Online Voltage Stability Assessment Using Artificial Neural Network considering Voltage stability indices

Abstract

The role of voltage stability indicators (VSI) is prominent in ascertaining the power system voltage stability. This paper illustrates an implementation of the artificial neural network (ANN) based on the Levenberg Marquardt (LM) technique to forecast the voltage instability of the system using voltage stability indicators. In this context, the performance of two indicators namely, fast voltage stability indicator (FVSI) and line stability factor (LQP) are compared. These indicators are used here to determine the vulnerability of the bus based on maximum loadability. Furthermore, the optimal static VAR compensator (SVC) location is obtained to combat the dynamic behavior of voltage instability. The neural network for forecasting is trained using these line indices. The real and reactive powers of different load buses are fed as input variables. The results demonstrate the feasibility of the techniques to voltage stability assessment using SVC and enhancement to variation of real and reactive power of loads with high precision. Newton Raphson(NR) method is used for a load flow analysis of the IEEE-14 bus standard system and the entire study is performed using MATLABO/NN Toolbox.