A surrogate model based calibration method for structural adhesive joint progressive failure simulations

Abstract

ABSTRACT An integrated adhesive material model calibration method is proposed for adhesively bonded joints’ deformation and progressive failure simulations under mixed loading modes. In this method, a surrogate model is trained to express the intrinsic numerical relationship between the key parameters (e.g., the yield normal stress and yield shear stress) and the simulated load-displacement curves of the bonded specimens. The parameter calibration process of the material model under multiple loading conditions is described as a multi-objective optimization problem. To minimize the load-displacement curve errors among the CAE simulation model and the experiment data, the model parameters are calibrated effectively based on the surrogate model using the genetic algorithm. The validity and efficiency of the proposed calibration method is verified by comparing the test data under various loading conditions. The better precision and efficiency indicates the potential of using this framework to effectively calibrate material properties without performing time-consuming CAE simulations.