Integrated Approach for Online Dynamic Security Assessment With Credibility and Visualization Based on Exploring Connotative Associations in Massive Data

An integrated approach for online dynamic security assessment based on exploring connotative associations in massive data is proposed; this approach consists of three stages and can give visual and credible dynamic security assessment results. The relationships of operation variables and transient stability margin are assigned scores by the maximal information coefficient and the Pearson-correlation coefficient. The connotative nonlinear functional relationships and linear relationships are explored by ranking the scores, and some highly ranked relationships are shown and curve fitted. In this approach, a processing unit is designed for dynamic security assessment based on the explored highly ranked relationships. The approach considers classification and prediction. This approach is tested in a 21-bus test system and a larger 1648-bus system. The robustness with respect to topology changes, variation of power distribution among generators/loads, and variation of peak load/minimum load are analyzed. The test results indicate that compared with the traditional methods, the proposed approach provides a relatively fast and accurate dynamic security assessment and is suitable for online use.