Engineering computing and data-driven for gating system design in investment casting

Abstract

The geometrical shape of the gating system and the casting process are vital to the occurrence of shrinkage porosity, and the effects of their combination directly determine the desired high-quality casting. Casting simulation plays a significant role in the deduction of the trial and error method for gating system design in investment casting, which is time-consuming. A breakthrough in optimization design often requires a comprehensive understanding of the geometrical structure of the gating system and the complex process parameter to generate a novel data-driven framework that improves the higher quality of casting and reduces the period of design. This paper proposed a data-driven framework to investigate the optimal gating system design integrated with complex casting processes through RBF optimization algorithm with the target of porosity-free casting as well as high casting yield percentage. The results demonstrate that the optimization of the gating design is more inspiring with high accuracy and high efficiency, and the casting yield is increased by 14.91%. Furthermore, the verification between simulation and experiments shows good agreement. The demonstrated data-driven approach is feasible and effective and can be extended not only incorporating the geometrical structure of the gating system and casting processes in investment casting optimization but also other potential types of casting design.