Chirplet-based imaging using compact piezoelectric array

This paper presents the implementation of a chirplet-based matching pursuit technique called excitelet for imaging. High frequency bursts are injected into a structure by a piezoceramic (PZT) actuator and measurement is conducted by a compact array of PZT sensors, located remotely from the damage. The matching pursuit algorithm is implemented with a dictionary of atoms obtained from dispersed versions of the excitation, where the parameters of each atom are the propagation distance and the mode. For a selected point in the scan area and a given mode, the measured signal is correlated with a given atom value for each propagation path in the array configuration. A round-robin technique is used to add the contributions of all these correlation values for each point in the scan area for imaging. Simulations are first conducted for a 1.5 mm thick aluminium plate with signals synthesized for A 0 mode propagating over distances corresponding to the location of a reflection or diffusion point in an area in front of an array of measurement points. The simulations show that the excitelet offers better localization of the reflection point, when compared with a group velocity-based, or time-of-flight (ToF) approach. The simulation results are validated experimentally using a 1.5 mm thick aluminium plate with a notch in the periphery of a hole. Bonded PZTs are used for both actuation and sensing of 2.5 cycles bursts at 300 kHz, 500 kHz and 850 kHz. Significant improvement of imaging quality is demonstrated with respect to classical imaging techniques.