Experimental investigation and optimization on machinability aspects of aluminium alloy AA7068 in dry turning using deep learning technique

This research addresses the machinability of aluminium alloy (AA) 7068 in turning with TiAlN-TiN coated carbide inserts under dry environment, which has not been previously investigated. Cylindrical turning involving twenty-seven trials (as per full factorial design of experiments) were performed by considering cutting variables, viz. feed, depth of cut, and cutting speed to assess various technological characteristics (chip morphology, crater wear, surface roughness, flank wear, cutting force, sound intensity, tool vibration, cutting temperature, and chip thickness). A combined approach of deep learning models (Deep neural network and Adaptive LeNet5) and meta-heuristic-based glowworm swarm optimization (GWSO) was applied for prediction modelling, and multiple response optimization of machinability performance characteristics. Findings revealed that increasing cutting speed led to significant escalation in flank and crater wear, while chip thickness and surface roughness decreased. Conversely, high feed rates and depth of cut corresponded with increased values of all responses. Notably, reduced speed, feed, and depth of cut yielded opposite trends. Additionally, an increase in machining parameters resulted in heightened tool vibration, cutting temperature, power consumption, and sound intensity. According to the deep learning based GWSO approach, the ideal parameters for dry turning of AA7068 using TiAlN-TiN coated carbide tool were identified as a feed of 0.1 mm/rev, speed of 220 m/min, doc of 0.075 mm.