基于局部积分均值分解方法的电能质量检测与分析

发布时间：2018-03-29 12:24

本文选题：经验模态分解　切入点：局部积分均值分解　出处：《中国矿业大学》2017年硕士论文

【摘要】：HHT作为现如今较为常用的时频分析方法之一,与FFT分析、小波变换以及S变换等处理方法相比较,在非线性、非稳态信号的处理以及自适应性等方面建立了一定的优势。但HHT在信号处理分析过程中依旧存在一些弊端,因此本文的工作主要是对HHT方法存在的缺点进行分析和改进,并将改进的HHT方法运用到电能质量扰动检测与分析领域。首先,针对HHT方法存在的均值曲线拟合、模态混叠、虚假分量以及端点效应等问题,提出了一种新的自适应时频分析方法——局部积分均值分解(Local Integral Mean Decomposition,LIMD)方法。该方法针对单分量信号的特点,重新定义了具有瞬时物理意义的单分量信号,提高了分解固有模态函数的有效性。针对EMD方法插值拟合效果差以及耗时时间长的缺点,改进了均值曲线的拟合方式,将积分中值定理引入到均值曲线的拟合过程中,运用极值点间的全部数据作为特征时间尺度,且只采用一次插值拟合的方式。通过仿真实验表明该方法提高了算法的计算效率以及在端点效应和虚假分量的抑制上取得了一定的改进效果。其次,针对LIMD方法在模态混叠效应以及分辨率低等问题还需进一步改进,提出一种自适应集总局部积分均值分解(Adaptively Ensemble Local Integral Mean Decomposition,AELIMD)方法。该方法分析不同频率形式噪声对极值点分布的影响,确定加噪频率采用高频辅助分解的优势,并以极值点分布特性作为评价指标自适应选择最优加噪频率。通过对EEMD加噪准则的研究,推导出加噪幅值和分解次数采取固定值:0.01 SD和2次,且以正负成对的形式加入到原始信号中。通过仿真实验验证了所提方法的自适应性和计算性能。最后,将改进的HHT方法(AELIMD方法与Hilbert变换)运用到电能质量扰动检测分析中。运用AELIMD方法对含噪的电能质量信号进行去噪分解,采用二阶导数的方式求得模极大值点,提高了通过模极大值点定位扰动时刻的准确性。针对高频复合扰动采取两次AELIMD分解方法去除噪声与虚假分量有效提取出扰动成分;针对稳态扰动提出先去除谐波再提取闪变包络的检测方法;针对未知复合扰动给出了基于HHT方法的检测思路。仿真结果表明了所提方法的有效性和可行性。并搭建了电能质量扰动实验平台,通过实验装置模拟产生真实的电网扰动故障,采用实测数据验证所提方法的检测效果,综合检验了AELIMD方法在电能质量扰动检测工程领域具有一定优势。
[Abstract]:Compared with FFT analysis, wavelet transform and S transform, HHT is one of the most commonly used time-frequency analysis methods. Some advantages have been established in the unsteady signal processing and adaptive analysis, but there are still some drawbacks in the signal processing and analysis of HHT, so the main work of this paper is to analyze and improve the shortcomings of the HHT method. The improved HHT method is applied to the field of power quality disturbance detection and analysis. Firstly, aiming at the problems of mean curve fitting, modal aliasing, false component and endpoint effect of HHT, etc. A new adaptive time-frequency analysis method, Local Integral Mean decomposition method (LIMD), is proposed, which redefines a single component signal with instantaneous physical meaning according to the characteristics of a single component signal. The efficiency of decomposing inherent mode function is improved. Aiming at the disadvantages of poor interpolation effect and time-consuming time of EMD method, the fitting method of mean value curve is improved, and the integral mean value theorem is introduced into the fitting process of mean curve. Using all the data between extremum points as the characteristic time scale, The simulation results show that the proposed method improves the computational efficiency of the algorithm and improves the endpoint effect and the suppression of false components. Secondly, the simulation results show that the proposed method improves the efficiency of the algorithm. Aiming at the problems of modal aliasing effect and low resolution of LIMD method, further improvement is needed. An adaptive Ensemble Local Integral Mean decomposition (AELIMD) method is proposed, in which the influence of noise in different frequency forms on the distribution of extreme points is analyzed, and the advantages of high frequency auxiliary decomposition are determined. The optimal noise frequency is adaptively selected by using the distribution of extreme points as the evaluation index. Through the study of the EEMD noise adding criterion, the amplitude and decomposition times of the noise adding are deduced by the fixed value: 0. 01 SD and 2 times. The proposed method is added to the original signal in the form of positive and negative pairs. The self-adaptability and computational performance of the proposed method are verified by simulation experiments. Finally, The improved HHT method and Hilbert transform are applied to the detection and analysis of power quality disturbance. The noised power quality signal is de-noised by AELIMD method, and the modulus maximum is obtained by the second-order derivative. The accuracy of locating disturbance time by modulus maximum is improved. The disturbance components are effectively extracted from noise and false components by twice AELIMD decomposition method for high frequency complex disturbance. Aiming at steady state disturbance, a detection method of flicker envelope is proposed, which removes harmonics first and then extracts flicker envelopes. The detection method based on HHT method is presented for unknown complex disturbances. The simulation results show that the proposed method is effective and feasible. A power quality disturbance experimental platform is built to simulate the real power grid disturbance faults. By using the measured data to verify the detection effect of the proposed method, the AELIMD method has some advantages in the field of power quality disturbance detection engineering.
【学位授予单位】：中国矿业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM711;TM930

【参考文献】