Probabilistic neural network for vulnerability prediction on a practical power system

Vulnerability prediction of power systems is important so as to determine its ability to continue to provide service in case of any unforeseen catastrophic contingency. It is considered one of the vital concerns due to the continual blackouts in recent years which indicate that the power system today is too vulnerable to withstand a severer disturbance. The objective of this paper is to investigate and compare the performance of two vulnerability indices used for assessing the vulnerability of power systems when subjected to various contingencies. The Probabilistic Neural Network (PNN) based on power system loss and possible loss of load will be used to speed up the assessment technique. In this study, contingency analyses were carried out on a practical 87 bus test system and the vulnerability indices were calculated using the MATLAB program. Results presented show that PSL index is more accurate for analyzing the impact of contingencies on a practical power system from the view point of power system loss considering the loss of power during contingencies.