Online Assessment of Power System Static Security Using Artificial Neural Network

Abstract

Artificial Neural Networks (ANN) using Pattern Recognition technique for Static Security Assessment (SSA) of electric power system is presented. ANN using Pattern Recognition (PR) is a promising methodology for SSA. It classifies the power system states as secure/insecure, subject to bus voltage and transmission line thermal constraints for varying load conditions. Conventional methods of security assessment involves of performing continuous load flow by simulation program. This is highly time consuming and infeasible for on-line application. Hence ANN using Pattern Recognition is proposed to perform classification of the states into secure and insecure classes, a much desirable feature for on-line security analysis. The ANN models designed are tested on IEEE 14 bus and 39 bus New England test systems for each contingency. The performance of Static Security Assessment for varying loads conditions and for (n-1) contingency is evaluated with ANN using PR and the results are compared with Newton-Raphson (NR) method. The results illustrate that, Static Security Assessment with ANN using PR technique having very less computation time and high accuracy is preferred for on-line applications.