UC San Diego

Nondestructive Testing (NDT) is an important tool to increase the service life of critical structures. Infrared Thermography is an attractive NDT modality because it is a non-contact technique, and it has full-field defect imaging capability. Typically, two problems are relevant in the domain of NDT of structures. The first problem deals with the detection of defects, while the second problem is based on estimating the defect parameters like size and depth. This dissertation aims at enhancing the capability of Infrared Thermography NDT vis-a-vis defect detection and defect quantification in metal and composite structures. In this dissertation, detection of defects was performed using two active thermography techniques - Pulsed Thermography and Lock-In Thermography. A two-stage signal reconstruction approach based on Pulsed Thermography was developed to analyze raw thermal data. In the first stage, the data was low-pass filtered using Wavelets. In the second stage, a Multivariate Outlier Analysis was performed on filtered data using a set of signal features. The proposed approach significantly enhances the defective area contrast against the background. In the second part of the study, a Lock-In Thermography technique was developed to detect defects present in a 9m CX-100 wind turbine blade. A set of image processing algorithms and Multivariate Outlier Analysis were used in conjunction with the classical Lock-In thermography technique to counter the "blind frequency" effects and to improve the defect contrast. Receiver Operating Characteristic curves were used to quantify the gains obtained. Following detection of defects, the determination of defect depth and size was addressed. The problem of defect depth estimation has been previously studied using 1D heat conduction models. Unfortunately, 1D heat conduction based models are generally inadequate in predicting heat flow around defects, especially in composite structures. A novel approach based on virtual heat sources was proposed in this dissertation to model heat flow around defects accounting for 2D axisymmetric heat conduction. The proposed approach was used to quantitatively determine the defect depth and size in isotropic materials. Further, the approach was extended to model 3D heat conduction in quasi-isotropic composite structures. The approach used the excess temperature profile that was obtained over a defective area with respect to a sound area to estimate the defect dimensions and depth. Using coordinate transformations, the anisotropic heat conduction problem was reduced to the isotropic domain, followed by separation of variables to solve the resulting partial differential equation. Relevant experiments were performed to validate the proposed theory