Thermoacoustic Tomography Modeling with Spectral Element Method

Abstract

Summary form only given. In this work, we propose a spectral element method (SEM) to solve the thermoacoustic tomography (TAT) forward problem. With the use of Gauss-Lobatto-Legendre (GLL) polynomials as basis functions and GLL points as quadrature integration points, the SEM can achieve the same accuracy as the FEM with a much lower sampling density, and therefore the number of unknowns can be greatly reduced. A perfectly matched layer (PML) is used at the boundary of the computational domain to absorb outgoing waves. Numerical results are shown to validate the SEM, and to demonstrate the significant advantages of the SEM over the FEM for large-scale realistic TAT problems