Transmission line fault-cause identification method for large-scale new energy grid connection scenarios

The accurate fault-cause identification for overhead transmission lines supports the operation and maintenance personnel in formulating targeted maintenance strategies and shortening the time of inspecting faulty lines.With the goal of achieving“carbon peak and carbon neutrality”,the schemes for clean energy generation have rapidly developed.Moreover,new energy-consuming equipment has been widely connected to the power grid,and the operating characteristics of the power system have significantly changed.Consequently,these have impacted traditional fault identification methods.Based on the time-frequency characteristics of the fault waveform,new energy-related parameters,and deep learning model,this study proposes a fault identification method suitable for scenarios where a high proportion of new energy is connected to the power grid.Ten parameters related to the causes of transmission line fault and new energy connection scenarios are selected as model characteristic parameters.Further,a fault identification model based on adaptive deep belief networks was constructed,and its effect was verified by field data.

Optimal transmission lines assignment with maximal reliabilities in multi-source multi-sink multi-state computer network

The optimal transmission lines assignment with maximal reliabilities (OTLAMR) in the multi-source multi-sink multi-state computer network (MMMCN) was investigated. The OTLAMR problem contains two sub-problems: the MMMCN reliabilities evaluation and multi-objective transmission lines assignment optimization. First, a reliability evaluation with a transmission line assignment (RETLA) algorithm is proposed to calculate the MMMCN reliabilities under the cost constraint for a certain transmission lines configuration. Second, the non-dominated sorting genetic algorithm II (NSGA-II) is adopted to find the non-dominated set of the transmission lines assignments based on the reliabilities obtained from the RETLA algorithm. By combining the RETLA and the NSGA-II algorithms together, the RETLA-NSGA II algorithm is proposed to solve the OTLAMR problem. The experiments result show that the RETLA-NSGA II algorithm can provide efficient solutions in a reasonable time, from which the decision makers can choose the best solution based on their preferences and experiences.

Analysis and Identification of Corona Parameters on Overhead Power Lines in Case of Direct Lightning Strikes

This paper presents a method for determining specific models of overhead power lines with presence of corona phenomenon. The obtained models provide stable numerical solutions for computer simulation of transients caused by direct lightning strikes. The corona non- linear charge-voltage characteristics obtained from experimental tests are used for identification of the corona parameters based on System Identification Toolbox implemented in Matlab package. Different transfer functions, which give the same waveshapes of overvoltages are de- termined using two parametric models. A circuit representation of the obtained transfer functions is proposed and the corona model is implemented in the EMTP-RV as a hierarchical structure composed of a overhead power line divided into sections with corona branches. Some computer simulations of lightning overvoltages propagated in a typical 220 kV power line due to direct lightning strikes to a line tower are presented. The proposed method and the model implemented in EMTP-RV are still valid for multi-conductor lines and for higher voltages of power lines but new corona nonlinear charge-voltage characteristics are required as an input parameter for the identification procedure.

STPGTN-AMulti-Branch Parameters Identification Method Considering Spatial Constraints and Transient Measurement Data

Transmission line(TL)Parameter Identification(PI)method plays an essential role in the transmission system.The existing PI methods usually have two limitations:(1)These methods only model for single TL,and can not consider the topology connection of multiple branches for simultaneous identification.(2)Transient bad data is ignored by methods,and the random selection of terminal section data may cause the distortion of PI and have serious consequences.Therefore,a multi-task PI model considering multiple TLs’spatial constraints and massive electrical section data is proposed in this paper.The Graph Attention Network module is used to draw a single TL into a node and calculate its influence coefficient in the transmission network.Multi-Task strategy of Hard Parameter Sharing is used to identify the conductance ofmultiple branches simultaneously.Experiments show that themethod has good accuracy and robustness.Due to the consideration of spatial constraints,the method can also obtain more accurate conductance values under different training and testing conditions.

