TABLE OF CONTENTS This chapter starts with a brief overview of the state-of-the-art of expert system applications to electric power systems. Several knowledge engineering techniques that were motivated by power system applications are reviewed: (1) identification of relations among rules or chains of rules, (2) estimation of the worst case processing time of rule-based systems, and (3) the equivalence class method for validation and verification of rule-based systems. The first issue, relation checking, is considered the most practical among the three and, therefore, the subject is discussed extensively in this chapter. A general relation checking algorithm developed at the University of Washington is described. A representation of rule-based systems in the attribute space is proposed. This representation is used to define several relations among rules. The relations defined are cause-effect, mutual exclusion, redundancy, conflict, subsumption and implication. A relation between a rule and a chain of rules is either complete, i.e. the relation holds for all instantiations of the rules, or partial, i.e. the relation holds only for some instantiations of the rules. An algorithm to detect relations between a new rule (to be added to the rule base) and rules in the rule base is developed. Example applications of this algorithm to rule-based systems are provided.
For more than a decade, researchers and practical engineers in many countries have been developing expert systems (or knowledge-based systems) for power system applications. The results of these research and...
