【作者】 刘守亮；

【导师】 肖先勇；

【作者基本信息】 四川大学 ， 电力系统及其自动化， 2006， 硕士

【摘要】 现代电力系统负荷构成的重大变化以及电力市场化进程的深入，凸现了电能质量问题的重要性。解决电能质量问题所需要进行的主要工作是对其进行监测控制和考核规范。而电能质量监测则是其他各项研究工作的前提。近些年来，这一问题在国内外已经形成了研究热点。其中，短时电能质量扰动作为暂态电能质量问题的主要研究内容，因其发生的频繁性、随机性以及对敏感设备的危害性而备受关注，已经成为电能质量监测分析的研究重点。探讨和研究如何从海量电能质量监测数据中准确地检测出扰动和正确地进行分类识别的方法是其核心内容，也是构建完整的电能质量监测系统的必要步骤和进一步进行电能质量控制的前提条件。 本文分析了基于不同方法的短时电能质量扰动研究现状，并着重讨论了多种短时电能质量扰动特别是电压凹陷的特性；在此基础上，分别提出了基于复小波变换的短时电能质量扰动检测方法，基于S变换的扰动分类识别方法和电压凹陷分类分析专家系统。 小波变换因具有良好的时频分辨特性，适合于分析具有暂态、突变特性的非平稳信号，因而在短时电能质量扰动分析尤其是检测分析中得到了广泛的应用。但是小波变换容易受噪声影响，实际应用中对检测方法的实时性要求较高，目前基于小波的检测方法都难以同时获得满意的噪声鲁棒性和实时性。针对这一问题，本文对正交紧支复小波用于抑制噪声和快速检测扰动进行了研究，提出了一种Daubechies复小波的完整、实用的生成方法，并结合复小波提供的特有的相位信息构造了多种新型复小波复合信息形式，应用Mallat算法，在不加去噪手段的情况下，实现了对短时电能质量扰动的快速准确检测。 S变换是小波变换的一种发展，它一方面继承了小波变换适合于分析非平更多还原

【Abstract】 Power quality (PQ) has been an important issue to the electric power utilities for the changes of loads composition in modern power systems and with the development of power market. The main problems in power quality are monitoring, control and evaluation, and power quality monitoring, which has been a significant part in power quality study, is the foundation for other research and work. In PQ monitoring analysis, short duration power quality disturbances (SDPQD) are the primary aspect, and they are known for their frequent and random occurrences, and harms to the sensitive loads. So with the increasing amount of measurement data from power quality monitoring devices, it is desirable that SDPQD detection and classification can be performed automatically and accurately.This paper reviews the method used in SDPQD analysis firstly, and analyzes the characteristics of different SDPQD in detail. Then the novel SDPQD detection and classification method based on complex wavelet and S-transform respectively is proposed, and the S-transform-based expert system for voltage dips (sags) classification is presented.Wavelet transform (WT) has been applied widely in SDPQD analysis, especially in SDPQD detection, because of its excellent ability in time-frequency resolution that makes WT suitable for analyzing non-stationary signals like SDPQD. However, WT is vulnerable by noise, and the current WT-based detection method has difficult in satisfying the noise robustness and real time at the same time in the practical application. Aiming at these problems, this paper study the feasibility using更多还原