LAUSR

Abstract The process of roller levelling is essential to ensure flatness of sheet metals and to reduce residual stresses within the material. Besides process parameters like machine geometry and roll positions, the process is sensitive to variations of material properties. In order to detect variations of the material properties, an inverse identification of material parameters from measured process data is desirable. Therefore, a semi-analytical levelling model based on bending theory that is capable to describe material behavior under low cyclic loading is introduced. This model calculates the levelling forces depending on given material data. Employing the model within an optimization loop, the model allows for an inverse determination of material data during levelling by comparing calculated and measured levelling forces. To validate the model, accompanying finite element (FE) simulations, levelling experiments as well as cyclic bending tests are performed. According to the results, the semi-analytical model is capable of calculating levelling forces in good agreement with FE results and much lower computation times. Furthermore, material data obtained directly from levelling experiments are in the same range as data determined from cyclic bending tests.