A New Principle of Fault Identification of on the Same Tower Based on Traveling Wave Reactive Powers

In order to improve the reliability of fault identification of the double-circuit transmission lines on the same tower, a new algorithm for fast protection of double-circuit transmission lines on the same tower based on the reactive powers of traveling wave is proposed. With the implementation of S-transform, the initial traveling wave reactive powers are calculated and the change characteristics of reactive power under different fault conditions are studied. The protection criterion is constructed by analyzing the ratio of the reactive powers of the same end on double-circuit transmission lines and the ratio of the reactive powers at both ends on the same line. According to the ratio of reactive power on the same side of the line and both ends of the same line, it is possible to identify whether the faults of the double-circuit line of the same tower occurred in or out of the protection zone. A large number of simulation results show that the protection performance is sensitive and reliable, and quick to respond. The criterion is simple and is basically not affected by fault initial angles, fault types, and transitional resistances.

基于小型移动机器人的高压架空输电线路融冰技术

高压架空输电线路在冬季会受到大量冰雪的覆盖,天气条件的多变会改变冰雪的形态和质地,导致以回波信号为基础的自动融冰方式,在存储回波干扰下,存在明显弊端。为此,研究基于小型移动机器人的高压架空输电线路融冰技术。采用搭载双目视觉传感器的小型机器人,实时捕捉线路的视频图像,提取视频图像中线路外观纹理特征,输入到神经网络当中,识别覆冰状态。利用相似三角形原理推导覆冰线路的三维空间坐标。利用比例积分微分控制(proportional-integral-derivative,PID),实现机器人机械臂越过障碍物后,利用携带的融冰装置铲除覆盖在电力线路上的冰层,完成融冰作业。结果表明:在融冰前输电线路的外观呈现模糊和不规则状态,融冰后输电线路区域变得清晰可见,露出了线路本身的颜色和纹理,表明融冰作业取得了显著的效果,证明了技术的有效性。

考虑暂态过电压特性的架空输电线路雷击故障时空预警方法

针对运行环境复杂多变,故障率持续居高不下,雷击故障占比较大,并具有随机性与无法预知性的问题,提出考虑暂态过电压特性的架空输电线路雷击故障时空预警方法。深入解析雷击机理与雷电流,确定雷击(感应雷与直击雷)输电线路暂态过电压计算方式,通过小波多分辨率分解算法处理输电线路暂态信号,评估暂态过电压在信号频段中的能量比,识别累计故障,判断雷击点与故障点是否一致,利用单端检测确定雷击点的位置,完成架空输电线路雷击故障的时空预警。实验数据显示,在不同实验组别背景下,当阈值pΔ取值为9.3时,雷击故障识别精度达到最大值92%,雷击点定位精度高达80%。可以获得精准的雷击故障识别结果与雷击点定位结果,充分证实了提出方法具备较好的雷击故障时空预警效果。

基于格拉姆角场与ResNet的输电线路故障辨识方法

针对如何利用实际故障录波数据,提取和放大故障特征差异,开展故障类型与故障原因辨识的问题,提出了基于格拉姆角场与迁移学习-ResNet的输电线路故障辨识方法。首先,统计分析了输电线路故障类型和故障原因的分布特征,用于指导构建适用于类不平衡问题的故障分类器。然后,利用格拉姆角场变换将采集得到的故障电压、电流时序信号转化为格拉姆角场图像,放大故障特征差异,作为故障分类器的输入。进一步,将生成的图像集输入搭建好的故障分类器进行网络训练和测试,输出输电线路故障类型和故障原因。最后,完全采用真实故障录波数据开展了算例分析。结果表明:所提方法对故障类型的辨识准确率达到了97.51%,对故障原因的辨识准确率达到了94.23%。并且将训练的故障辨识网络迁移至其他地区时,仍然具有较好的故障辨识效果和泛化性能。所提方法为基于暂态波形数据驱动的故障辨识提供了新方法,可以用于实际电网的输电线路故障辨识。

深度学习下直流输电线路绝缘子破损识别仿真

输电线路绝缘子由于架设在野外,受天气因素、环境污染等影响,使得绝缘子发生破损,导致输电线路供电中断。为避免输电线路发生故障,最大程度保障电网安全运行,提出基于深度学习的直流输电线路绝缘子破损识别方法。分析航拍图像噪声来源,利用中值滤波算法确定像素中值,经过非线性平滑处理去除噪声。分析常见的绝缘子破损类型、破损表现及原因,建立随机森林决策树。通过深度学习,选出明显的破损特征作为识别依据;构建Alex Net卷积神经网络模型,计算损失函数,确定最佳学习速率;通过学习训练,输出识别结果。实验结果显示,所提方法能够增强图像细节信息,且绝缘子破损识别正确率在0.8以上、收敛速度快。

