基于投影能量函数和Pin-SVM的电力系统暂态稳定评估

提出一种基于Pin-SVM的电力系统暂态稳定评估方法。首先,采用系统指标(如平均机械功率、初始加速度和系统冲击等)和投影能量函数指标(如投影角速度、投影角加速度和投影动能PKE)构建暂态稳定指标的原始特征集,通过最大相关最小冗余特征选择方法对暂态指标集进行特征压缩,寻找对电网暂态变化敏感度高的特征子集;然后,基于Pin-SVM思想将特征子集映射到高维空间,实现非线性暂态稳定评估问题的线性转换,进而引入分位数改变系统稳定类与不稳定类之间的最近点位置,将暂态稳定分类问题转换为在Pin-SVM中寻找最优分位数距离问题,以减小边界干扰样本的影响,提高电力系统暂态评估方法的评估准确率和稳定性。最后,以IEEE-39节点系统、IEEE-145节点系统和某实际算例进行仿真计算,计算结果验证了该方法的有效性和准确性。     

基于深度学习的电力系统暂态功角与暂态电压稳定裕度一体化评估

随着面向高比例可再生能源新型电力系统的转型,系统运行特性日趋复杂。暂态功角稳定(transientangle stability,TAS)与暂态电压稳定(transientvoltagestability,TVS)问题相互耦合且频发,为系统安全稳定评估带来严峻挑战。研究首先采用变步长二分法通过调用PSASP从时间维度上构建了暂态电压与暂态功角的稳定边界。研究了不同故障位置、感应电动机占比、负荷率对稳定边界的影响并依托边界确定主导失稳模式。其次提出一种基于注意力机制与一维卷积神经网络融合的电力系统功角稳定及电压稳定裕度评估的新方法。该方法直接面向测量数据,将节点稳态与暂态运行的电压幅值、有功功率、无功功率数据作为输入特征,节省了数据处理时间。通过一维卷积神经网络构建输入特征与极限切除时间的映射,利用注意力机制进一步提高了模型预测效果。通过新英格兰IEEE39节点系统进行分析验证,结果表明该方法可以实现暂态安全裕度的快速评估且具有较高的预测精度。     

基于卷积神经网络的暂态电压稳定评估及风险量化

提出了一种暂态电压稳定性评估及其风险量化方法.首先,探讨卷积神经网络(CNN)与暂态电压稳定评估的关联性和匹配性,建立了基于CNN的暂态电压稳定评估模型.其次,在可信度框架下引入四元评估结构,可有效解决CNN在稳定边界识别上对时域仿真依赖的难题.然后,通过评估结果获取暂态电压稳定裕度,并将其与可信度相结合来构建风险函数,从而实现对暂态电压稳定的风险量化分级.实际电网算例分析结果验证了该方法的有效性.     

基于深度学习的暂态稳定评估与严重度分级

提出一种安全域概念下的堆叠降噪自动编码器和支持向量机集成模型相结合的暂态稳定评估方法。将故障前的潮流量作为输入,利用堆叠降噪自动编码器对输入量进行多层抽象表达,使用提取的各层特征训练支持向量机;建立支持向量机集成分类模型进行暂态稳定评估,对评估结果进行可信度分析,将输入空间划分为稳定区、边界区和失稳区;利用效用理论结合所提出的暂态稳定裕度指标对运行方式进行严重度分级。算例结果表明,所提暂态稳定评估方法具有更高的评估准确率和一定的泛化能力;所提严重度分级方法能够直观表现不同运行方式的危险程度。     

基于信息熵的属性离散化算法在暂态稳定评估中的应用

针对人工神经网络用于大规模电力系统暂态稳定评估时所面临的训练样本集庞大、训练负担重，以及收敛性差的问题，提出一种用于训练样本集压缩的属性离散化方法。在矢量空间聚类的基础上，将每个聚类在各属性轴上投影的边界设为候选离散断点，采用基于信息熵的正交化增益函数选择最终断点进行离散，并对离散后的数据进行压缩处理；然后，采用一种改进的反向传播神经网络对离散压缩后的训练样本集进行学习训练和分类测试。在新英格兰10机39节点系统和某省电力系统中的应用表明，在保证暂态稳定分类精度的前提下，所提出的属性离散化和样本集压缩方法可有效地压缩训练样本集的规模，减轻神经网络的训练负担，提高收敛速度。     

