基于统计学习的模糊暂态稳定事故筛选与排序

暂态稳定事故筛选与排序的目的是针对一组预想事故集合挑选出严重事故或滤除掉无害事故,以减少待分析的预想事故数目,满足在线动态安全分析的需要.文中提出一种基于统计学习的模糊暂态稳定事故筛选与排序方法,该方法采用了反映事故严重程度的10个性能指标,并通过样本学习,在每个指标集合上具体定义了系统稳定性的模糊隶属度函数;然后综合运用这10个性能指标的稳定性模糊隶属度得到的平均稳定性模糊隶属度,对预想事故集合按严重性进行了排序;接着根据排序结果,结合所设定的稳定阈值,将无害事故过滤掉.最后,用新英格兰10机39节点网络验证了该方法的有效性.     

电力系统暂态安全的观测与控制

电力系统的暂态安全问题主要是由大扰动引起的。暂态安全防御需要对大扰动引起的功角失稳进行有效的观测和控制。本文分析了紧急状态下电力系统的运动特征,并为观测和控制的配置提出了秩判据。提出了基于主导不稳定平衡点和临界割集的能观性与能控性分析方法,用于定量分析不同的地点对于故障失稳的不同的观测和控制效果。最后以10机39节点新英格兰系统为例选择了观测和控制的地点。这为设计电力系统的暂态安全防御奠定了基础。     

基于核主成分分析和深度置信网络的暂态稳定评估

针对电力系统暂态稳定评估实时性较差以及错误率较高的问题,提出了一种核主成分分析结合深度置信网络的暂态稳定评估方法。首先,构造了一组反映电力系统暂态稳定的特征向量;然后,基于核主成分分析法对特征向量集进行特征提取,降维特征向量维数以及过滤冗余特征,将降维后的特征向量传输至深度置信网络;最后,进行训练分析,训练过程包括预训练和微调,优化网络参数,提升深度置信网络评估精度。新英格兰10机39节点系统仿真结果表明,该方法可以有效降低输入数据的维数,去除冗余特征,降低暂态稳定性评估的错误率和测试时间,能准确、快速地判断电力系统的稳态状态。     

A GENETIC-BASED POWER SYSTEM STABILIZER

ABSTRACT

A Genetic-based Power System Stabilizer (GPSS) is presented in this paper to improve power system dynamic stability. The proposed GPSS parameters are optimized using Genetic Algorithms (GA). The main advantage of the proposed GPSS is that far less information than other design techniques is required without the need for linearization process. Time domain simulations of a synchronous machine subject to major disturbances are investigated. The performance of the proposed GPSS is compared with that of conventional lead-lag power system stabilizer (CPSS) to demonstrate the superiority of the proposed GPSS. The effect of parameter changes on the proposed stabilizer performance is also examined. The results show the robustness of the proposed GPSS and its capability to enhance the system damping over a wide range of operating conditions and system parameter variations.     

CORRESPONDENCE: DISCUSSION OF “GENERALIZED ANALYSIS AND PERFORMANCE OF RELUCTANCE MOTORS”

Abstract

Based on the boundary theorems developed in [1], a method of voltage stability analysis referred to as the Eigen-Structure Analysis method, is developed that does not require the complicated nonlinear programming calculations for evaluation of the closest boundary condition corresponding to a given power network operating point with voltage controlled and PQ buses. An algorithm for determining the stability margin and its sensitivity to bus voltages and bus injections is proposed which is capable of handling large scale power systems by utilizing the sparse matrix techniques for saving computation time and memory space. The Eigen-Structure Analysis method is applied to a number of test system models. The simulation results confirm the theory and show that the proposed stability margin decreases monotonically to zero when the system approaches voltage collapse. Simulation results for 30-bus New England Test System are included in this paper.     

ON-LINE ESTIMATION OF SYNCHRONIZING AND DAMPING TORQUE COEFFICIENTS USING AN ARTIFICIAL NEURAL NETWORK BASED APPROACH

ABSTRACT

This paper presents an artificial neural network based on-line approach to assess the dynamic stability of a single machine connected to an infinite bus system. The proposed approach is based on estimating the synchronizing and damping torque coefficients of the synchronous machine from voltage and power on-line measurements. Test results show that the proposed approach is very encouraging for fast evaluation of the dynamic stability.     

IMPROVED CONTINGENCY SELECTION ALGORITHM FOR VOLTAGE SECURITY ANALYSIS

ABSTRACT

A simple and efficient method for optimal selection of weights, along with higher order performance indices for voltage contingency selection, has been suggested in this paper. Some of the existing performance indices for voltage contingency selection has been critically reviewed. The proposed performance indices are able to eliminate misranking and masking effects. The post-outage quantities have been computed, using a new set of distribution factors, defined in terms of pre-outage real and reactive power flows in transmission lines/ outputs of generators. The proposed algorithm has been tested on IEEE 14-bus system and a practical 75-bus Indian system.     

