Bayesian emulation for forecasting modal frequencies under multivariate environmental variability with data association metric and incremental updating

Le-Le Zhang, Wang-Ji Yan, Ka-Veng Yuen, Costas Papadimitriou, Wei-Xin Ren

Abstract: Highlights•Data association metric is used to screen out multiple correlated EOPs according to the MIC rank.•Bayesian emulator provides a nonlinear mapping between modal frequencies and multivariate EOPs.•Bayesian emulator offers capacity of quantifying uncertainties for the predicted values of frequencies.•Bayesian predictive model can be updated automatically to adapt the growing data over monitoring period.•The scheme holds the potential for distinguishing the variation of frequencies due to damage and EOV.

External IDs:doi:10.1016/j.jsv.2025.119206