Title : Comparison of Short - time Fourier Transform and Wavelet Transform of Transient and Tone Burst Wave Propagation Signals For Structural Health Monitoring

1 Advanced signal processing approaches are widely used for nondestructive evaluation, structural health monitoring and damage detection. In this research, signals related to damage in a thin plate specimen were collected by piezoelectric wafer active sensors. Damage is detected and evaluated by using advanced signal processing approaches, including the short-time Fourier transform and the wavelet transform for the two dimensional time-frequency analysis. In addition, a denoising process is included before the signal analysis to improve the detectability. Advantages and disadvantages of each approach are investigated based on the results of simulation. Comparisons are concluded at the end of this paper. 1. Introduction Damage detection can be used to monitor the system performance, detect damage occurrence and give prognosis or diagnosis, by which the operators can make corresponding maintenance and repair decisions. Typically, damage detection consists of two parts: first, a sensing system to collect the signals related to system performance; second, a signal processing and analysis algorithm to interpret the signals. According to Staszewski (2002), many of the advances in damage detection are related to the development of signal processing techniques. The objectives of signal analysis for damage detection, nondestructive evaluation (NDE), and structural health monitoring (SHM) is to eliminate or reduce the noise in the original signals, to analyze the frequency composition, and to evaluate the damage if it occurs. In damage detection, signal analysis is used to extract changes in a signal related to damage and then characterize the changes. Hence, there are two things that need to be considered: determining which signals to measure for monitoring, and choosing efficient signal analysis approaches to best interpret the measurements. Considering the disturbance coming from the environment and the measuring process, it is necessary to erase or reduce the disturbance before signal analysis. Most practical signals in engineering are non-stationary signals, whose