Fault diagnosis method of modular multilevel inverter based on state observation

Abstract

The invention discloses a novel fault diagnosis method of a modular multilevel inverter based on extended state observation. In the stage of fault detection, an extended state variable observed value which is magnified and filtered is compared with an extended state variable calculated value which is filtered in a hypothetical normal working condition by an output matching module, and whether the modular multilevel inverter has faults is determined according to comparative results; in the stage of determining fault positions, extended state variable calculated values which are filtered in hypothetical fault positions are respectively compared with the extended state variable observed value which is output by an observer module and is magnified and filtered by the output matching module, and the fault positions of the modular multilevel inverter are determined according to comparative results. The novel fault diagnosis method aims to determine whether the modular multilevel inverter has the faults by utilizing an extended state observer and reversed switch state variables, and if the faults occur, the fault positions can be quickly and accurately determined.

Description

technical field [0001] The invention belongs to the field of fault diagnosis of modular multilevel inverters, and in particular relates to a novel fault diagnosis method for modular multilevel inverters based on state observation. Background technique [0002] A Modular Multilevel Converter (MMC for short) is a kind of multilevel converter emerging in recent years. The structure of the modular multilevel inverter was first proposed by Professor R. Marquardt in 2001. Compared with the traditional multilevel inverter structure, the outstanding advantages of MMC include: 1) High modularity, making this The inverter can realize the application of any voltage level by cascading modules; 2) The harmonic characteristics are good, due to the cascading of the same module, the number of levels is large, and the AC output side does not need to use a filter; 3) Redundancy and fault tolerance Strong capability, the internal structure of each module is the same, when a large number of mo...

AI Technical Summary

Problems solved by technology

[0006] However, although the above several fault diagnosis methods can all have the function of MMC fault diagnosis, they all have their own defects.

For example, in the synovium observer method, multiple physical quantities need to be observed and compared, which increases the complexity of the algorithm; while in the Kalman filter method, although there is only one physical quantity to be observed and compared, compared with the synovium observation The device method, the algorithm is simple, but because it only determines the fault location by comparing the measured value of the capacitor voltage of each module, it can only determine the fault module, but cannot specifically determine the fault switch device

The defects of the above methods have affected their use in fault diagnosis

Images