An Identification and Location Method for Power Quality Disturbance Sources in MMC Converter Based on KNN Algorithm

With the application of new energy and DC transmission, various of electronic devices have been widely used and brought great economic and social benefits. At the same time, new power quality problems have also been produced by these electronic devices. In order to solve power quality problems fundamentally, identification and location for power quality disturbance sources is needed. This paper focuses on MMC converter, and proposes an identification and location method based on K nearest neighbor (KNN) algorithm. First, featured parameters of the fault are obtained by analyzing the power quality characteristics of the fault of MMC. Second, the parameters are used for designing and training the KNN classifier. Third, based on IEEE30 bus system with MMC inverters paralleled, simulations are applied and power quality parameters are acquired. Then the parameters are analyzed by the trained KNN classifier, by which the results of identification and location of the fault are obtained. Eventually, by comparing with other machine learning algorithms, the accuracy of proposed method is further verified to be more accurate and more effective.