摘要
为了降低特高压交流变电站内变压器、电抗器等设备对电晕噪声检测的影响,根据变压器、电抗器噪声与电晕噪声的差异化时频特征,提出了一种电晕噪声提取方法,利用小波分析方法对原始电晕噪声信号进行多频带分解,采用阻带梳状滤波器对重构后的电晕噪声低频部分进行滤波处理,从中消除变压器本体及电抗器噪声,根据电晕噪声的短时脉冲性以及变压器冷却装置噪声的平稳性,利用基于谱减法的语音增强技术从电晕与冷却装置的混合噪声信号中提取出纯净电晕声信号。结果表明,利用该方法能够剔除带噪电晕声信号中的变压器、电抗器本体噪声以及冷却装置噪声,并且不改变电晕信号的时域特性,从而表明了该方法的可行性及有效性,能够为特高压交流变电站电晕噪声现场检测提供技术支持。
To suppress the influence of power transformer and reactor noises on corona noise measurement in UHV AC substations,a corona noise extraction method is proposed according to the noise differences in frequency-and time-domain.The wavelet analysis method is used to divide the frequency band of the original noise signal into several segments.The stop-band comb filter is designed to process the low frequency part of the reconstructed noise, eliminating the body noises of the power transformer and reactor.According to the short-time impulsive characteris- tic of the corona noise and the steady characteristic of the cooling device, the spectral subtraction based speech en- hancement technique is employed to extract the pure corona noise from the mixed noise of the corona and the cooling device.The result shows that the body noises of the transformer and reactor are removed from the noisy corona acoustic signal and the corona noise keeps consistent before and after processing,which validates the feasi- bility and effectiveness of the proposed method.This method may provide technical support for field corona noise measurement in UHV AC substations.
出处
《高压电器》
CAS
CSCD
北大核心
2017年第11期78-84,共7页
High Voltage Apparatus
基金
国家电网公司科技项目(5216A015001T)~~
关键词
特高压
电晕噪声
小波分析
语音增强技术
UHV
corona noise
wavelet analysis
speech enhancement technique