Bayesian Calibration of Material Model Parameters for High-Velocity Impact Problems through Ensemble Kalman Inversion

In this work, we employ ensemble-based data assimilation (DA) to calibrate material model parameters under high-velocity impact conditions. DA integrates experimental observations with numerical models. By iteratively updating simulation inputs based on observational data, DA minimizes errors arising from both numerical approximations and experimental inaccuracies. Specifically, we use smoothed particle hydrodynamics (SPH) simulations as the dynamic system. The discrete-time nature of SPH simulations makes them particularly well-suited for sequential DA methods. We apply the ensemble Kalman filter (EnKF), a robust DA technique, to refine material model parameters by reconciling discrepancies between experimental observations and simulation results. This work signifies a substantial progression towards integrating DA techniques into high-strain-rate material modeling and demonstrates the potential of combining experimental data with advanced numerical methods to address challenges in high-velocity impact applications. The expected outcome is an improved methodology for calibrating material models and estimating model parameters, yielding more accurate and reliable high-velocity impact simulations.