Hybrid forming process of AA 6108 T4 thin sheets: modelling by neural network solutions

The highly non-linear deformation processes occurring in most dynamic sheet metal forming operations cause large amounts of elastic strain energy to be stored in the formed material and massive related springback phenomena. Therefore, this paper investigates how effective a laser source is in reducing the extent of springback in mechanical contact forming operations. The hybrid forming process investigated was composed of using a high-power diode laser to induce local heating of mechanically bent AA 6108 T4 thin sheets in order to minimize the extent of the springback. In particular, experiments were carried out to assess the influence of the leading process parameters such as laser source power, scan speed, and starting elastic deformation of the mechanically bent sheets. It was found that the trends in the experimental response of residual deflection were always consistent with the operating parameters. Artificial intelligence techniques were then used to model the hybrid forming process. The extent of the springback in the hybrid forming process of AA 6108 T4 thin sheets was predicted by using different neural network models and training algorithms. Lastly, the reliability of the best neural network solutions was checked by comparing these solutions with experimental results and by developing an ad hoc first approximation technical model.