Intelligent Computational Approaches to Signal Processing and Damage Detection

Abstract

AbstractThis chapter presents intelligent computational methodologies towards signal processing and damage detection of a structural system, considering uncertainties such as noise, incompleteness, and variability in sensed data and computational model. The proposed methodology is based on the adept integration of dynamic artificial neural network, wavelet signal analysis, and Bayesian probabilistic assessment. A dynamic fuzzy wavelet neural network model is employed to perform the multiple-input-multiple-output nonparametric system identification of the structure subjected to external excitation using incomplete sensor data under healthy condition. The trained model is used to predict dynamical responses of the structural system under unknown conditions. Both predicted and sensed time history data are decomposed into multiple time-frequency resolutions using a discrete wavelet packet transform method. The wavelet packet component energy is computed in terms of the decomposed coefficients and used as signal feature to detect damage in a structural system. The effectiveness of the selected features is assessed using both time and frequency domain metrics. The Bayesian probabilistic assessment method is developed to incorporate possible uncertainties in both multivariate sensor data and model prediction and provide a quantitative measure of confidence of a computational model or structural status. These methodologies are illustrated for application to an aerospace structure thermal protection system panel and a four-story benchmark building frame structure, representing two different disciplines.KeywordsBayesian statisticsWavelet packetNeural networkDamage detectionHypothesis testing