Clustering algorithm based open-circuit failures diagnosis for modular multilevel converters

Based on the abnormal change of capacitor voltage caused by the fault of modular multilevel converter (MMC) sub-module (SM) switching device, a fast fault detection and diagnosis method is proposed in this paper. To enable effective consistency evaluation of numerous SM voltages, the density-based spatial clustering of applications with noise (DBSCAN) algorithm, which is an ideal analyzing method for the clustering analysis of big data, is applied to SM failure diagnosis in this paper. The DBSCAN is used to analyze the capacitor voltages of all sub-modules in the bridge arm, so as to realize the rapid detection and location of faulty SMs at the same time. The proposed method can be applied to any fault types of faults, including single fault, multiple faults in the same bridge arm, and multiple faults in different bridge arms. This method is not affected by the uncertainty of system parameters, and at the same time, has stronger applicability and faster diagnostic rate than the traditional methods. Experimental results demonstrate the validity of the proposed diagnosis strategy.