【作者】 曾凌云；

【导师】 黄纯；

【作者基本信息】 湖南大学 ， 电力系统及其自动化， 2007， 硕士

【摘要】 随着大规模敏感性电力电子器件在电力系统中的广泛应用,暂态电能质量扰动问题已成为众多领域关注的热点。对电能质量扰动信号进行检测与分析,实现扰动特征的提取与分类识别是对电能质量扰动进行监测和治理的必要前提;与此同时,为了记录信息一般采用高频采样率,造成信号存储和传输成本增加;另外,采集到的扰动信号中叠加的噪声会影响信号检测分析的效果。本文基于上述问题做了如下工作:指出了电能质量问题的定义、产生原因、分类以及评价标准,对五种常见的暂态电能质量扰动现象作了详尽而细致的描述。对傅立叶变换、短时傅立叶变换和小波变换这三种信号分析方法的基本原理及其在电能质量分析领域的应用做出探讨和比较。根据电能质量扰动信号的非平稳特性,利用小波变换良好的时频局部化特性,信号非零点的奇异性可以通过小波变换模极大值来表征。根据Mallat算法,通过信号的多分辨率分解提取信号奇异点的小波变换模极大值,实现电能质量扰动信号的检测。提出了一种基于小波变换低频系数模值差检测信号过零点扰动的方法。利用小波变换下有效信号和噪声在多尺度空间中不同的模极大值传播特性,提出一种基于小波变换的改进的阈值函数,实现噪声和不必要信息的滤除。比较了在选择不同阈值和阈值函数时电能质量扰动信号的去噪效果,对去噪方法的可行性和有效性进行了仿真效验,结果表明,该方法在去噪的同时,能有效地保留信号的奇异点的信息。更多还原

【Abstract】 With the wide application of large-scale sensitive electronic devices in the power system, power quality disturbances phenomenon have already become the focus that numerous fields have paid close attention to. The power quality disturbances are detected and analyzed accurately, as to realize the characteristic extraction and identification,which is the essential presupposition of monitoring and managing power quality disturbances phenomenon. Generally speaking, the sampling frequency of the monitor may be very high in order to analyze the transient phenomena,which will bring a large data storage and communication cost. Simultaneously,sampling data are usually overlapped by noise which will impact the analysis results. Some research is processed here on the problems above as follow:The definition, cause, classification, and evaluation criterion of power quality problems are discussed, and five common power quality transients are described detailedly.Three methods, including Fourier transform, short-time Fourier transform and Wavelet transform are discussed. This paper also explains the theories and how to use the theories in power quality analysis, and indicates their advantages and disadvantages.According to the non-steady characteristic of power quality disturbances, the good time-frequency localization makes the singularity of signal on non-zero points can be signified by the wavelet transform modulus maximum. Using Mallat algorithm, the wavelet transform modulus maximum of singularity is extracted through multiresolution decomposition, by which the accurate detection of fault signal can be realized. A method based on the difference of low frequency coefficient modules of the decomposed signal detects disturbance on zero-cross points.Proposed a new threshold function based on wavelet transform by utilizing the different characters of evolution of the wavelet transform modulus maximum across scale of efficient signal and noise. The de-noising and data compression results of different threshold and different threshold function are analyzed to verify the feasibility and effectiveness of the new threshold function. The information of singularity points can be reserved well and the de-noised and compacted signal is a good estimation of the original signal.更多还原