Abstract

An alarming pace of increase in worldwide energy consumption is being caused by population expansion and economic development, particularly in emerging market countries. Due to their efficiency and reproducibility in accomplishing high-precision machining, CNC machine tools are widely employed in most metal machining processes. The use of simple machining features to search for the assignment of cutting parameters and the machining process on the output variables is limited since a part, in reality, can contain complex interacting features. Therefore, this study focuses on pocket/groove features by integrating grey relational analysis (GRA) and hybrid PSO-ANN and ANFIS algorithms to optimize and predict surface quality, cost and energy consumption (QCE). Taking into account the population size of the swarm (pop) and the number of neurons (n) in the hidden layer, a parametric study was carried out to find the best prediction using the hybrid algorithm PSO-ANN. This study reported the highest trained correlation values (R2) for all output variables (greater than 0.97%). The study shows that the assignment of machining strategies and sequences on energy consumption can reach 99.25% between the minimum and maximum values. The mean square error (MSE) data demonstrates that the PSO-ANN model is effective. Indeed, an MSE improvement of 99.84%, 99.87%, and 97.62% has been demonstrated in terms of Etot, Ctot, and Ra, respectively, of the PSO-ANN model compared to ANFIS. This study reveals the potential of the PSO-ANN hybrid for multi-criteria prediction (quality, cost, and energy: QCE) by comparing it with the ANFIS model.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call