DIAGNOSING CONTINUOUS SYSTEMS WITH QUALITATIVE DYNAMIC-MODELS

This paper presents several innovations in the development of model-ba sed diagnostic systems for diagnosing faults in continuous dynamic phy sical systems. The approach utilises recent developments in qualitativ e simulation techniques to cope with the inherent lack of modelling kn owledge and to provide a qualitative description of the dynamic behavi our. In particular, techniques for the synchronous tracking of the mod el-based predictions and the evolution of the physical system between equilibria are developed. A discrepancy metric is defined that allows for the continuous degradation of the system behaviour from normal to faulty to be detected. And, most fundamentally, a method for iterative ly searching through the space of possible model variations is present ed. This provides explicit feedback from detected discrepancies to mod el adjustments and has the important advantage of reducing the sensiti vity to modelling errors and approximate fault models. In the limit, n o fault models are required. However, if available these can be used t o initialise the search. An example is included which outlines the bas ic approach discussed in this paper.