrinkage dueto welding itself, and the inteface element was introduced to simulate the assembly process. The proposed method was appliedl0 study the in/luence of welding sequence on the buckling distortion of the large thin plate panel structure during assembly.Key words welding distortion, buckling, finite element metod , interface element, welding sequence0 Introductionined. In the proposed method, the local shrinkage is de-Application of welding process in fabricating largescribed by the inherent strain , and the gap, misalignmentstructures ofers several advantages over mechanical join-and welding sequence are modeled using interface ele-ing methods such as improved structural performance ,ment!2). Using the proposed numerical method, the influ-flexibility of design, weight reduction and cost savingsence of welding sequence on welding distortion of thin-etc. In ship and automobile industry,according to currentplate panel structure is studied.trend in design, to achieve weight reduction, large struc-tures are made up of relatively thin section of high strength1 Method of analysissteels. However, buckling is a common problem occurring.1 Interface elementin the welding fabrication of thin-plate panel structure.The assembly of the structure can be regarded as theBuckling distortion causes loss of structural integrity, di-retition of the positioning of new parts and welding. Themension control and increased fabrication costs due to poorparts are physically free before positioning. Through posi-fit-up between panels. Several factors that infuence dis-tioning, the relative position of the part, the gap and thetortion control strategy can be categorized into design-relat-misalignment are corrected within tolerable limits. Byed and process -related variables. Significant design relat-welding, full bonding between the parts is formned. Toed variables include welded joint details ,plate thickness,nodel the real assembly process, the change of physicalsifener spacing etc. Important process-related variablesstate must be modeled. For this purpose, the interface el-are welding heat input, welding sequence ,positioning,ement is introduced. The change of physical state betweentack weld and restraint condition etec"].parts can be described through the bonding stress-relativeIn this research, a new elastic finite element methoddisplacement relation.to predict the distortion of thin-plate panel structure duringFig. 1 shows the interface element defined betweenassembly by welding is proposed and the influence oftwo narts ( or two elements belonging to the parts to bewelding sequence on the geometrical accuracy is exam-中国煤化工nts btween the two partsYHCNMHG●Luo Yu and Jjang Xiaoling, Shanghai Jiaotong University, Shanghai, 200030. E-mail:!uoyu@ sjtu. edu. cn (Luo Yu) Deng De'an, Chongqing University , Chongqing, 4044.154CHINA WELDINC Vol. 14 No. 2 November 2005are normal opening δw,transverse opening δr,longitudi-sitioning and welding. The property of the interface ele-nal sliding δ and relative rotation δgx. Similarly, thement is mostly defined by the scale parameter r。and bond-stresses acting between the parts are denoted as σv, σr,ing strength y. The scale parameter To gives the precisionσl and σax,respectively.in positioning and a small enough value must be selectedto ensure the required precision at the positioning stage.Except for the positioning stage, large value can be usedfor r。. On the other hand,the bonding strength γ must besmall when the parts are free. At the positioning stage, arelatively small value such as 10-* MPa corresponds toweak positioning which is not strong enough to close thegap. For the strong positioning which is strong enough to .close the gap, a large value such as 10° MPa can be giv-Fig. 1 Types of gap described in the interface elementn. In this research, the value is assumed to be 10-1MPa. As shown in Fig. 3, the precision of the positioningThe relationships between the stress σ and the dis-can be controlled through adjusting the scale parameter ro.placerment δ are schematically shown in Fig. 2a andWhen the full bonding formed by welding, a large valueFig. 2b. Due to the symmetry of the shear deformation andlike 10'0 MPa can be used. The gap between parts remai-rotational deformation, the relation between displacermentning after the positioning stage, δc, is fixed and intro-and stress is symmetric as shown in Fig. 2b. While that induced into the function describing the state after weldingnormal direction is not symmetric as shown in Fig. 2a.as shown in Fig. 4.。°vY~|_0170.1δ(mm)8Fig.3 Relationship between bonding stress and(a) Normal directiondisplacement during positioning叶, Oex .n..20x10%Y&.8, δox一)I8cv8N(b) Shear and rotational direction中国煤化工r for welded joint with gapFig.2 Relationship between bonding stress and displacementYHCNMHG1.2 Procedure of analysisThe physical relation between parts during the assem-The finite element method proposed in this study isbly process can be separated into three , namely, free, po~ developed to predict the distortion of the large thin-plate.Prediction of welding distortion during assembly process of thin plate structures155panel structures by considering the details of the assemblyber of element is 2274, and that of node is 2 248. Theprocess,such as welding sequence, the positioning, gapmodel is fixed at three points as shown in Fig. 5 to preventcorrection , tack welding and joining by welding. The partsrigid body motion.are subdivided into a finite element mesh using 4-pointrectangular plate elements. The state of joining is de-2.2 Welding conditionscribed by the interface element that is introduced betweenIn this research, the joint type between skin plateparts to be joined. The deformnation due to the weldingand siffener is double fllet weld, and that between thethermal cycle is taken into account