Neural network based transient stability assessment of electric power systems

In this paper a systematic procedure to design an intelligent neurocontroller to assess the dynamic security of interconnected power systems is presented. This approach focuses on the integration of modern control theory together with the adaptive networks to determine critical clearing time for power systems.     

基于卷积神经网络的电力系统暂态稳定预防控制方法

为了更好地实现电力系统暂态稳定预防控制,提出了基于卷积神经网络(Convolutional Neural Network, CNN)的电力系统暂态稳定预防控制方法。通过CNN模型输出变量灵敏度选择控制发电机并确定控制量,然后采用CNN和时域仿真相结合的暂态稳定评估方法进行控制方案校核,得到使系统在预想故障下稳定的控制方案。采用某省级电网算例进行预防控制效果验证。结果表明,采用所提出的预防控制方法,可以找到使系统恢复稳定的预防控制策略。     

基于GPS同步量测量的模糊神经网络用于暂态稳定预测研究

利用GPS同步时钟获得系统各机组的功角或系统内最大摇摆角 ,然后通过模糊神经网络进行暂态稳定性预测 ,充分利用了模糊系统和神经网络的优点 ,一方面考虑了专家的经验 ,另一方面又通过样本集进行学习 ,能自动提取模糊规则、优化隶属函数等 ,因此具有较高的模式分类正确率和函数逼近精度。对 6机 2 2节点的算例表明 ,所提方法是有效的。     

基于模糊神经网络的交流伺服系统

提出了一种基于神经网络的模糊控制交流伺服系统,将神经网络与模糊逻辑结合起来,输入信号先模糊化,然后通过构建的神经网络,在线调整其权值和变化的控制参数,使系统的输出具有更好的动、静态性能,提高了系统的鲁棒性。仿真实验证明了这种基于神经网络模糊控制方法在交流伺服系统中应用的可行性和可靠性。     

A two stage model for rotor angle transient stability constrained optimal power flow

Transient stability constrained optimal power flow (TSCOPF) is a nonlinear optimization problem with both algebraic and differential equations. This paper utilizes the Imperialist Competitive Algorithm (ICA) as an evolutionary optimization algorithm and Artificial Neural Network (ANN) to develop a robust and efficient two stages scheme to solve TSCOPF problem. In the first stage an Artificial Neural Network is constructed to predict the rotor-angle transient stability margin, and is then incorporated in the TSC-OPF as the transient stability estimator. To solve the proposed TSC-OPF problem the ICA is used as the optimizer. The performance of the proposed method is verified over the WSCC three-machine, nine-bus system under different loading conditions and fault scenarios.     

基于模糊高斯基函数神经网络控制的交流伺服系统

将模糊控制与神经网络相结合,用神经网络来实现模糊推理,提出了一种把高斯基函数作为隶属函数的模糊神经网络,并将之用于交流伺服系统的控制中     

模糊神经网络光伏功率调节系统

光伏功率调节系统(PVPCS)集光伏并网发电与无功补偿为一体,以此来提高电能质量和减少电网功率损耗.在分析了系统的工作原理和控制策略的基础上,提出了基于模糊神经网络(FNN)的智能控制策略,构成了具有双FNN模型结构的光伏功率调节系统,能够稳定直流侧电容电压,优化对电网谐波、无功的补偿效果,而且具有更强的鲁棒性和适应性.仿真结果在调整系统功率的同时使谐波含量从4.61％下降到4.18％,证实了所提策略的可行性.     

模糊J-K触发器神经网络降低传感器交叉敏感方法

用模糊J-K触发器作为神经元的作用函数,构建了模糊J-K触发器神经网络系统,并将其应用于消除压阻式压力传感器干扰电流对目标参量的影响.实验结果表明,该方法结合了模糊技术和J-K触发器的优势,具有结构简单、训练速度快的优点.同采用常规S型函数作为神经元作用函数的神经网络相比,模糊J-K触发器神经网络训练次数少,均方根误差...     

