Application of artificial neural network and genetic algorithm to predict and optimize load and torque in T-section profile ring rolling. (September 2019)

• Record Type: Journal Article
• Authors: Parvizi, Ali; Rohani Raftar, Hamid Reza
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 233:Number 17(2019)
• Page Start: 5966
• Page End: 5976
• Publication Date: 2019-09
• Subjects: Profile ring rolling -- finite element simulation -- artificial neural network -- genetic algorithm -- load -- torque; Mechanical engineering -- Periodicals; 621.05
• DOI: 10.1177/0954406219861989
• Languages: English
• ISSNs: 0954-4062

Abstract

Artificial neural network is implemented to predict the required load and torque in T-section profile ring rolling process for the first time in this study. Moreover, an optimal condition of T-section profile ring rolling process for specific limit of input factor is acquired using genetic algorithm technique. Various three-dimensional finite element simulations are carried out for different collections of process variables to obtain initial data for training and validation of the neural network. Besides, the finite element model is verified via comparison with the experimental results of the other investigators. The back-propagation algorithm is utilized to develop Levenberg–Marquardt feed-forward network and the optimum architecture is achieved by estimating the performance considering different number of hidden layers and neurons. It is concluded that results of artificial neural network predictions have an appropriate conformity with those ones from simulation and experiments. Moreover, a reasonable accuracy is obtained from the implemented model by which the prediction of ring rolling load and torque in different conditions can be achieved.