9360217 Sabbagh The ability to solve inverse problems lies at the heart of electromagnetic nondestructive evaluation (NDE). These problems are difficult to solve for two reasons: they are mathematically ill-conditioned, and they are nonlinear, which means that they involve considerable computational effort, often requiring supercomputer capabilities. In this proposal, we describe an inversion algorithm for electromagnetic (eddy-current) NDE that will run on a wide class of machines, from personal computers and workstations to minisuper-computers and larger machines. Furthermore, we expect an improved robustness that will mitigate the effects of ill-conditioning. The algorithm employs modern mathematical ideas of set theoretic estimation and robust statistical analysis. The technology that we are developing in this project will be applicable to nondestructive evaluation of materials and structures, electromagnetic tomography, computer vision, and manufacturing science and technology.
