摘要
总有机碳含量(Total Organic Carbon,TOC)是评价页岩油气潜力的关键参数,利用测井数据预测TOC,能刻画整段储层的TOC变化,对于明确地质-工程“甜点”意义重大.和尚塬地区延长组长7段陆相页岩由于沉积环境频繁交替变化,页岩层系内大量发育粉砂质泥岩条带.针对岩性非均质性强的特点,将页岩储层细分为页岩和粉砂质泥岩两种岩性,分别建立TOC预测模型.在比较现有方法的基础上,结合研究区的测井资料实际,选用△lg R改进法和机器逐步回归学习法分岩性建立预测模型.经过实测数据的精度检验,机器逐步回归学习法的预测精度更高.通过对单井预测值与实测值的误差分析,印证了细分岩性机器逐步回归学习预测模型的可靠性,证明该方法对预测以常规测井为主的陆相泥页岩TOC是有效的.在此基础上,应用该方法对研究区101口井长7段泥页岩进行预测,获得了TOC的空间展布.
TOC is a key parameter to evaluate shale oil and gas potential,TOC prediction by using logging data can depict the change of TOC in the whole section of the reservoir,which is of great significance for clarifying the geological-engineering sweet spot.Due to the frequent alternation of depositional environments,a large number of siltstone mudstone bands are developed in the shale of member 7 in Yanchang formation in Heshangyuan area.In view of the characteristics of strong lithologic heterogeneity,TOC content of the shale reservoir was predicted by distinguishing the two lithologies of shale and siltstone.Based on the comparison of the existing methods and the actual well logging data in the study area,the improved method of¢lgR and the machine stepwise regression learning method were selected to establish the prediction model by lithology.After testing the accuracy of the measured data,the results show that the machine stepwise regression learning method has higher prediction accuracy.The error analysis between the predicted and measured values in a single well also confirms the reliability of the machine stepwise regression learning method by subdivision lithology,proving that the method is effective in predicting TOC content of mud shale,which is mainly based on conventional logging.On this basis,TOC content and its spatial distribution is obtained by the logging interpretation of 101 wells’shale of member 7 in Yanchang formation.
作者
秦晓艳
王震亮
程昊
赵晓东
QIN Xiaoyan;WANG Zhenliang;CHENG Hao;ZHAO Xiaodong(College of Civil Engineering,Shaanxi Polytechnic Institute,Xianyang Shaanxi 712000,China;Department of Geology,Northwest University,Xi’an Shaanxi 710069,China;China National Logging Corporation Geology Research Institute,Xi’an Shaanxi 710077,China;No.12 Oil Production Plant,PetroChina Changqing Oilfield Company,Heshui Gansu 745400,China)
出处
《新疆大学学报（自然科学版中英文）》
CAS
2024年第5期620-628,共9页
Journal of Xinjiang University(Natural Science Edition in Chinese and English)
基金
国家自然科学基金面上项目“延安地区下古生界碳酸盐岩天然气的成藏过程和机理”(41172122)
陕西工业职业技术学院“青年科技创新团队研究项目”(KCTD2022-01),“基于地质-工程一体化的湖相页岩评价方法研究”(2024YKYB-023)。
关键词
总有机碳含量预测
测井资料
陆相页岩
机器学习
total organic carbon content prediction
logging data
continental shale
machine learning
