Numerical simulation and process optimization for producing large-sized castings

Research & DevelopmentAugust 2008Numerical simulation and processoptimization for producing large-sizedcastings*Wang Junqing', Sun Xun', Guan Yang', Wang Penghua', Li Hailan', Bai Limei,Sun Xinzhi?P. R. China)Abstract: 3-D velocity and temperature fields of mold fling and slidification processes of large-sized castingswere calculated, and the efficiency and accuracy of numerical calculation were studied. The mold flling andsolidification processes of large-sized stainless steel, iron and aluminum alloy castings were simulated by using ofnew scheme; their casting processes were optimized, and then applied to produce castings.Key words: numerical simulation; mold fling; soldification process; process optimization; stainless stee; alloyed ironand aluminum castingsCLC number: TP391.9Document code: AArticle ID: 1672-6421(2008)03-179-07( ince the 1960s there are 3 progressive periods for modeling most of the mould filling processes, Lipinski, Y ang and) and simulation of mold flling and solidification processes Sun considered the effects of turbulent flow on velocity andof castings. First, in 1988 3-D temperature fields were calculated temperature fields, and solved more equations than thosesuccessfully during solidification of steel casting . Second, .considered laminar flow only 5，but it needed quick andin 1995 3-D velocity fields of metal flow in mold fillingaccurate calculating schemes of 3-D velocity and temperaturewere calculated successfully by B. Sirrell of Birmingham fields. Powerful software was developed to increase efficiencyUniversity [5-9. Third, since 1990 a new task to model andand assure accuracy of numerical calculation. 24 million tonsof castings were manufactured annually in China, therefore,underway. In order to calculate nucleation in a macroscopic mesh,proper numerical simulations by using of appropriate softwareundercooling at each microscopic cell and the random numberare needed eagerly in order to produce sound and large-sizedof nuclei by Monte-Carlo method, the 3-D velocity, temperaturecatings eficiently.and concentration fields in the meshes of casting should becomputed. Some researchers tried to simulate 3-D microstructure1 Modeling equationsof superalloy and ductile iron castings 10-41. Meanwhile, foundryWhen turbulent flow effects on processing parameters aretechnicians are more interested in simulating mould flling, taken into account and time averaged method is being used,solidifiation process of castings and optimizing the processing the governing equations, including continuity, momentum,parameters to produce sound castings.volume fraction, energy, turbulent kinetic energy, dissipateSince Re Number≥2300 is the practical value in therate of turbulent kinetic energy equations, are as following:a(pu,)/ax, =0(1)o(pu,)Ot + 0(pu,u,)/ax, = a(u.0u,/ox )ax, - op/ax, +(u。0u,/ax,)/ax, +pg,(2aF/Ot +8(Fu,)ex, =0(3)E(pT)/8t + a(pu,T)/ax, = 0(u,8T/ax,)/ax, +qr .(4)a(pK)/8t + 0(pu,K)/ax, =(u4kOK/x,)/ax, +G-pε(5)a(p8)/ot + 8(pu,e)/ax, = 8(u20e/ar,)/ax, +e(fCG- f2C2pe)/K(6)*Wang JunqingMale, borm in 1942, Ph.D, Chief Engineer. Research interests:Some variables of中国煤化工fllowinnumerical simulation of casting process, foundry metallurgy andμ。MHCNMH Gengineering.μr =μ/P, +μ,/στ(8)Received: 2008-04-07; Accepted: 2008-06-10) 79CHINA FOUNDRYVol.5 No.3μK =μ+ μ,/σk(9)2 Program for the calculation andM。=μ+μ,/σ:(10)display of 3-D velocity andtemperature fieldsμ,=f,C,pK2/ε(11)2.1 Pre-processing and post-processingG=从, (Ou,/ax, + 0u./ax,)/(Ou,/ax,) (12)Pre-processing and post-processing programs were developed.fu =exp(- 3.4/(1+R./50)(13)It increases the enmeshment speed by 13 times by using newR。