Loading [MathJax]/extensions/MathMenu.js
A Fault Diagnosis Method for Open-Circuit Faults Based Voltage Vector Trajectory Distortion Characteristics in Three-Level T-Type Inverters | IEEE Journals & Magazine | IEEE Xplore

A Fault Diagnosis Method for Open-Circuit Faults Based Voltage Vector Trajectory Distortion Characteristics in Three-Level T-Type Inverters


Abstract:

Three-level T-type inverters are widely used due to their low losses and harmonic distortion. The increased number of switching devices raises failure probability, reduci...Show More

Abstract:

Three-level T-type inverters are widely used due to their low losses and harmonic distortion. The increased number of switching devices raises failure probability, reducing inverter reliability. Hence, an accurate fault diagnosis method is essential for reliability enhancement. Based on the analysis of voltage vector trajectory under a single-phase single-switch fault in a three-level T-type inverter, voltage vector trajectory distortion characteristics of double-switch open-circuit faults are obtained through vector rotation and combined superposition. Furthermore, the relationship between the voltage vector trajectory distortion characteristics and the locations of open-circuit faults is deduced. Consequently, a diagnosis method is proposed based on the distortion characteristics of voltage vector trajectory. The angle of the distorted voltage vector trajectory is utilized to locate the fault phase, and the average voltage Park vector (AVPV) transformation is employed to obtain two diagnostic features: the normalized mean of phase voltage and the mean error of direct-axis voltage, which enables an accurate location of fault switches. An adaptive threshold is adopted. The method has low complexity and fast speed, with the diagnostic time ranging from 0.05 to 0.55 fundamental cycles. This method can realize the same-phase single-switch and double-switch fault diagnosis, reducing load sensitivity and misdiagnosis.
Page(s): 1 - 1
Date of Publication: 05 March 2025

ISSN Information:

Funding Agency:


Contact IEEE to Subscribe