Expert System for Power Plant Operator Performance Evaluation

The aim of this work is to design an intelligent system for evaluating knowledge level of power plant operators who have completed training program. To build the evaluation system, skilled plant operator performance during fault conditions is simulated through expert system techniques. Rule based expert system developed, stores domain knowledge as set of rules. The system is integrated with Supervisory Control And Data Acquisition (SCADA) System of the power plant. User input to the system is a set of SCADA alarm names. Through backward knowledge inferencing process all reason codes related to particular alarms are extracted by the expert system. The responses obtained from trainees are then compared with inferred knowledge and alarm configuration values obtained from data base. Score calculating programme running at the back ground prepares cumulative scores which is used to classify the knowledge level of employees. Traditional method of performance evaluation by operational experts requires their man-hours and results differ between experts, which prevents a systematic analysis. The proposed scheme follows a prototypical methodology, which provides means for comparing different operators in a systematic manner in less time.