Multi-field Coupling Models And Deformation Prediction Of Aluminum Alloy Large Complicated Workpieces During Heat Treatment

Al-Si series are the most widely used casting aluminum alloy. ZL114A alloybelongs to Al-Si series and it has been widely applied in aerospace, automotiveindustries and other aspects because of its excellent castability, good corrosionresistance and high strength-to-weight ratio. ZL205A belongs to Al-Cu series and itis another alloy that has been widely applied in aerospace industries. The castingaluminum alloy workpieces are large, integration and complex due to the rapiddevelopment of aerospace industries. Heat treatment is very important because it isthe last operations for hot processing of metals. Prediction and control thequenching residual stress and distortion of large complicated workpieces areurgent problems that need to be solved in aerospace industries. The factors thatinfluence the quenching residual stress and quench distortion include the structurefeatures and size of workpieces, the thermo-physical properties and mechanicalproperties of materials, the cooling performance of the quenching medium, thepatterns of workpieces immerge into quench tank, boundary conditions andhuman influence, etc. But nowadays a comprehensive systematic research on thequenching distortion of large workpieces is seldom reported.The solid solution temperature of ZL114A and ZL205A alloy was determinedusing differential scanning calorimetry analysis and metallographic analysis. Thesolid solution temperature of ZL114A and ZL205A alloy are535-550℃and540-550℃, respectively. Stress-strain curves of these two alloys were obtainedthrough thermal simulation deformation experiments. At constant strain rate, theflow stress decreases with increasing the deformation temperature. For thedeformation temperature range of300-500℃, at constant temperature, the flowstress increases with increasing the strain rate. While at low temperatures(＜300℃),at constant temperature, the influence of strain rate on flow stress is not obvious.According to the stress-strain curves, the constitutive equations of these two alloyswere established at temperature range of350-500℃. Explicit expression for yieldstress was given and four variables (SYIELD, HARD (1), HARD (2) and HARD (3))were programmed in the subroutine UHARD with ABAQUS. The constitutivemodel was complied on ABAQUS platform.Based on the quenching medium cooling characteristics curve, using inverseheat conduction method, the heat transfer coefficients of the medium were obtained. Through compare the simulation data and experimental results, the accuracy of theheat transfer coefficients was verified.The heat transfer coefficients of the mediumare the function of temperature, and the peak coefficients appear in the entiretemperature range. Temperature is the sensitive factor for heat transfer coefficientsof water and5%UCON-A, but it is not sensitive for No.20machine oil.Based on the FLUENT software, the simulation models of two types of quenchtank were established. As for the tank without mixing system, the velocity ofmedium in middle zone of tank increases with increasing the inlet velocity. But thecurve peak does not appear regularity. As for the tank with mixing system, roundedtransition in the outlet of draft tube can significantly reduce the loss of medium flowvelocity. The medium flow velocity increases with increasing the rotation speed ofstirrer propeller, but the uniformity of flow field decreases with increasing therotation speed. The flow velocity increases with increasing the propeller diameter.The flow velocity and flow field uniformity can be adjusted through moving theposition of guide plate. On the basis of verification for accuracy and reliability ofquenching tank modeling and simulation accuracy, the methods of FLUENTsimulation combine with artificial neural network (ANN) was proposed. Thepropeller speed, the position of guide plate and the position of x-coordinate inquench tank were used as the input variables, while y-velocity of medium was takenas the output. The results show that the mean relative error between the predictedy-velocity of the ANN model and the FLUENT simulation data is4.24%, whichdemonstrates that the ANN model is able to predict the y-velocity of medium withhigh precision.Based on ABAQUS software and according to the structural features ofaluminum alloy typical axsymmetric part, the boundary and initial conditions wereset, and the finite mesh elements were divided. The coupled thermo-mechanicalsimulation of axsymmetric part during heat treatment process was established. Thetemperature field of axsymmetric part during heating, quench transfer andquenching stage were simulated. The residual stresses of axsymmetric part afterquenching were analyzed. Through using the quenched stage of axsymmetric part asthe initial stage of tensile deformation, the finite element (FE) model of tensiledeformation was established. The coupled thermo-mechanical models of threealuminum blocks from publications were established to simulate the quenchingprocess of blocks, then the FE model of tensile deformation of the quenched blockswere established. By comparing the simulation results from this paper and thepublications data. The accuracy and reliability of coupled thermo-mechanical for simulation of heat treatment of aluminum alloy was verified. The feasibility of FEmodel for simulation of metal small deformation was verified.The heat treatment process of aluminum large workpieces was calculated usingmulti-field coupling simulation method. Influences of some factors on quenchedresidual stress and distortion were investigated. These factors include boundaryconditions, solid solution temperature, wall thickness, quench medium and theirstemperatures, time of workpiece immerse into quench tank and stage of quenchmedium. Boundary conditions are sensitive factor for quenched distortion of largeworkpieces. Solid solution temperature is not a sensitive factor for quencheddistortion. The quenched residual stresses of workpieces decrease with increasingthe quench medium (water or5%UCON-A) temperature. The influences of wallthickness on maximum residual tensile stress are more obvious than the maximumresidual compression stress. Time of workpiece immerse into quench tank is asensitive factor for quenched residual stress and distortion of large workpieces. Themaximum residual tensile stress of x-and y-component increase with increasing thestirrer rotation speeds. When the rotation speeds is0-600rpm, the rotation speed isnot a sensitive factor for the maximum residual tensile stress of z-component, whilethe rotation speed is900rpm, the maximum residual tensile stress of z-component islower than the other rotation speed conditions.Finally, the simulation quenched distortion of actual large complicatedaluminum workpieces were verified by the actual measurement. For the1st actualworkpiece, when water is25and80℃, the average relative error between thesimulated distortion and measured values on two characteristic positions ofworkpiece is13.4%. For the2nd actual workpiece, when water is25and45℃, theaverage relative error between the simulated distortion and measured values ofcharacteristic position is7.6%. For the3rd actual workpiece, when water is60℃,the relative error is8.8%. As for the2nd and3rd actual workpiece, the absoluteerror between the simulated distortion and measured values is approximate0.5mm.In summary, the numerical simulation model of large complicated aluminumworkpieces during heat treatment process is able to predict the quench distortionwith high precision.