Error Bounds for Inverse Electromagnetic Problems in Soil Mechanics

This paper explores the possibility of using electrical resistance tomography for the identification of buried objects. The technique requires the solution of both direct and inverse problems, which for noncanonical geometries must be solved by numerical techniques like finite element methods and soft computing. The unknown is the perturbation field, a small fraction of the source field. Therefore, simulation and error estimation tools adopted in the solution process must be both accurate and reliable. This paper shows that complementary formulations readily allow determination of error bounds for global quantities like the electric resistance.