Modeling the Influence of Antimony on the Mechanical Properties of Carbidic Austempering Ductile Iron Using an Artificial Neural Network

To simulate the impact of antimony powder microquantity on the mechanical and tribological characteristics of Carbidic Austempering Ductile Iron (CADI), artificial neural networks (ANNs) were utilized. The results indicate that the hardness, impact toughness and wear resistance increase with increase in the antimony alloy. The ANN's simulation findings demonstrate an outstanding agreement with the experimental results, with Mean Square Error values ranging from 1.0692 to 1.2987 and acceptable R-values of 0.81 for wear resistance, 0.91 for hardness values, and 0.97 for impact toughness values. Microstructurally, it was seen that the antimony content succeeded in modifying the coarse carbide to granular carbide and refined the graphite particle. Statistical package for social science (SPSS) was used to determine the importance of the contributing experimental parameters to the overall improvement of mechanical properties of CADI. It was seen that the three independent variables (austempering temperature, austempering time and antimony) actually contributed to the overall improvement of the mechanical properties of the CADI but their contributions are not of the same percentage. It was observed that antimony contributed about 60%, austempering temperature is 26% and austempering time contributed 14%. These could have been responsible for the overall improvement in the mechanical and tribological properties of the CADI. This shows that ANN can successfully and suitably predict the mechanical properties of CADI modified with antimony alloy.