COMBINING DESIGN AND OPERATIONAL KNOWLEDGE TO ENHANCE GENERATOR PLANTDIAGNOSTICS

An approach to the construction of an intelligent system for performin g generator plant diagnosis in which a model of diagnostic expertise i s constructed prior to actual system development is considered. The us e of the expertise of both plant designers and operators has allowed a n enhanced model of generator plant diagnostics to be constructed whic h can identify a fault from a set of observed symptoms in addition to predicting the likely operational impact of the identified fault, This will assist the operator in making better decisions when under severe time pressure. Furthermore, it will allow the operator to assess an a ppropriate course of remedial actions. The paper discusses the feature s of design and operational knowledge and how these can be used to der ive enhanced diagnostics, A methodology for creating the model of diag nostic expertise is then described. Once the model has been constructe d an intelligent system can be developed. An architecture for such an intelligent system is discussed in the paper. An advantage of the mode lling approach is that it allows flexibility in the actual implementat ion. The use of a number of potential intelligent system technologies for implementing the system is discussed. A hybrid power station imple mentation is described which utilises many intelligent system technolo gies to further extend the decision support available to operators.