Measurement and analysis of harmonic distortion in power distribution systems

Power distribution systems often exhibit different harmonic behaviours at different locations. It is not usually possible to completely characterise the distortion at a particular site, however, due to insufficient data about the measured system and instrumentation inability to evaluate or register complete set of measurement quantities. In this paper existing power quality standards are analysed and new harmonic distortion indices are derived. A new index describing the total harmonic distortion in current measurements, TRD, is suggested in order to simplify and standardise harmonic measurements. Another index representing the increase of harmonic voltage distortion with TRD, equivalent harmonic impedance z_eq, is proposed as an efficient way of characterising a particular site. The framework described here can be used to predict future harmonic behaviour at any previously measured point. Finally, extensive harmonic measurements of a real distribution system were performed to prove the usefulness of the proposed indices.     

Wide-area measurement-based voltage stability sensitivity and its application in voltage control

This work proposes a measurement-based voltage stability index, namely wide-area measurement-based voltage stability sensitivity based on L index. The sensitivities of L index with respect to nodal reactive power (Q) and real power (P) injections are first derived. The derived L-Q and L-P sensitivities analyze the impact of nodal injection to nodal voltage stability and can help determine the reactive power compensation and emergency load shedding amount for voltage stability control. To improve the computational efficiency, a simplified L index, L′, along with its sensitivities with respect to nodal reactive and real power injection (L′-Q and L′-P sensitivities) are derived which makes the proposed approach suitable for the practical large-scale systems. Moreover, a control strategy for voltage stability is proposed based on the L-Q, L-P, L′-Q, and L′-P sensitivities. The proposed sensitivities and control strategy are tested on the New England 39 bus system and the IEEE 118 system. Test results on both systems verify the proposed sensitivities and the control strategy by demonstrating their accuracy and robustness in voltage stability assessment and control. In conclusion, the proposed measurement-based sensitivities can be applied to voltage stability assessment and control by using the wide-area measurements.     

A decentralized control system for stabilizing a longitudinal powersystem using tieline power flow measurements

A decentralized control system for stabilizing a longitudinal power system is studied. The system is designed by a design methodology proposed earlier for decentralized control. In the process of the design, it is found that only the control information of phase angles from other areas is indispensable for maintaining the system stability. New measurements of the tieline power flow are substituted for the indispensable information. Simulation studies demonstrate that the resulting decentralized control system using tieline power flow measurements has as excellent performance in enhancing the stability of a longitudinal power system as a centralized control system     

Improving large-disturbance stability through optimal bifurcation control and time domain simulations

This paper presents an approach for improving, through practical assumptions, the stability of power systems subjected to large disturbances. The proposed method is composed of two steps, solved iteratively. The first step solves an optimal bifurcation control problem that guarantees the small-signal stability of the equilibrium point. The proposed optimal bifurcation control addresses saddle-node and Hopf bifurcations. The second step is an N − 1 contingency analysis computed through time domain simulations. The second step guarantees the large-disturbance stability of the equilibrium point. The WSCC 9-bus and New England 39-bus systems are used to illustrate and test the proposed technique.     

基于PMU和混合支持向量机网络的电力系统暂态稳定性分析

提出了一种基于粗糙集理论的混合网络模型,结合简单的计算将同步相量测量单元(PMU)获得的故障后短时间窗内各发电机的功角信息作为输入,首先利用粗糙集和自组织特征映射(SOFM)网络分别对原始输入进行特征提取和预分类,然后对那些不能直接利用SOFM网络进行稳定性判断的样本采用提取后的特征量并利用支持向量机(SVM)优良的统计特性进一步寻找其各自的最优分类面,以确保对所有样本进行正确分类。结合新英格兰10机系统的计算结果从预测精度和训练时间两方面对多种SVM模型进行了比较,结果表明了利用文中所提模型进行电力系统暂态稳定性分析的有效性,该模型可以提高训练效率以及分类的准确性。     

Optimal Control of Superconducting Magnetic Energy Storage Unit

ABSTRACT

A study of power system transient stability enhancement through effective use of superconducting magnetic energy storage unit is presented. Simultaneous optimal control of active and reactive power is formulated as the control which forces the fastest dissipation of power system energy involved by a fault in transmission network. Optimal control strategy is obtained from direct Lyapunov method. It requires control signals based on state variables locally not available at the bus where energy storage unit is installed. It is proved that these optimal control signals can be replaced by the time derivative of two electric quantities derived from local measurements. Control algorithm based on these quantities gives smooth and very effective regulation, tests clone for two-machine system have confirmed validity and robustness of the proposed new control algorithm.     

考虑风电-光伏功率相关性的因子分析-极限学习机聚合方法

针对风电和光伏时序数据单独聚合改变风电-光伏序列相关性的问题,提出因子分析-极限学习机聚合方法.首先,将z-score标准化的风电-光伏原始日场景集分解为水平分量日场景集和波动分量日场景集.其次,对水平分量日场景集进行近邻传播聚类,得到K类场景簇,再通过分层抽样获取2n天的水平分量日场景集.在标准化的风电-光伏原始日场景集中选取对应2n天的原始日场景集,分别得到n天的训练集和n天的测试集.然后,通过极限学习机获取水平分量日场景集和原始日场景集之间的映射关系,输出拟合功率日场景集.最后,通过反标准化,分别得到n天的风电和光伏功率的聚合序列,并通过概率统计指标、相关系数、仿真计算结果验证所提方法的准确性和可行性.     

