Stability of guided wave signals from bonded piezoelectric sensors

Abstract

This study investigates the stability of guided wave signals generated and recorded from bonded piezoelectric sensor packages proposed for use in structural health monitoring systems. The study considers the effects of both actuation cycling and thermal exposure on the piezoelectric sensors. The tests are performed using polyimide film encapsulated piezoelectric sensors bonded to aluminum plates and a titanium wing attachment lug, with the testing specifically designed to avoid causing any damage to the host structure. Stability is quantified by computing a correlation coefficient between a reference signal and each test signal. The reference signal is recorded under the initial healthy condition, so any potential changes in the correlation coefficient value are attributed to aging effects. The effects of possible timing differences causing decreases in the correlation coefficient values are reduced by a jitter correction algorithm. The first set of experiments uses four aluminum plates held at constant temperature; four piezoelectric actuators on each plate transmit to a centrally located piezoelectric sensor. To investigate the effect of accumulated actuation cycles on the transducers, different numbers of actuation signals are applied to each set of four actuators. Each test is conducted for 150 blocks, with each of the four actuators producing either 1000, 500, 250, or 100 signals per block. Results from the testing are mixed, with some excitation/sensing paths remaining stable over the duration of all blocks while other paths show substantial changes, including clear trends of decreasing correlation coefficient values. In a second experiment, sensors on a titanium wing attachment lug are exposed to relatively benign levels of thermal exposure. Each test starts with the temperature at a selected baseline value of 120°F. A series of ten elevated temperature exposures are applied with the exposure temperature increasing in 10°F increments to 220°F. The reference and test signals are collected after returning to the baseline temperature. As in the first set of experiments, results from the testing are mixed. Some excitation/sensing paths remain stable over the duration of the test, while others are substantially degraded. For both tests, the exact mechanism causing the instabilities remains unknown. However, the mixed outcomes suggest that the signal changes observed over the course of the collections may be due to flaws within the piezoelectric or electrode material of a specific sensor, or involve the adhesive bond between a particular sensor and the structure.