Gathering of Process Data through Numerical Simulation for the Application of Machine Learning Prognosis Algorithms

In recent years, FEA simulation of forming processes has increasingly developed as a good alternative to complex experimental work in the determination of process parameters and product properties. However, detailed material data (e.g. flow curves) are necessary for the execution of FEA simulations, which are often not available to manufacturers and users in the early stages of the product development. In this paper, a method is shown by which application it is possible, that only on the basis the general mechanical properties (e.g. tensile strength, sheet thickness) and the use of data-based prognosis models of supervised machine learning to predict directly a result regarding suitable process parameters as well as expected forming result properties. Thereby an extensive technological database was generated for the joining by forming process self-pierce riveting (SPR) by means of numerical simulation. Subsequently, different learning algorithms are trained using these numerical data and their prediction quality is compared.