| Structural health monitoring (SHM) is the periodic interrogation of man-made structures to detect damage and characterize structural integrity. The motivation for performing SHM is to identify defects or damage in a structure before they become problematic, either through a degradation in performance or catastrophic failure. SHM plays a key role in condition-based maintenance, wherein parts and equipment are repaired or replaced on an as needed basis. Condition-based maintenance offers significant cost savings over more traditional time-based maintenance and obsolescence, which replaces parts based on time in service, frequently resulting in the replacement of good parts.;Ultrasonic guided waves are able to propagate over long-distances with minimal loss and are capable of interacting with both surface and subsurface defects. As such, many SHM research efforts are exploring the use of ultrasonic guided waves for the interrogation of large, plate-like structures, such as aircraft skins, ship hulls, bridge gusset plates, and storage tanks. Of these, spatially distributed arrays of permanently attached, inexpensive transducers are of particular interest since they offer an economical solution that can be made completely automated and available to interrogate the structure at any time. The research presented here uses such a distributed array of permanently attached transducers to produce useful images of a large, plate-like structure with ultrasonic guided waves.;The work is largely split into five areas: minimum variance imaging, parameter estimation, parameter compensation, array configuration performance, and damage characterization. Minimum variance imaging, which involves the incorporation of minimum variance techniques into conventional delay-and-sum imaging, is shown to significantly improve resolution and reduce artifacts with only minor increases in computational complexity. Parameter estimation is achieved through the model-based parameter estimation algorithm (MBPE). The MBPE algorithm was developed to adaptively estimate dispersion relations, transducer-specific transfer functions, transducer spacings, and propagation loss. Additional improvements in guided wave imaging are then demonstrated through deconvolution and dispersion compensation using MBPE parameter estimates rather than nominal parameters. A preliminary investigation into the factors that affect distributed array imaging performance is conducted. The imaging algorithm employed, excitation function, number of sensors, sensor arrangement, array aperture, and array location are all considered. Finally, damage characterization, which represents a new capability for distributed array imaging, is performed by leveraging the inherent sensitivity of minimum variance imaging to scattering behavior.;This dissertation consists of the following contributions:;· Adaptation of the MVDR algorithm for guided wave imaging.;· Development of a model-based algorithm for adaptively estimatingwave propagation parameters with minimal a priori information.;· Incorporation of adaptively estimated parameters into guided wave imaging algorithms through the use of distance domain signals.;· A methodology for quantitatively characterizing the ability of an array to detect and locate damage throughout a structure.;· A methodology for characterizing defects or damage using guided waves generated from a spatially distributed array. |
