Switching decision using a knowledge based system for distribution networks

A real time knowledge-based system is developed to assist switching decisions in open loop operated distribution networks. This system applies the logical ways and interlocking rules to avoid wrong switching decisions which may result in severe conditions. The developed system features a composite knowledge base of rules and frames, and a backward chaining inference engine. It manages to keep an up-to-date view of the entire network status in its knowledge base by dynamically updating its model for switching states in real time whenever a switching request is honored. Frame-based knowledge is used to capture both network topology and the technical and operating states of the bays and the switch gear. Alternatively, the logical and interlocking aspects are represented using rule-based knowledge. For different system operation conditions, the inference engine determines and applies the appropriate strategy for taking a switching decision. A prototype has been implemented and run on a personal computer using Turbo Prolog. Test results and frame representations for a simple network are demonstrated.