Verification of an advanced power system restoration support system using an operator training simulator

This paper describes an advanced restoration support system for regional load dispatching and local supervisory control centers. A model-based diagnostic task shell is applied to diagnose network faults, and an inference algorithm not dependent on a specific power system has been developed exclusively for restoration plan determination. With this system, adjustments for network fault diagnosis, necessitated by power system facility expansion or alteration, can be made simply by adding or changing models. Restoration knowledge can be partially revised in the case of restoration policy change. Connected on-line to an operator training simulator, this system´s restoration support performance has been verified through daily restoration training. In this process, several problems of the system were detected and corrected. Consequently, it has become possible for the system to support restoration almost as well as a trainee, thereby confirming that connecting this system to an operator training simulator is an effective method for both verifying and improving restoration support system performance