Multi-Objective Optimization of Casting Parameters for Mn70Ni25Cr5 Alloy Using ProCAST Simulation and Response Surface Methodology

Authors:
Open Access
Publisher website
Google Scholar
Abstract
To simultaneously suppress shrinkage-related defects and refine the solidification microstructure of Mn70Ni25Cr5 alloy ingots, ProCAST simulation was combined with Box–Behnken response surface methodology to optimize pouring temperature, filling time, and mold temperature. Porosity in the ingot body and secondary dendrite arm spacing (SDAS) were selected as the response variables, and quadratic regression models were established for both responses. The optimized casting parameters were determined using analysis of variance, response surface analysis, and the desirability function approach. The porosity and SDAS models were both statistically significant, with non-significant lack-of-fit terms and R2 values of 0.9906 and 0.9901, respectively. The optimal parameters were a pouring temperature of 1220.74 °C, a filling time of 6.34 s, and a mold temperature of 294.47 °C, corresponding to a predicted porosity of 0.426% and a predicted SDAS of 47.51 μm. A supplementary simulation and a validation casting experiment were then performed using practical process settings derived from the optimized solution. The supplementary simulation indicated that shrinkage-related defects were concentrated mainly in the riser, while metallographic examination revealed no large continuous shrinkage-porosity region in the examined ingot-body sections. The overall measured SDAS across the center, half-radius, and edge positions was 48.54 μm, differing from the response-surface prediction by approximately 2.2%. These results support the applicability of the combined ProCAST–RSM approach for simulation-assisted optimization of Mn70Ni25Cr5 alloy casting parameters within the investigated process range.
Funding Information
  • Scientific and Technological Project of Yunnan Precious Metals Laboratory (YPML-20240502014)

This publication has 43 references indexed in Scilit: