利用S变换局部奇异值的同调机组识别方法

文中提出一种利用S变换局部奇异值的同调机组识别方法.将广域测量系统(Wide Area MeasurementSystem,WAMS)采集到的系统发电机功角信息进行S变换,得到每台发电机的时频信息模值矩阵,将矩阵分割成块,计算各个子块的最大奇异值,利用时频信息模值矩阵中各个子块最大奇异值构造机组特征矩阵,采用分布聚类法对特征矩阵进行聚类分群.IEEE-39节点系统算例表明,该方法能够有效提取功角信息特征,具有很强的抗噪性,能够在不同故障类型下准确识别同调机组.     

利用小波能量特征的增长型自组织神经网络同调机组分群方法

提出了一种利用小波变换多尺度空间能量分布特征的自组织神经网络同调机组分群方法。首先改进了同调机群识别判据,然后利用小波变换的多尺度空间能量分布特征提取方法,对机组功角摇摆曲线提取特征,将时域特征、频域特征及小波能量特征构成的综合向量,作为增长型自组织神经网络的输入,通过调节阈值λ,得出不同精度的分群结果。最后在IEEE-39节点系统上对只考虑时频域特征和同时考虑小波能量特征、时频域特征的同调机组识别结果进行了对比分析,最终表明同时考虑小波能量特征、时频域特征的分群结果具有更高准确性。     

基于希尔伯特-黄变换（HHT）的机组同调研究

概述了利用功角曲线进行同调识别所面临的困难,阐述了希尔伯特-黄变换(HHT)在非线性、非平稳功角信号处理中的优势,提出了一种基于HHT的同调识别新方法。该方法利用经验模式分解方法将功角信号进行分解,得到剩余分量和各固有模态信号,逐次对这些分量进行比较得到机组的同调特性。该方法不受系统模型、故障场景限制,能随着不同的故障自适应调整数据时间窗的长度,并能克服小波分析、prony分析等算法难以处理非平稳信号的不足。测试结果表明了该算法的有效性。     

基于WAMS和曲线相似的同调机群识别方法*

提出了一种基于曲线相似的电力系统同调机群辨识新方法。首先获取WAMS实时监测得到的功角轨迹曲线作为基础数据,根据离散曲线相似性的定义,将各发电机的轨迹曲线分段,借助遗传算法确定分段轨迹之间的最优相似距离,保存了轨迹段的局部特性。然后整合各段相似距离得到各发电机之间的整体相似度并以此作为聚类指标,再利用层次聚类法实现多机系统同调机组分群。该方法简单可行,不受系统模型参数和故障类型的限制。最后对IEEE39节点系统进行分析计算,该仿真实验佐证了所提算法的有效性。     

基于发电机同调性的相量测量单元优化配置

采用发电机同调性辨识的方法实现相量测量单元（PMU）最优布点以实现电力系统动态监控。首先将模糊聚类原理中的基于模糊图论的最大树法应用到发电机同调识别中,将发电机进行同调分群;然后用中国电力科学研究院的《电力系统分析综合程序》PSASP中EPRI_36节点纯交流系统进行仿真验证,结果证明了方法的有效性及可行性;最后在每一同调群中配置一个PMU以实现对全网稳定性的监测和控制。     

Coherency identification by the method of weak coupling

A computationally efficient method for identifying groups of slowly coherent machines in an interconnected power system is presented. The method is based on the concept of weak coupling and uses a linearized representation of the total system. An algorithm is developed that rearranges the synchronous machines of the power system in order of their relative coupling. The machines are grouped into slowly coherent areas corresponding to the point of least coupling using a ‘coupling graph’. The method is illustrated by a case study on an 18-machine system. The results are validated by comparison with swing curves obtained from a transient stability simulation of the total system. It is shown that the proposed method identifies coherent groups that are independent of fault location.     

Coherency identification and equivalents for transient stability studies

This paper presents techniques for identifying coherent generators without conducting a transient stability study and for developing stability equivalents of a group of coherent generators. Identification is based on the property of equal acceleration for coherent machines and it requires only examination of the elements of the Jacobian matrix of the linearized power system model. After the coherent groups have been identified, the stability equivalents are developed on the basis of the power invariance condition at the tie buses and at the internal buses of the coherent generators. These stability equivalents are used in a transient stability study. The techniques are illustrated with a 12-machine system.These techniques greatly improve the efficiency of calculation of coherency based equivalents. They have potential applications in planning studies of large power systems.     

