Reasoning Applied to Substation Data Extracted from Relays and Circuit Breakers

Abstract- The paper describes how rough set theory can extract decision rules from the data acquired by relays and circuit breakers. The methodology correctly estimates the fault section and can help operators in their decision making process. Developing a knowledge based diagnostic system is always costly and time consuming. Extra and superfluous conditions in a knowledge base may lead to serious inconve-niences especially in rules maintenance. Reducing the size of the knowledge base and improving the quality of knowledge acquisition, benefits the operating behaviour of the power system. The proposed technique not only induces the deci-sion rules, it also reduces the size of the knowledge base with-out the loss of useful information. Results can be used by an expert system or a neural network to generate supervisory automation and to support operators during an emergency situation. The former includes the generation of HMI alarms and the latter, the diagnosis of the type and cause of the event and suggestions for network restoration and post emergency repair. The PSCAD/EMTDC simulator has been used to in-vestigate the effect of faults and switching actions associated with a typical distribution network. The fundamental ideas of rough set theory are discussed, followed by a rule assess-ment method that is outlined using an illustrative example