Optimization of vacuum counter-pressure casting process for an aluminum alloy casing using numerical simulation and defect recognition techniques

Vacuum counter-pressure casting (VCPC) process is an effective forming process of producing components with complex thin-walled characteristics. In this study, a surrogate-based optimization method with five process parameters is used to reduce the shrinkage porosity defect during the VCPC process of a complicated thin-walled aluminum alloy casing. The prediction of the shrinkage porosity defect is implemented by using the numerical simulation of the VCPC process. In general, performing the optimization directly based on the casting simulation may not be practical due to the computational and time constraints. To mitigate this problem, the response surface model (RSM) constructed via computer experiments is employed to accurately approximate the functional relationship between the process parameters and the response variable which is defined as the area of defects in the defect image generated from the casting simulation. To measure the area of defects in the image, a histogram-based threshold segmentation approach is proposed to recognize and quantify the defects from the image. The optimization problem is subsequently formulated using a desirability function-based optimization method, and the objective function is estimated using the built RSMs. The final results indicate that a set of optimal process parameters with minimum defects can be obtained.