基于改进生成对抗网络的谐波状态估计方法

传统的基于最小二乘法的谐波状态估计受到量测装置少、精确的谐波阻抗获取难、网络拓扑结构复杂以及电网运行方式变化等因素的限制,造成量测方程欠定、系统非全局可观以及节点间耦合关系难以准确提取等问题。文中提出了一种基于改进生成对抗网络的谐波状态估计方法。该方法基于pix2pix谐波状态估计网络拟合监测节点与目标节点间的耦合关系,利用采集的历史谐波数据,对模型进行批量训练,通过训练之后的生成网络估算目标节点谐波电流、谐波电压幅值,实现基于数据驱动的谐波状态估计。在加噪环境下对模型进行测试,仿真结果验证了所提方法的有效性。     

电网谐波状态估计中的可观性分析

谐波状态估计的一个难点,在于量测数目通常小于状态量数目,导致量测—状态方程欠定,系统不完全可观。针对该问题,进行了3个方面的研究:首先,结合数学分析和实际电网分析,从减少状态量的角度,实现方程由欠定变为恰定或超定,在保证系统状态完全可观的前提下,大大减少量测设备的数目;其次,将奇异值分解法(SVD)应用在系统可观部分的判断和计算中;然后,针对可观性分析中传统代数方法不足,结合矩阵条件数,提出一种实用的线性无关排序法,并将该方法用于指导不可观系统量测设备的调整和添加,实现系统良好的可观性。最后以菏泽—仿山220kV—110kV—35kV配电系统为例,对SVD和线性无关排序法进行说明,验证了方法的可行性。     

Power quality meters placement using seeker optimization algorithm for harmonic state estimation

Seeker optimization algorithm (SOA) is employed to present a new method for optimal placement of power quality (PQ) meters for harmonic state estimation (HSE). Using the same SOA algorithm, HSE was also improved during the proposed placement process. The placement is carried out by minimizing the weighted sum of maximum relative errors obtained from HSE for all existing harmonic orders. The improved HSE performs estimation for both real and imaginary parts of each harmonic by minimizing the errors between the measured values from PQ meters and those estimated. The simulation studies were performed on an IEEE 14-bus test system. The simulation results indicate that the proposed algorithms are effective in performing harmonic state estimation and optimal meter placement.     

基于LS-SVM的谐波源定位及谐波源注入电流分解方法

为了能够有效检测出电力系统中存在的多谐波源，通过谐波敏感因子确定谐波监测点的数量及分布，利用基于最小二乘支持向量机（LS-SVM）的谐波源定位估计器判断嫌疑母线的范围，在给定的范围中，通过求解网络状态方程进一步确定谐波源母线位置，并准确计算出谐波源注入的谐波电流大小。仿真结果表明，这种先定性缩小嫌疑母线范围，再定量确定谐波源母线位置以及注入谐波电流大小的方法对于配电网中多谐波源的检测是有效的。     

配电网谐波量测装置最优配置方法

为满足油田节能减排与工艺调整要求,变频器等非线性电力电子设备的应用日益广泛,油田配电网谐波污染日趋严重。谐波在线监测系统为测量和分析油田配电网谐波分布特征奠定了基础;但受应用经济性约束,谐波量测终端安装数量有限,有必要对其安装位置进行优化。本文综合考虑谐波状态估计可观测度与冗余度,提出了一种量测终端优化定址方法。采用谐波状态估计可观性逻辑判断法,利用节点关联矩阵计算不同谐波量测终端配置方案下系统可观测度与冗余度;在保证系统可观测度最大的前提下,以冗余度最大为目标优化量测终端安装位置。实现在终端数量有限的情况下提高油田配电网谐波状态估计的可观测度与精度,建模仿真结果证明了该方法的有效性。     

基于相量量测的电力系统谐波状态估计算法的研究

基于相量量测单元的电力系统谐波状态估计技术,可选取电力网络的节点谐波电压为状态量,支路电流、节点注入电流和节点电压相量为量测量,从而使谐波状态方程成为线性方程.本文根据这个特点,再结合量测方案的不同情况,提出了在冗余量测和基本量测时,采用便于工程实际应用的乔累斯基算法来求解线性谐波状态方程,其计算量只是高斯消元法的一半,且不用考虑选主元,大大减少了计算的时间;在量测矩阵欠定时,采用奇异值分解算法,不仅可以为欠定的状态方程提供稳定的最小二乘解,而且该算法本身具有可观性分析的功能,不需要采用其它算法进行可观性分析,从而极大地简化了谐波状态估计分析的难度.最后通过仿真程序验证了这两种算法在谐波状态估计求解问题中的有效性和可靠性.     

A distribution system state estimator accommodating large number of ampere measurements

This paper presents a three-phase state estimation method for distribution network which can accommodate large number ampere measurements and voltage measurements. In this method, the power on sending end of branch and the square of branch current magnitude are chosen as the state variables. Therefore, measurement functions of ampere and voltage measurements are significantly simplified and no measurements transformation is needed in this method. Furthermore, a novel robust state estimation model is presented, which can suppress bad data automatically. The observability analysis indicates the proposed method can be used in distribution system with very limited measurements. Numerical tests on radial and weak-meshed network show the real-time branch current measurements can significantly improve the accuracy of estimation results.     

