180 / 2022-03-14 16:44:02

Open circuit fault diagnosis method based on ARTIFICIAL intelligence algorithm MMC sub-module

modular multilevel converter,fault diagnosis,open circuit,artificial intelligence algorithm

A fault detection method is proposed, which does not need to collect fault samples and only trains the classification model according to normal samples. The open circuit fault of MMC sub module will cause the distortion of fault bridge arm current and abnormal charge and discharge of capacitor of fault sub module. This subject plans to use the bridge arm current as the fault parameter.The self-organizing mapping neural network belongs to unsupervised learning, supports online learning and visualization, trains the normal value of bridge arm current, calculates the distance between bridge arm eigenvalue and threshold for fault detection. After the fault arm is diagnosed, the capacitance voltage slope of the sub module is used as the fault parameter. We regard fault location as a classification problem. The capacitance voltage change rate of the fault sub module can be regarded as an abnormal value when the fault characteristics appear. Because the outliers are outliers far away from all normal values, the outliers are regarded as one class and the normal values as one class. The detection of outliers is realized only by calculating the distance, so as to realize the positioning of fault sub modules.