An integrated scheme for static voltage stability assessment based on correlation detection and random bits forest

Abstract To make static voltage stability assessment (VSA) more efficient for practical operation of power systems, an integrated scheme is proposed to rapidly predict the voltage stability margin (VSM) based on correlation detection (CD) and random bits forest (RBF). A feature selection framework is designed to select the representative features strongly related to the VSM with lower redundancy based on CD algorithms. By using the pivotal feature set and the corresponding VSM, the training of the RBF-based prediction model can be achieved. Once the real-time operation information of systems is received, the trained model will provide the corresponding result rapidly. The proposed scheme is examined on the IEEE 30-bus system and a practical 7917-bus system, and the encouraging prediction performance is verified. Moreover, the influences of the uncertainty of load increase direction (LID), variation of generation distribution (GD) and topology change on VSM prediction are analyzed.