ARTIFICIAL NEURAL NETWORK BASED VOLTAGE STABILITY EVALUATION

摘要

Voltage stability has been a major concern for the power system utility, because of several events of voltage collapse in past decade. It becomes therefore, necessary to develop tools for fast determining the voltage stability state of the power system. With the developments of FACTS devices it has been possible for the power system to operate closer to its operational limit. In this paper, loadability margin, which is defined as the distance from operating point to voltage collapse point, is calculated using artificial neural network for the power system with Static Var Compensator (SVC). Multi layer feed forward neural network is proposed. To reduce the size of neural network, input features are selected by employing system entropy method. All ac limits are considered. The proposed method is applied to IEEE-30 bus and IEEE-118 bus system. The results obtained are promising in terms of accuracy and speed.