Heat Treatment Process Study and ANN-GA Based Multi-Response Optimization of C45 Steel Mechanical Properties

The main purpose of the present investigation is to achieve an optimization of the heat treatment process along with improving the mechanical properties of Medium-carbon steel C45 illustrated by the tensile strength (Ts), the hardness (HV) and the elongation (A). This is performed through varying three necessary factors represented by the quench temperature, the holding time and the cooling medium (air, oil and acid solution). An experimental design illustrated by the performance of 27 (\($$$$3^3\)) tests was adopted for the planning of the experiments. The Response Surfaces Methodology (RSM), the Neural Networks Approach (ANN) and the Genetic Algorithm (GA) were selected for the modeling and optimization of the heat treatment process. Analysis of Variance (ANOVA) results show the cooling medium to have the greatest influence on the mechanical properties of C45 steel with contributions of 47.34% to 74.3%. The experimental results demonstrate that heating at a temperature of \(T = 800\,^\circ {\rm C}\) for 30 minutes and then performing a quenching in the acid solution gives large values of both Ts and HV, and achieves \(Ts = 1060 {\rm MPa}\) and \(HV = 825\).