=K2/vε(14)algorithm; for example, 1.93 million meshes can be generatedf2=1-3. exp(R2)(15)in 5 seconds. The new program can read STL format file ofWhere:automatically; the inflow boundary condition can be set upRe - turbulence pulse Reynolds number;and changed by mouse indicator; post-processing may showρ - liquid density, kg/m';flling pattern, 3~D velocity vector variation, 3-D temperature划xj - coordinate, m;distribution and result of ll, p and T on any section of thell, u,- velocity, m/s;casting, and operation is quick and convenience. It is able tochoose 19 different kinds of materials to represent differentv - molecular kinematic viscosity, m'/s;attribution of meshes during numerical calculation of moldμ- turbulent viscosity, Pa .s;flling and solidification process of casting, which brings moreμe, lur, lIx, μe - generalized diffusion cofficient, Pa .s;practicality for engineering analysis 8.1- time, s; .gi - gravity acceleration, m/s';2.2 Fluid flow and heat transfer calculation byT- temperature, K;high resolution MINMOD schemep - pressure, N/m';In this study, high resolution MINMOD mesh with 5 pointsF- volume fraction, 0≤F≤1;was adopted to make the discretization of momentum andq。- heat source term, kg . K/s . m';energy conservation equations on unequal size of meshesK- turbulent kinetic energy, m2/s";improved VOF method was used to deal with the evolutionP,- molecular Prandl number;of free surface and SOLA technique was used to calculate theε - dissipate rate of turbulent kinetic energy, m'/s';velocity of fluid flow. Results showed that numerical accuracyG - the producing term of turbulent kinetic energy, Pa/s;coefficients: στ =0.9-1.0, σx=1.0, σ。=1.3, C =1.44,was increased, as shown in Fig. 1, false diffusion and virtualphysical phenomenon that generated from the numericalC2=1.92, C,=0.09,f= 1.0Based on these equations, with appropriate initial andcalculation were eliminated. The symmetrical distribution ofboundary conditions, 3- D velocity and temperature fields cantemperature was reasonable during mold flling process ofaluminum alloy plate casting.be calculated.Fig. 1 Mould flling simulation of plate casting by high resolution MINMOD scheme3 Exploration for increasing efficiencyis times cycles to get a new evolution of fuid). Figure 2 showedthe simulation results of Benchmark test by using of the newof numerical simulationalgorithm with Courant Number equal to 4, where C=uOt/Ox.Since large portion of computing time was consumed in the The simulated results were in agreement with the Benchmarksimulation of velocity field during mold flling, so manytest.efforts were devoted to increase the calculation efficiency.Figure 3 showed the velocity field simulation results ofVelocity field during mold flling was calculated by using ofa plate casting with inflow velocitv of 50 cms'. Time costa new algorithm named predictor-two step corrector-VOF, for the whole sim中国煤化工and 317 min,which applied the implicit algorithm and large time step δ onwhen Courant Nu1YHC N M H G0, respectively.calculation of velocity, pressure and temperature; and explicitComputing time can be shortened by 19% if Courant Numberalgorithm with a small time step δ/1n over the free surface range (n of 8.0 was used instead of 4.0. Although the adoption of a180Research & DevelopmentAugust 2008large Courant Number can lead to good simulation eficiencyTherefore, a suitable Courant Number should be selectedand the general tendency of fAuid flow is similar, some obvious carefully in order to speed up the simulation without much lostdifferences exist. The details can be found in Figs. 3 (b), (b)of engineering accuracy.and (c), (").(a) 0.52s(b) 0.77 s(c) 1.01 s(d) 1.49s(e) 1.74sFig. 2 Mould flling simulation of the Benchmark test by predictor-two step corrector-VOF scheme with C=4(a)0.44 s, 10%(b) 1.73 s, 40%(c) 3.45 s, 80%(a") 0.45 s, 10%(b') 1.73 s, 40%(c) 3.46 s, 80%Fig. 3 Velocity fields of mould flling by predictor-two step corrector-VOF scheme for plate casting (a)-(c): C=4.0, (a")-(c"): C=8.0Fig. 4 showed the temperaturefields result by using ofpredictor-two step corrector-Number is 4.0; the temperaturedistribution is symmetrical andreasonable. New algorithmpromotes calculation eficiencyand favors engineering中国煤化工application.MYHCNMH G_(a) 0.44 s, 10%(b) 3.45 s, 80%Fig. 4 Temperature fields of mould flling for plate casting181 .CHINA FOUNDRYVol.5 No.34 Numerical simulation andprocess optimization of large-sized castings4.1 Dual-phases stainless steel castingAim at producing dual -phases stainless steel impellercasting of 2,667 kg, numerical simulations for 3different casting process techniques were carried out.Figure 5 showed the solid geometry of the impeller Lcasting. Figure 6 showed the chill design and the(a) Chill design(b) Solidified time fieldsoldified time field of casting process 1. Figure 7showed the increased chill size to