Knowledge discovery and data mining for power system contingency analysis

Abstract

Power system contingency analysis is an important task to discover underlying problem in system operations and planning. Due to tremendous amount of data involved in the process, it is not easy for power system operators and engineers to interpret the security assessment results in a quick and intuitive manner. Instead, some approximate approaches such as generated security constraints in the form of generic constraints are used by some power companies and system operators. However, in many cases, those generated generic constraints are mostly based on experiences following numerous simulations. There lack a systematic way in obtaining system security constraints yet. In this paper, we present an approach for extracting security constraints in the form of rules from the security assessment results by using rule extraction algorithms with neural network theory. Extracted security constraint rules represent security boundaries of the power system operation and hence provide useful and comprehensive security constraints for system operations and planning studies.