A Hybrid Mechanism-Based and Data-Driven Approach for the Calibration of Physical Properties of Ni-based Superalloy GH3128.

Hao Sun,Lin Zhou,Shengqiang Zhao,Teng Zhang,Fangyu Peng,Rong Yan
DOI: https://doi.org/10.1109/m2vip49856.2021.9665158
2021-01-01
Abstract:Within the finite element simulation model of metal manufacturing processes, the setting of physical properties of the materials will significantly affect the final simulation results. In this paper, a hybrid approach to integrate mechanical models and experimental data is proposed to identify the constitutive parameters and tool-chip friction coefficient of superalloy GH3128. Firstly, mechanical model is established and the relation between the constitutive parameters and the measured shear flow stress is determined. Combined with Bayesian inference, the fast inverse identification of the above parameters is realized. Then, a finite element surrogate model and Bayesian framework are fused for calibrating tool-chip friction coefficient. Finally, the effectiveness of the proposed approach is verified through milling experiments. The prediction accuracy of the prediction model are significantly improved by considering the uncertainties of physical properties. The relative error of the model in direction and direction are 11.4681% and 4.2989%, respectively.
What problem does this paper attempt to address?