Time-Frequency Analysis Using Stockwell Transform with Application to Guided Wave Structural Health Monitoring

In damage detection strategies it is of importance to use as straightforward and accurate as possible methodologies to detect the defects in structures. This paper aims at health monitoring of defected isotropic plates applying a modified Stockwell transform on the guided waves received from sparse arrays of active PZT transducers. Two damage detection strategies, namely the reconstruction algorithm for the probabilistic inspection of damage and the delay-and-sum algorithm, are utilized. First a numerical analysis is carried out using commercial ABAQUS/EXPLICIT to excite and receive lamb wave. Numerical results showed that using features based on Stockwell transform leads to better detection performance than traditional methods. S-Transform is basically a short-time Fourier transform with a special window function. To demonstrate the capability of S-Transform in real applications an experimental setup is prepared to acquire the propagation of guided waves in an aluminum plate. It is then concluded that the proposed features obtained from S-Transform can estimate the damage position with sufficient accuracy. It is straightforward, easy to implement and well suits the problems with high-dimensional data.