Neural Network based Process Design of Cold Forged Shafts with Differnt Work Materials

The process design of cold forged shafts made of different work materials is carried out in this paper through an intelligent computation approach. A neural network based intelligent system was devised for decision making support in the generation of working sequences for multiple-step cold forging of multidiameter shafts, starting from a single-diameter bar and a choice of different work materials. The input to the system is given by the properties of the work material, the diameter of the initial bar, and the geometrical description of the final shaft. The system output consists in the corresponding technologically feasible working sequences to be submitted to the process designer for consideration. The performance of differently trained neural network systems in providing working sequences acceptable from a technological point of view is critically assessed.