Intelligent Process Modeling and Optimization of Porosity Formation in High-Pressure Die Casting

In this paper, are presented design and implementation issues of predictive models developed for improving the quality of aluminum die castings by minimizing scrap due to porosity. A predictive model for porosity of casting parts is created using fuzzy systems optimized by genetic algorithm and simulated annealing. High-pressure die casting is a complex process that is affected by a large number of process parameters with influence on casting defects such as porosity. In this study, porosity of casting parts is expressed as a function of counter-pressure, first phase velocity, first phase length, second phase velocity, first cooling period, and second cooling period. It was found that the developed GA- and SA-based fuzzy systems have great predictive capability of porosity in die castings. The second objective of this work was to obtain a group of optimal process parameters leading to minimum porosity in high-pressure die casting using genetic algorithm and simulated annealing as optimal solution finders. The optimal parameters were validated experimentally, and the castings with minimum percentage of porosity were achieved.