Finite Element Modeling of TIG Welding for 316L Stainless Steel Plate using Sysweld

Transcription

1 Volume-5, Issue-2, April-2015 International Journal of Engineering and Management Research Page Number: Finite Element Modeling of TIG Welding for 316L Stainless Steel Plate using Sysweld V. V. Narayanareddy 1, D. Srinivasa Rao 2, M. N. V. Krishnaveni 3, M. Amareswarireddy 4 1,2,3,4 Assistant Professor, Department of Mechanical Engineering, Anits Engineering College, Visakhapatnam, Andhra Pradesh, INDIA ABSTRACT In this present work numerical simulation of thermal phenomena during welding of 316L stainless steel plate. Different bead-on-plate weld trials conducted on 3 mm thick plate with Amp heat input at a different welding velocity from mm/min. On the basis of experimental results the influence of welding current and speed on the shape and size of fusion zone in bead-on-plate is compared with simulation. A three dimensional Double ellipsoidal heat source model and temperature dependent thermo physical properties of 316L steel are used for thermo-mechanical analysis of TIG welding in SYSWELD software. Based on bead-on-plate experiment and prediction correlation an optimized parameter for 3 mm thick used for final weld plate. During welding of final weld plate thermal cycles are measured using thermocouple further compared with finite element analysis results. Amount of heat power density distribution is predicted using SYSWELD is compared with analytical density in MATLAB. An isotherm curve distribution in molten zone pool for different weld velocity for a heat input variation is predicted. Keywords L stainless steel, thermal analysis, Fusion zone, SYSWELD, heat power density, TIG welding. I. INTRODUCTION The process of welding is an integral part of manufacturing of nuclear components and has a direct influence on the integrity of the components and their thermal and mechanical behaviour during service. In fusion welding process due to localized heating and cooling, residual stresses and distortion are inducing in the welded component [1-2]. Many investigators have developed numerical models for studying the effect of welding process on the welding characteristics 316L stainless steel is a regular choice in power plant and super critical nuclear application. Many researchers published work on thermo mechanical analysis of stainless steel welding like, D. Deng proposed finite element model for welding residual stress analysis in ABAQUS [4,5]. A.H. Yagi measured residual stress and thermal data for thick cylinder using multi pass welding by element birth technique in ABAQUS.From literature it is observed that number of researchers have been working on analytical and computational welding analysis of GTA welding process for thick pipe and plate austenitic steel but for thin plate of austenitic stainless steel material still need some serious attention. During welding the interaction [6,8,9] of the heat source and the parent metal leads to rapid heating and melting with vigorous circulation of molten metal in the weld pool. This thermal cycle plays an important role in determining weld geometry, HAZ dimensions, weld metal and, residual stress and deformation. Undesirable affects from thermal cycles of welding may occur in the zone of solidified metal (weld metal) or in the heat affected base metal (Heat affected zone). Localized melting or amount of fusion penetration in to a joint being welded is controlled mainly by heat input, but other subtle features like weld speed and arc length also can exert an effect. The peak temperature reached by the weld area can determined the extent of grain growth in the heat affected zone, as well as that amount of softening that may take place in the heat affected zone of work-hardened base metal[7,11]. During GTA Welding weld plate interacts with arc for a short period of time only and it comes under very high heating and cooling cycles, during this short time plate is subjected complex thermal cycles. This condition makes welding process in to more complicated for getting desired solution using analytical modelling. Therefore numerical technique is a preferable option for studying complex process with small increment time interval. The problem of distortion, residual stresses, and reduced strength of a structure in and around a welded joint are a major concern of the welding industry for decades. These problems primarily result directly from the thermal cycle caused by localized intense heat input in arc welding process [13,14.15]. Hence for the development of effective simulation strategy for weld analysis, the accurate prediction of 390 Copyright Vandana Publications. All Rights Reserved.

