摘要
提出了一种从混合有色背景噪声环境提取变压器局部放电信号的方法——自适应高斯基表示。自适应高斯基表示是一种参数化最优联合时频分析工具 ,它能够有效地去除交叉项干扰 ,滤除有色噪声。系统运行中存在的服从高斯或非高斯分布的有色噪声 ,具有不满足细节小波分解系数趋于零的特性 ,不能有效地被小波分解滤除。基于信号匹配原则 ,通过自适应估计最优高斯基分解参数 ,将信号分解为最优高斯基函数的线性组合 ,滤除噪声 ,无限逼近原信号 ,弥补了局部放电信号检测精确性的不足。大量的仿真结果表明自适应高斯基表示是一个有效的局部放电检测方法 ,它将在电力系统局部放电信号检测中得到广泛的应用。
The adaptive Gaussian representation(AGR), which has been used to extract par tial discharge (PD) signals from colored noises(CN), is proposed in this paper . Because AGR is a type of parametric joint Time-Frequency (T-F) analysis, whi ch uses more flexible T-F analytical windows, it can be immune from cross-term interference and filter out CN availably.AGR, which overcomes a shortage of con stant Q filter frame, can infinitely approximate to signals by conversation of energy. The power spectral density of CN existing in power system submits to Gaussian or non-Gaussian distributing, so it can not be filtered by wavelet tra nsform. Whereas AGR based on adaptive matching can approximate the signal closel y,it gets rid of disturbances effectively to make up the inaccuracy in signal de tection. The results of simulation show that AGR is an effective solution. The A GR will be widely used in detecting PD signals of power system.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2005年第1期31-35,共5页
Proceedings of the CSU-EPSA
关键词
自适应高斯基表示
局部放电信号
有色噪声
adaptive Gaussian representation
partial dischar ge signals
colored noises