基于GD-FNN的特高压直流输电暂态稳定控制

特高压直流输电输电容量大、控制灵活迅速,在我国“西电东送”战略中扮演了重要角色,其对受端交流系统暂态稳定性的影响也随输电容量的增大而变大。提出采用广义动态模糊神经网络（GD-FNN）控制来提高系统暂态稳定性,通过附加控制信号动态调节输送功率从而给系统提供足够阻尼。根据系统选择合适的调制信号以及控制维数,对GD-FNN系统训练后通过系统误差和模糊规则的 -完备性作为判据来优化系统结构,同时对隶属度函数的参数进行修正,从而保证了控制器具有紧凑的结构和良好的泛化能力。仿真结果表明,所设计的暂态稳定控制器在保持系统稳定方面具有优越的性能,并且鲁棒性较好,可有效保障机组和电网的安全稳定运行。     

基于边界强化混合采样的两阶段电力系统暂态稳定评估

受制于样本固有的不平衡性,基于数据挖掘的暂态稳定预测方法不易用于工程实践,为此,提出一种基于边界强化混合采样的两阶段暂态稳定评估模型。在第1阶段,利用预训练的级联卷积神经网络模型确定边界和非边界样本集,利用条件生成对抗网络合成边界集失稳样本,并对非边界集稳定样本进行欠采样,以实现边界强化;在第2阶段,利用混合采样后的重构样本集再训练卷积神经网络模型,以更好地挖掘失稳样本的隐含特征,并采用改进后的焦点损失函数加强模型对边界集样本的学习能力。新英格兰39节点系统与南方某省级电网的仿真结果表明,所建模型有效降低了对失稳样本的漏判率,提高了整体预测精度,在样本极不平衡的情况下仍有良好的评估性能。     

基于模糊神经网络算法的交流伺服系统DSP控制

将神经网络与模糊逻辑控制结合起来,设计模糊神经网络控制器应用于交流伺服系统中的转速调节器,克服交流调速系统中参数漂移、非线性和耦合等因素的影响.针对模糊神经网络控制器运算量大、收敛慢的特点,硬件采用数字信号处理器(DSP)作为控制器运算单元,并在DSP上实现模糊神经控制算法,提高了系统实时性.实验结果表明,采用该控制器的调速系统具有较快的响应速度、较高的稳态精度和较强的鲁棒性.     

一种神经网络变结构电力系统稳定器的设计

在变结构控制理论的基础上 ,设计了一种基于神经网络的电力系统稳定器 (NNPSS)。利用BP神经网络对系统的状态进行辨识 ,通过对网络的训练 ,使神经网络能对不同的运行状态及扰动产生相对应的附加励磁控制。仿真实验证实了神经网络电力系统稳定器的可行性 ,所设计的NNPSS有效改善了电力系统的稳定性。     

Online critical clearing time estimation using an adaptive neuro-fuzzy inference system (ANFIS)

This paper describes an approach using an adaptive neuro-fuzzy inference system (ANFIS) for the assessment of online critical clearing time (CCT). The ANFIS can integrate neural networks and fuzzy logic principles, and has a potential to combine the advantages of both in a single framework. In this paper, the ANFIS is applied for the prediction of CCT by varying load levels and fault locations in buses and transmission lines. The IEEE 39-bus system and 9-bus western system coordinating council are tested and implemented in this study. All machines of the IEEE 39-bus system are considered as the classical model without considering any generator’s exciters. While three machines in the 9-bus western system coordinating council are considered as detailed models, forth-order differential equation is described for all machines by considering the excitation system controller. CCT values obtained by the time domain simulation method using step-by-step calculation are used as the benchmark. The power world version 17 is used for transient simulation, and the ANFIS is implemented using MATLAB version 2014B. The results obtained from the ANFIS approach are quite satisfied with high accurate solutions and much lower computation time. Finally, the graphical user interface in MATLAB is applied for the online CCT estimation of two test power systems by using appropriate ANFIS models obtained from simulations.     