2 thermal history is of key importance. Finite element analysis gives capability for modelling of temperature distribution nearer to welding and heat affected zone in terms of temperature and molten zone. II. NUMERICAL SIMULATION The weld plate model is considered as a solid body. A moving heat source model is made in the meshing file to present the heat generated by the torch in the TIG welding process. The Goldak s double-ellipsoid heat source model is adopted to calculate volumetric heat flux distributions as heat input around the welding pool. The heat source distribution as shown in Fig 4 combines two different ellipses, i.e. one in the front quadrant of the heat source and the other in the rear quadrant. The power densities of the double-ellipsoid heat source, q f (x, y, z) and q r (x, y, z), describing heat flux distributions inside the front and rear quadrant of the heat source.[3,10] Gaussian distribution of power density (W/m ) in an ellipsoids amount of heat applied per unit volume. The heat density Q(x, y, z) at an arbitrary point within the front half ellipsoid and the rear half ellipsoid is described by the following equation 6 3fQ f Qxyzt (,,,) = e e e a bcπ π f [ τ ] 2 f 3 x 2 / a 2 3 y 2 / b 2 3 z+ v( t) / c fQ r r z v t c Qxyzt (,,,) = e e e a bcπ π f [ τ ] x 2 / a 2 3 y 2 / b ( ) / 1 Where a f, a r, b, c are the ellipsoidal heat source parameters, a f is front length of molten pool, a r is rear length of molten pool, c is depth of penetration in all cases and b is half width as shown in fig 1. Q is the arc heat input, η is welding efficiency, V and I are arc voltage and current, f f and f r proportional coefficient at front and rear ellipsoid of the heat source respectively, such that f f + f r =2, and V is the welding speed. For all weld analysis double-ellipsoid parameters are measured at 75% welding efficiency (η). III. EXPERIMENTAL WORK Fig 2 shows the final weld plate experimental setup with backup plate same set up is used for all beadon-plates and final weld plate welding. Temperature history nearer to weld line is recorded using K type thermo couple with sampling length 2 cycle placed at 5, 10, 12, 15 and 18 mm away from weld line as shown in Fig 2. During welding thermal cycles history is recorded in data logger. Experimental heat input paramters used for bead-onplate and obtained weld length and width are given in table1. Bold paramters are used for final weld plate. Experimental work on Automatic GTA Welding machine for 3 mm thick 316L stainless steel bead-onplate is carried out at a different heat input for a different weld speed. After welding, weld sample is prepared using cutting machine and roughly polished with flexible abrasive paper to remove surface impurities and irregularities, further cleaned with acetone. Fig 2 Weld Plate Experimental Setup 391 Copyright Vandana Publications. All Rights Reserved.

3 S.no Current (amps) Voltage (volts) Welding Speed (mm/min) Weld length(mm) Width b(mm) Arc length(mm) Table 1 Heat source fitting parameters for different heat input. Fig 3 shows bead-on-plate nugget profile for a different heat input at weld speed of 140 mm/min, Heat source fitting parameters for double ellipsoidal model (Goldak) are weld length, depth of penetration and weld width for each bead-on-plate measured using optical microscope as shown in Fig 3 (c). (a) 160 Amps (b) 130 Amps (C) 140Amp (d 120Amp Fig 3. Bead On plate for different current at 140 mm/min weld velocity 392 Copyright Vandana Publications. All Rights Reserved.

4 IV. HEAT SOURCE FITTING The first part of the sequentially coupled elastic-plastic thermo-mechanical analysis is discussed here in the present study and the effects of various heat source and welding process parameters on transient temperature distributions, boundary and shape of FZ and HAZ are presented in detail. Weld profile obtained by FEA is compared with result by optical microscope for each bead-on-plate with minimum error as shown in Fig 4.The shape of molten zone profile is almost same with bead shape profile observed in experiment. It is evident that there is good agreement between these two sets of results with minimum error. Fig 4 Experiment Weld profile with Simulated Result Fig 5 (a-d) shows the cross section of weld profile comparison for experimental and finite element analysis result at Amp current range for welding speed 140 mm/ min at 1.5 mm arc length, there was a good agreement between experimental and Fea model. (a) 160 Amp (b) 140 Amp (c) 130 Amp (d) 120 Amp Fig 5 Comparison of Experimental Weld profile with Simulated Results at different current V. HEAT POWER DENSITY Fig 6 shows the heat power density comparison of weld nugget profile as obtained using finite element analysis, MATLAB (analytical) and experimental results. Heat power density is the amount of heat concentrated on welded plate for a particular heat input. For checking accuracy of heat power density by finite element analysis using Goldak s heat distribution model as shown in Fig 6 (a) is compared with analytical results in Fig 6 (b) for 160Amp current at 140 mm/min weld speed. The heat power density is calculated by self developed MATLAB code based on simple heat input power calculation using double ellipsoidal model. Heat power density is estimated to be 45w/mm 3 by MATLAB and 47w/mm 3 by finite element analysis. Goldak s Heat source distribution model Fig 6. Heat Power density at 160 Amp 393 Copyright Vandana Publications. All Rights Reserved.

