基于物理信息同步学习的高频传输线电压预测研究

打开文本图片集

中图分类号：TP391 文献标识码：A 文章编号：2096-4706（2025）17-0022-06

Abstract： High-frequency transmissonlines play a vital role in power systems and are widely used in the propagation andinteractionofelectromagneticwaves.Therefore，acuratepredictionoftheirvoltageisofgreatsignificanceforinformation acquisition.Howee，existingnumeicalthodsavecertainlitationsincomputatioaliencyTotisend，theoltage predictionmethodofhigh-frequencytransmisionlinesisstudied，andamethodbasedonPhysicallyInformedSynchronous Learning （PISL）isproposed.FirstlyaNeuralNetwork forpredictingvoltageisonstructed，anditrandomlysamples tobtain the spars trainingdatasetandcolocationpointsetwithout labels.Secondly，adata-physics informationhybridloss function is constructedtotrainthenetwork，whichconsidersdatalossandphysics-informedlosssythetically.Finallytheearson corelationcoeffcientandRootMean SquareErrorareusedas evaluationcriteria toverifytheeffectivenessof the proposed method through experiments.Meanwhile，a comparativesensitivityanalysisof therelevant network parameters isappliedto verify the effectiveness and robustness of the method.

Keywords：data-physics information hybridloss function; Physicaly Informed Synchronous Learing; voltage prediction; Neural Network

0 引言

电力系统是支撑社会运作的关键基础设施，也是推动经济和科技发展的核心动力。（剩余7098字）