NDE of Discontinuities in Thermal Barrier Coatings with Terahertz Time-Domain Spectroscopy and Machine Learning Classifiers

Abstract

Internal discontinuities are critical factors that can lead to premature failure of thermal barrier coatings (TBCs). This paper proposes a technique that combines terahertz (THz) time-domain spectroscopy and machine learning classifiers to identify discontinuities in TBCs. First, the finite-difference time-domain method was used to build a theoretical model of THz signals due to discontinuities in TBCs. Then, simulations were carried out to compute THz waveforms of different discontinuities in TBCs. Further, six machine learning classifiers were employed to classify these different discontinuities. Principal component analysis (PCA) was used for dimensionality reduction, and the Grid Search method was utilized to optimize the hyperparameters of the designed machine learning classifiers. Accuracy and running time were used to characterize their performances. The results show that the support vector machine (SVM) has a better performance than the others in TBC discontinuity classification. Using PCA, the average accuracy of the SVM classifier is 94.3%, and the running time is 65.6 ms.