基于小波分析的输电线路故障识别检测

在输电线路故障识别的过程中,容易受到高频干扰信号影响,导致降低识别检测结果的准确性。因此,该文提出了基于小波分析的输电线路故障识别检测方法。分析输电线路故障信号奇异性,计算故障暂态信号的小波包能量,构建小波窗函数,以突变特性信号下的过电流为识别判据,检测输电线路故障情况,识别冗余信号,避免检测结果受到高频干扰信号影响。实验结果表明,该方法检测的输出电流波动范围与实际电流波动范围一致,说明识别检测精准度较高。

一种基于CNN与FFT‑ELM的输电线路故障识别与定位方法

及时、准确地检测输电线路故障类型与位置是提高电力系统可靠性最重要的问题之一,为此提出一种基于卷积神经网络(convolutional neural networks,CNN)与基于快速傅里叶变换(fast Fourier transform,FFT)的极限学习机(extreme learning machine,ELM)分类模型并行的输电线路故障识别及定位方法。首先,以故障电压时序图作为输入,构建CNN;然后,利用FFT将时域故障电压数据分解,提取各频段的电压峰值与相角作为故障特征样本;接着,以提取的故障特征样本集作为输入,构建ELM网络;最后,通过特征融合层将2个神经网络进行融合,输出故障类型和定位结果。实验结果表明,此方法对输电线路故障识别的准确率为99.95%、故障定位误差在500 m以内、平均误差为263.5 m,可靠性优于其他模型。

电网架空输电线路绝缘子缺陷识别算法研究

电网架空输电线路绝缘子长期处于外部暴露环境中,易产生侵蚀、开裂、破碎等问题,导致电网运行的可靠性降低。为此,对电网架空输电线路绝缘子缺陷识别算法进行了研究。通过滤波去除采集的绝缘子图像乘性噪声,利用不同尺度高斯掩膜进行卷积处理,并结合图像梯度方向提取表面缺陷特征。将构建的深度前馈神经网络作为基本架构,设置三层网络结构的自编码器。在前馈网络的优化目标函数中融入惩罚因子,以构建稀疏性限制条件。通过遗传算法优化自编码器初始权值,经网络迭代得到表面缺陷识别结果。试验结果表明,所提算法对绝缘子缺陷的识别准确率始终高于90%、平均识别用时为1.72 s,具有显著的可行性。该研究对监控安防、图像分析等领域具有一定的启示意义。

基于卷积Transformer的支路参数辨识

为了解决在电力系统支路参数辨识任务中传统的单支路特征参数辨识只能对单一目标进行识别,无法充分利用电力系统的历史信息的问题。在Transformer模型基础之上采用一种卷积自注意力,帮助模型更好地捕获输入特征之间的关系,将局部上下文更好地纳入注意力机制;同时采用卷积门控循环单元进行位置编码,保证内容和位置的一致性,从而降低训练损失,进一步改善预测的精确度。仿真结果表明,该算法预测精度较高,明显优于其它的机器学习算法和深度学习算法。

雷击线路电磁暂态响应干扰下电缆避雷器故障传感检测

交流输电电缆终端的避雷器故障特征与雷击线路电磁暂态过响应过程的信号及其类似。当前,从单端行波信息中的检测方法很难区分雷击特征与真实故障行波的特征,提出了一种雷击线路电磁暂态响应干扰下的电缆避雷器缺陷故障传感检测。利用传感器收集电缆避雷器的故障行波数据,分析出避雷器故障时,与雷击线路电磁暂态过响应不同,其初始行波的波头下降沿存在转折、突变或者后续存在振荡过程等。运用特征提取和选择技术,从大量的数据中提取出与这种故障模式相关的有效特征,建立相应的故障诊断方法,并通过训练和优化来实现准确的故障智能诊断。实验结果表明,所提方法能够有效地区分故障特征与雷击线路电磁暂态过响应特征,误差监测数值最大为0.148 mA,具有较好的可行性。

