基于遗传算法-深度神经网络的分布式光纤监测工作面矿压预测-豆柴文库

您所在位置：网站首页 / 基于遗传算法-深度神经网络的分布式光纤监测工作面矿压预测.docx / 文档详情

在线预览结束，喜欢就下载吧，查找使用更方便

5 金币

下载文档

/ 2

下载提示

如果您无法下载资料，请参考说明：

1、部分资料下载需要金币，请确保您的账户上有足够的金币

2、已购买过的文档，再次下载不重复扣费

3、资料包下载后请先用软件解压，在使用对应软件打开

文本预览

基于遗传算法-深度神经网络的分布式光纤监测工作面矿压预测
Title:PerformancePredictionofMining-inducedPressureatDistributedFiber-opticMonitoringWorkingFaceBasedonGeneticAlgorithm-DeepNeuralNetwork
Abstract:
Mining-inducedpressureinundergroundcoalminingposesasignificantthreattothesafetyandstabilityofworkingfaces.Tomitigatethisrisk,accurateandtimelypredictionofmining-inducedpressureiscrucial.Inthispaper,weproposeanovelapproachcombininggeneticalgorithm(GA)anddeepneuralnetwork(DNN)topredictthemining-inducedpressureatdistributedfiber-opticmonitoringworkingfaces.TheGAisutilizedtooptimizetheDNNmodelinordertoimprovethepredictionaccuracy.
1.Introduction:
Mining-inducedpressureatworkingfacesisanimportantfactoraffectingthesafetyandstabilityofundergroundcoalmining.Accuratepredictionofmining-inducedpressurecanhelpengineerstakeproactivemeasurestopreventpotentialrisksandensurethesmoothprogressofminingoperations.Traditionalpredictionmodelsoftenstruggletocapturethecomplexnon-linearrelationshipbetweeninputparametersandmining-inducedpressure.Toaddressthischallenge,weproposeahybridapproachthatintegratesGAandDNN.
2.LiteratureReview:
Previousstudieshaveemployedvarioustechniquestopredictmining-inducedpressure,includingtraditionalstatisticalmodels,machinelearningmethods,andnumericalmodeling.However,thesemethodshavelimitationsinaccuratelycapturingthenon-linearrelationshipsinthedataoraccuratelypredictingdynamicchangesinmining-inducedpressure.Recentadvancementsindeeplearningtechniques,particularlyDNNs,haveshownpromisingresultsinvariousfields.However,theoptimizationprocessofDNNsisoftentime-consumingandcomputationallyexpensive.Inthisstudy,weproposetheuseofGAtooptimizetheDNNmodelforefficientpredictionofmining-inducedpressure.
3.Methodology:
3.1DataCollectionandPre-processing:
Acomprehensivedatasetofdistributedfiber-opticmonitoringdata,includingstress,strain,temperature,andotherrelevantparameters,wascollectedfromaworkingface.Thedatawaspre-processedtoremoveoutliersandnormalizetheinputparameters.
3.2GeneticAlgorithmOptimization