Artificial Intelligence/Expert System (AI/ES) Pre-processor for Computational Electromagnetics (CEM)

This proposed effort is for an unclassified research and exploratory development program to adapt AI/ES software technologies for the design and demonstration of a smart pre-processor for Computational Electromagnetics (CEM). Despite their inherent power, most CEM tools can challenge even the most proficient user, let alone the novice. One method of alleviating difficulties is by adapting AI/ES technologies to establish a user-friendly pre-processor. This innovative approach employs a knowledge or rule base to assist CEM code users during input modeling tasks, particularly for the detailed definition of geometries, and the automated selection of optimum physics formalisms and numerical solutions. This will facilitate efficient use of CEM codes and will readily allow a wide range of problems to be addressed. While it will alleviate the burdensome modeling process often associated with sophisticated CEM codes, it also will establish a foundation for intelligent CEM post-processing. It will have direct relevance to commercial/private sector applications and will be marketed as a dual-use technology. To determine the technical feasibility of developing an enhanced capability for various CEM codes in Phase II, Phase I focuses on the design and demonstration of a prototype tool called the Intelligent CEM Expert System (ICEMES) Pre-processor, for the U.S. Air Force code called the General Electromagnetic Model for the Analysis of Complex Systems (GEMACS). ICEMES will convert CAD system model data into a valid EM structure model, or validate user-created models, and store such in a syntactically-correct input format compatible with GEMACS.