Optimisation of material properties for the modelling of large deformation manufacturing processes using a finite element model of the Gleeble compression test

The finite element modelling of manufacturing processes often requires a large amount of large plastic strain flow stress data in order to represent the material of interest over a wide range of temperatures and strain rates. Compression data generated using a Gleeble thermo-mechanical simulator is difficult to interpret due to the complex temperature and strain fields, which exist within the specimen during the test. In this study, a non-linear optimisation process is presented, which includes a finite element model of the compression process to accurately determine the constants of a five-parameter Norton–Hoff material model. The optimisation process is first verified using a reduced three-parameter model and then the full five-parameter model using a known set of constants to produce the target data, from which the errors are assessed. Following this, the optimisation is performed using experimental target data starting from a set of constants derived from the test data using an initial least-squares fit and also an arbitrary starting point within the parameter space. The results of these tests yield coefficients differing by a maximum of less than 10% and significantly improve the representation of the flow stress of the material.