Modelling of the clearance effects in the blanking process of CuZn30 sheet metal using neural network ？ a comparative study

Clearance effects on the product quality and blanking force in sheet metal blanking process are initially investigated experimentally, and then modelled through neural network (NN) approach. Using eleven different clearances from 8% to 18% with a sampling rate of 1%, blanking processes are applied to sheet material CuZn30 with a thickness of 1mm using a modular template. During the experiments, blanking force, smooth sheared/fractured rate and burr height for the resulting products are measured for each of clearance value and a certain portion of them are taken as example patterns to train the developed feedforward NN. Several NN-based estimation results are presented which verify that a satisfactory neural network model is attained for the concerned parameter estimations. Moreover, comparisons with a recent study that benefits from fuzzy logic as an estimator tool are also presented for the same system. We realize that estimating performance is improved using the NN and a significant contribution of our proposal is that its design is much simpler than that of its counterpart which requires proper and sufficient expert knowledge for tuning of characteristic parameters such as numbers and shapes of membership functions, linguistic control rules