through the inherentlongtudinal siffeners and the transverse stifeners is cross-strain introduced to the elements facing the welding line.shaped joint. The heat inputs for each welding line are as-Thus,the detailed assembly process can be simulatedsumed as follows:through controlling the parameters involved in the interfaceTransverse siffener/Skin plate: 600 J/ mmelement and the inherent strain according to the assemblyLongitudinal siffener/ Skin plate: 600 J/mmstage.Transverse siffener/ Longitudinal siffener: 400 J/mm2 Thin-plate panel structure2.3 Welding sequenceIn order to study the influence of welding sequenceTo clarify the influences of welding sequence andon welding distortion, a three dimensional thin-plate panelheat input on welding distortion of a mild steel thin-platewelded structure made up of mild steel with three longitu-structure, five cases were investigated in this study. Thedinal siffeners (Lstiffeners) and two transverse siffeners( T-tiffeners) attached to the skin plate as shown in Fig. 5welding sequence and heat input of the five cases are sum-was used for the analysis Lsifeners. The panel is 3 000marized in the Table 1.mm long and 1 500 mm wide with 5 mm thickness. Theheight of the longitudinal siffeners is 300 mm, and that ofTable 1 Welding sequencethe transverse siffeners is 125 mm. The thickness of allWelding caseWelding sequencestiffeners is 5 mm.Tack weld all the parts, and then weld themCase Asimultaneously.①Weld the longitudinal siffeners and the transverseCase B stfener. ②Position the steners and the skin plate.③Weld the stfeners and the skin plate.①Weld the transverse sifeners and the skin plate.②Position the longitudinal sifleners and the skin plate.CaseC③Weld the longitudinal stfeners and the skin plate.④Position the Lstiffeners and the T-tifeners.Fig,5 Thin-plate panel structure⑤Weld the Lstifeners and the Tsifeners.2.1 Mesh division and boundary condition①Wweld the longitudinal siffeners and the skin plate.②Positin the transverse sifeners and the skin plate.The thin-plate panel structure model is subdivided in-"eners and the skin plate.toa finite element mesh as shown in Fig. 5. The skin中国煤化工:ners and T-sifners.plate, the longiudinal siffeners and the transverse siffen-TYHCNMHGandtheTffenere.ers are divided into 30x39=1 170,3 x39x3=351,2x30 x9 =540 elements, respectively. The number of inter-CaseEWelding sequence is the same as Case C, butthe heat input is 20% smaller than that of case C.face element between welded parts is 213. The total num-.156CHINA WELDING Vol. 14 No. 2 November 20053 Simulated results3.1 : Influence of welding sequenceThe final welding distortions of the four cases areshown in Fig. 6 - Fig.9, respectively. From these compu-ted results, it can be clearly seen that a small local buck-ling distortion occurred at the edge of skin plate along thelongitudinal direction in case A and case B, respectively ,but an obvious global buckling distortion happened at thefull skin plate in case C and case D, respectively. In oth-er words, even through the same heat input is employedFig.9 Welding distortion of case Dthe final welding distortion is different because of diferentwelding sequences. From Fig. 6 and Fig.7, it can be seenbuckling modes of case C and case D is same, but the de-that there is very small difference between the two cases.From Fig.8 and Fig.9, it can be understood that theflection distributions of the two cases are different.Fig. 10 shows the deflection distributions along line 1of each case. The deflection of case A is the smallest. Thedellection of case B is a litle larger than that of case A.From this figure, it can be clearly seen that the deflectionsof case C and case D are much bigger than those of casesA and B.Fig.6 Welding distortion of case AFig. 10 Delection on the line 1 of cases A ,B,C and DFig.7 Welding distortion of case BFrom the observations above , the infuence of weldingsequence on buckling distortion in thin-plate structure hasbeen clarified in this study.3.2 Infuence of heat inputIn this section, the influence of heat input on buck-中国煤化工1 shows the final welding:YHC N M H Ghe welding sequence wassame as that of case C, but the heat input was 20% smal-ler that of case C. It is clearly that welding distortion inFig.8 Welding distortion of case Ccase E is much smaller than that of case C. In otherPrediction of welding distortion during assembly process of thin plate structures157words, even though the same welding sequence is used,(1) Welding sequence strongly afets distortion dur-because of diferent heat input, the final elding distotioning the assembly of large thin-plate panel structure. Fromis much different.the viewpoint of buckling prevention, adoping an appro-priate welding sequence with a relatively high sifness candecrease tendency of buckling occurrence.(2) The welding residual stress of skin plate is main-ly governed by longitudinal and transverse shrinkage inweld zone. Because the shrinkage is nearly in proportionto heat input, to the extent possible, by minimizing thewelding heat input, the buckling distortion can be preven-ted.Fig. 11 Welding distortion of case EReferences[1] Tsai CL, Park s C, Cheng W T. Welding distortion of thin-4 Conclusionsplate panel stucture. Welding Joumal, 199, 78(5): 156sThe numerical investigation presented here demon-- 165sstrates the effectiveness of using numerical analysis to sim-[2] Deng D, Murakawa H, Ueda Y. Theoretical prediction ofulate the infuence of welding sequence on buckling distor-welding distortion considering positioning and gap betweention for large thin-plate panel structure. The conclusionsarts. Intemational Joumal of Offshore and Polar Engineer-that can be drawn from this investigation are as follows :ing, 2004, 14(2): 138 - 144中国煤化工MYHCNMHG

论文截图
版权:如无特殊注明,文章转载自网络,侵权请联系cnmhg168#163.com删除!文件均为网友上传,仅供研究和学习使用,务必24小时内删除。