针对干扰频率多和信噪比低这两大局部放电窄带干扰抑制中难以解决的问题,提出一种复小波变换一自适应滤波算法.首先,利用复小波变换分频特性将信号分解到不同的频段内,使各个频段内的窄带干扰数少且频率接近,然后进行逐层自适应滤波;采用一种实部与模平方乘积的新复合信息以提高对低信噪比信号的提取能力以及对信号形状的保持能力.研究结果表明;复小波变换一自适应滤波算法在信噪比δ0=0时能对4种局部放大脉冲模型分别获得信噪比δ为22.22,16.16,17.16和21.75和波形相似度η为0.909 2,0.942 7,0.668 8和0.657 0的滤波效果,与传统的小波去噪和自适应滤波相比有较大改进.%Because there are too many different interference frequencies and low signal-noise-ratio(δ) in suppressing partial dischange's narrow bandwidth noise, an algorithm of complex wavelet transform-adaptive filter was proposed.First, the signals were decomposed into different bands with the property of frequency domain division of complex wavelet transform, which can reduce the number of interferences in each band. Secondly, the adaptive filter level-by-level was adopted. Finally, a new combining information was utilized to greatly enhance the ability of extracting PD signals from large noise and the maintenance of signal's waveform. The results show that results of signal to noise ratio δ=22.22,16.16, 17.16, 21.75 as well as wave form similarity η=0.909 2, 0.9427, 0.668 8, 0.657 0 can be achieved when original signal to noise ratio δ0=0, which shows great improvement compared with traditional adaptive filtering.
展开▼
