节点文献
电力系统动态同步相量估计算法及其应用研究
Dynamic Synchronized Phasor Estimator for Power Systems and Its Applications
【作者】 麦瑞坤;
【作者基本信息】 西南交通大学 , 电力系统及其自动化, 2010, 博士
【摘要】 同步相量测量技术为电网的大面积实时监控提供了先进的信息技术保证,而同步相量测量技术的核心及基础是同步相量估计算法的设计,算法的估计精度将直接影响到广域测量系统(Wide Area Measurement System, WAMS)其它高级应用的效果,如状态估计、自适应保护以及故障测距等。现有的商用同步相量估计算法在静态条件下具有较高的测量精度,而在动态条件下(如低频振荡、频率偏移等情况下)却受到系统动态特性的影响而导致算法精度急剧下降。因此研究在动态条件下仍然具有较好估计精度的同步相量估计算法,对于判断电网状态、预测电网发展方向以及采取正确、快速的保护措施都具有较大的现实意义。论文在动态条件下的频率跟踪算法、同步相量估计、同时消除直流衰减和动态影响方面进行了深入研究,提出了对应的算法;在此基础上,利用PMU的同步相量估计值来近似表示信号的动态特性,提出一种考虑信号动态特性的新型故障测距算法。在频率跟踪方面,论文中根据全部三相电压信号的信息来在线估计电压信号的噪声含量以及动态特性,并通过动态特性方差和噪声方差两个参数来量化系统的实时状态,从而调整ⅡR滤波器的遗忘因子来使频率跟踪算法不但在动态条件下具有快速跟踪的特性而且在静态条件下也具有较好的抗噪特性。由于递归算法的应用,简化了计算,本算法易于在重计算负荷的DSP中实现。通过不同的工况下的仿真研究(包括:相位突变测试,频率跳变测试以及PSCAD/EMTDC仿真)表明无论在动态还是在静态条件下,该频率跟踪算法都具有比其它传统算法更为优越的性能,如高抗噪声性能以及快速跟踪等。在同步相量估计方面,利用泰勒级数对电力信号的时变相量进行建模,并利用相邻采样数据窗的相量测量值来表示相量导数,以此相量导数来修正传统傅立叶变换的相量估计值从而减弱或消除系统动态特性对相量测量精度的影响。理想信号数字仿真以及实时数字仿真的分析研究表明:新算法能够在增加少量额外的计算负荷的前提下大大地提高同步相量测量在多种动态条件下(如低频振荡和频率偏移等情况)的精度。此外,还可以利用同步相量估计值的泰勒导数来实现在低频振荡下的故障检测。在消除直流分量影响方面,论文提出了一种同时考虑信号动态特性和衰减直流分量的同步相量测量新算法。首先对电力信号的基波信号和衰减直流分量进行建模,再通过利用一个基波滤波器来对相量的动态特性进行估计和一个直流滤波器来对衰减直流分量进行估计,然后将两个滤波器的估计结果代入到动态模型中进行迭代运算,最终得到精确的同步相量测量结果。对理想信号(包括衰减直流分量、动态信号以及两都的综合)以及PSCAD/EMTDC建模仿真信号的大量仿真结果表明,此方法在一定程度上提高了同步相量测量算法在功率振荡下发生接地故障时的相量估计精度。在同步相量估计的应用方面(动态故障测距),论文提出一种利用信号的动态特性对基于分布参数模型的故障测距算法进行改进和扩展的新型故障测距算法。该算法不但从空间上认为输电线路上不同位置的电流、电压信号是不同的,而且从时间上认为输电线路上的电流、电压信号的幅值和频率是随时间而变化的,从而扩展了传统的故障测距算法,使其具有同时描述电气信号的空间特性和时间特性的能力。算法通过输电线路两端的同步相量估计值的导数形式来表示信号的动态特性,再利用牛顿迭代法对故障距离进行测定。论文利用数字仿真软件对一个230kV输电系统进行建模,并对各种工况进行仿真验证,如不同的故障位置、故障过渡电阻和故障类型。仿真结果表明,所提出的故障测距算法不但具有传统双端测距算法的优点,而且能对发生低频振荡时的故障进行精确定位。总的说来,本论文最终形成一个包括改进数据采集系统的频率跟踪算法、提高精度的同步相量测量算法以及基于同步相量测量结果的动态故障测距算法的理论和应用体系。
【Abstract】 Synchrophasor measurement has provided a novel method to monitor the power grid widely and the design of synchrophasor measurement technique is one of the most important essential elements. The accuracy of measurement has direct bear on the efficiency of other applications of Wide Area Measurement System (WAMS), such as state estimation, adaptive protection and fault location, etc. Although the accuracy of phasor measurement of the commercial Phasor Measurement Units (PMUs) can meet the most of the requirement of real-time applications in the steady state, the error of phasor measurement arises when a power system is under dynamic condition such as power oscillation and frequency deviation due to the impact of the dynamic characteristic of supplied signals. Therefore, it is necessary to develop a new synchrophasor measurement algorithm and a new frequency estimation algorithm, which can have accurate estimation under both steady state and dynamic condition. it is very important that can offer accurate phasor and frequency information across the whole power grid to help the high-level applications predict the develop direction of power system and make a rapid decision to protect the power system correctly.A great deal of effort has been put to the research on frequency tracking algorithm and synchrophasor measurement algorithm in presence of decaying dc components and dynamic characteristitcs. After a new dynamic frequency tracking estimator and a dynamic phasore estimator have been proposed, a dynamic fault-location estimator considering power systems’dynamic chariteristics expressed by the phasor measurement from PMUs is presented.Three-phase voltage signals are employed to estimated the density of the noise and the dynamic characteristic online, and then the real-time state of power system can be measured by introducing two variables, noise density variance and dynamic characteristic