基于相量量测的电力系统谐波状态估计(Ⅱ)--可观性、质量评估与算例研究

主要介绍了电力系统谐波状态估计技术的可观性分析理论、质量评估算法以及实际应用情况.谐波状态估计问题的可观性依赖于测量值的性质、数量、位置以及网络的拓扑结构.文章不仅将传统潮流状态估计问题的可观性分析算法应用到谐波状态估计领域,而且还根据量测配置的特点和性质提出了基于相关性理论的可观性分析算法与逻辑判断法,大大简化了算法的难度.另外,文章调查了谐波状态估计问题的统计特性,并在此基础上提出了评价估计质量的具体算法.最后,利用一个虚拟的算例对有关的算法给予了说明.     

A reduced model for power system observability: analysis andrestoration

An algorithm for observability analysis and restoration in power system state estimation is presented. The problem of P-δ, Q-V and complete observability is addressed. In the proposed algorithm, groups of buses are represented by supernodes and the problem of observability analysis and restoration is examined with a reduced network. The proposed algorithm is easily implemented and it is independent of the state estimation solution algorithm. In case of unobservability, the proposed methodology identifies directly the maximal islands and determines the number and the placement of the required pseudomeasurements for observability restoration. Results from several test cases are presented     

基于深度神经网络伪量测建模的交直流混合配电网交替迭代法状态估计

当前配电网状态估计面临的一个突出问题是实时量测数量不足,难以实现全网可观性。为了给配电管理系统提供准确的基础数据,提出一种基于深度神经网络伪量测建模的交直流混合配电网交替迭代状态估计方法。首先,建立电压源换流器的稳态模型和混合配电网的实时量测模型;然后,利用历史量测数据对深度神经网络进行离线训练,建立负荷节点注入功率的伪量测模型;最后,对交流区域和直流区域进行交替迭代状态估计,在交替过程中区域间交换VSC支路状态量的估计值,保证了边界状态量的一致性。算例测试结果表明,所提方法能在实时量测覆盖率低的情况下,准确估计混合配电网的状态值。     

基于深度神经网络伪量测建模的交直流混合配电网交替迭代法状态估计

当前配电网状态估计面临的一个突出问题是实时量测数量不足,难以实现全网可观性。为了给配电管理系统提供准确的基础数据,提出一种基于深度神经网络伪量测建模的交直流混合配电网交替迭代状态估计方法。首先,建立电压源换流器的稳态模型和混合配电网的实时量测模型;然后,利用历史量测数据对深度神经网络进行离线训练,建立负荷节点注入功率的伪量测模型;最后,对交流区域和直流区域进行交替迭代状态估计,在交替过程中区域间交换VSC支路状态量的估计值,保证了边界状态量的一致性。算例测试结果表明,所提方法能在实时量测覆盖率低的情况下,准确估计混合配电网的状态值。     

Implementation of the hypothesis testing identification in powersystem state estimation

The authors consider the online implementation of a general, reliable and efficient bad-data analysis procedure for power system state estimation. It is based on hypothesis testing identification, which was previously proposed and subsequently improved by the authors. The procedure involves a sequential measurement error estimator along with adequate sparsity programming techniques. Both make the procedure easy to implement on any state estimator. A criterion for multiple noninteracting bad-data identification is also proposed, which is applicable to any bad-data analysis method. Simulations are reported on systems of up to 700 buses. A thorough comparison with classical methods is also included     

Fault-location Observability Analysis on Power Distribution Systems

Most of the fault-location methods developed in the past employed measurements obtained from a limited number of meters installed in a power system. Optimal meter placement in power systems is to make the best use of a limited number of meters available to keep the entire network observable. This article presents the fault-location observability analysis for distribution systems and proposes a novel optimal meter placement algorithm to keep the system observable in terms of fault-location determination. First, the observability of fault location in power systems is defined. Then analysis of the whole system is performed to determine the fewest number of meters needed and the best locations to place those meters to achieve fault-location observability. Case studies based on a 16-bus distribution system have been carried out to illustrate the proposed algorithms.     

A new numerical algorithm to check the observability and bad-data processing ability of the metering scheme

This paper provides a unified approach for the optimization of measurements placements employed for power system online monitoring through state estimation. The proposed methodology, which can be suitable for the mixed measure system, preserves state estimation observability and bad-data processing capability by employing numerical algorithms for observability checking, critical measurements and critical couple identification. First, node injection radix measurements and measurement categories are defined. According to the above definitions, the coefficient matrix can be solved. The analysis on the column vectors of the coefficient matrix can determine each measurement classification. Furthermore, the numbers of each measurement class contains can determine bad-data processing capability. The observability can be checked by the type number of measurements. The proposed method is illustrated with the IEEE39-bus system and the IEEE118-bus system. Results from the case studies are presented to demonstrate that the approach adequately fulfills the desired properties related to observability, bad-data processing, cost, and robustness.     