Generator Coherency Using the Hilbert–Huang Transform

Coherency between generators is tracked, by examining the instantaneous phase differences among interarea oscillations and swing curves in disturbed multiarea power systems. Huang's empirical mode decomposition is applied to extract dominant oscillatory modes from interarea oscillations/swing curves. Hilbert transform on these modes yields their instantaneous phase. The coherency is thus revealed by the temporal evolution of the phase difference among interarea oscillations/generator swing curves. The technique is illustrated using simulations on a test system, as well as actual wide-area measurements.      

Security constrained economic dispatch using nonlinearDantzig-Wolfe decomposition

This paper proposes a new approach to the security constrained economic dispatch based on a nonlinear version of the Dantzig-Wolfe decomposition principle. The dispatch problem is formulated using truly nonlinear unit cost functions and a new, detailed representation of reserve curves. The solutions obtained by the method have the equal λ property. The proposed decomposition scheme, is more efficient as the number of generating units in the problem is increased. Numerical results obtained on three test cases are presented     

基于滑差同调等值的空调群负荷建模

随着空调负荷比例的迅速增长,对系统的暂态电压稳定逐渐带来威胁。为更好地描述含空调启动的空调群负荷特性,文中将同调等值法扩展到空调群的负荷建模中,提出了基于滑差同调等值的负荷建模方法和模型,用阻尼最小二乘法对含空调启动特性的空调群综合负荷进行建模,通过仿真验证了建模方法的有效性。     

基于Prony分析特征提取的同调机组分群方法

同调机群识别在电力系统的动态等值、主动解列控制中具有重要意义。提出一种基于Prony分析特征提取的同调机组分群方法。首先针对Prony分析受噪声干扰严重的缺点,利用集成经验模式分解(Ensemble Empirical Mode Decomposition, EEMD)方法对含噪声的信号降噪。然后对降噪后的功角信号进行Prony分析,提取功角信号的幅值、频率和阻尼特征值,形成每台机组的特征向量。最后将系统中所有机组特征向量组成的特征矩阵输入到自组织神经网络进行聚类,从而实现同调机组分群。EPRI-36节点系统和华北电网系统算例表明,所提方法可以很好地降低噪声影响,充分提取功角曲线特征,准确识别同调机组。     

电力系统同调机群的识别方法

在对电力系统中同调机群的识别方法进行研究的同时，并对以往的各种同调机群的识别方法也进行了归纳总结，针对这些方法所存在的问题，给出了一种新的较为实用的同调机群识别方法。     

Coherency and model reduction: state space point of view

Analytical connections are determined that relate coherent stability behavior of generators to redundancy in the controllability and observability gramians of the linear model of the power system. This redundancy is removed from the gramians of the reduced-order model generated by aggregating the coherent machines. Model reduction in large power systems is shown to be better if the order estimation information in the controllability and observability gramians is weighted properly by balancing the system. Perfect coherency of generators is related to the parameters of the linear model of the power system. Waveforms for generating the RMS coherency measures are compared. An algorithm for order reduction based on balancing is outlined and justified by testing on a 240-generator, 2500-bus system     

下垂并网逆变器网络的慢同调与聚合方法

近年来,同调等值方法被广泛应用于大型系统的动态等值中,得到的降阶模型能够准确反映系统的动态响应,其中,代表性的慢同调算法具有对故障位置和严重程度不敏感等优点.然而,由于模型限制及准确性差的问题,使得传统慢同调算法难以应用于分布式发电网络中.为此,提出了一种适用于下垂控制并网逆变器网络的慢同调与聚合方法,依据奇异摄动理论...     

