A machine learning approach for simulation of multi-stage laser forming process

Bending angle obtained in laser forming process is much smaller than that conventional die pressing or roller bending processes. Therefore, multi-stage necessary to form deep drawn shapes and/or complex by forming. However, irradiation varies material properties, posture, and stiffness of metal plates, those influences are accumulated during the consecutive stages process. Thus, thermo-elasto-plastic deformation models theoretical analysis cannot predict final shape precisely. Regarding this problem, authors propose a simulator, which employs artificial neural network correlate parameters study. Teaching data for machine learning collected through experiments with stainless-steel plates high-power diode laser. The trained used simulate plate demonstrate feasibility proposed method. And structure on investigated, conditions discussed aspect prediction accuracy. Trained acquired relationship between irradiating work as simulator formed bent at scanning paths. Prediction accuracy was same experiments.