UC San Diego

Ultrasonic evaluation technologies have made considerable advancements throughout time. The research work presented in this dissertation aims to apply the well-studied ultrasound wave propagation properties with novel data processing frameworks to enhance the robustness and performance of ultrasonic nondestructive evaluation systems, namely rail inspection technologies.

First, an ultrasonic non-contact rail defect detection prototype exploiting the properties of guided surface waves is presented. The system uses air-coupled transducers to generate and detect the surface waves. Because of the challenges (mainly low SNR) in this application of air-coupled transducers, numerical simulations results aided in the prototype hardware and software developments. This system utilizes a multivariate outlier classifier framework to distinguish a defective rail from a healthy rail. The statistical outlier analysis algorithm takes advantage of an adaptive baseline that allows for a higher rate of true detection, while minimizing false positives. A field test was conducted last October 2014 at the Transportation Technology Center in Colorado, proving the effectiveness of the rail defect detection prototype.

Next, an initial internal flaw 3D imaging prototype utilizing advanced synthetic aperture focusing (SAF) frameworks is proposed. In this dissertation, the mode structure of the longitudinal and shear reflected waves are explored and considered as adaptive weights in existing SAF frameworks to increase the gain of the ultrasonic array without physically increasing its number of elements. In addition, a novel set of Global Matched Coefficients (GMC) is implemented to further decrease unwanted noise and artifacts in the images. Results from these new additions are presented, and their applications to rail defect imaging is proposed. The proposed defect imaging prototype reconstructs a 3D volumetric image from multiple planar ultrasonic images. An early defect rail defect imaging prototype is development and preliminary results are presented. The results show that the system is of high potential to be an advancement in solids internal defect imaging, particularly rail flaw characterization.