A new optimization method for sheet metal forming processes based on an iterative learning control model

Abstract

Aiming at improving the efficiency of process optimization for industrial application, a new intelligent optimization method for sheet metal forming processes is proposed based on an iterative learning control model, which is constructed to imitate the die trial process in production. In this model, process parameters are adjusted according to a learning updating law, in which the deviations between the practical forming quality and the desired forming quality are used to determine input parameters in the next cycle, in combination with experience learned from preceding trials. Each learning process leads to a forming quality deviation smaller than the latest one, and the desired forming quality is obtained when the forming quality deviation becomes small enough to be neglected after several times of learning. To apply this model to actual processes, the forming quality is quantified firstly by accumulating the forming quality of the related elements with their contributions considered. Afterwards, the learning updating law is designed by multiplying the forming quality deviation with a varied learning gain. In the end, the intelligent updating strategy of the learning gain is proposed, which guarantees the proper parameter variation in each trial and gives rise to a fast convergence speed. By using this method, a new optimization algorithm of the drawbead restraining force (DBRF) is proposed, in which the scheme of drawbead segmentation can be decided automatically. The rapidity and practicability of the algorithm are verified by numerical experiments of automobile covering panels.