基于神经网络的快速汽门暂态稳定控制器的研究

提出一种应用单隐层前向神经网络决定电力系统暂态稳定快速汽门控制规律的方法。将实现控制规律的神经网络控制器分为两个子神经网络：子神经网络 1，用来确定系统的暂态稳定情况，同时给出相应汽门最小开度和汽门持续关闭的时间；子神经网络 2，用于确定不稳定情况下汽门的最终开度。该神经网络控制器用于受扰动后的三机系统中，其仿真结果令人满意。     

一种用于半监督BP算法的实用结束判据及其应用

针对用于BP网络训练的半监督学习算法，提出了一种新的实用的训练结束判据，并将其应用于电力系统暂态稳定分类中。根据暂态稳定输入特征的空间分布，提出一个新的表征样本混杂度的指标，用于判断类别的可分离度。在新英格兰l0机39节点系统中的应用表明，该训练结束判据用于半监督BP算法对BP网络进行训练，可以有效地避免基于神经网络暂态稳定评估的误分类，提高稳定评估结果的可靠度。     

基于复合神经网络的电力系统暂态稳定评估和裕度预测

提出一种基于复合神经网络的暂态稳定评估与故障临界切除时间(CCT)裕度预测新方法,它将概率神经网络(PNN)和径向基函数(RBF)网络组合使用,充分利用两者各自的优点,以提高暂态稳定评估能力和CCT裕度预测能力。该方法首先利用PNN进行暂态事故场景分类,分类时充分考虑了相邻故障样本类型重叠的影响;进一步采用RBF网络对分类结果进行裕度预测;最后,通过自检和校正以提高预测精度。利用New England 39节点系统,通过与反向传播(BP)神经网络、RBF神经网络等方法的比较,证明了本文方法的优越性。     

基于深度学习的电力系统暂态稳定评估方法

在机器学习领域,暂态稳定评估问题被定义为通过大量故障样本来估计稳定边界的二分类问题。该文提出了一种深度学习方法来解决这个二分类问题。该方法包含4个步骤:首先,利用样本数据构建原始输入特征来描述电力系统动态特性;然后,采用变分自动编码器(variational auto-encoders,VAE)对原始输入特征进行无监督学习实现特征抽取,从而获得高阶特征;之后,对卷积神经网络(convolution neural network,CNN)进行有监督学习训练得到高阶特征与电力系统暂态稳定性之间的映射关系;最后,将训练得到的模型应用于电力系统在线暂态稳定评估。在新英格兰39节点测试电力系统的仿真试验表明,所提出的暂态稳定评估(transient stability assessment,TSA)模型具有评估精度高、不稳定样本评估错误率低、抗噪声干扰能力强的特点,适合基于广域测量信息的准实时在线暂态稳定评估。     

基于一维卷积神经网络的电力系统暂态稳定评估

系统遭遇暂态故障的过程是随时间发展的过程,基于传统机器学习的暂态稳定评估方法通常难以捕捉其时间维度信息,限制了评估性能的提高。针对该问题,提出了一种基于一维卷积神经网络(1D-CNN)的暂态稳定评估方法。该方法直接面向底层量测数据,凭借其特有的一维卷积和池化运算特性,能自动提取出暂态过程所蕴含的时序特征,从而达到对系统暂态稳定状态准确刻画的目的。设计了一种适用于暂态稳定评估的四卷积层1D-CNN模型,实现了端到端的"时序特征提取+暂态稳定性分类",并通过调整模型关键参数以提高失稳样本查全率,增强了评估结果的可靠性。新英格兰10机39节点测试系统的仿真实验表明,相较于传统机器学习暂态稳定评估方法,所提方法能以更短的响应时间做出更准确的暂态稳定性判断,满足在线暂态稳定评估准确性与快速性的要求。     

