Application of fuzzy inference systems for evaluation of failure rates of power system components

Reliability parameters, such as the failure rates of power system components, are vital in evaluating power system reliability. This paper summarizes the research of the authors in using fuzzy inference systems to infer the failure rates of transmission lines in the power systems affected by hurricanes. The emphasis is on using fuzzy clustering methods to build fuzzy inference systems automatically. Here, two fuzzy clustering methods, subtractive clustering and fuzzy c-mean clustering, are adopted and compared in details. Besides, adaptive neuro-fuzzy inference system (ANFIS) is used to improve the performance of subtractive clustering. Then, the obtained results are compared to those of fuzzy c-mean clustering. Finally, possible future research on this topic is proposed. The proposed approaches were applied to the modified IEEE reliability test system (RTS). The numerical results show that the proposed approaches are efficient and are flexible in their applications.