基于无人机机载激光雷达的输电线路单端故障自动化识别系统

为了解决输电线路单端故障识别时易出现误判或漏判的问题,该文设计基于无人机机载激光雷达的输电线路单端故障自动化识别系统。该系统将激光雷达系统搭载于无人机中,采集输电线路点云数据,粗提取输电线路点云,聚类粗提取结果,实现输电线路点云的细提取。将得到的结果作为极端学习机的输入,利用核函数优化极端学习机,输出输电线路单端故障自动化识别结果。实验结果表明,所设计系统能够自动识别金属性接地等不同类型的输电线路单端故障,且识别精度高。

基于AI技术的输变电脆弱线路自动识别与控制应用

输变电系统是电力输送的重要纽带,为了降低运维成本和提高识别效果,提出一种基于AI技术的输变电脆弱线路自动识别与控制应用方法。利用人工智能技术(AI)的决策树贪心算法(ID3)度量线路不同运行特征之间的信息量,得到线路特征,引入互信息特征选取(MIFS)优化决策树,利用调节系数和惩罚项平衡线路运行特征间的冗余度,并且引入加权度数,衡量节点权重并重点考虑所有连接线路的有功总量。在综合分析电纳介数和加权度数等指数前提下得到综合的脆弱线路自动识别指标,建立识别模型,最后改变输变电线路的潮流传输及负载变化实现控制。实验证明所提方法能够准确识别出脆弱线路位置,方法有效可行,其控制应用也有效降低了运维成本。

面向危险施工机械智能识别的改进YOLOv7算法

为了实现输电线路上危险施工机械(挖掘机、推土机和打桩机)的智能识别,该文提出了一种改进的YOLOv7算法。通过爬虫获取复杂条件下的危险施工机械图像并建立数据集,结合Mixup和Mosaic等多种图像增强方法预处理数据集。采用K-means++算法改进YOLOv7,提高算法的识别效率。实验结果证明,K-means++算法提高了YOLOv7的模型收敛速度;所提出算法能有效进行危险施工机械的识别,识别效果优于YOLOv7,准确率提高了3.70%。

基于无人机的架空输电线路缺陷智能识别

为了提高智能识别效果,设计一种基于无人机的架空输电线路缺陷智能识别方法。建立图像信息提取模型,在图像信息卷积输出结构中获取卷积核的输入输出值以及激活函数,得到函数的回归损失函数值。生成架空输电线路缺陷判别网络,得到不同采样点的目标函数,建立离散微分算子的矩阵信息,计算双边拟合误差,匹配特征点的函数值。设计线路缺陷智能识别算法,完成架空输电线路缺陷的智能识别方法设计。实验结果证明,设计方法可以以最少的训练次数获取最小的损失值,得到损失值的最小值约为0.35,识别效果更佳,具有一定应用价值。

基于连续波SAR成像模型的输电线路悬垂绝缘子风偏角识别方法

为了在恶劣环境中保证输电线路悬垂绝缘子风偏角识别结果可信,提出基于连续波合成孔径雷达(SAR)成像模型的输电线路悬垂绝缘子风偏角识别方法。基于连续波SAR成像模型,持续发射脉冲信号,通过基于位置标定的绝缘子风偏角识别方法,定位绝缘子目标区域中绝缘子串2个端点的像素坐标,生成绝缘子SAR图像。采用基于改进频率变标算法的连续波SAR成像处理方法,缓解绝缘子SAR图像失焦现象,进行SAR图像分割,获取生成绝缘子SAR图像中绝缘子目标区域,并结合“两点确定一条直线”标准识别绝缘子串的风偏角。实验验证:所提方法在恶劣环境中像素拟合效果较佳,对绝缘子风偏角识别结果和实际风偏角的风偏度信息高度匹配。

基于GIS的海宁市输电线路及通道管理系统

针对众多的输电线路,如何快速了解线路的路径走向、电压等级等台账信息;如何快速识别输电线路通道内的鸟巢、树障、人工建筑物等危险点信息;如何把输电线路信息高效、安全共享给线路设计等部门;如何快速更新输电线路通道状态信息。针对上述问题,利用GIS技术、信息技术和无人机、地物智能判别等融合手段,构建基于GIS的海宁市输电线路通道管理系统,通过此系统可极大提高输电线路管理效率和实现通道内危险物的自动识别、预警,以保证通道安全。