《电子技术应用》2023年第49卷第4期5GVerticalIndustryApplication5G垂直行业应用基于AI算法的5G无线通信设备能耗建模方法研究陆赟,丁薇,王梓丞,尹以雁,杨丹,吕沛锦,杨晓康(中国移动通信集团云南有限公司,云南昆明650000)摘要:基于5G网管系统中可采集到的大量能耗、性能指标和基站配置数据,利用机器学习算法对现网主流AAU设备建立能耗测算模型,并对模型的准确性进行验证。测试结果表明,利用该方案对现网几款主流AAU设备建立的能耗测算模型精度都达到97%以上,充分证明该能耗建模方法具有很高的实用性和推广价值。关键词:5G基站;AAU能耗;建模;机器学习算法中图分类号:TN925+.1文献标志码:ADOI:10.16157/j.issn.0258-7998.223378中文引用格式:陆赟,丁薇,王梓丞,等.基于AI算法的5G无线通信设备能耗建模方法研究[J].电子技术应用,2023,49(4):7-10.英文引用格式:LuYun,DingWei,WangZicheng,etal.Researchonenergyconsumptionmodelingmethodfor5Gwirelesscom‐municationequipmentbasedonAIalgorithm[J].ApplicationofElectronicTechnique,2023,49(4):7-10.Researchonenergyconsumptionmodelingmethodfor5GwirelesscommunicationequipmentbasedonAIalgorithmLuYun,DingWei,WangZicheng,YinYiyan,YangDan,LvPeijin,YangXiaokang(ChinaMobileCommunicationsGroupYunnanCo.,Ltd.,Kunming650000,China)Abstract:Basedonalargeamountofenergyconsumption,performanceindicatorsandbasestationconfigurationdatathatcanbecollectedinthe5Gnetworkmanagementsystem,thispaperusesmachinelearningalgorithmstoestablishanenergyconsump‐tioncalculationmodelforcommonAAUequipmentontheexistingnetwork,andverifiestheaccuracyofthemodel.Thetestre‐sultsshowthattheaccuracyoftheenergyconsumptioncalculationmodelsestablishedbythisschemeforseveralcommonAAUdevicesintheexistingnetworkhasreachedmorethan97%,whichfullyprovesthattheenergyconsumptionmodelingmethodhashighpracticabilityandpromotionvalue.Keywords:5Gbasestation;AAUenergyconsumption;modeling;machinelearningalgorithms0引言随着5G移动通信系统的大规模部署,网络提供更快的速率、更大的容量和更广泛的连接的同时,通信设备功耗问题给网络建设和运维带来了极大的困扰。一方面,由于5G使用的频率更高,这使得所需要的5G基站数量相比...