Performance investigation of SVR for evaluating Voltage Stability Margin in a Power utility

Abstract

Voltage stability margin is one of the key parameters to ensure the credibility of the power system in real time. Voltage stability margin indicates the maximum loading capacity of the system i.e, further increase in load will collapse the system voltage, and the system becomes unstable. Hence in real time, the evaluation of voltage stability margin is crucial and requires more significance. In recent times, reinforcement learning algorithms attracts a great deal of interest for determining the voltage stability margin. In this paper, voltage stability analysis on the power system has been accomplished using the Continuous power flow method with Spider SVM regression algorithm. The analysis was carried out in the IEEE 118 bus system, with the help of PSAT software in MATLAB. The result obtained from SVM regression model satisfies the expected model with the best accuracy for many cases.