A Strategy for Qualitative Model Based Diagnosis

Abstract A novel approach to time-constrained model based fault diagnosis of ill-defined dynamic systems is presented. Fuzzy qualitative simulation techniques are utilised to generate qualitative predictions of the dynamic behaviour of the faulty plant. These are then compared to the observed behaviour to identify the correct qualitative parameter values. The precision of the diagnosis is determined by the precision of the fuzzy model. This allows the precision of the solution to be part of the problem specification and, hence, traded-off against time-constraints; less precise models take less computation time. Experimental results for a simplified 3rd order dynamic system are presented.

[1]  Q. Shen,et al.  Qualitativeness does not imply fuzziness , 1994, Proceedings of 1994 IEEE 3rd International Fuzzy Systems Conference.

[2]  Mike J. Chantler,et al.  MYCROFT: a framework for qualitative reasoning , 1994 .

[3]  Daniel S. Weld,et al.  Query-directed approximation , 1993 .

[4]  Qiang Shen,et al.  Fuzzy qualitative simulation , 1993, IEEE Trans. Syst. Man Cybern..

[5]  Edmund H. Durfee,et al.  Approximate Processing in Real-Time Problem Solving , 1988, AI Mag..

[6]  Rolf Isermann,et al.  Process fault detection based on modeling and estimation methods - A survey , 1984, Autom..

[7]  Benjamin J. Kaipers,et al.  Qualitative Simulation , 1989, Artif. Intell..