Design of an Intelligent Adaptive Control with Optimization System to Produce Parts with Uniform Surface Roughness in Finish Hard Turning

Abstract

In this paper, a real-time intelligent adaptive control with optimization methodology is proposed to produce parts with uniform surface roughness in finish turning of hardened AISI D2. Unlike traditional optimization approaches, the proposed methodology considers cutting tool real condition. Wavelet packet transform of cutting tool vibration signals followed by neural network was used to estimate tool flank wear. Intelligent models (artificial neural networks and genetic programming) were utilized to predict surface roughness and tool wear during machining process. Particle swarm optimization algorithm determined optimum feed rate that resulted in desired surface roughness. Performed confirmatory experiments indicated that the proposed adaptive control method not only resulted in parts with acceptable uniform quality, but also decreased the machining cost up to 8.8% and increased material removal rate up to 20% in comparison with those of traditional CNC turning systems.