应用于电力系统低频振荡模式估计的组合算法  

A Novel Combination Method Used for Measurements of Power System Low Frequency Oscillation Mode Parameters

作　　者：陈荔[1] 

机构地区：[1]中国能源建设集团广东省电力设计研究院,广州510663

出　　处：《电测与仪表》2013年第3期64-68,80,共6页Electrical Measurement & Instrumentation

摘　　要：提出了一种基于复香浓小波变换(Complex Shannon Wavelet Transform,CSWT)和粒子群算法(ParticleSwarm Optimization,PSO)的低频振荡模式参数测量新算法。通过调整复香浓小波的带宽参数,CSWT能具备较好的动态特性和频域分辨率,不但能准确提取振荡波形中同时存在的多种主导模式的瞬时振荡频率和时域信息,而且对非平稳的振荡信号具有很好的适应性。利用最小二乘法对CSWT提取的时域信息进行拟合能获取相应的振荡幅值和阻尼因子。PSO具有优秀的并行搜索性能,能高效估计低频振荡模型中的直流分量幅值和相位参数。MATLAB仿真合成信号、四机两区域系统仿真和基于广域测量系统(WAMS)的实测数据三个算例分析结果表明,本文提出的组合方法能够准确测量振荡波形中同时存在的多种主导振荡模式的频率、阻尼因子、幅值和相位,而对噪声具有较好的抗干扰能力。This paper presents a novel combination method based on the complex Shannon wavelet transform （CSWT） and particle swarm optimization （PSO） for measurements of power system low frequency oscillation mode parameters. By controlling the bandwidth parameters of complex Shannon wavelet （CSW）, the CSWT can achieve desirable dynamic response and mode resolution. The CSWT can not only extract oscillation frequencies and time domain information of multiple oscillation modes, but also be applied to non-stable oscillations. Then, the least square estimation is deployed to estimate oscillation amplitude and attenuation factors from the acquired time domain information. The PSO has excellent concurrent searching capability, and can estimate DC component and oscillation initial phase parameters based on the built low frequency oscillation signal model. The effectiveness and correctness of the proposed method have been verified by MATLAB simulation signal, four machines and two areas power system simulation and practical verification using real WAMS data. The results show that the proposed method achieves high accuracy in measurements of oscillation mode parameters especially in the presence of noise.

关 键 词：复香浓小波变换 粒子群算法 低频振荡 动态特性 频域分辨率 WAMS 

分 类 号：TM71[电气工程—电力系统及自动化]