Realization of Fault Detection Platform for Voltage Source Converter

Various types of faults could occur in AC and DC sides of voltage source converters due to external interference, and the transient-state and steady-state information of these faults are difficult to recognize and distinguish. However, current research is mostly oriented to off-line simulation while real-time detection is difficult to implement. This paper aims to propose a fault detection platform based on PC+FPGA architecture to solve above problems. The high-frequency wavelet coefficients and low-frequency scale coefficients obtained by Mallat decomposition are used as criteria to distinguish transient state from steady state. For steady-state signals, an improved FFT algorithm is proposed for harmonic analysis. For transient-state signals, the disturbance duration is obtained through the modulus maximum point, and the modulus maximum value is used as the threshold to determine whether fault occurs or not. The wavelet entropy is calculated using high-frequency wavelet coefficients to determine the phase of the fault. Experimental results show that the platform can detect AC and DC faults and identify the fault line with high accuracy.