CAJ | 학술논문

在取心、岩石物理实验的基础上，采用统计方法对雷64区的岩性进行研究，将其分为砂质细砾岩和不等粒砂岩两大类。利用直方图法优选出识别这两大类岩性的敏感测井曲线，并且依据优选出来的敏感测井曲线，建立了用于识别砂质细砾岩和不等粒砂岩的测井响应交会图法、决策树模型、系统聚类模型。通过对上述三种模型进行精度分析，发现决策树模型的精度要高于测井响应交会图法和系统聚类模型。同时利用所建立的两种决策树模型对雷71井进行岩性识别，将识别结果与实际取心对比，发现综合判断准确率均大于85％。结果表明，给出的决策树模型可很好地用于该地区砂砾岩储层的岩性识别。
재취심、암석물리실험적기출상，채용통계방법대뢰64구적암성진행연구，장기분위사질세력암화불등립사암량대류。이용직방도법우선출식별저량대류암성적민감측정곡선，병차의거우선출래적민감측정곡선，건립료용우식별사질세력암화불등립사암적측정향응교회도법、결책수모형、계통취류모형。통과대상술삼충모형진행정도분석，발현결책수모형적정도요고우측정향응교회도법화계통취류모형。동시이용소건립적량충결책수모형대뢰71정진행암성식별，장식별결과여실제취심대비，발현종합판단준학솔균대우85％。결과표명，급출적결책수모형가흔호지용우해지구사력암저층적암성식별。
Based on coring and petrophysics experiment, the lithology in Lei 64 were studied by statistical methods and then divided into sandy micro-conglomerate and inequigranular sandstones.The sensitive log was optimized by column diagram, and the models of well log-ging cross-plot, decision tree and systematic clustering were developed to identify the two categories lithology based on the sensitive log.Through the study of the accuracy of the above three models, it was found that the accuracy of decision tree was higher than well logging cross-plot and systematic clustering.Finally two models of decision tree were used to iden-tify lithology of Lei 71 well.The integrated identification accuracy rate was more than 85%compared with actual coring.The results showed that the model of decision tree could be well used to identify lithology in sand-conglomerate reservoirs of this district.