Using particle swarm optimization based neural network for modeling of thrust force drilling of PA-6/ Nanoclay Nanocomposites

This paper presents a newly approach for modeling thrust force in drilling of PA-6/ Nanoclay Nanocomposites materials, by using Particle Swarm Optimization based Neural Network (PSONN). In this regard, advantages of statistical experimental algorithm technique, experimental measurements, particle swarm optimization and artificial neural network are exploited in an integrated manner. For this purpose, numerous experiments for PA-6 and PA-6/ Nanoclay Nanocomposites are conducted to obtain thrust force values by using drill of high speed steel with 118° point angles and 2mm in diameter. Then, a predictive model for thrust force is created by using PSONN algorithm. Also, the training capacity of PSONN is compared to that of the conventional neural network. The results indicate that nanoclay content on PA-6 polyamide significantly decrease the thrust force. Also, the obtained results for modeling of thrust force have shown very good training capacity of the proposed PSONN algorithm with compared to that of a conventional neural network (BPNN).