Research On The Micro-Modelling Of Metal Solidification Process

Simulation of Microstructure Formation (SMF) is also called Micro-Modelling. It is relative to Macro-Modelling during solidification. To be specific,it means the simulation of solidification process at the scale of grain.Two models have been employed to calculate the concentration fields:a) Modified Scheil model. That is to modify the solute equilibrium partition coefficient k with a coefficient ft to take account of the convection in the front of S/L (Solid/Liquid) interface,b) Solute model based on Micro-Modelling. The solute diffusion is simulated by DFDM,and the solute redistribution in the S/L front is calculated by actual solute partition coefficient model. The S/L phase change is judged according to the results of Micro-Modelling directly. The segregation in grain and boundary can be simulated well by this model.Simulate the grain nucleation and growth by using Monte Carlo (MC) method. MC method is based on the law of the lowest energy and the effects of bulk free energy and interfacial energy are taken into account here. In addition,the anisotropy intensity factor is used to take account of the preferred growth direction of grain. Just by so,the results of the simulation are more agreement with that obtained experimentally.A modified Cellular Automaton (CA) model has been developed based on the assumption:grain shape is a quadrangle and has the preferred growth direction. By judging the relation of position between dot (element) and quadrangle (grain),develop a simple evolution rule of grain growth,which overcome the drawback of Rappaz's model. Here,the dentritic tip velocity can be calculated by KGT model.By coupling macro-thermal transfer and Micro-Modelling,combining the deterministic model and probabilistic model,develop a novel CAMC model for the first time. CAMC model exploits the fully advantages of MC methodand CA model. The skeleton of grain is first described by using CA model,and then,the growth inside of grain is corrected by MC method. The evolution of amount,dimension and appearance of grains can be described quantitatively by using CAMC model. Therefore,it is very successful to simulate the microstructure formation during metal solidification by CAMC model.It's the first time to develop Local Area Simulation (LAS) method to reduce the amount of calculation of Micro-Modelling by only choosing local area to simulate. The LAS method includes three parts:mapping the simulation data belonging to macro-scope onto the local area,dealing with the boundary of the local area and magnifying the graphical results of the local area. In addition,it's possible to do Micro-Modelling of the larger ingot by using LAS method with paste method.By tracking the S/L front,GET (Columnar to Equiaxed Transition) interface can be captured easily. In addition,Random Sampling based on Database Technique (RSDT) is used hi this dissertation. RSDT is to choose a number by a random sampling process,and the number,which corresponds to a certain calculation cell,is just a primary key index in database. Therefore,the application program can exchange data with database through the unique number. RSDT can effectively save computer memories because that data are not stored in memory but in database.Experimental study on the solidification process of the Al-4.5%(mass)Cu alloy is used to validate the models of this dissertation. The experiment parameters are the pouring temperature and the cooling water flux. The results of simulation are compared to the experimental results. It shows that the models used to Micro-Modelling can meet the targets.