A diagnostic method for open-circuit faults of switching tubes in sub-modules of modular multilevel converters

Abstract

The invention discloses a method for diagnosing an open-circuit fault of a sub-module switch tube of a modular multilevel converter, belonging to the technical field of flexible direct current transmission and distribution. Based on the sub-module capacitor voltage distribution characteristics of a modular multilevel converter under an open-circuit fault, a machine learning anomaly detection model is constructed to realize fault diagnosis. First, the voltage raw data is segmented using a sliding window, and then the time-domain features are extracted from each data segment to construct a model data set. The model parameters and probability density threshold based on the multivariate Gaussian distribution are obtained by using the data set training. Comparing the sub-module with the probability density threshold can determine the working status of the sub-module. The present invention does not need to add additional sensors or reconstruct sensors; the present invention is realized based on actual data, and does not need to build an accurate modular multilevel converter mathematical model, and has strong robustness; and the present invention is more robust than the classification machine learning model It has higher fault diagnosis accuracy and lower false alarm rate.

Description

technical field [0001] The invention belongs to the technical field of flexible direct current transmission and distribution, and in particular relates to a method for diagnosing an open-circuit fault of a sub-module switching tube of a modular multilevel converter. Background technique [0002] Due to the advantages of high modularity, high robustness and high controllability, modular multilevel converters have recently attracted extensive attention in HVDC transmission systems. Despite its many advantages, safety remains one of the most important challenges because a modular multilevel converter is constructed of many cascaded sub-modules, each of which is a potential point of failure. [0003] According to the fault condition of the switch in the sub-module, the sub-module fault can be divided into open circuit fault and short circuit fault. As for short-circuit faults, they are usually handled by the gate driver due to their high destructiveness and short time scale due...

AI Technical Summary

Problems solved by technology

However, the generalization ability of multi-classifiers is limited when encountering novel types of faults not included in the classification, which leads to misdiagnosis

Images