The Determination of Spring-back in L Bending Operation Using Neural Networks

Bending operations are frequently used to produce sheet metal components. Spring-back is an issue in sheet metal forming processes and some times is a principle problem for producing precise components. This is more important if some kind of cuttings exists on the bending surfaces. Experiments show that punched holes on bending surfaces affect the value of the spring-back. In this paper the spring-back of components with oblong holes on bending surfaces are studied experimentally and an artificial neural network is applied to predict the springback in L-bending shape dies. Based on 40 case experiments, a neural network is trained. Sheet metal material type, the size of the hole, blank holder force, the ratio of die clearance to sheet thickness, die and punch radius is used as input and the final angle of the bending is the output of the neural network. The trained neural network was tested by experiment results of 30 new cases. It was observed that the neural network gives close predictions of final angles.