Prediction of Ultrasonic Guided Wave Propagation in Fluid–Structure and Their Interface under Uncertainty Using Machine Learning

Abstract

Structural health monitoring (SHM) systems use nondestructive testing principles for damage identification. As part of SHM, the propagation of ultrasonic guided waves (UGW) is tracked and analyzed for the changes in the associated wave pattern. These changes help identify the location of a structural damage, if any. We advance the existing research by accounting for uncertainty in the material and geometric properties of a structure. The physics model employed in this study comprises a monolithically coupled system of elastic and acoustic wave equations, known as the wave propagation in fluid–structure and their interface (WpFSI) problem. Because the numerical simulation of the WpFSI problem becomes computationally extremely expensive for many realizations of the uncertainty, we developed an efficient algorithm in this work that employs machine learning techniques like Gaussian process regression and convolutional neural networks to predict UGW propagation in a fluid–structure and their interface under uncertainty. First, a small set of training images for different realizations of the uncertain parameters of the inclusion inside the structure is generated using the computationally costly physics model. Next, Gaussian processes trained with these images are used for predicting the propagated wave with convolutional neural networks for further enhancement to produce high-quality images of the wave patterns for new realizations of the uncertainty. The results indicate that the proposed approach provides an accurate prediction for the WpFSI problem in the presence of uncertainty.