UC San Diego

A combination of passive and active sensing technologies is proposed as a structural health monitoring solution for several applications. Passive sensing is differentiated from active sensing in that with the former, no energy is intentionally imparted into the structure under test; sensors are deployed in a pure detection mode for collecting data mined for structural health monitoring purposes. In this thesis, passive sensing using embedded fiber Bragg grating optical strain gages was used to detect varying degrees of impact damage using two different classes of features drawn from traditional spectral analysis and auto-regressive time series modeling. The two feature classes were compared in detail through receiver operating curve performance analysis. The passive detection problem was then augmented with an active sensing system using ultrasonic guided waves (UGWs). This thesis considered two main challenges associated with UGW SHM including in-situ wave propagation property determination and thermal corruption of data. Regarding determination of wave propagation properties, of which dispersion characteristics are the most important, a new dispersion curve extraction method called sparse wavenumber analysis (SWA) was experimentally validated. Also, because UGWs are extremely sensitive to ambient temperature changes on the structure, it significantly affects the wave propagation properties by causing large errors in the residual error in the processing of the UGWs from an array. This thesis presented a novel method that compensates for uniform temperature change by considering the magnitude and phase of the signal separately and applying a scalable transformation