Title:  Optimization and Design of Coils for Electromagnetic Induction Systems

Committee:

Dr. Waymond Scott, ECE, Chair , Advisor

Dr. Andrew Peterson, ECE

Dr. Gregory Durgin, ECE

Dr. Morris Cohen, ECE

Dr. Samer Naif, EAS

Abstract: A new method of optimizing coils for electromagnetic induction (EMI) systems that are used in subsurface sensing is presented. The optimization method represents coils using stream functions rather than attempting to optimize wire coils directly, which allows a more general optimization than is possible with conventional wire coils. A new set of normalized metrics was developed to take into account important coil characteristics, such as target sensitivity, soil sensitivity, power dissipation, coil coupling, thermal noise, etc. The metrics are applicable to both conventionally-defined wire coils and to the stream-function representation and allow comparison between wire coils and stream function coils. Even with various techniques to reduce the number of unknowns, the problem would be intractable without exploiting the partial convexity of the problem statement. A bi-convex optimization procedure was developed that allows coils to be optimized in hours instead of the unfathomable number of years a brute-force technique would require. The bi-convex stream-function optimization method was then used to optimize the coils that are represented as stream functions. The coils that were optimized using the new method have new, interesting wire paths and are shown to perform significantly better than those optimized using a conventional representation, especially when the performance in magnetic soils is considered.