Elaboration of the SEPT expert system as the coupling of a simulator and a diagnostician

ABSTRACT

We present the SEPT expert system which has been elaborated to resolve one of EDF's most crucial problems, namely the rapid processing of complex analysis for the repair and maintenance of EHV (Extra High Voltage) substations. In the first step, we implemented a prototype, which we called diagnostician, to validate the expert system approach. In the second step, a functional model of the system, known as the simulator, was developed to reinforce the diagnostician. The original nature of our particular diagnosis problem led us to opt for a type of coupling that does not rely on the bases usually evoked in literature on the second generation expert systems. The two prototypes are coupled for two reasons: first, to completely validate the diagnostician together with the simulator (inputs and outputs of both prototypes will be connected), and, second, to develop later an explicative expert system for training future operators. This paper outlines the reasons behind our choices.

REFERENCES

• BRE 88a: Brizillon P., P. Fauquembergue & A. Hertz, "SEPT, an expert system approach for the monitoring of EVH substation control equipment", In- Symposium on Expert Systems Application to Power System, Stockholm- Helsinski, August 22-26, 1988, p.6-9, 6-13.
• BRE 88b: Brizillon P., Fauquembergue P., "SEPT: un syst~me expert pour la Surveillance d'Equipements dans un Poste ~t tr~s haute Tension", EDF Bulletin de la Direction des Etudes et Recherches, S6rie B, R6seaux Electriques, Mat6riels Eleetriques, 1988, 3' 5-10.
• BRE 89a: Br6zillon P. & D.Y. Bau, "Elaboration du syst~me expert explicatif SEPT: choix du couplage entre le simulateur et le diagnostiqueur", Rapport de Recherche N~ 508 du LRI, Universit6 PARIS XI, Orsay, France, Aofit 1989.
• BRE 89b: Br6zillon P."Interpretation and rule packets in an expert system", Proc. of the IASTED International Symposium Expert Systems Theory & Applications, Ziirich, Switzerland, June 26-28, 1989, Acta Press, Anaheim CA, USA, pp.38-40.
• BRE 89c: Patrick Brezillon, Interpretation and rule packet in expert systems. Application to the SEPT expert system, Proceedings of the international conference on Knowledge based computer systems, p.78-87, August 1990, Bombay, India
• BRE 89d: Br6zillon P., "METAL: un moteur d'inf6rences bas6 sur la gestion dynamique de paquets de r~gles et d'interpr6tations", Rapport de Recherche du L.R.I. , Universit6 PARIS XI, Orsay, France, (to appear).
• BRE 90a: Br6zillon P. & D.Y. Bau, "Elaboration du syst~me expert SEPT: II Un simulateur de poste ~t THT en METAL", Rapport de Recherche LRI, N~ 548, F6vrier 1990.
• BRE 90b: Br6zillon P., "Elaboration du syst~me expert SEPT: III Un diagnostiqueur en METAL pour la surveillance des 6quipements dans un poste h TILT", Rapport de Recherche LRI, N~ 550, F6vrier 1990.
• CHA 83: Chandrasekaran B. & S. Mittal, "Deep versus compiled knowledge approaches to diagnostic problem-solving", Int. J. Man-Machine Studies, 19 (1983), p. 425-436.
• CHA 87: Chandrasekaran B. & W.F. Punch, "Data validation during diagnosis: a step beyond traditional sensor validation", Proc. of AAAI-87, Seattle, Washington, July 1987, p.778-782.
• CLA 86: William J Clancey, From Guidon to Neomycin and Heracles in twenty short lessons, AI Magazine, v.7 n.3, p.40-60, August 1986
• CLA 89: William J. Clancey, Viewing Knowledge Bases as Qualitative Models, IEEE Expert: Intelligent Systems and Their Applications, v.4 n.2, p.9-15, 18-23, June 1989 [doi>10.1109/64.24918]
• DAV 83: Davis R., "Reasoning from first principles in electronic troubleshooting", Int. J. Man-Machine Studies, 1983, 19: 403-423.
• KNO 88: Knowledge Craft, User's Manual, Carnegie Group Inc., 1988.
• FAU 88: Fauquembergue P., Br~zillon P., Harmand Y., "Synthesis of events in an EHV substation' an expert system approach", IFAC Symposium on Power Systems, Modelling and Control Applications, Brussels, Belgium, 1988, pp. 16.6.1,I6.6.5.
• FAU 89: Fauquembergue P., Perrot L., "An expert system in fault analysis of electric power system", In' Symposium on Expert Systems Application to Power System, Seatle, August 22- 26, 1989.
• MUL 85: Mulet-Marquis D., Gondran M., "Un langage pour les syst~mes experts' ALOUETTE" Bulletin de la Direction des Etudes et Recherches d'EDF, S6rie C, Math6matiques et Informafique, 1985, 3/4: 5-42.
• ROU 88: Rousset M.C., "On the consistency of knowledge bases: the COVADIS system", Proceedings of the ECAI 88 Conference, Munich, Aofit 1988.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.1 Applications and Expert Systems
          Subjects: Industrial automation

Additional Classification:
  C. Computer Systems Organization
  C.4 PERFORMANCE OF SYSTEMS
      Subjects: Reliability, availability, and serviceability

  J. Computer Applications
  J.2 PHYSICAL SCIENCES AND ENGINEERING
      Subjects: Electronics

General Terms:
Design, Reliability