Adaptable Modeling of Electrical Systems

Much research in model-based reasoning has concentrated on the use of a single, usually qualitative, level of modeling. This is less than ideal for many engineering applications, where the amount of knowledge available about a product design increases during the design process. This paper describes a framework that enables a modelbased system to use more information about an electrical design as it becomes available. It can provide common results at different stages through the design process, while using more detailed design information to resolve situations that are ambiguous in the qualitative case. The paper concentrates on the modeling of electrical systems containing electronic and mechanical components, specifically in vehicles, but considers the more general lessons to be learned from this work.