含噪电能质量扰动信号分析的前提是准确找到突变点信息，对信号进行去噪的同时，又必须保留突变点特征。针对此问题，选取平稳小波变换分解信号，并利用提出的改进阈值函数对信号进行去噪。将含噪的电能质量扰动信号进行多层平稳小波变换，逐层估计平稳小波变换细节系数中噪声的均方差σj ，计算各层阈值σj 2 ln k 并根据信号、噪声的小波系数在不同尺度上的分布特点，通过ln ( j+1)对各层阈值进行修正，结合改进的阈值函数分别对各层小波系数进行处理。利用尺度系数和处理后的小波系数进行重构，得到去噪后的信号。仿真结果表明，改进的阈值函数去噪方法能够较好地滤除噪声并保留突变点特征，从处理后的小波系数中可以清晰地观察到扰动的起止时刻，并能够分辨出暂态振荡与谐波干扰。%The accurate detection and localization of mutation point information is the premise for analyzing power quality (PQ) disturbance signal with noise. But the mutation point feature must be retained during the signal de-noising. Accordingly, the stationary wavelet transform (SWT) is selected to decompose signals, and then the improved threshold function is proposed to de-noise signals. In addition, the noise mean squarejσ(j is decomposition level) is estimated and the thresholdkσjln could be calculated at each decomposition scale adaptively. According to the fact that the wavelet coefficients of signal and noise distribute on different scales, the proposed method exploitsln(j+1) to amend the threshold. Finally the improved threshold function is used for PQ signal de-noising. Simulation results show that the proposed scheme can suppress the noise of PQ signal while keeping the mutation points well. The starting and ending time of the disturbance can be clearly observed from the treated wavelet coefficients, the transient oscillations and harmonic interference can also be distinguished.
展开▼