enlarge the coolingFig.6 Casting technique 1ability and the simulated result using casting process2. In both technique 1 and technique 2, there are someisolated regions near blade bottoms or middle parts ofthe impeller casting, therefore shrinkage or porositymay be formed in the corresponded regions. Thiscannot meet the requirements of high strength and nodefects in the connecting part of the blade root andthe main body of the impeller, so techniques 1 and2 were denied. Figure 8 showed the solid geometryand enmeshment of impeller casting with in-gate andLand enhancing feeding ability from inside of mainbody. Figure 9 showed the flow pattern, free surfaceFig. 7 Casting technique 2evolution and temperature distribution during moldflling.(a) Solid geometry of casting with feeders(b) EnmeshmentFig. 5 Solid geometry of impeller castingFig. 8 Casting technique 3中国煤化工(a) 5.52 s, 25%(b)16.55 s, 75%| CNMHGFig. 9 Flow patterns and temperature fields during mould flling of impeller casting for casting technique 3182Research & DevelopmentAugust 2008solidification process. From these results it can be sure that was 22 s and solidifying time was 3.26 h, both of these timesliquid metal flls the mold cavity steadily. This solidificationwere suitable for foundry operation. Figure 11 showed the sideprocess can meet the requirement of sequential solidification, and top views of the produced sound casting.(a) 352 s(b) 1,996 s(c) 4,344s .Fig.10 Solidified time field during solidification process of impeller casting for casting technique 34.2 Alloyed iron castingNumerical simulation of JT30-54alloyed iron air compressor blockcarried out to calculate temperaturefields under 3 different castingprocesses. Two typical results forthe casting technique 1 and 3 areshown in Figs. 12 and 13. For thecasting technique 2, the simulatedresult was between the casting ;technique 1 and 3.(a) Side view(b) Top view .Fig. 11 Impeller casting(a) Solid geometry(b) Riser and chills distribution(C) Defects predictionFig. 12 Numerical simulation and defect prediction under casting technique 1Figure 12 shows the solid geometry, riser and chill positions Large area of shrink中国煤化工，be formedin the mold, and also defect prediction of the casting underin this region. LeakagYHCNMHGeproducedcasting technique 1. As shown in Fig 12 (C), there exists a largecasting was under pressurizea (sung. Il uemonstrated thatrange of isolated solidified region in the middle part of casting. the simulated results were reliable. Figure 13 shows the solid183CHINA FOUNDRYVol.5 No.3R8r(a) Solid geometry(b) Riser and chills distibution(C) Defects predictionFig. 13 Numerical simulation and defect prediction under casting technique 3geometry, riser position, chill distributions in the mold and casting, as shown in Fig. 14, was machined and pressurizedalso defect prediction of the casting under casting technique 3.at 0.6 MPa, no leakage was detected. The casting satisfied theCompared with those of casting technique 1, casting technique requirements in specification.3 enhanced the cooling ability by increasing both the quantity4.3 Aluminum castingand size of the chills, and thus reduced the isolated volumesignificantly. The isolated volume distributed separately anNumerical simulation of large thin walled ZL1 14A aluminumno linked porosity defects can be formed. The produced blockmould temperature of 25 C (no-bake furan resin bonded sand)and infow velocity 20- -50 cm-s' was carried out to calculatevelocity and temperature fields under 2 different castingprocesses by using of low pressure casting method. The solidgeometry and enmeshment of aluminum casting with 15.03million meshes are shown in Fig. 15 (a). There was a misrundefect if low inflow velocity of 20 cm:s' was used, as shownin Fig. 15 (b). No misrun was appeared if inflow velocity of50 cm's' was used, as shown in Fig. 15 (c), in this case thetemperature fields during solidification process in which thewhole casting solidified in about 24 min. So final castingprocess parameters adopted were that the time of increasingpressure was 44 s, inflow velocity at elevated liquid metal tubewas 50 cms', holding pressure time was 24 min. By usingthese parameters, sound casting was produced, as shown inFig. 14 Block casting of an air compressorFig. 15(f).