Structural model updating using adaptive multi-response Gaussian process meta-modeling

Finite element model updating utilizing frequency response functions as inputs is an important procedure in structural analysis, design and control. This paper presents a highly efficient framework that built upon Gaussian process emulation to inversely identify parameters through sampling. In particular, multi-response (MRGP) meta-modeling approach formulated can accurately construct the error surface, i.e., discrepancies between predictions actual measurement. order reduce computational cost of repeated finite simulations, adaptive sampling strategy established, where search unknown guided by surface features. Meanwhile, information previously sampled corresponding errors utilized additional training data refine MRGP meta-model. Two stochastic optimization techniques, particle swarm simulated annealing, are employed train meta-model for comparison. Systematic case studies conducted examine accuracy robustness new updating.