5 VI. RESULT AND DISCUSSION 6.1Thermal Analysis of Final weld plate Double ellipsoidal model parameters measured by experiment for 160 Amp current at 140 mm/min weld speed are a f = 2.8 mm, a r = 3.8 mm, b = 5.3 mm, c = 3 mm. Using these parameters, thermal analysis of final weld plate is presented in Fig 13 at various stages of welding. It is evident from the figure that the temperature around the torch reaches 1510 o C suggesting melted material in the fusion zone (FZ). Next to fusion zone, the presence of higher temperatures indicates the presence of heat affect zone. A small area ahead of the heat source also shows comparatively higher temperature because of the front parameter of heat source model. The heat input from the heat source to the weldments gradually transferred to the rest of the base plate in all directions due to conduction, convection and radiation phenomenon. The peak temperature obtained during welding simulation is 1640 o C. Fig 7 (a-b) shows the higher temperature gradient is located at the welding line. The boundary of molten pool is defined by isotherm representing 1640 o C. (a) Initial Stage of weld at 18 sec (b) Middle Stage of weld at 41 sec Fig 7. Finite elment Thermal analysis of welding Fig 8 Temperature distribution in transverse direction Fig 9 Position of thermocouple (a) 18 mm (b) 15 mm (c) 12 mm 394 Copyright Vandana Publications. All Rights Reserved.

6 (d) 10 mm (e) 5 mm Fig 10 Temperature evolution at 5 characteristics location from weld line. Thermal history for Final weld plate is recorded using five thermo couple at a marked characteristic location during GTA welding process. Position of five thermocouples on welding plate is shown in fig 9. Thermal history during welding process is recorded up to room temperature using five thermocouple placed at 5 mm, 10 mm, 12 mm, 15 mm and 18 mm from weld line on the top of weld plate. Fig 10 (a-e) shows the comparison of measured thermal cycles using thermocouple with finite element thermal analysis results. The peak temperature obtained by thermocouple at 5 mm distance from weld line is1052 o C and thermocouple reading is 1039 o C temperature. The peak temperature for 10 mm distance from weld line is by thermocouple is 714 o C and FEA is 704 o C, only 10 o C difference is observed. Similarly for 12, 15 and 18 mm distance peak temperature of simulated and experimental have good agreement. 6.2 Cooling rate Comparison Finite element analysis of cooling rate for Final weld plate at 160 Amp current for 140 mm/min weld speed is shown in Fig 11. The maximum heating and cooling rate is taken in the quasi-steady state condition. From the temperature rate the heating and cooling rate observed are 222 C/s, and C/s respectively. It has no much change during quasi-steady state and it can be considered as maximum in the welding process. The heating and cooling rate obtained by experimental work is compared through FE analysis result for 5 to 18 mm distance from weld line. Fig 11 Cooling rate at 160 Amp 6.3Isotherms for different weld speed Fig 12.(a c) shows the isothermal lines of the weldment surface temperature for 120 Amp at 120 mm/min, 140 mm/min, 160 mm/min and 180 mm/min welding speeds. Increase in weld speed is reducing the peak temperature obtained during welding. Weld pool size, weld width, weld depth and the heataffected zone are also decreased for constant welding parameters. Welding pool becomes large at welding speed 120 mm/min, the bottom layer remelts in welding upper layer because of lower welding speed and local overheat. The welding pool is small and shallow with incomplete penetration for 180 mm/min weld speed, so 140 mm/min is chosen for final weld plate. It can be concluded that a few changes of welding speed have great impact on welding pool formation; temperature gradient and distribution in weldment. 395 Copyright Vandana Publications. All Rights Reserved.

