Optimal feature selection for islanding detection in distributed generation

• Published in: IET smart grid 2018-10, Vol.1 (3), p.85-95
• Main Authors: Chandak, Sheetal, Mishra, Manohar, Nayak, Subrat, Rout, Pravat Kumar
• Format: Article
• Language: English
• Subjects: Accuracy; Algorithms; anti-islanding techniques; Artificial neural networks; B0260 Optimisation techniques; B8120J Distribution networks; B8120K Distributed power generation; C1180 Optimisation techniques; C6170K Knowledge engineering techniques; C7410B Power engineering computing; Computing time; Datasets; designed IEEE 13 bus system; Distributed generation; distributed generators; distributed power generation; distribution system; Electric power systems; Evolutionary algorithms; Evolutionary computation; feature evaluation; feature extraction; Feature selection; IEEE 1547 standards; IEEE standards; invertors; islanded condition; islanding detection; Islanding technique; kernel-based extreme learning machine classifier; learning (artificial intelligence); Machine learning; Methods; multiobjective differential evolution algorithm; optimal feature selection; Optimization; optimum features; particular detection feature; pattern classification; power distribution faults; power engineering computing; selected optimal features; standard objective functions; system protection; teething problems; wrapper feature selection approach
• ISSN: 2515-2947; 2515-2947
• DOI: 10.1049/iet-stg.2018.0021

The integration of the distributed power generation into a distribution system comes with several system problems. One of the teething problems related to system protection is islanding detection. Various anti‐islanding techniques based on feature evaluation were proposed in the recent past. However, they overlook the need for justifying the selection of a particular detection feature among all the possible measures. In this study, a wrapper feature selection approach is proposed where a modified multi‐objective differential evolution algorithm is coupled with a kernel‐based extreme learning machine classifier. To select the optimum features, five standard objective functions have been considered, such as dependability, security, accuracy, F‐measure, and the number of features. About 1864 cases have been generated from the designed IEEE 13 bus system to extract the sensitive features. IEEE 1547 standards have been considered while designing and testing the IEEE 13 bus system against islanding. The selected optimal features detect the islanded condition decisively for both synchronous and inverter‐based distributed generators. The features also validate their performance under noisy environment accurately with lesser computational time.