Damage detection using self-powered wireless sensor data: An evolutionary approach

A major limitation in implementation of self-powered wireless sensors pertains to considerable loss of the sensed information. Consequently, interpretation of the limited but valued data generated by the self-powered wireless sensing technology becomes a challenging problem. To tackle this issue, this study presents an evolutionary computational approach for structural damage detection using the self-powered wireless sensor data. The proposed data interpretation system is based on the integration of a robust evolutionary technique, called gene expression programming (GEP), and finite element (FE) method. Several damage indicator variables are extracted upon the simulation of the compressed data stored in memory chips of self-powered sensors. For the analysis, the complicated case of gusset plate of bridge is considered. Different damage scenarios are introduced to the plate and for each scenario and sensor configuration, a damage detection model is derived. Bases on a logistic regression analysis, probabilities are assigned to each model to find the most probable damage state. The damage detection models are presented as MATLAB and Visual Basic codes for further analysis. An uncertainty analysis is performed through the contamination of the damage indicator features with different noise levels. The results indicate that the proposed method is efficiently capable of detecting different damage states in spite of high-level noise contamination.