基于神经网络暂态稳定评估方法的一种新思路

This paper presents a new framework for arificial neural networks (ANN) based transient stability assessment(TSA).The ANN-based TSA problems may be treated as a two-pattern classification problem separating the stable class from the unstable class.The cases close to the classification boundary are often liable to be misclassified.This paper proposes to train a back-propagation ANN using a novel semisupervised learning algorithm for deriving a continuous-spread stability index.The derived stability index is used to indicate the relative stability degree and define a boundary zone between the two different classes.A new classification scheme is hence proposed to group the boundary-zone cases into an extra indeterminate class to avoid misclassifications.A nonlinear mapping for data structure analysis is employed as a convenient tool to observe the separability of the input space.Two well-studied power systems are employed to demonstrate the validity of the proposed approach.     

人工神经网络和短时仿真结合的暂态安全评估事故筛选方法

结合人工神经网络（ＡＮＮ）和短时数字仿真提出一个用于在线暂态安全评估的事故筛选方法,将３层ＢＰ网络作为模式分类器,用来建立稳定评估结果和所选特征量之间的映射关系,在故障切除时刻终止的短时数字仿真被用来生成ＡＮＮ的输入量,每个ＡＮＮ处理一个特定的事故状态,使用一个半监督学习算法,ＡＮＮ可产生一个能够指示相对稳定度的连续分布的暂态稳定指标,基于这个连续分布的稳定指标,设置一个相对保守的分类门槛值,避免     

基于正则化投影孪生支持向量机的电力系统暂态稳定评估

提出一种基于正则化投影孪生支持向量机的暂态稳定评估方法。将基于传统支持向量机进行暂态稳定评估的高维二项式优化问题转化为两个低维二项式优化问题,并在投影孪生支持向量机的目标函数中引入正则项来改善评估稳定性。首先,构建由系统特征和投影能量函数特征组成的初始样本集,通过特征选择对初始特征进行压缩,获取可有效表征暂态稳定性的最优特征集。然后,基于正则化投影孪生支持向量机的思想将暂态稳定状态分成稳定类与不稳定类,寻找各稳定状态的最佳投影坐标轴,使稳定类投影到稳定类投影超平面上后尽可能地聚成簇,而不稳定类投影到稳定类投影超平面上后尽可能远离稳定类聚成的簇,降低暂态稳定评估的计算时间,同时借助遗传算法进行参数选择以提高准确率。最后,通过IEEE-145和南方电网算例的仿真分析,验证了所提方法的有效性和准确性。     

ANN-based novel fault detector for generator windings protection

In this paper, an artificial neural network (ANN) based internal fault detector algorithm for generator protection is proposed. The detector uniquely responds to the winding earth and phase faults with remarkably high sensitivity. Discrimination of the fault type is provided via three trained ANNs having a six dimensional input vector. This input vector is obtained from the difference and average of the currents entering and leaving the generator windings. Training cases for the ANNs are generated via a simulation study of the generator internal faults using Electromagnetic Transient Program (EMTP). A genetic algorithm is employed to reduce training time. The proposed ANN algorithm is compared with a conventional differential algorithm. It is found to be superior regarding sensitivity and stability     

Modern approaches for protection of series compensated transmission lines

Series compensation has been employed to improve power transfer in long-distance transmission systems worldwide. However, this in turn introduces problems in conventional distance protection. The complex variation of line impedance is accentuated, as the capacitor's own protection equipment operates randomly under fault conditions. This paper proposes two approaches based on travelling waves and artificial neural networks (ANN) for fault type classification and faulted phase selection of series compensated transmission lines.A modal transformation technique, which decomposes the three-phase line into three single-phase lines, is used for this purpose. Algorithms based on two different modal transformations are developed for phase selection and fault classification. Each algorithm is derived from a corresponding truth table. The truth tables are constructed for different types of faults with different faulted phases and different transformation bases.The proposed ANN topology is composed of two levels of neural networks:

• In level-1, a neural network (ANN_F) is used to detect the fault. In level-2, four neural networks (ANN_A, ANN_B, ANN_C and ANN_G) are used to identify faulted phase(s), and activated by the output of ANN_F if there is a fault.
• System simulation and test results, which are presented and analyzed in this paper indicate the feasibility of using travelling waves and ANN in the protection of series compensated transmission lines.

