基于机器学习算法的深部页岩储层物性预测及有利勘探区优选

Prediction of physical properties of deep shale reservoirs and optimization of favorable exploration areas based on machine learning algorithms

以四川盆地泸州区块龙马溪组页岩为例,在岩心物性测试与测井数据基础上,对比分析了支持向量回归算法(SVR)、梯度提升决策树(GDBT)和极端梯度提升算法(XGBoost)3种机器学习算法预测结果并优选出适宜的评价模型,实现了深部页岩储层物性预测与有利勘探区优选。结果显示:1)3种热门学习算法在深部页岩储层物性预测方面均具有良好的应用效果,经过对比,认为GDBT为预测研究区深部页岩储层孔隙度最适宜的算法,XGBoost为预测研究区深部页岩储层渗透率最适宜的算法;2)基于上述优选模型,泸州区块龙马溪组页岩孔隙度为2.67%~9.67%,平均孔隙度为4.88%,渗透率为3.22~28.63μD,平均渗透率为11.34μD;3)依据龙马溪组页岩储层物性与含气量,可划分出7个I类有利区,3个Ⅱ类有利区和4个Ⅲ类有利区。该研究成果基于研究区实际情况,可为研究区与类似区块页岩储层评价及有利区预测提供参考。

To achieve the prediction of the physical properties of deep shale reservoirs and the selection of favorable exploration areas,Longmaxi shale in Luzhou Block of Sichuan Basin was taken as an example.Based on the physical property test of core samples,predicted physical properties of the reservoirs were compared by three machine learning algorithms:Support Vector Regression(SVR),Gradient Boosting Decision Tree(GDBT),and Extreme Gradient Boosting(XGBoost),and the appropriate evaluation model was selected.Then,the physical properties of deep shale reservoirs were predicted,and the favorable exploration areas were selected.The results showed that:1)Among the three popular machine learning algorithms,GDBT was the most suitable algorithm to predict the porosity of deep shale reservoirs in the study area,and XGBoost was the most suitable algorithm to predict the permeability of deep shale reservoirs in the study area;2)Based on the above-selected models,the porosity of Longmaxi shale in Luzhou Block ranges from 2.67%to 9.67%,with an average of 4.88%,and the permeability ranges from^(3).22μD to 28.63μD,with an average of 11.34μD;3)Based on the physical properties and gas content of Longmaxi shale reservoir,7 Class I favorable areas,3 Class II favorable areas,and 4 Class III favorable areas were identified.Based on the actual situation of the study area,this achievement can provide a reference for the evaluation of shale reservoirs and the prediction of favorable areas in the study area or similar areas.