A non-direct method for contingency ranking and classification in power systems using neural networks

摘要

Power system security is assessed in three stages: contingencies prediction, contingencies ranking and severe contingencies evaluation. The most important in these is contingency ranking. In this stage contingencies are ranked based on their seventies. In on-line control only the most severe contingencies are investigated. For contingency ranking many methods are presented in two categories: direct and in-direct methods. This paper presents an effective non-direct contingency ranking method based on neural network. The approach has been tested and evaluated on WEE Reliability Test System (IEEE-RTS) with satisfactory results.