     

A new power system transient stability assessment method based on Type-2 fuzzy neural network estimation

Transient stability assessment (TSA) of large power systems by the conventional method is a time consuming task. For each disturbance many nonlinear equations should be solved that makes the problem too complex and will lead to delayed decisions in providing the necessary control signals for controlling the system. Nowadays new methods which are devise artificial intelligence techniques are frequently used for TSA problem instead of traditional methods. Unfortunately these methods are suffering from uncertainty in input measurements. Therefore, there is a necessity to develop a reliable and fast online TSA to analyze the stability status of power systems when exposed to credible disturbances. We propose a direct method based on Type-2 fuzzy neural network for TSA problem. The Type-2 fuzzy logic can properly handle the uncertainty which is exist in the measurement of power system parameters. On the other hand a multilayer perceptron (MLP) neural network (NN) has expert knowledge and learning capability. The proposed hybrid method combines both of these capabilities to achieve an accurate estimation of critical clearing time (CCT). The CCT is an index of TSA in power systems. The Type-2 fuzzy NN is trained by fast resilient back-propagation algorithm. Also, in order to the proposed approach become scalable in a large power system, a NN based sensitivity analysis method is employed to select more effective input data. Moreover, In order to verify the performance of the proposed Type-2 fuzzy NN based method, it has been compared with a MLP NN method. Both of the methods are applied to the IEEE standard New England 10-machine 39-bus test system. The simulation results show the effectiveness of the proposed method in compare to the frequently used MLP NN based method in terms of accuracy and computational cost of CCT estimation for sample fault scenarios.     

一种应用模糊神经网络对完全电机定子进行故障保护的新方法

该文研究一种由模糊逻辑和人工神经网络(ANN)组成的，对发电机定子线圈进行故障保护的综合方法。该方法只采用发电机机端电压、电流信号，并对这些信号进行特征提取，然后用到FNN的故障诊断和定位上。该技术由两个阶段组成：基于人工神经网络(ANN)的故障类型分类和基于一个包括模糊逻辑以及人工神经网络的综合网络的进行定子绕组故障精确定位。     

Two approaches to transient stability-constrained optimal power flow

Formulation of transient stability-constrained optimal power flow (TSC-OPF) and finding a practical solution for the problem have gained much attention recently. In this paper, two approaches to include transient stability constraints in the OPF problem considering detailed dynamic models for generators and their controls are introduced. The first method is based on the maximum relative rotor angle deviation (MRRAD) of generators which suits systems that have specific requirements on MRRAD. The second method represents the transient stability margin of the system based on generators output power (GOP) and hence does not rely on MRRADs. The transient stability boundary can be represented by a nonlinear function of GOP. The Artificial Neural Network (ANN) curve-fitting tool is used to derive a mathematical formulation for the transient stability boundary (TSB). The closed form representation of the TSB is then inserted in the OPF problem as a new constraint. The proposed method is examined using the WSCC 9-bus, the New England 39-bus and the IEEE 300-bus test systems. The results indicate that the proposed methods lead to lower computational time in solving TSC-OPF which has been a serious challenge for this problem.     

基于稳定域边界理论的暂态稳定指标及其应用

基于故障后电力系统稳定域边界的隐式方程,提出了隐式的暂态稳定指标,并利用稳定域的二次近似方法给出了该暂态稳定指标的一个近似估计.进一步,将所得的暂态稳定指标应用于临界切除时间计算和故障排序,其中临界切除时间由求解暂态稳定指标和切除时间的拟合二次方程得出,故障的严重程度排序由比较该暂态稳定指标直接得出.最后,在IEEE 3机9节点和新英格兰10机39节点系统中进行的仿真结果验证了所提出的暂态稳定指标的准确性.