基于Daubechies小波的多分辨分解在电压闪变信号分析中的应用

Application of wavelet multiresolution decomposition using Daubechies in voltage flicker analysis

电压闪变是低频时变的非平稳信号,传统的傅里叶变换在分析非平稳信号方面存在很大的局限性,而小波变换具有时频局域化性质,是分析这类信号的有力工具。提出了利用小波多分辨分析提取电压闪变信号的方法,并根据小波函数的特点和分析的目的,选用不同N值(消失矩阶数)的Daubechies小波对闪变信号进行特征提取、定位及去噪处理。根据调幅波的时频信息,获得闪变信号的频率和幅值;通过检测小波变换模极大值,实现了对电压闪变发生、恢复时间的精确定位;并采用软阈值方法消去电压闪变信号中的噪声。仿真结果表明,不同N值的Daubechies小波和多分辨分析的结合在信号分析中可以取得良好的效果。

Voltage flicker is a low-frequency and time-varying non-stationary signal. Traditional Fourier transform has large limitation when it analyzes non-stationary signal. However, wavelet transform has outstanding time-frequency localization characteristics. Consequently, it is a powerful tool to analyze such signal. The paper proposes a method which uses wavelet multiresolution analysis to extract characteristics of voltage flicker. Furthermore, different N (number of vanishing moments) of Daubechies wavelet basis are selected to extract the characteristics,localize the time, and de-noising according to the property of wavelet function and the analyzing object. The signal frequency and amplitude can be achieved using the time-frequency information and wavelet transform modulus maxima are detected to localize the start time and end time of voltage flicker and de-noising by soft-threshold respectively. The simulations show that the methods can get better results.