Cyclical Variational Bayes Monte Carlo for efficient multi-modal posterior distributions evaluation

Multimodal distributions of some physics based model parameters are often encountered in engineering due to different situations such as a change environmental conditions, and the presence types damage nonlinearity. In statistical updating, for locally identifiable parameters, it can be anticipated that multi-modal posterior would found. The full characterization these is important methodologies structural condition monitoring structures frequently comparison damaged healthy models structure. using state-of-the-art sampling techniques require large number simulations expensive run physics-based models. Therefore, when limited run, occurs engineering, traditional not able capture accurately multimodal distributions. This could potentially lead numerical errors assessing performance an structure under uncertainty.