多元非线性矿物解释模型在页岩分析中的应用.pdfVIP

多元非线性矿物解释模型在页岩分析中的应用

摘要该研究以泸州地区五峰组一龙马溪组已钻井的测井资料、岩心测试数据和全岩衍射数据为基础，通过数据集人

工智能训练、多参数优选和多元非线性拟合，探讨了计算矿物体积分数的敏感参数，利用常规测井GR,KTH,CNL,AC,

DEN等5种数据建立了伊利石、石英、方解石和黄铁矿等矿物体积分数的预测解释模型）其中：伊利石与方解石体积分数与

KTH,DEN,CNL相关性强；石英体积分数与CNL,KTH,AC相关性强；黄铁矿体积分数与GR,CNL,AC相关性强）对比发

现，利用多元非线性矿物解释模型解释的伊利石、方解石、石英和黄铁矿等4种矿物体积分数与元素测井矿物体积分数吻

合度好，标准偏差和残差均小于10%-在该模型基础上进行岩性的识别，可以解决深水陆棚沉积储层复杂岩相划分的问

题，为识别沉积微相储层划分提供了快捷高效的方法，对于页岩气开发具有重要意义-

关键词测井数据；岩性评价；五峰组；龙马溪组；泸州地区

Applicationofinterpretationmodelsofmultivariatenon-1inearmineralinshale

formationanalsis

Abstract：Basedowellloggingdata,coretestdataandwholerockdiffractiodataofdrilledwellsiLongmaxi-Wufeng

FormatioiLuzhouAreaithispaper,thesensitiveparametersthataffectmineralscontentarediscussedbyAItrainingofdata

setmulti-parameteroptimizatioandmultivariatenonlinearfitting,andpredictiomodelsofmineralscontentincludingillite,

quartz,calciteandpyriteareestablishedbyusingconventionalloggingGR,KTH,CNL,ACandDENcurves.Amongthem,the

illiteandcalcitecontentshowstrongcorrelatiowithKTH,DENandCNLcurves,thequartzcontentshowsstrongcorrelatiowith

CNL,KTHandACcurves,andthepyritecontentshowsstrongcorrelatiowithGR,CNLandACcurves.Itisfoundby

comparisothatillite,calcite,quartzandpyritecontentsinterpretedbyconventionalloggingareconsistentwiththatofelement

capturedlogging,andtherelativeerrorofeachmineralcontentislesstha10%.Theidentificatiooflithologybasedothe

modelscabeusedtosolvetheclassificatioofcomplexlithologiesideep-watershelfsedimentaryreservoirs,andprovidesafast

andefficientmethodfordivisioofsedimentarymicrofacies,whichisofgreatsignificanceforshalegasdevelopment.

Keyw