改善接入地区电压稳定性的风电场无功控制策略

针对大规模风电场接入电网带来的电压无功问题,提出一种改善接入地区电压稳定性的变速恒频风电场无功控制策略,在系统发生扰动时,以接入点为电压控制点,扰动前的稳态电压为控制目标,动态调节风电场输出无功功率,维持接入点电压水平。实际应用时,该策略利用系统部分雅可比矩阵推导风电场的电压无功灵敏度信息,并根据风电场的无功输出能力计算风电场无功调整量,同时通过设置控制死区和延时,避免了风电机组的频繁调控。仿真算例表明,采用所提策略能够充分发挥变速恒频电机风电场的快速无功调节能力,有效抑制风速扰动、负荷变化等因素引起的电压波动,维持接入地区电网的电压稳定性。     

基于轨迹特征根的暂态稳定实用判据

基于轨迹特征根在大扰动后系统分析中的适用性,尝试用轨迹特征根方法来分析和预测系统的暂态稳定性。通过大扰动后逐个断面的泰勒展开,将原非线性系统由一族带扰动项的线性系统来近似。通过线性系统的稳定性理论及电力系统物理特征,给出电力系统暂态失稳判据,进而基于该判据,设计暂态稳定在线预测算法。最后,在新英格兰系统和国内实际电网上的算例测试,表明了暂态失稳判据和在线预测算法的准确性和适用性。     

Stability Control of Smart Power Grids with Artificial Intelligence and Wide-area Synchrophasor Measurements

Abstract—Environmental concerns due to emissions from nuclear and fossil fuel based power plants have triggered widespread utilization of renewable energy-based small- and large-scale distributed generation technologies. These technologies have been transforming the energy market towards a deregulated and dispersed entity. To cope with these transformations, and ensure appropriate grid monitoring and control, the conventional power grids across the globe have been enduring a paradigm shift towards a smart grid that is empowered with cutting edge technologies. The operational stability of these emerging smart power grids necessitates sophisticated real-time monitoring and control technologies. This article analyzes various stability concerns in smart power grids pertaining to distributed generations and proposes novel methodologies for ensuring operational stability. The proposed methodologies entail real-time stability monitoring and stability control with the use of wide-area synchrophasor measurements and artificial intelligence methods. The efficacy of the proposed methodologies has been verified through simulation studies conducted on an IEEE 14-bus system. Results of this research validate the necessity of coordinated control for maintaining stability of smart grids incorporating distributed generation technologies.     

Adaptive Power System Stabilizer Design Using Optimal Support Vector Machines Based on Harmony Search Algorithm

Abstract—This article presents the application of support vector machines to adaptive power system stabilizer design in a multi-machine power system based on the harmony search algorithm. Data from a multi-machine power system are the input features of the support vector machines. Support vector machine parameters and power system features are simultaneously optimized by harmony search based on the k-fold cross-validation technique. The proposed algorithm is trained by the optimal support vector machine parameters and optimal power system features. Power system stabilizer parameters produced by the proposed algorithm can be adapted by various operating conditions when the power system operates either inside or outside the training ranges. Simulation studies in the IEEJ Western Japan ten-machine power system demonstrate that the proposed algorithm is far superior to conventional power system stabilizers with fixed parameters and those designed by a robust coupled vibration model under various operating conditions and severe disturbances.     

基于支持向量机和决策函数的暂态稳定评估方法

暂态稳定评估是保证电力系统安全稳定运行的关键点,为解决应用机器学习进行暂态稳定评估保守性不足的问题,提出了一种基于支持向量机和决策函数的暂态稳定评估方法。该方法以故障前潮流量为初始特征集,结合暂态安全稳定量化评估和统计理论方法,提取输入特征;通过支持向量机训练暂态稳定评估模型,得出评估模型的决策函数,并依据支持向量的决策值确定门槛值,保证评估结果保守性。新英格兰10机39节点测试系统和实际系统算例验证了所提方法的可靠性和实用性。     

一种基于脆弱性指标的电力系统动态安全评估模式

为了实现有效的电力系统动态安全评估目标,提出了一种基于脆弱性评估指标的电力系统动态安全评估模式。该评估模式基于系统非故障潮流状态运用三类脆弱性指标来评估电力系统的动态安全,即求取电力系统在非故障潮流变化过程中的脆弱支路集合;评价系统在任一潮流状态下经受故障冲击的脆弱程度指标;评价系统在当前潮流状态下发生连锁故障的可能路径集合及其发生概率。文章对此三类脆弱性指标的建立进行了初步探讨。在该评估模式中,系统当前的潮流状态通过广域测量得到,这可以充分发挥广域测量技术的特点,使得动态安全评估能够反映系统的实际场景,同时,本评估模式依据系统非故障潮流状态中评价系统在当前运行场景下的脆弱性,及早地发现电力系统存在的问题,可为运行人员提前消除危机提供分析依据,因而具有一定的应用前景。