基于随机森林算法的BP神经网络模型在坝基渗压水位预测中的应用

打开文本图片集

关键词：渗压水位；随机森林算法；BP神经网络；精度；白鹤滩水电站

中图分类号：P731.34 文献标志码：A

doi：10.3969/j.issn.1000-1379.2026.01.024

Application of BP Neural Network Model Based on Random Forest Algorithm in Predicting Seepage Pressure Water Level of Dam Foundation

WANG Zhuoqun，WANG Jianxin，WANG Huimin， SHENG Jinchang，FENG Jun1 （1.School of Water Resources and Hydropower，Hohai University，Nanjing 21O98，China; 2.Power China Huadong Engineering Co.，Ltd.，Hangzhou 31Oooo，China）

Abstract：Inordertoimprovethepredictionacuracyofsepagewaterlevelofhydropowersationdamfoundation，aBPneuraletwork modelbasedonrandomforest（RF-BPmodel）wasproposed.Taking Baihetanhydropowerstationasanexample，thedataof18seepage measurementpointsatthedamfoundationfromAugust1，2O21toFebruary23，2O23wereanalyzed.TheGA（GeneticAlgorith）-B，PSO （Particle SwarmOptiization）-BP，RF，STM（LongShortTermMemory）-BPmodelswereselectedtocomparethepredictonaccuracy withtheRF-BPmodel.Consideringthattereasacertaincoelationbetweenthesepagewaterlevelandthreservoirwaterlevel，the Pearsoncorelationcoeficientofthetwowascalculated.Teresultsshowthat theRF-BPmodelhastesmalestMAE，RMSEandMAPE andthehighestpredictionaccuracat tetypicaleasurementpointsofOH-W1-1，OH-WM1-2andOH-WM5-3，hichighlightsthe significanteffctofadomforesalgoiinotigseleiofctosTestrogerhorelationbetnteseepgewatelevelad thereservoiratevelatteaseet，grdictucfF-odeicatitato twen the sepage water level and the reservoir water level has an important impact on the prediction accuracy.

Key words：seepage water level；random forest algorithm；BP neural network；accuracy；Baihetan hydropower station

0 引言

随着计算机技术的发展，神经网络模型展现出优秀的自我调整能力和函数近似能力。（剩余6641字）