Informed boundary condition and constitutive parameter calibration via Identifiability Classes

This paper presents an enhanced method for sequential identification of boundary condition and constitutive parameters, incorporating advanced measurement uncertainty-based sensitivity analysis through Identifiability Classes. The approach is further optimized by implementing an a priori selection of regularization weight, significantly accelerating the identification process. Additionally, the use of an elastic (linear) model is explored for initial boundary condition parameter estimation, enabling rapid approximate results. The methodology is applied to a previously studied tensile test, demonstrating its robustness through the use of intentionally offset constitutive parameters. The results confirm the method effectiveness in achieving accurate identification while maintaining computational efficiency. The combination of Identifiability Classes with measurement uncertainty-informed regularization provides a robust framework for developing models that accurately match experimental observations.