基于希尔伯特—黄变换和小波包去噪的暂态电能质量扰动定位与分析

发布时间：2018-06-13 05:55

本文选题：暂态电能质量 + 小波去噪　；参考：《宁夏大学》2016年硕士论文

【摘要】：评估一种检测方法优劣的关键所在就是看这种方法能否对暂态电能质量扰动起、止时刻进行准确、快速的定位,本文旨在对暂态电能的五类典型扰动(包括电压骤升、电压骤降、暂态振荡、电压中断、电压闪烁)的起、止时刻的定位分析进行实现。首先通过小波包去噪法对染噪扰动信号进行去噪处理,其次,通过希尔伯特-黄变换(Hilbert-huang Transform,简称HHT)对各扰动的起、止时刻进行提取。文中包含了对小波去噪和小波包去噪两种方法的去噪效果对比、筛选,也涵盖了小波变换检测法和希尔伯特-黄变换检测法的定位效果对比,通过对比,选择出检测性能稳定且准确的检测方法,并结合Matlab实现扰动定位的建模仿真过程。本文的主要内容和研究成果如下：(1)本文利用小波包去噪方法和小波去噪方法分别对染噪信号进行去噪,选取分解层数为3层、5层、7层进行仿真,随着分解层数的递增,去噪效果呈现差-优-差的趋势变化,且小波包去噪的效果比小波去噪的好：两种方法均在分解层数为5层时去噪效果达到最佳。(2)为了进一步确定小波包去噪方法中哪类函数的去噪效果更好,我们选取db1、db3、db5、 db7、sym1、sym3、sym5、sym8八种函数进行去噪效果对比,仿真发现效果较好的是db3、db5、 sym5、sym8四种函数,我们又对这四类函数的去噪效果进行了对比。综合(1)、(2)两点,最终得出在分解层数为5层时,sym5函数的去噪效果最好。(3)本文使用了两种方法来对暂态电能质量的扰动进行定位。第一种是根据小波算法原理进行检测,即小波变换检测法。第二种是根据希尔伯特-黄变换检测算法进行检测,即为希尔伯特-黄变换检测法,分别用这两种方法检测扰动的起、止时刻,对检测数据进行对比。对含噪信号进行检测时,小波算法对电压闪烁、暂态振荡、电压骤降三类扰动不能准确定位,对于电压闪烁扰动设置的四个扰动点直接检测不出来,故该检测方法在实验中会导致暂态电能质量扰动检测的准确率下降。(4)本文使用的希尔伯特-黄变换检测算法对电压骤升、电压骤降、电压闪烁、暂态振荡和电压中断这五类典型的暂态电能质量扰动信号进行了定位分析,从仿真结果看,这种检测方法能够得到扰动发生处的准确数据。说明希尔伯特-黄变换检测算法是切实可行的。并且在定位之前,先使用了小波包去噪算法对信号进行去噪,进一步提高了扰动检测的准确性。也为电力系统扰动治理提供了更加准确、可靠的依据。
[Abstract]:The key to evaluate a detection method is to see if the method can accurately and quickly locate the transient power quality disturbance. This paper aims at five types of typical transient power disturbance (including voltage surge). Voltage sag, transient oscillation, voltage interruption, voltage flicker. Firstly, wavelet packet denoising method is used to Denoise the noisy disturbance signal. Secondly, Hilbert-huang transform (HHT-HHT) is used to extract the disturbance. This paper includes the comparison of the denoising effect of wavelet denoising and wavelet packet denoising, screening, and the comparison of localization effect between wavelet transform detection method and Hilbert-Huang transform detection method. The stable and accurate detection method is selected, and the modeling and simulation process of disturbance location is realized with Matlab. The main contents and research results of this paper are as follows: 1) in this paper, wavelet packet denoising method and wavelet denoising method are used to Denoise the noisy signals, and the decomposition layers are selected as three layers and five layers and seven layers for simulation. The effect of de-noising shows a trend of poor, excellent and bad. The effect of wavelet packet denoising is better than that of wavelet denoising: both of the two methods achieve the best denoising effect when the number of decomposition layers is 5 layers.) in order to further determine which function of wavelet packet denoising method is better than wavelet packet denoising method, the denoising effect of wavelet packet denoising method is better than wavelet packet denoising method. We compare the denoising effects of the eight functions, db1, db3, DB7, and db7, and we compare the denoising effects of the four functions. The results of simulation show that the four functions are better, db3ndb5 and Sym5sym8, and we also compare the denoising effects of these four functions. In this paper, two methods are used to locate the disturbance of transient power quality when the decomposed layer number is 5 layers, and the denoising effect of the sym5 function is the best. The first is based on the principle of wavelet algorithm, namely wavelet transform detection method. The second is based on Hilbert-Huang transform detection algorithm, that is, Hilbert-Huang transform detection method, using these two methods to detect the beginning and end of the disturbance, the detection data are compared. In the detection of noisy signals, the wavelet algorithm can not accurately locate three kinds of disturbances, such as voltage flicker, transient oscillation and voltage drop, but it can not directly detect the four disturbance points set by voltage scintillation disturbance. Therefore, the accuracy of transient power quality disturbance detection will be decreased in the experiment.) the Hilbert-Huang transform detection algorithm used in this paper is used to detect voltage sudden rise, voltage drop, voltage flicker. Five typical transient power quality disturbance signals, transient oscillation and voltage interruption, are located and analyzed. From the simulation results, the accurate data of the disturbance can be obtained by this method. It shows that Hilbert-Huang transform detection algorithm is feasible. Before locating, wavelet packet denoising algorithm is used to Denoise the signal, which further improves the accuracy of disturbance detection. It also provides a more accurate and reliable basis for power system disturbance control.
【学位授予单位】：宁夏大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM711

【参考文献】