A Stratified Method for Large-Scale Power System Transient Stability Assessment Based on Maximum Relevance Minimum Redundancy Arithmetic

Abstract

A stratified method to transient stability assessment in large-scale power systems based on mutual information theory and artificial intelligence algorithm is proposed in this paper. A set of inter-complementary dynamic stability features are picked up one by one through the maximum-relevance minimum-redundancy (MRMR) algorithm. Besides, multiple extreme learning machines (ELMs) are trained based on the generated feature datasets. Because of the high requirement of evaluation speed in practical application, in order to balance the contradiction between assessment speed and accuracy, a hierarchical assessment structure is adopted in the final assessment process. Different ensemble classifiers with different response times are trained to construct different layers. The performance of the proposed technique is tested in the IEEE-39 bus system and a practical 1648 bus system provided by PSS/E. The experimental results indicate that, compared to other traditional methods, the proposed hierarchical method can give a more accurate result in a shorter period of time. As an efficient method, it is suitable for on-line transient stability assessment.

Keywords

• Transient stability assessment
• maximum-relevance minimum-redundancy
• extreme learning machine
• ensemble learning
• artificial intelligence