基于相似度同调的双馈风电场聚合研究

文章提出了一种基于相似度同调的双馈风电场聚合方法。首先利用Prony算法提取各风电场中双馈风机故障下功率响应曲线的动态特征,基于相似理论确定双馈风机间振荡模式中相似部分,应用相似振荡模式中的频率、阻尼、幅值等信息定义了双馈风机同调判别的相似度量化指标并确定同调机群。然后,根据双馈风机的外特性,提出功率聚合算法,即以双馈风机输出有功功率为权重求取等值机的暂态内电势并确定等值母线及等值机参数。该方法从同调机群识别、同调母线聚合、同调机群参数聚合等方面,系统地解决了不同双馈风电场的同调聚合问题。参数聚合方法不仅解决了聚类分群的复杂性和局限性,而且有效保留原系统中电气与结构特征。最后以实际云南电网中某地区风电场为例进行仿真验证。     

Fully reference-independent LMP decomposition using reference-independent loss factors

The decomposition of locational marginal price (LMP) under the popular DCOPF framework generally depends on the choice of the reference bus. A previous work has achieved reference independence for the overall LMP and LMP congestion component, but not all individual LMP components. This paper proposes a method to obtain a truly reference-independent LMP decomposition such that all three components of LMP at each bus will be invariant w.r.t. the choice of the system reference bus. This is achieved with loss factors based on a new AC-based distribution factor model, which depends on the network topology and the present operating condition only, but not the system reference bus. This model gives reference-independent loss prices, which can serve for a better loss hedging financial transmission rights, since the choice of reference bus will not change the loss prices. Further, this paper uses the fictitious nodal demand (FND) model to obtain loss distribution factors (LDFs). FND gives more reasonable power flows since losses should be distributed in each individual line, rather than at load buses when the load-weighted LDFs are applied. Also, the proposed reference-independent distribution factors and loss factors may have great potentials in other areas of power system analysis.     

基于两步聚类法的受扰机群同调性分析

提出了一种基于两步聚类法的电力系统受扰机群同调性分析方法。首先通过BIRCH算法处理广域量测系统提供的受扰机群电角速度数据,构建受扰机群的聚类特征（clustering feature,CF）树来实现预聚类;然后采用层次聚类法处理预聚类结果并自动确定聚类数量来实现正式聚类;最后根据聚类结果进行受扰机群的同调性分析。通过对新英格兰10机39节点系统的算例仿真,对比不同的同调机组识别方法,验证了该方法的有效性以及能够实现自动确定同调机组数的特点。     

Finger-knuckle-print recognition using Gabor feature and MMDA

Recently, a new biometrics recognition, finger-knuckle-print, has attracted the interest of researchers. The popular techniques used in face recognition are not applied in finger-knuckle-print recognition. Inspired by the success of Gabor in face recognition, we propose a method that uses Gabor feature and a multi-manifold discriminant analysis (MMDA) method to identify finger-knuckle-print. The experimental results show that our proposed method can work well.     

基于广域量测信息相关性的电力系统同调辨识

提出一种基于广域量测信息相关性识别电力系统同调机群的新方法。针对同调发电机运动轨迹具有一致性或相似性的特征,借鉴基于电力系统模型确定同调机群的方向余弦法,引入广域量测信息相关性来辨识系统的同调机群,在计算系统中各发电机组间的相关系数基础上,提出了划分系统主导同调机群及主导同调机群内部子同调机群的相关同调判据。同时针对广域量测窗口长度对同调辨识结果的影响,进一步提出一种适用于辨识主导同调机群及主导同调机群内部子同调机群的实用鲁棒同调判据。最后将所提方法应用于典型两区域互联交直流系统和南方电网中,分析结果验证了该方法的有效性和可行性。     

考虑同调约束的电力系统主动解列断面的搜索方法

电力系统主动解列是避免发生大停电事故和系统崩溃的有效保护措施,需要及时、准确地确定解列方案。文中提出一种考虑发电机同调约束的最佳解列断面的搜索方法。首先基于故障后广域测量系统(WAMS)提供的实时动态响应信息,利用改进拉普拉斯特征映射算法在线识别同调机群,利用发电机同调虚拟节点聚合同调机群,减小解列断面的搜索空间,然后利用加权拉普拉斯分群策略得到初始的解列断面。最后考虑断面有功功率的方向,利用启发式变邻域搜索算法确定最佳的解列断面,使得解列后各孤岛的净不平衡功率最小。文中方法能充分考虑不同故障模式所带来的同调行为的差异性,使得主动解列控制具有选择性和鲁棒性。通过新英格兰39节点系统验证了所述方法的有效性和适应性。