基于SSA-CNN模型的煤储层含气量预测方法研究

——以鄂尔多斯盆地东部M区块本溪组为例

打开文本图片集

中图分类号：P618.13 文献标识码：A 文章编号：2097-5465（2026）01-0052-07

Abstract：Accuratepredictionofgascontentideepcoalreservoirsisofsignificantengneeingvaluefortheeficientdevelopent ofcoalbedmethane.However，relianceonasinglegeophysicalparameteroftenoverlooksthecontrolof interalstructuresand heterogeneityofcoalreservoirsongascontent，leading todeviationsbetwenpredictedresultsandmeasuredValues，whichfailto meettheprecisionrequirementsforeficientdeepcoalbed methane extraction.Basedonin-depth miningof geophysicalinformation， thisstudintroducesanitrisicmehanisticprameterofoalresevoirsshontentsasupplementtogologicalchaacteristics， therebycompensatingfortheinadequacyof geophysicalinformationandconstructingamultimodalfeaturesystemforgascontent prediction. Taking the No.8 coal reservoir of the Carboniferous Benxi Formation in Block M on the eastern margin of the Ordos Basinas thestudyobject，aconvolutionalneuralnetwork optimizedbythe SparrowSearch Algorithm（SSA-CNN）wasemployedto automaticallyextractspatialfeaturesof thedataandestablishahigh-precisiongascontent predictionmodel.Theresultsindicate that：1）ThroughSpearmannonlinearcorelationanalysis，sixfactorswithsignificantinflunceongascontentwereselectedand determined as inputs for the prediction model；2）The SSA-CNN model achieved a prediction accuracy of R2=0 .817 on the test set，with the mean absolute error reduced by 1.067% compared to the traditional CNN model. Practical research and analysis demonstratethattheSSA-CNNmodelcanbefectivelyappliedtohigh-precisiongascontentpredictionincoalreservoirsandhas promotional value for gas content prediction inregions with similar geological backgrounds.

Keywords;ashcontent；gascontent；BenxiFormation；sparowsearchalgorithm；convolutionalneuralnetwork；OrdosBasin

0 引言

近年来，随着勘探开发工作的深入，深部煤层气资源潜力日益显现。（剩余7910字）