AI-based Multi-Objective Prediction and Optimization of Control Parameters in Turning D2 Steel with Minimum Quantity Lubrication using Grey Relational Analysis (GRA)-Artificial Neural Network (ANN)-Particle Swarm Optimization (PSO) Approach

This study investigates the application of artificial intelligence (AI) for multi-objective prediction and optimization of control parameters in turning D2 steel, a vital material in the automotive industry for components that require high strength and wear resistance, such as gears, shafts, bearings, and dies. The research prioritizes achieving minimal surface roughness, cutting zone temperature, tool wear, and maximum material removal rate while utilizing Minimum Quantity Lubrication (MQL) for reduced environmental impact. A novel Grey Relational Analysis (GRA)-Artificial Neural Network (ANN)-Particle Swarm Optimization (PSO) approach is proposed. GRA is employed to transform the multi-objective responses into a single Grey Relational Grade (GRG). This GRG represents the desired machining performance. An ANN model is then trained to predict the GRG based on the control parameters. Finally, PSO utilizes the ANN predictions to optimize the control parameters, aiming for the highest GRG, which signifies optimization of response variables. Experimental turning trials validate the effectiveness of the proposed approach. This research offers a significant contribution to the automotive industry by enabling efficient and sustainable machining of D2 steel. By optimizing control parameters for MQL and utilizing the GRG for multi-objective evaluation, the approach promotes reduced environmental impact without compromising product quality through minimal surface roughness and extended tool life.