Control-oriented temperature modeling for multi-stage press hardening of thin workpieces

Abstract

State estimation and control of multi-stage press hardening processes require accurate and computationally efficient models. Thermo-plastic high-fidelity simulations provide valuable insights into the process, they are however often computationally very expensive and thus not suitable for online applications. To tackle this, a control-oriented modeling approach for the temperature evolution of thin workpieces is proposed, where first a linear surrogate model is designed to approximate the high-fidelity model, and secondly two different projection-based model order reduction techniques are applied. In simulative studies and evaluations using experimental data the model reduction approach is shown to accurately replicate the temperature evolution of the high-fidelity model, while the computation time is reduced massively.