In this paper, a gap-statistic-algorithm (GSA)-based data-mining technique is applied to enhance the data debugging in power system operations. In the proposed approach, the GSA technique is embedded into a neural network frame in anticipation of improving the detection capability of bad data. Thanks to the clustering capability exhibited by GSA in which the number of clusters can be optimally determined, the proposed approach becomes highly effective to localize the group of abnormal data. This proposed approach has been tested through the data collected from different scenarios made on an IEEE 30-bus system and 118-bus systems. Test results reveal the feasibility of the method for the data diagnosis applications.
All Science Journal Classification (ASJC) codes
- Energy Engineering and Power Technology
- Electrical and Electronic Engineering