A model-based approach to on-line process disturbance management

A methodology has been developed, which can be applied to the design of a real-time expert system to aid control-room operators in coping with process abnormalities. The approach encompasses diverse functional aspects that are required for an effective on-line process disturbance management: (1) intelligent process monitoring and alarming, (2) on-line sensor data validation, (3) on-line sensor and hardware (except sensors) fault diagnosis, and (4) realtime corrective measure synthesis. Accomplishment of these functions is made possible through the application of various models, goal-tree success-tree, process monitor tree, sensor failure diagnosis, and hardware failure diagnosis models. The models used in the methodology facilitate not only the knowledge acquisition process--a bottleneck in the development of an expert system--but also the reasoning process of the knowledge-based system. These transparent models and model-based reasoning significantly enhance the maintainability of the real-time expert systems, which is one of the primary concerns in practical applications of expert system techniques. The proposed approach has been applied to the feedwater control system of a nuclear power plant, and implemented into a real-time expert system, MOAS II, using the expert system shell, PICON, on the LMI machine. Performance of MOAS II was tested for a variety of transient scenarios of the target process, which were obtained by emulating occurrence of process faults in a dynamic simulation program. The test results show that the expert system successfully carries out its intended functions: timely recognition or detection of occurring disturbance, diagnosis of the disturbance cause, and presentation of optimal control advice to the operator. Therefore, the model-based technique lends itself to the development of a valuable operator aid for on-line process disturbance management.