一中国煤化工(a) Enmeshment of aluminum(b) Misrun in mould flling withYHCNMHGuldflingwthcasting with 15.03 millioninflow velocity 20 cm:s'inflow velocity 50 cm-s' 'meshes184Research & DevelopmentAugust 20085 Conclusions(1) Time averaged governing equations,include continuity, momentum, volumefraction, energy, turbulent kinetic energy,dissipate rate of turbulent kinetic energyequations were established. It is able tocalculate the velocity and temperature fieldsfor laminar and turbulent flow during moldflling and solidification process.(2) The developed computer program can(d) Temperature field at solidifed 20%(e) Temperature field at solidified 60%read STL file, enmeshing solid geometry,displaying velocity vector, 3-D velocity andtemperature fields, flling pattern, temperaturedistribution of mold flling and solidificationprocess.(3) The numerical simulation can be usedto predict the casting defects and to optimizethe casting process parameters to producelarge-sized sound stainless steel, alloyed ironand aluminum castings.References1] Morroen R E, Wilkes J 0 and Pehlke R D.Numerical Simulation of Solidification. CastMetal Research, 1970(6): 184- _192.2] Niyama E, et al. Predicting shrinkage inLarge Steel Castings from Temperature(f) Produced sound aluminum castingGradient Calculations. AFS, 1981(6 ):16- -22.3] Zhang Yi and Wang Junqing. NumericalFig. 15 Numerical simulation of large thin walled aluminum casting by lowSimulation of Solidification Process ofpressure cast methodCastings. Foundry, 1980(1): 10- -16. (in Chinese)[4] Giamei A F and Abbaschian G J. Hammer Casting Simulation12] Lesoult G Catro M and Lacaze J. Physical Modeling of the- Round Rubin. Proceedings of Modeling Casting andSolidification of Spheroidal Graphite lrons: Revision of theWelding and Advanced Solidification Processes IV, Palmcoupled zone concept. Proceedings of Modeling Casting andCoast, Florida, USA, 1988: 775- -861.Welding and Advanced Soldification Processes VII, USA,[5] Stoehr R A. Simulation in the Design of Sand Casting.[13] Catro M, et al. Prediction of the Soldifcation Microstructure1998: 479.Proceedings of Modeling Casting and Welding Processes,Rindge, New Hampshire, USA, 1980: 3 -18.of Hypo- and Hyper- eutectic Spherical Graphite Cast lrons.6] Stoehr R A and Hwang W s. Modeling the Flow of Molten MetalProceedings of Modeling Casting and Weldling and AdvancedHaving a Free Surface during Entry into Molds. Proceedings ofSolidification Processes VII, USA, 1998: 487.Modeling Casting and Welding I, USA, 1983: 47 -58.14] Zhang Yutuo, Wang Junqing et al. Simulation the Dendritic7] Stoehr R A, Wang C, Hwang W S and Ingeslev P. ModelingGrowth of Pure Material with the Phase Field Method.the Flling of Complex Foundry Mould. Proceeding of MdelingFoundry, 2003, 52(11): 1091-1095. (in Chinese)Casting and Welding Processes I, Santebarbara, California,15] Lipinski D M, Schseter W and Flender E. Numerical Modeling ofFlling Sequence and Soldification of Castings. Proceedingsof Modeling Casting, Welding and Advanced Slidification8] Wang Junqing, Hansen s F and Hansen P N. 3-D Modelingand Simulation of Mould Fllig Using PC'S. Proceedings ofModeling Casting and Welding Processes IV, Palm Coast,[16] Yang Bingqian et al. Equivalent Laminar Model for 3-D VelocityField of Turbulent Flow in Crytallizer of Continuous Casting.Florida, USA, 1988: 741-753.Sirrel B, Holliday M and Compbell J. The Benchmark Test 1995.Jourmal of Xi an Jiaotong University, 1996, 30(4): 79 -85.Proceedings of Modeling Casting and Welding and Advanced17] Sun Xun et al. Numerical Simulation of Mold Filling andSlidficaton Poesses VI, England, 195: 915 933Solidification Process for Permanent Mold Casting.10] Zhu P and Smith R W. Dynamic Simulation of Crystal Growthby Monte Carlo Method-I Model Description and Kinetics.Solidification Processes IX, Aachen, Germany, 2000: 389- -396.Acta Metal. Mater., 1992, 40(4): 683- -692.18] Guan Yang, et al. Research on Preprocessing and Processing11] Li Dianzhong, Su Shifang, Xu Xuehua and Wang Junqing.of Mold Flling and Slidifcation Process of Casing. Foundry,Modeling of Microstructure Formation for Nickel Based Super2004, 53(11): 905中国煤化工Alloy Turbine Blade Casting. Proceedings of 61st World [19] Tao Wenquan.ICN M H Gerical HeatTransfer. Beiing:GeFoundry Congress, Beijing, China, 1995: 63-74.185