| Structural Health Monitoring of critical structural parts is a vital activity for preventing structural failure and loss of human lives. In response to this need, the use of piezoelectric wafer active sensors (PWAS) array in which the local structural health can be monitored with the electro-mechanical (E/M) impedance method has been proposed. The goal of the research was to develop the scientific basis and engineering know-how for the extensive use of PWAS and the E/M impedance method in structural health monitoring with direct application to aging aircraft and civil engineering structures. PWAS were studied from both theoretical and practical aspects. For the first time, a PWAS model, which describes the dynamics of elastically constrained PWAS was derived in both 1-D and 2-D geometries. The model was validated with experimental results. Issues of PWAS fabrication, testing, and installation were also studied. In addition, for the first time, a method for PWAS self-diagnostics, using the imaginary part of the E/M impedance, was described.; A theoretical model for describing the sensor-structure interaction and explaining the sensing mechanism of the E/M impedance method was developed for 1-D and 2-D geometries. The solution predicts the E/M impedance spectrum, as it would be measured at PWAS terminals, and accounts for both sensor dynamics and structural dynamics. Both flexural and axial vibrations of 1-D and 2-D host structures were considered in the solution. The validation of theoretical results was performed experimentally using metallic beams and circular plate specimens.; The effect of damage on the E/M impedance spectra was studied using controlled experiments performed on a statistical set of calibrated specimens. Damage detection algorithms based on (a) statistical analysis; (b) overall-statistics damage metrics; and (c) probabilistic neural networks (PNN) were used to classify spectral data according to location of damage. It was observed that the use of the correlation coefficient deviation damage metric was the most appropriate for comparison of raw spectra. However, PNN was found to be the best classification algorithm for classifying spectra based on resonance frequencies data features.; The application of PWAS and the E/M impedance method for crack identification in aging aircraft panels was successfully demonstrated. Damage detection algorithm utilizing the PNN method was able to identify cracks not only in the field near PWAS, but also in the medium field. The in-field implementation of E/M method for SHM of composite retrofits installed on a civil structure is also presented. |
