Research on neural network genetic algorithm optimization in the preparation of CuSn10P1 semi-solid slurry with the fully enclosed melt-constrained cooling inclined plate

Abstract

The application of semi-solid metal forming technology to the production of high-quality copper alloy parts has attracted more and more attention. Accordingly, as the key and foundation of semi-solid metal forming technology, it is particularly important to stably and efficiently prepare copper alloy semi-solid slurry with high quality. The purpose of this study was to innovatively optimize the process parameters by combining multiple optimization algorithms to stabilize the preparation of high-quality CuSn10P1 semi-solid slurry. And the self-developed Fully Enclosed Melt-Constrained Cooling Inclined Plate (FEMCIP) device was taken as the experimental equipment. Based on the optimization method of neural network genetic algorithm, the parameters of the preparation process of CuSn10P1 semi-solid slurry were optimized by combining orthogonal experimental design, grey correlation analysis method and Latin hypercube sampling, and using the finite element analysis software ProCAST and data analysis software MATLAB with the average shape factor as the evaluation index. By comparing with the slurry before optimization, numerical simulation results and experimental verification results, it was found that the semisolid slurry prepared by the crucible with preheating temperature of 990°C, cooling water flow rate of 680.517 L/h, cooling channel angle of 45°, inclined plate length of 300 mm and quasi-isothermal time of 42.715 s had the best quality. In this research, the semi-solid slurry of CuSn10P1 alloy with uniform microstructure and high roundness of solid particles was prepared by using this optimization method, which provided a new optimization strategy for the preparation of semi-solid metal slurry by using inclined plate type equipment.