准噶尔盆地玛湖凹陷风城组陆相页岩油储层测井裂缝智能识别

Fracture Intelligent Identification Using Well Logs of Continental Shale Oil Reservoir of Fengcheng Formation in Mahu Sag,Junggar Basin

准噶尔盆地玛湖凹陷北部斜坡区风城组陆相页岩油储层为多物源混合沉积,多种岩性频繁互层,岩石力学层厚度小,导致其裂缝尺度小,裂缝常规测井响应弱,识别难度大.针对页岩裂缝测井识别的难题,应用集成学习中的极端梯度提升树方法,通过深度挖掘裂缝信息与测井数据之间的非线性关系,将多个弱分类器集成强分类器,降低裂缝识别的不确定性,以提高裂缝的识别能力.该方法将岩心裂缝描述和井壁成像测井裂缝解释结果作为标签,常规测井信息作为模型训练的输入数据,在异常点筛查、SMOTE过采样处理和特征优选的基础上,通过网格搜索方法获得裂缝智能识别模型的最优超参数.通过与目前常用的支持向量机和逻辑回归等机器学习方法对比,极端梯度提升树具有比其他两种非线性机器学习方法更好的裂缝识别效果,测试集识别准确率可达90%.A1井风3段识别结果反映了该段裂缝较为发育,且模型对于裂缝段与非裂缝段都具有较好的识别效果,与岩心观察结果符合率较高.表明极端梯度提升树具有较好的裂缝识别能力,能够为玛湖凹陷陆相页岩油储层的裂缝智能识别提供有效手段.

The continental shale oil reservoir of Fengcheng Formation in the northern slope area of Mahu Sag,Junggar Basin is a mixed deposition of multiple provenances,with frequent interbedding of various lithologies and small thickness of rock mechanics layer,resulting in small fracture scale,weak conventional logging response of fractures and great difficulty in identification.Aiming at the logging identification of shale fractures,in this paper it applies the extreme gradient boosting(XGBoost)method in ensemble learning to deeply mine the nonlinear relationship between fracture information and logging data,integrates multiple weak classifiers into strong classifiers,reduces the uncertainty of fracture identification,and improves the accuracy of fracture identification.In this method,core fracture description and fracture interpretation results of image logging are used as labels,and conventional logging information is used as input data for model training.On the basis of outlier screening,SMOTE oversampling and feature optimization,the optimal hyperparameters of fracture intelligent identification model are obtained through grid search method.Compared with the commonly used machine learning methods such as support vector machine and logical regression,the XGBoost has better fracture identification ability than the other two nonlinear machine learning methods,and the accuracy of test set identification reaches 90%.The identification results of the P1f3 of Well A1 reflect that the fractures in this section are relatively developed,and the model has a good identification ability for both fractured and non-fractured sections,with a high coincidence rate with the core observation results.It can provide effective means for intelligent identification of fractures in continental shale oil reservoirs in Mahu Sag.