第42卷第1期红水河Vol.42No.12023年2月HongShuiRiverFeb.2023基于PCA-BP神经网络的水电站库区边坡稳定性分析甘海龙(广西机电职业技术学院,广西南宁530007)摘要:为检验神经网络在水电站库区边坡稳定性预测的可行性,通过结合原始数据PCA及神经网络的模式识别特性,构建神经网络模型,对38组实测数据进行PCA处理,选取32组数据作为神经网络输入端数据、6组数据验证神经网络的工作性能。结果表明,神经网络对水电站库区边坡稳定性模式识别率达到83.3%。训练良好的神经网络可以用于工程实践中的水电站库区边坡稳定性预测。关键词:PCA-BP;神经网络;边坡稳定性;水电站;库区中图分类号:TP642.2文献标志码:A文章编号:1001-408X(2023)01-0122-05DOI:10.3969/j.issn.1001-408X.2023.01.025开放科学(资源服务)标识码(OSID):SlopeStabilityAnalysisofHydropowerStationReservoirAreaBasedonPCA-BPNeuralNetworkGANHailong(GuangxiTechnologicalCollegeofMachineryandElectricity,Nanning530007,China)Abstract:Inordertotestthefeasibilityofneuralnetworkinthepredictionofslopestabilityinthereservoirareaofhydropowerstation,aneuralnetworkmodelisconstructedbycombiningtheoriginaldataPCAandthepatternrecognitioncharacteristicsofneuralnetwork.PCAprocessingisperformedon38setsofmeasureddata,and32setsofdataareselectedastheinputdataofneuralnetworkand6setsofdatatoverifytheworkingperformanceofneuralnetwork.Theresultsshowthattherecogn...