7 (a) 120 mm/min (b) 140 mm/min (c) 180 mm/min Fig 12. Isotherm Lines at 120 Amp Current input VII. CONCLUSIONS In this study, a systematic procedure followed to perform the thermal analysis during the autogenous TIG welding of 316L stainless steels. A Gaussian distributed moving heat source model based on Goldak s heat source model is implemented in SYSWELD and experimentally validated for the 3D finite element simulations of arc welding process. Based on the research work it is conclude that the variation in total heat energy input to the metal i.e. variation of welding process parameters primarily welding current and weld speed significantly affects both the boundaries (magnitude) and shape (heat distribution) of FZ and HAZ along with affecting the peak temperatures in both the zones. Other important investigation here is that the peak temperature which is depended upon the total heat input is also varied. In Study the effects of welding process parameters on the temperature profile, fusion and heat affected zones. Different heat inputs are used by varying the welding process parameters. In the numerical model the parameters are calibrated based on the beadon-plate experiments. Calibrated weld fusion zone profile by optical microscope has been compared with heat source fitting tool with a good agreement. Numerically observed fusion zone isotherm profile is validated by thermocouple result which has good agreement. Measured thermal cycles has been compared with Finite element thermal analysis result at 5 mm, 10 mm, 15 mm and 18 mm distance form weld line. It is noticed that the temperature follow a non-linear trend which is caused by the non-linear material properties used in the process. The heating and cooling rates has been validate at various instance in the quasi-steady state condition. Changes in heating and cooling rate have been compared for both with and without preheat welding. REFERENCES [1] Anca,A.Cardona,Risso,Fachintti,V.D.Finite element modelling of welding process. Applied Mathematical Modelling 2011,35, [2]Edwin raja Dhas,Kumanan S Modeling of residual Stress in butt welding.material and manufacturing processes 2011,26, [3] Shaodong Wang, John Goldak, Jianguo Zhou, Stanislav Tchernov, Dan Downey Simulationon the thermal cycle of a welding process by space time convection diffusion finite element analysis International Journal of Thermal Sciences 48 (2009) [4] Hidekazu Murakawa, Dean Deng, Ninshu Ma, Jiangchao Wang Applications of inherent strain and interface element to simulation of welding deformation in thin plate structures Computational Materials Science 51 (2011) [5] H. Long, D. Gery, A. Carlier, P.G. Maropoulos Prediction of welding distortion in butt joint of thin plates Materials and Design 30 (2009) [6] Dean Deng, Hidekazu Murakawa Influence of transformation induced plasticity on simulated results of welding residual stress in low temperature transformation steel Computational Materials Science 78 (2013) [7] Liam Gannon Yi Liu, Neil Pegg, Malcolm Smith Effect of welding sequence on residual stress and distortion in flat-bar stiffened plates Marine Structures 23 (2010) [8] Olivier Asserin, Alexandre Loredo, Matthieu Petelet, Bertrand Iooss Global sensitivity analysis in welding simulations What are the material data you really need? Finite Elements in Analysis and Design 47 (2011) [9] M.J. Attarha, I. Sattari-Far Study on welding temperature distribution in thin welded plates through experimental measurements and finite element simulation Journal of Materials Processing Technology 211 (2011) [10] N. T. Nguyen, A. Ohta, K. Matsuoka, N. Suzuki, Y. Maeda, 1999, Analytical Solutions for transient temperature of semi-infinite body subjected to 3-D moving heat sources, Welding Journal, pp.265s-274s. [11]. Wang JC, Ma Ns, Murakawa H. Prediction and measurement of welding distortion of a spherical structure assembled from multi thin plates. Mater Des 2011;32: [12].Numerical simulation of welding distortion in large structures Dean Deng a,*, Hidekazu Murakawa b, Wei Liang b Comput. Methods Appl. Mech. Engrg. 196 (2007) Copyright Vandana Publications. All Rights Reserved.

8 [13] Deng, D., Murakawa, H., Liang, W., 2007b. Numerical simulation of welding distortion in large structures. Comp. Methods Appl. Mech. Eng. 196 (45 48), [14].Nishikawa, H., Serizawa, H., Murakawa, H., Actual application of large-scaled FEM for analysis of mechanical problems in welding. Quart. J. Jpn. Weld. Soc. 24 (2), [15].Ueda, Y. Yamakawa, T, Analysis of thermal elastic-plastic stress and strain during welding by finite element method. Trans. Jpn. Weld. Soc. 12 (2), Copyright Vandana Publications. All Rights Reserved.