Application of the distributed point source method to rough surface scattering and ultrasonic wall thickness measurement

Abstract

The distributed point source method is commonly used to predict the complex acoustic field emitted by ultrasonic transducers. In this paper, it is presented as an alternative to conventional approaches often used when solving rough surface scattering problems. Surface shadowing and multiple scattering effects are inherently included in the mesh-free semi-analytical simulation method through matrix manipulation making it very efficient and simple to implement. Results are presented which illustrate the improvement in accuracy gained over the Kirchhoff approximation and the decrease in computational load over the finite element method, culminating in greater than an order of magnitude decrease in required simulation time. The method is applied to the practical problem of online wall thickness monitoring within corrosive environments, illustrating the variability in reflected pulse shape that could be expected from rough surfaces with similar statistics. Three commonly implemented time-of-flight algorithms are used to analyze a large number of simulated signals from which it is concluded that those based on first arrival time are more stable under increasing roughness conditions than those which are based on reflected pulse shape.