一种基于实测数据的谐波源定位方法

为了准确划分各谐波源责任以便采用技术及经济手段治理谐波问题,提出了一种基于实测数据的谐波源定位方法。该方法以系统母线的谐波电压幅值及相位作为量测量,考虑一定的建模及测量误差,根据系统各节点的谐波功率与母线阻抗之间的符号关系对可疑母线进行分类。提出定位指标对各节点进行排序,确定最为可能的谐波源位置,减少后期不必要的勘测工作。最后,在ETAP 12.6中建立了IEEE 14节点的算例模型并进行计算验证,计算结果证明了提出方法的有效性。     

面向新型配电网的能观性分析与精细化协同规划方法

现代控制理论和电力系统状态估计中的能观性定义难以满足配电网态势感知的需求,配电网规划模型中也缺乏对能观指标的考虑,为此,提出面向新型配电网的能观性分析与精细化协同规划方法。对配电网的变量与状态空间模型进行定义,并根据传统能观性定义对配电网中的传统/新型变量进行能观性分析;基于传统能观性定义的缺陷提出新的配电网能观性和能观指标定义,并建立考虑能观性的配电网精细化协同规划模型。算例分析结果表明,所提规划方法考虑了实际量测误差的影响,实现了一次网架规划和二次态势感知设备配置的综合最优,达到了经济性和能观性的平衡。     

State estimation and bad data processing for systems including PMU and SCADA measurements

Conventional state estimators (SE) are based on real-time measurements, consisting of bus voltages and active and reactive power flows and injections, and estimate the voltage phasors of the network buses. Until recently, these measurements were obtained only through SCADA. With the advent of GPS synchronized measurements obtained by phasor measurement units (PMU), effective techniques are required to incorporate the extremely accurate PMU measurements into state estimation, in order to improve its performance and observability. This paper develops a non-linear weighted least squares estimator by modeling the current phasor measurements either in rectangular or in polar coordinates and compares the two approaches. Any numerical problems arised at flat start or for lightly loaded lines, are resolved. The error amplification, due to the current phasor measurement transformation from polar into rectangular coordinates, is also investigated. The normalized residual test is used to effectively identify any bad data in the conventional and phasor measurements. The proposed techniques are tested with the IEEE 14-bus system.     

A Novel Topological Genetic Algorithm-Based Phasor Measurement Unit Placement and Scheduling Methodology for Enhanced State Estimation

Phasor measurement units are emerging as a potential tool for on-line power system state estimation. Incorporation of phasor measurement units to the existing power system's monitoring system is impeded by various physical and economic constraints. This article proposes a novel topological genetic algorithm for optimal placement of phasor measurement units along with existing conventional measurement units such that state estimation can be achieved with enhanced accuracy and immunity against power grid contingencies. The proposed algorithm optimally places phasor measurement units so that complete observability of the power system is achieved through them and enhanced redundancy in measurement can be accomplished through conventional measurement units. Since practical phasor measurement unit placements are accomplished in multiple horizons, intelligent sorting and phase optimization methodologies have been presented to attain maximum observability during phasing periods. Placement of phasor measurement units with multiple channel limits has also been studied in this article. The efficacy of the proposed topological genetic algorithm for optimizing the number of phasor measurement units and enhancing state estimation under various operating conditions has been validated through extensive simulation studies conducted in IEEE standard bus systems. Practical case studies have been performed in the western and southern region Indian power grids.     

An incremental measurement placement algorithm for state estimation

Measurement configuration and redundancy significantly impact the quality of state estimation solutions. Practical and efficient algorithms are needed to suggest additional measurement locations in order to increase local redundancy and strengthen network observability. In this paper, an incremental measurement placement algorithm is proposed to enhance the quality of state estimation solutions. The algorithm identifies buses with low accuracy based on state estimation error variances. A list of candidate measurements is suggested to reduce these variances. Certain heuristics are applied to discourage introduction of the possible leverage point measurements. The algorithm has been implemented in the PG&E EMS environment and tested using the real-time data     

基于特征集重构与多标签分类模型的谐波源定位方法

传统基于谐波状态估计的谐波源定位方法需要专门的同步相量量测装置,工程应用受到限制。为此,基于电能质量监测装置所采集的非同步量测数据,提出了基于特征集重构与多标签分类模型的谐波源定位方法。利用监测数据的充分统计量来挖掘量测时段的谐波信息,同时利用标签特定特征学习算法重构特征集,从而消除冗余特征以及无关特征对于谐波源定位精度的影响;提出基于邻接矩阵以及灵敏度分析的测点配置方法,结合电路网络拓扑信息实现测点的优化配置;提出基于改进极限学习机的谐波源定位方法,该方法以重构特征集为输入,建立多标签分类模型,实现谐波源定位。通过仿真与算例分析,验证了所提方法的可行性及有效性。