Material Parameter Identification from Machining Simulations Using Inverse Techniques

Machining is a complex process during which the material undergoes large deformations at high strain-rates with large variations in temperatures. One of the difficulties faced during the simulation of machining is that of determining appropriate material parameters which are valid for such large ranges of strains, strain-rates and temperatures. An inverse method of material parameter identification from machining simulations is proposed in this paper. An error function is defined that takes into account the chip overlap error and the cutting force difference at different frames of observation. The two components are suitably weighted so that the contribution of each is rendered almost equally. A two stage optimisation process is employed for the minimisation of the error function where the Levenberg-Marquardt algorithm is used in the first stage for faster convergence and the Downhill Simplex algorithm in the second stage in order to navigate through the noisy error landscape. A wide range of cutting conditions is used and the method is shown to work also for non-adiabatic simulations. However, the converged parameter sets are found to be non-unique.