文档名:基于改进DDAE的风电场集电线单相接地故障测距
摘要:为解决风电场混合接线的集电线短路后难以精确定位的问题,提出基于改进深度去噪自编码网络的故障测距方法.分析集电线故障零序电流可知,暂态电流值、稳态电流幅值、稳态电流相位与故障距离呈现强非线性关系,借助深度学习挖掘这一复杂关系以实现集电线精确定位.在深度自编码框架上添加距离回归输出端口,采用联合训练以提升定位网络的准确性、抗噪性和鲁棒性.其过程为:借助PSCAD/EMTDC搭建集电线模型,将给定时窗内故障零序电流序列和对应距离作为故障样本,仿真不同情况故障生成样本集;在训练集上训练改进深度自编码网络,得到最优网络用于精确测定故障距离.借助各测点零序电流幅值关系可先确定故障区域,将故障信号送入已训练好的网络即可确定故障所在精确位置.文中方法对集电线多分支、混合短线路有着良好的适应能力,定位性能明显优于传统机器学习算法,且受过渡电阻、采样率、噪音、故障相位角影响较小.
Abstract:Inordertosolvetheproblemthatitisdifficulttoaccuratelylocatethecollectorlineaftershort-circuitofhybridconnectioninwindfarm,afaultlocationmethodbasedonimproveddeepdenoisingauto-encoder(DDAE)networkispresentedinthispaper.Byanalyzingthezero-sequencecurrentoncollectorlinefaults,itisknownthatthetransientcurrentvalue,steady-statecurrentamplitude,steady-statecurrentphaseandfaultdistancearestrong-lynon-linear,andthepreciselocationofcollectorlineisachievedbydeeplearningminingthiscomplexrelation-ship.Adistanceregressionoutputportisaddedtothedeepdenoisingauto-encodernetwork,andjointtrainingisusedtoimprovetheaccuracy,noiseresistanceandrobustnessofthepositioningnetwork.Firstly,thehubmodelisbuiltwithPSCAD/EMTDC,andfaultzero-sequencecurrentsequenceandcorrespondingdistanceinagiventimewindowareusedasfaultsamplestosimulatedifferentfailurecasestogeneratesamplesets.Then,animproveddeepauto-encodernetworkistrainedonthetrainingsettoobtainanoptimalnetworkforpreciselymeasuringthefaultdistance.Withthehelpofthezero-sequencecurrentamplituderelationshipofeachmeasurementpoint,thefaultareacanbedeterminedfirst,andthepreciselocationofthefaultcanbedeterminedbyfeedingthefaultsam-plesintothetrainednetwork.Thismethodproposedinthispaperhasagoodadaptabilitytomulti-branchandhy-bridshortlinesofcollectorlines.Locationperformanceissignificantlybetterthantraditionalmachinelearningalgo-rithms,aswellaslessaffectedbytransitionresistance,samplingrate,noise,faultphaseangle.
作者:朱永利 刘富州 张翼 郑艳艳Author:ZHUYongli LIUFuzhou ZHANGYi ZhengYanyan
作者单位:华北电力大学电气与电子工程学院,河北保定071003
刊名:电测与仪表 ISTICPKU
Journal:ElectricalMeasurement&Instrumentation
年,卷(期):2024, 61(5)
分类号:TM416
关键词:深度去噪自编码 风电场 集电线路 故障定位
Keywords:DDAE windfarm collectorline faultlocation
机标分类号:TM774TM835.2TK774
在线出版日期:2024年5月27日
基金项目:国家自然科学基金基于改进DDAE的风电场集电线单相接地故障测距[
期刊论文] 电测与仪表--2024, 61(5)朱永利 刘富州 张翼 郑艳艳为解决风电场混合接线的集电线短路后难以精确定位的问题,提出基于改进深度去噪自编码网络的故障测距方法.分析集电线故障零序电流可知,暂态电流值、稳态电流幅值、稳态电流相位与故障距离呈现强非线性关系,借助深度学习挖...参考文献和引证文献
参考文献
引证文献
本文读者也读过
相似文献
相关博文
基于改进DDAE的风电场集电线单相接地故障测距 Single-phase grounding fault location of wind farm collector based on improved DDAE
基于改进DDAE的风电场集电线单相接地故障测距.pdf
- 文件大小:
- 7.03 MB
- 下载次数:
- 60
-
高速下载
|
|