改进的小波阈值去噪法在省级电网低频振荡分析中的应用

Application of improved wavelet thresholding denoising algorithm in low-frequency oscillation analysis of provincial power system

Prony分析的精度受信号噪声影响很大,为此,一个有效的去噪算法就成为应用Prony方法的关键。本文引入新的阈值选取和阈值函数对传统的小波阈值去噪算法进行改进,利用改进的小波阈值去噪算法对采样信号进行预处理。新的阈值选取能够保证小波系数随尺度的分布而变化,同时贴合不同类型信号中噪声在各层的实际分布情况,进而确保充分的去噪能力而不使重构的信号失真。新的阈值函数相比于常用的软阈值函数减小了其中存在的固定偏差,而相比常用的硬阈值函数确保了函数的连续性,从而消除了重构信号可能出现的振荡。并构造信号仿真实验,通过Matlab仿真实验与软阈值法、硬阈值法等方法去噪效果及辨识结果的对比,验证了利用本文提出的改进算法的有效性。

The accuracy of Prony analysis is greatly affected by signal noise.Therefore,an effective denoising algorithm becomes the key to apply Prony method.In this paper,a new threshold selection and threshold function are introduced to improve the traditional wavelet threshold denoising algorithm,and the improved wavelet threshold denoising algorithm is used to preprocess the sampled signal.The new threshold selection can ensure that the wavelet coefficients change with the distribution of scale,and fit the actual distribution of noise in different types of signals in each layer,so as to ensure the sufficient de-noising ability without distortion of the reconstructed signal.Compared with the commonly used soft threshold function,the new threshold function reduces the fixed deviation,while compared with the commonly used hard threshold function,it ensures the continuity of the function,thus eliminating the possible oscillation of the reconstructed signal.A signal simulation experiment is constructed,and the effectiveness of the improved algorithm is verified by comparing the results of MATLAB simulation with soft threshold method,hard threshold method and so on.