电能质量扰动检测与识别研究

发布时间：2018-01-10 21:36

本文关键词：电能质量扰动检测与识别研究　出处：《广西大学》2017年硕士论文　论文类型：学位论文

【摘要】：近年,随着现代工业技术的快速发展,电力电子技术也发展迅速,越来越多的新能源装置和冲击性负荷被接入电网,使得电网的电能质量扰动问题日益变得严重,给国民生产和生活带来一系列问题,造成了巨大的经济损失。因此必须对这些电能质量扰动问题进行分析和治理,而快速、准确检测出这些电能质量问题以及准确识别出这些电能质量问题的种类是提高电能质量的关键。本文将电能质量问题的检测和识别分别进行研究,分析了目前已有的检测和识别的方法,着重对S变换的原理和计算方法进行了分析,在此基础上了提出了一种改进S变换方法。在电能质量扰动检测方面,利用改进S变换实现了对谐波、电压暂升以及电压暂降等电能质量扰动的检测,并与利用S变换的检测结果进行了比较分析,实验证明了基于改进S变换的检测方法的有效性,且检测结果更准确。在电能质量扰动识别方面,首先利用改进S变换对谐波、暂升、暂降、振荡、脉冲、含谐波的暂升以及含谐波的暂降等7种扰动进行了分析,选择了信号改进S变换模矩阵中提取的幅值包络曲线、基频幅值曲线以及时间幅值平方和均值曲线等三种曲线作为特征曲线,然后选用两种方法分别进行识别分析。其一,利用二分类支持向量机对上述扰动进行识别,构建了二分类支持向量机树,实现了扰动分类;其二,利用极限学习机进行分类,用均值、标准差、偏度、峭度以及均方根值分别去刻画三种特征曲线,得到15个特征量作为极限学习机输入量,得到了良好的分类效果。以上两种分类识别结果证明了基于改进S变换提取的电能质量扰动特征量的有效性。
[Abstract]:In recent years, with the rapid development of modern industrial technology, power electronics technology has also developed rapidly, more and more new energy devices and impact load are connected to the power grid. The power quality disturbance problem of power network becomes more and more serious, which brings a series of problems to the national production and daily life, resulting in huge economic losses. Therefore, it is necessary to analyze and deal with these power quality disturbance problems. The key to improve power quality is to detect these power quality problems quickly and accurately and to identify the types of power quality problems accurately. In this paper, the detection and identification of power quality problems are studied separately. The existing methods of detection and recognition are analyzed, and the principle and calculation method of S-transform are analyzed. Based on this, an improved S-transform method is proposed to detect the disturbance of power quality. The improved S-transform is used to detect the power quality disturbances, such as harmonic, voltage rise and voltage sag, and the results are compared with those of S-transform. Experiments show that the detection method based on improved S-transform is effective and the detection results are more accurate. In power quality disturbance identification, the improved S-transform is first used to detect harmonics, suspensions, sags, oscillations and pulses. Seven kinds of disturbances, such as the rise of harmonics and the sag of harmonics, are analyzed, and the amplitude envelope curves extracted from the improved S transform modulus matrix are selected. The fundamental frequency amplitude curve and the time amplitude square sum mean curve are taken as characteristic curves, and then two methods are selected to identify and analyze them. The disturbance is identified by two-classification support vector machine, and the tree of two-classification support vector machine is constructed, and the disturbance classification is realized. Secondly, using the extreme learning machine to classify, using the mean, standard deviation, deviation, kurtosis and root mean square value to describe the three characteristic curves, get 15 characteristics as the input of the learning machine. The above two classification and recognition results prove the effectiveness of the power quality disturbance feature extraction based on the improved S-transform.
【学位授予单位】：广西大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM711;TM930

【参考文献】