基于改进变分模态分解和小波阈值法的单相接地故障电流降噪

Denoising of Single-phase Grounding Fault Current Based on Improved VMD and Wavelet Threshold

针对电力系统输配电线路发生单相接地故障时,电气设备间的电磁环境复杂,现场环境干扰严重导致故障录波装置采集到的故障零序电流信号含有大量噪声,影响后续选线准确率的问题,提出了一种改进变分模态分解(variational mode decomposition,VMD)和小波阈值法联合的单相接地故障的零序电流降噪方法,通过北方苍鹰优化算法优化改进VMD对零序电流信号分解,引入自适应相关阈值对分解后的分量进行筛选,对噪声分量进行小波阈值法降噪,最后将信号进行重构。通过搭建模型进行仿真实验,所提算法比传统VMD降噪算法信噪比提高了5.52%~35.99%,均方根误差降低了12.78%~30.88%,与小波阈值降噪方法、经验模态分解(ensemble empirical mode decomposition,EEMD)-小波阈值降噪方法、完全自适应噪声集合经验模态分解(complete ensemble empirical mode decomposition with adaptive noise,CEEMDAN)-小波阈值降噪方法相比,也都有明显的优势,并且在标准测试信号Heavy Sine信号和Bumps信号中进行实验,验证了算法的适用性。

When a single-phase grounding fault occurs in power system transmission and distribution lines,the electromagnetic environment between electrical equipment is complex,and the severe field environmental interference causes a lot of noise in the fault zero-sequence current signal collected by the fault recorder,which affects the accuracy of subsequent line selection.A denoising method for single-phase grounding fault zero sequence current based on improved variational mode decomposition(VMD) and wavelet threshold method was proposed.By optimizing and improving VMD for zero sequence current signal decomposition through the northern eagle optimization algorithm,adaptive correlation thresholds were introduced to filter the decomposed components,and wavelet threshold method was used to denoise the noise components.Finally,the signal was reconstructed.By building a model to conduct simulation experiments,the proposed algorithm improves the signal-to-noise ratio by 5.52%~35.99%,and reduces the root mean square error by 12.78%~30.88% compared with the traditional VMD denoising algorithm.Compared with the wavelet threshold denoising method,ensemble empirical mode decomposition(EEMD) wavelet threshold denoising method,and complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN) wavelet threshold denoising method,the proposed algorithm also has obvious advantages,and the algorithm's applicability is verified by experiments in standard test signals Heavy Sine signal and Bumps signal.