电压源换流器实时故障检测平台实现

摘要

电压源换流器交、直流侧易受外界干扰产生各类故障,这些故障的暂、稳态信息不易于区分和识别.目前多数研究面向离线仿真而实时检测难以兼顾.本文针对上述问题,提出了PC+FPGA架构的故障检测平台,运用Mallat分解得到的高频小波系数和低频尺度系数作为区分暂态与稳态的判据.针对稳态信号,提出改进型快速傅里叶变换FFT算法进行谐波分析.针对暂态信号,采用模极大值点求出扰动持续时间,模极大值作为判断故障是否发生的阈值,高频小波系数进行小波熵计算,判断故障发生的相.实验结果表明,该平台可判断交、直流故障,识别故障线路,具有较高准确度.%The AC and DC sides of voltage source converter are vulnerable to external interference,which results in var-ious types of fault.In addition,the transient and steady state information of these faults are not easy to distinguish or identify. At present,most researches in this filed are based on off-line simulation,which is difficult to be combined with real-time detection.To solve this problem,a fault detection platform based on PC+FPGA architecture is proposed in this paper.The high-frequency wavelet coefficients and low-frequency scale coefficients obtained by Mallat decompo-sition are used as criteria to distinguish between the transient and steady states.According to the steady state signal,an improved fast Fourier transform(FFT)algorithm is proposed for harmonic analysis.According to the transient signal, disturbance duration is obtained by using the modulus maximum point,and the modulus maximum value is adopted as a threshold to judge whether a fault occurs.Moreover,high-frequency wavelet coefficients are calculated by wavelet entro-py to judge the fault phase.Experimental results show that the proposed platform can be used to determine the AC and DC faults,and identify fault lines with high accuracy.