Application of Machine Learning to Generate a Contingency Ranking for Power System Security

Recent grid expansion has included several renewable and nonrenewable power sources in an effort to meet increased demand. Transient load variations compound the issues for generators, transmission lines, and distribution networks. Predicting the system's characteristics requires investigation into the most suitable load modeling. A ZIP load model is used in this research strategy because it produces accurate and durable representations of loads over long time periods when combined with contingency criteria and constant-impedance, constant-current, and constant-power loads. Purpose. The proposed technique, infers the single transmission line outage for obtaining the contingency ranking using Machine learning application to ensure the system behavior. Methods. The proposed mathematical model with Particle swamp optimization algorithm has been used for observing the stability analysis with and without considering STATCOM and Unified Power Flow Controller along with cost analysis. Results. The power system involves many unpredictable conditions or contingency conditions like single transmission line outage, double transmission line outage, generator outage and load variations. This paper mainly focuses on the single transmission line outage and generator outage for obtaining the contingency ranking. Practical value. The recommended stability analysis has been very beneficial in establishing a secure transmission power system.