Effect of Thermomechanical Processing on Forming Limit Diagrams Predicted by Neural Networks

Abstract

In the present work, an artificial neural network (ANN) model was developed for predicting the effect of thermo-mechanical processing on the forming limit diagram (FLD) of low carbon steels. The model introduced here considers the content of carbon, the hot finishing temperature, the degree of cold work, the work hardening exponent, the initial yield stress and the ASTM grain size as inputs; while, the predicted FLDs are presented as outputs. The results show that the predicted FLDs by the ANN model are very accurate exhibiting the maximum error of 9% over the whole strain region. The model predicted that with increasing the degree of cold rolling before annealing, the drawability is increased; whereas, the stretchability is decreased. With decreasing the cold reduction, both the predicted and experimental FLD0 values shift to higher strain limits. Besides, the higher the hot finishing temperature, the higher the strain limits of the predicted FLDs will be.

Keywords:

• Forming limit diagrams
• Low carbon steels
• Manufacturing process
• Materials manufacturing modeling
• Neural network modeling
• Thermomechanical processing