A Fault Diagnosis Method of PWM Converter Based on Insulation Monitoring Device and Chaotic Detection Circuit

Abstract

The invention discloses a PWM converter fault diagnosis method based on an insulation monitoring device (IMD) and a chaotic detection circuit. The method comprises: (a), a positive electrode and a negative electrode of an IMD injection source based on a direct-current injection method are connected with a neutral line and a ground of an earth-free distribution system with a PWM converter circuit respectively; (b), because of a difference of IMD injection signal directions, a far-side voltage to earth of a Rs without the effect of an interference source Uc is measured; (c), on the basis of a basic formula of an insulation resistor of the earth-free distribution system, cooperation with cyclic switching of low, medium, and high voltage gears of the IMD is carried out, so that a resistance value Rt; (d), a fault of a switching device of a PWM converter of the earth-free distribution system is diagnosed; and (e), a control fault of the PWM converter of the earth-free distribution system is diagnosed. According to the invention, with utilization of the insulation monitoring function of the IMD and the chaotic detection circuit, features of the fault of the switching device and the control fault of the PWM converter are extracted; and when a fault occurs at the converter or a fault degrades, the diagnosis can be carried out rapidly and system maintenance can be carried out conveniently.

Description

technical field [0001] The invention relates to a fault diagnosis method for a grid-connected interface converter of a microgrid and an ungrounded power distribution system, in particular to a fault diagnosis method for a PWM converter based on an insulation monitoring device (IMD) and a chaotic detection circuit. Background technique [0002] The faults of PWM converters mainly include hardware faults and control faults. Hardware failure refers to switching device (such as GTO, IGCT, IGBT) failure, and control failure refers to PWM signal failure. The switch device itself is broken down or the insulation layer of the bridge arm is damaged, etc., which will cause a short circuit of the switch device, resulting in a rapid rise in current and a sudden drop in conduction voltage drop. In severe cases, the device will be damaged and the system will be paralyzed; When a switching device is poorly connected or burnt out due to overcurrent, it will cause the switching device to op...

AI Technical Summary

Problems solved by technology

[0004] The spectrum analysis method is to extract the time-domain signal of the fault. Usually, the Fourier transform is used to change the time-domain signal of the fault into the frequency domain for analysis. However, this method does not consider the process of the gradual deterioration of the switching device, and only directly opens the circuit of the switching device. The direct detection method judges whether the switching device is faulty by detecting the terminal voltage of the switching device or the current on the bridge arm, but this method requires more voltage and current sensors, which virtually increases the cost and cost of the system. Complexity, the system may have to pay a higher price for the failure of the sensor: the dictionary diagnosis method requires a large number of numerical simulations and experiments to obtain the fault value and eigenvalue, which is difficult to achieve in the actual system; the expert system method and the neural network fault Diagnosis methods are all methods with strong calculation ability and artificial intelligence simulation ability, but the shortcomings of its training samples are not easy to obtain, the diagnosis ability is not strong, and the expression form of network weight is vague, which leads to the limited practical application range of this method.

Images