Predicting the Size and Location of a Cavity in a Solid Half-Space from the Scattered Ultrasonic Fields Using Genetic Algorithms

Abstract

Genetic algorithms are used to predict the size and location of an elliptical cavity within a solid half-space. The scattering of ultrasonic waves in the solid half-space with a cavity is modeled using Distributed Point Source Method (DPSM). DPSM which is a semi-analytical technique that utilizes Green’s function is used for modeling because this technique is more efficient than popular but not so efficient Finite Element Method (FEM). FEM is very inefficient for modeling ultrasonic wave propagation problems at high frequencies and for solving an inverse problem one needs an algorithm that can solve the forward problem efficiently. The inverse problem is solved by applying a genetic algorithm to the forward problem to determine the optimum solution. The optimum population size and number of generations are determined. Results and analysis are performed for 3 cases of unknown variables.