Gaussian process NARX model for damage detection in composite aircraft structures

Abstract This paper demonstrates the Gaussian process regression model's applicability combined with a nonlinear autoregressive exogenous (NARX) framework using experimental data measured PZTs' patches bonded in composite aeronautical structure for concerning novel SHM strategy. A stiffened carbon-epoxy plate regarding healthy condition and simulated damage on center of bottom part stiffener is utilized. Comparing performance terms simulation errors made to observe if identified models can represent predict waveform confidence bounds considering confounding effect produced by noise or possible temperature variations assuming dataset preprocessed principal component analysis. The results GP-NARX model have attested correct classification reduced number false alarms, even uncertainties propagation damaged conditions.