variance. Therefore, the new frequency-tracking algorithm can have fast tracking characteristic under dynamic condition and good anti-noise characteristic under high noise density condition by changing the forgetting factor in the Infinite Impulse Response (IIR) filters according to the estimation of noise density variance and dynamic characteristic variance. Besides, the algorithm proposed can be employed with ease when limited Digital Signal Processor (DSP) bandwidth is available as the the computation burden has been sharply reduced by applying a recursive algorithm. A few simulations under different conditions including phase shift, frequency step change and PSCAD/EMTDC simulation allows us to make a conclusion that the performace of proposed algorithm is much better than that of traditional one under both high-density noise condition and dynamic conditions.With the aspect of synchrophasor estimation, by considering the dynamic characteristic of supplied signals, the time-varying phasor of supplied signals are modeled by Taylor expansion and the Taylor derivatives are described by the adjoining phasor estimations of different data windows. The synchrophasor estimation from Discrete Fourier Transform(DFT) can be reassigned by using the derivatives in order to improve the accuracy of phasor estimation. Computer-generated signals tests and simulations by Real Time Digital System draw a conclusion that the proposed algorithm can dramatically improve the performance under variety of dynamic conditions with the cost of minus computational burden increase. Besides, the derivative of Taylor expansion can also be applied to detect the fault under power oscillation condition.It is necessary to consider dynamic characteristic and decaying DC component simultaneously and reduce their affect on phasor estimation. First, a dynamic signal model considering the decaying DC component is applied to model the fault current. Two filters, fundament component filter and DC component filter, estimate the fundamental phasor and decaying DC component. Finally, better fundamental phasor estimation is obtained by using a recursive algorithm. Ideal signal test, including decaying dc component, dynamic characteristics and both of them, and PSCAD/EMTDC based simulations allows us to draw a conclusion that the proposed algorithm can improve the accuracy of phasor measurement under power oscillation condition when a earth fault happens.A new algorithm extended from traditional fault location algorithms is proposed. This approach not only can express the space characteristic of supplied signals but also considers the supplied signals as time-variable signals whose magnitude and frequency are changing against time so that it has the ability of describing space property and time property of signals. Then, after using the adjoining phasor measurement to express the dynamic characteristic in term of derivatives, the accurate fault location can be attained via a Newton iteration method. PSCAD/EMTDC simulation based on a typical 230kV transmission line under different conditions, such as different fault location, fault resistance and fault type is performed. The simulation results show that the proposed algorithm not only have the merit of trandition two-terminal fault-location algorithm, but also can have a precise estimate of fault location under power oscillation condition.Overall, a theoretical system based on dynamic characteristics of power systems is achieved. This system includes a frequency-tracking algorithm that improves the performance of data acquisition system, a dynamic phasor estimator and a dynamic fault-location estimator.