基于模糊神经网络高压电气设备绝缘状况评价方法研究

提出了利用模糊逻辑-神经网络分析高压电气设备绝缘参数变化趋势的方法。根据在线监测获得的实时和历史数据确定绝缘参数变化的趋势法,应用隶属函数实现对变化趋势的描述,建立了相应的数学模型。基于模糊神经网络构造了评价绝缘状况的“正常”、“观察”和“检修”的模糊评价体系,把绝缘参数变化归为“减小”、“平稳”和“增大”3个模糊集合,得到评价设备绝缘状况的变化趋势准则。该方法可用在高压电气设备绝缘在线监测系统中,保证在线监测的准确性和可靠性。     

基于FFT 和神经网络的高精度谐波分析

为了精确分析整数和非整数次谐波,提出了基于快速傅里叶变换(FFT)和神经网络的谐波分析方法,该方法的特点是采用基函数参数可调的神经网络。具体是先把信号进行FFT处理,得到谐波个数和精度不高的谐波幅值、相位、谐波次数;其次根据谐波个数设定神经元的个数,根据预处理后得到的幅值、和相位、谐波次数设定神经网络权值和基函数参数迭代的初始值;最后对人工神经网络进行训练,便可实现整数和非整数次谐波的精确分析,同时能将频率相近的非整数次谐波分离。仿真结果验证了该方法的有效性与易实现性。     

基于XGBoost-EE的电力系统暂态稳定评估方法

深度学习在暂态稳定评估中发挥着越来越重要的作用,然而电网规模逐渐扩大导致数据出现维数灾难,这对模型的性能提出了更高的要求.目前,暂态稳定特征构建需要依靠人工经验,具有主观性;深度学习的模型在设计和训练上耗时、耗力.针对以上两点,结合极限梯度提升(XGBoost)算法和实体嵌入(EE)网络,提出了一种基于XGBoost-...     

基于模糊输入的BP-ART2混合神经网络在电力变压器故障综合诊断中的应用

根据模糊理论和神经网络理论,提出了变压器故障诊断的新方法。根据DGA(dissolvedgasanalysis)法、电气试验法及外部故障特征法,建立了基于模糊输入的BP ART2混和神经网络对电力变压器故障进行综合诊断。仿真结果表明本方法能有效提高变压器故障诊断正确率。     

An estimation method of available transfer capabilities from viewpoint of power system transient stability under deregulated environment

To conduct electric power transactions effectively and to operate a power system efficiently while maintaining reliability under the deregulated environment, it is required that ATC (Available Transfer Capability) be calculated at high speed and with reasonable precision. In order to address this issue, in this paper, an Artificial Neural Network‐based estimation method for evaluating Maximum Transmission Capability (MTC), which is a key step but also a highly time consuming process in ATC, is proposed. It is confirmed through simulation studies that the proposed method is capable of estimating MTC (ATC) with high speed and sufficient precision. Furthermore, the authors examined the reduction of calculation time at learning by using the transient stability index. © 2009 Wiley Periodicals, Inc. Electr Eng Jpn, 167(1): 66–73, 2009; Published online in Wiley InterScience ( www.interscience.wiley.com ). DOI 10.1002/eej.20781     

基于灵敏度分析和时域仿真的暂态稳定预防控制优化方法

根据临界切除时间灵敏度将暂态稳定约束条件线性化,简化了考虑暂态稳定约束的最优潮流(TSCOPF)问题的求解过程,为暂态稳定预防控制优化问题提供了一种新的解决方案。针对大系统临界切除时间灵敏度计算费时的实际问题,采用线路有功功率灵敏度筛选待调发电机的方法和并行计算技术以提高计算效率。WSCC 9节点系统标准算例和SGCC 9 177节点电网算例的仿真结果表明,所提方法具备良好的优化效果和工程实用性。     

Study of a new method for power system transients classification based on wavelet entropy and neural network

The detection and classification of transient signals are widely applied in many fields of power system. The study of transient signal detection and classification is a sustaining focus of researchers as well as a difficult issue. There are still many problems needed to be solved in this area. Based on the wavelet transform (WT), the idea of entropy and weight coefficient is introduced, and the wavelet energy entropy (WEE) and wavelet entropy weight (WEW) are defined in this paper. The distribution picture of WEE and WEW along with scales are presented for the first time. PSCAD/EMTDC models for six types of transients, namely breaker switching, capacitor switching, short circuit fault, primary arc, lightning disturbance and lightning strike fault, are constructed. With WEE and WEW, the eigenvectors for the six transients are established and a model which uses the eigenvectors as the input of the BP (back-propagation) neural network is set up to realize the classification of these transients. The simulation has been executed based on a 500 kV transmission line model in China and the results show that feature extraction based on WEE and WEW can effectively discover the useful local features. With the help of neural network classifier, it has effective classifying result. This method is applicable in the power system.