电力系统低频振荡辨识及相关算法研究

发布时间：2018-01-16 11:57

本文关键词：电力系统低频振荡辨识及相关算法研究　出处：《天津大学》2014年硕士论文　论文类型：学位论文

【摘要】：近年来,电力系统规模不断扩大,电网间互联不断增多,同时,发电机多使用快速励磁系统却多缺乏彼此协调,这都为低频振荡的发生埋下了隐患。系统弱阻尼、电网间互联的弱联系、快速励磁系统的应用以及系统重负荷都成为低频振荡发生的原因。国内外也对低频振荡进行了深入的研究,WAMS/PMU系统可在线提供电力系统运行中发电机、线路、变压器、负荷等元件的各项运行数据,为低频振荡的在线分析提供了帮助。Prony算法假设任何时间序列均由一系列具有一定振幅、相位、频率和衰减因子的指数函数的线性组合构成,能够直接提取对应振荡信号的特征,为振荡模式和阻尼分析提供了可能,已在电力系统低频振荡、负荷动态模型、阻尼控制等方面得到了广泛研究。但Prony算法在对信号进行分析处理时对噪声较为敏感,这是限制Prony算法发展的重要问题。本文研究的目的,是希望利用高阶累积量技术,来提高已有Prony算法在进行低频振荡分析时的抗噪声能力。高阶累积量是现代信号处理的一种重要且有效的数学工具,是高阶统计量中的一种。由于任何高斯过程的高阶累积量均等于零,高阶累积量在理论上可完全抑制高斯噪声,比传统的基于功率谱或相关函数的方法具有更好的抗噪性能。针对Prony算法抗噪性能差和高阶累积量能够抑制高斯噪声的特点,本文提出了一种基于高阶累积量的改进Prony算法,用于电力系统的低频振荡分析中,即将高阶累积量与Prony算法相结合来提高后者进行低频振荡分析时的抗噪能力。通过典型含噪信号和New England 39节点算例系统的仿真验证了所提方法在含有噪声条件下的低频振荡在线分析中的正确性和有效性。
[Abstract]:In recent years, the scale of power system continues to expand, the interconnection between power grids is increasing, at the same time, the generators often use fast excitation system, but more lack of coordination with each other. These are hidden dangers for the occurrence of low frequency oscillations. Weak damping of the system and weak interconnection between power grids. The application of fast excitation system and the heavy load of the system have become the cause of low frequency oscillation, and the low frequency oscillation has also been deeply studied at home and abroad. The WAMS/PMU system can provide the running data of generators, lines, transformers, loads and other components in the power system. For the on-line analysis of low frequency oscillation, the. Prony algorithm assumes that any time series is composed of a series of linear combinations of exponential functions with certain amplitude, phase, frequency and attenuation factor. The characteristics of corresponding oscillation signals can be extracted directly, which provides the possibility for oscillation mode and damping analysis, and has been used in power system low-frequency oscillation and load dynamic model. Damping control has been widely studied, but the Prony algorithm is sensitive to noise in signal analysis and processing, which is an important problem limiting the development of Prony algorithm. It is hoped that the high-order cumulant technique will be used to improve the anti-noise ability of the existing Prony algorithm in low-frequency oscillation analysis. High-order cumulant is an important and effective mathematical tool for modern signal processing. Because the higher order cumulant of any Gaussian process is equal to zero, the high order cumulant can completely suppress Gaussian noise in theory. Compared with the traditional methods based on power spectrum or correlation function, the Prony algorithm has better anti-noise performance and higher order cumulant can suppress Gao Si noise. This paper presents an improved Prony algorithm based on higher-order cumulants for low frequency oscillation analysis of power systems. The high order cumulant is combined with the Prony algorithm to improve the anti-noise ability of the latter in the analysis of low frequency oscillation. Typical noise-containing signals and New England. The simulation of 39-bus system verifies the correctness and validity of the proposed method in the on-line analysis of low-frequency oscillation with noise.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM712

【参考文献】