Assessment of the Risk of Voltage Collapse in a Power System Using Intelligent Techniques

This paper describes the implementation of a fast and easy-to-use, intelligence-based algorithm to assess the risk of voltage collapse when risk is defined as the product of the event likelihood and a severity function. In the event likelihood, the effect of weather is taken into account; the failure rate of each transmission line under different weather conditions is calculated using real historical outage data. A new severity function model that utilises the voltage collapse prediction index is proposed in this paper. Two intelligent techniques, i.e., support vector machines and a generalised regression neural network are studied, and their performances are evaluated using mean absolute and mean square error. The proposed methodology has been applied in a real power system network. Simulation results show that a generalized regression neural network provides the lowest mean absolute and mean square error.