Optimization of turning nose radius for Haynes 230: A hybrid RSM and deep learning approach

Abstract

Haynes 230 is a nickel-based superalloy recognized for its strength and high-temperature performance, making it vital in aerospace, automotive, and energy sectors. However, its hardness and low thermal conductivity pose machining challenges. This research investigates the impact of nose radius (NR) on the machinability of Haynes 230 during turning, focusing on material removal rate (MRR) and surface quality to find the optimal nose radius for both. The study uses response surface methodology (RSM) with an orthogonal array for experiments, creating quadratic models for surface roughness and MRR. Optimal parameters are validated through a multilayer perceptron (MLP) deep learning model, showing a mean absolute error of 0.37 and mean squared error of 0.26 for regression. The classification achieved a training accuracy of 94.44% and a testing accuracy of 90%, ensuring reliability. The findings indicate that larger nose radii improve the material removal rate (MRR), while smaller nose radii improve the machining surface quality. This optimized compromise aligns with Industry 5.0, where AI-driven smart manufacturing enhances productivity and quality. Deep learning integration ensures accuracy, enabling efficient machining of high-performance materials like Haynes 230.