Prediction of line voltage stability index using supervised learning

Abstract

In deregulated environment, stability issues have become dominant. Reliability of the power is essential for successful operation of the power system. Often high and dynamic loading conditions present new challenges in terms of decision of the control strategies to the system operator at energy management centre. For the achievement of voltage stability, identification of weak buses is very important. Line stability indices are important predictors of the weak buses in the over loaded system. Identification of the weak buses is the first step of control strategy. This paper presents an effective methodology based on Artificial Neural Network (ANN) to predict the Fast Voltage Stability Index (FVSI). Comparative analysis of different topologies of ANN is carried out based on the capability of the prediction of FVSI. Results are validated through offline Newton Raphson (NR) simulation method. The proposed methodology is tested over IEEE-14 and IEEE-30 test bus System.