A CASE-BASED INTELLIGENT TRAINING AGENT WITH AN APPLICATION TO POWER SYSTEM RESTORATION

Abstract

A speedy restoration of a power system can minimize the consequences of a total or partial blackout. Due to the increase in system reliability, major disturbances occur less frequently and consequently system operators receive little experience in restoration. Power system operators seldom get working knowledge of restoration of a whole system from their daily work. However, system operators can be trained to handle the necessary steps of restoration with the help of a simulation program. This research attempts to address this need. In this research, a simulation program named as Intelligent Training Agent (ITA) has been developed to train system operators the steps of power system restoration. An object-oriented approach has been utilized to develop the graphical user interface of the simulation program. A case-based reasoning approach has been applied to form the required knowledge base of the simulation program and to perform the reasoning task. Several numerical analysis algorithms have been used to verify the restoration actions. By using the case-based reasoning mechanism, the proposed intelligent agent can provide a restoration plan when the initial system conditions are known and the goal state of the system is submitted to the ITA. System operators, therefore, can interact with the agent through the object-oriented graphical interface. The interface will enhance the operator's learning process. The structure of the agent consists of an object-oriented graphical interface block, a case-based reasoning block, and an action check block.