Advanced parallel classifier model for dynamic stability diagnosis in power system

Abstract

Abstract Data for dynamic stable state in power system are very large. This leads to making difficult for a single classifier to learn boundaries of classes. This paper proposes a procedure to build Advanced Parallel Classifier Model (APCM) for dynamic stability diagnosis in power system with the aim of improving diagnosis accuracy. To build APCM, the Sequential Forward Floating Search is applied to select feature subset, and the Hybrid K-means is applied to choose data reduction and partition subsets. To find feature subset and sub-dataset, K-nearest neighbour classifier is employed to evaluate the classification accuracy. Then, the APCM is built by three kinds of classifiers as follow: the Multi-layer Perceptron Network, the Generalize Regression Neural Network, and the Support Vector Machines. The study is implemented on IEEE 39-bus power system network. The simulation results showed that the proposed the APCM can achieve classification accuracy higher than that of the single classifier.