In this paper, a self-diagnosis technique is introduced for structural health monitoring based on the timefrequency analysis of electromechanical impedance (EMI) signatures. It is first shown that the EMI signature is essentially a pulse-echo signal represented in the frequency domain, and it can be converted into time domain using the convolution theorem. The time-frequency plot is then generated from the recovered time domain signals to cover a wide range of excitation frequencies and provide a more comprehensive damage detection capability. Presenting the EMI signal in the time and the time-frequency domains provides the physical insights that explain how different factors influence the EMI signature. As such, the time domain signal acquired from the EMI is divided into “resonant phase” and “echo phase”. The resonant phase includes the immediate response of the sensor to the excitation and is used to monitor the sensor bonding layer condition, while the echo phase only includes wave reflections from structural damages and boundaries, and is implemented for the structural damage detection. Finally, the proposed method is implemented on a beam structure to detect and localize structural damage in the presence of a damage in the sensor bonding layer.
|