A diagnostic algorithm based on models at different level of abstraction

TLDR: A novel model-based diagnostic method, whose distinctive features make it practical for diagnostic problem solving in automated systems for monitoring continuous processes, and makes use of models at different levels of abstraction, qualitative and quantitative.

Abstract: The difficulties encountered in applying knowledge-based system technology to complex industrial environments have made the need for representing and using deep knowledge about physical systems increasingly clear to system designers. A rather large number of approaches to modeling and reasoning with deep knowledge have been experimented, but the impact of these new techniques, often referred to as model based reasoning, on real applications is still poor. This paper presents a novel model-based diagnostic method, whose distinctive features make it practical for diagnostic problem solving in automated systems for monitoring continuous processes. The method we introduce makes use of models at different levels of abstraction, qualitative and quantitative. In particular, we discuss an algorithm based on a quantitative, real-valued algebraic model, and a qualitative causal model that can be easily derived from the former in an automated way. The causal model is used for candidate generation, and the real-valued model for validation/rejection of candidates.