Prediction of Required Torque in Cold Roll Forming Process of a Channel Sections Using Artificial Neural Networks

One of the most important issues in the review of cold roll forming process of metals is estimation of required torque. The optimum production line can be designed by determining the effective parameters on torque. Some of these parameters are sheet material and thickness, bending angle, lubrication conditions, rolls rotational speed and distance of the stands. The aim of this study is to predict amount of required torque considering the factors influencing torque, including thickness, yield strength, sheet width and forming angle using artificial neural network. So the forming process was 3D simulated in a finite element code. Simulation results showed that with increase of yield strength, thickness and forming angle, applied torque on rolls will increase. Also the increase in sheet width -assuming constant web length- will decrease the torque needed for forming. The effects of thickness and sheet width were experimentally investigated which verified the results obtained by finite element analysis. A feed-forward back-propagation neural network was created. The comparison between the experimental results and ANN results showed that the trained network could predict the required torque adequately.