Development of an Online Static Power System Security Assessment Module Using Artificial Neural Networks in 118- Bus Test System.

Abstract

Contingency analysis is an important task in today’s power system. Fast and accurate contingency analysis is some of the major issues. In this paper two types of Artificial Neural Network (ANN) viz. Multilayer feed forward neural network (MLFFN) and Radial basis function network (RBFN) are used to implement online static security assessment. Newton Raphson (NR) method is done on an IEEE 118-test bus system and Composite Security Index (CSI) is calculated. Loads are varied from the base case values and for each load condition, line flow and bus voltages are calculated using a model based on the NR load flow method for training an ANN with the help of back propagation algorithm. Expected range of load variation and randomly selected 20-contingencies are tested in the training ANN model. The results obtained by the above ANN methods are matched with NR methods. The CSI is found out for various loads and contingencies in MLFFN and RBFN. The computation time required for MLFFN and RBFN is compared with NR method and found that RBFN is using less computation time average of 35.67291s