Optimal Frequency selection for vibration based damage detection using pattern recognition

Abstract Damage detection methods of structural components have been extensively evaluated in theoretical and experimental research studies in the last few years. As the research objectives for structural health monitoring (SHM) are broadly diversified, this study focuses on the signal interpretation techniques. The acquired data from a sensor network should be optimally evaluated in consideration of environmental effects. In this context, pattern recognition techniques are widely used to evaluate the health state of structures from acquired data. This work assesses the dependency of various excitation frequencies in guided-wave based SHM and the performance of damage detection. As a result, a solution for an efficient and optimized signal interpretation model is provided. The needed vibration response to detect damage states is applied by a piezoelectric sensor network, which is used to apply guided-waves through the structure and to measure the vibration response. Finally, the pattern recognition approach is evaluated experimentally using beam and sandwich structures to expose differences in the frequency domain. The important outcome of this study is to improve the efficiency and performance of SHM systems by optimizing the excitation frequency using pattern recognition approaches.