摘要
为了快速预测鄂尔多斯盆地白豹-南梁地区长3、4+5低渗透储层的产能级别并明确主控因素,基于42口井76层的常规测井曲线、测井解释砂体结构、试油结果及相关沉积研究成果,建立主要测井参数与产能级别的配置关系,利用相关性较高的测井参数和砂体结构,通过多层感知器网络分析方法,构建低渗透储层产能级别的预测模型,并确定出各测井参数与低渗透储层产能级别相关程度的大小次序为:砂体结构>电阻率>密度>自然伽马>声波时差>中子.应用确定的多层感知器模型对研究区78口井2341层的产能级别进行预测,解释出Ⅰ类储层109层、Ⅱ类储层194层、Ⅲ类储层2038层;Ⅰ类、Ⅱ类储层以块状砂岩为主,约占81%、68%,电阻率、孔隙度较高;Ⅲ类储层则以互层状为主,达到85.7%,电阻率、孔隙度相对较低.规模应用结果表明,方法预测结论与试油结果具有较高的吻合度,正确率达到82.6%,判断失误的储层均是相邻级别,没有出现跳级现象.通过多层感知器网络分析方法,协同测井-地质信息,可以快速、有效地评价预测低渗透储层产能级别,能为有利建产目标优选及开发方案调整提供可靠依据.对其他盆地低渗透-超低渗透储层产能评价亦具有较好的参考和借鉴意义.
To quickly predict the productivity level of Chang 3 and 4+5 low permeability reservoirs in Baibao, Nanliang areas of Ordos Basin and identify the main controlling factors, the relationships between main logging parameters and productivity level were established based on the conventional logging curves, the structure of sand bodies interpreted by logging, the results of oil testing and the related sedimentary research results of 76 layers of 42 wells, and obtained the main well logging parameters and sand body structure with high correlation. On this basis, multilayer perceptron network analysis methods were used to establish the quantitative prediction models of productivity level of low permeability reservoirs and analyze the correlation degree between logging parameters and productivity level of low permeability reservoir. The order of correlation degree from big to small is sand body structure, resistivity, density, natural gamma, interval transit time, and neutron. The established multi-layer perceptron fast prediction methods of productivity level of low permeability reservoirs were applied to 2341 layers of 78 wells at Chang 3, 4+5 of Baibao and Nanliang areas in Ordos Basin. The results shows that 109 layers are class Ⅰ, 194 layers are class Ⅱ and 2038 layers are class Ⅲ. There are 81% block sandstones with higher resistivity and porosity among class Ⅰ and 68% in class Ⅱ. While in class Ⅲ, there are 85.7% interbedded sandstones with lower resistivity and porosity, few block sandstones. On the whole, a large number of applications show that the results of the method interpretation are in good agreement with the actual test results, and the correct rate is 82.6%. Meanwhile, the productivity level of low permeability reservoirs with misjudgment are all adjacent levels, and there is no jumping phenomenon. Therefore, the prediction error of the method will not have a great impact on the application effect,and will not bring fatal errors. By means of multilayer perceptron network analysis method and synergistic logging-geological information,the productivity level of low permeability reservoirs can be evaluated and predicted quickly and effectively,and reliable basis can be provided for optimizing production targets and adjusting development schemes. It also has good reference and reference significance for productivity evaluation of low-permeability and ultra-low-permeability reservoirs in other basins.
作者
张兆辉
廖建波
李智勇
郑希民
邸俊
余平辉
ZHANG Zhao-hui;LIAO Jian-bo;LI Zhi-yong;ZHENG Xi-min;DI Jun;YU Ping-hui(Key Laboratory of Petroleum Resources,Gansu Province/Key Laboratory of Petroleum Resources Research,Chinese Academy of Sciences,Lanzhou 730020,China;University of Chinese Academy of Sciences,Beijing 100049,China;Petro China Research Institute of Petroleum Exploration and Development-Northwest,Lanzhou 730020,China)
出处
《地球物理学进展》
CSCD
北大核心
2019年第5期1962-1970,共9页
Progress in Geophysics
基金
甘肃省自然科学基金课题(17JR5RA313)
中国科学院油气资源研究重点实验室开放基金课题(KFJJ2016-02)
甘肃省油气资源研究重点实验室“十三五”科技创新基金课题(135CCJJ20160524)联合资助
关键词
多层感知器
砂体结构
低渗透储层
产能级别
鄂尔多斯盆地
Multilayer perceptron
Sandbody structure
Low permeability reservoirs
Productivity level
Ordos basin
作者简介
第一作者:张兆辉,男,1982年生,陕西渭南人,博士研究生,主要从事储层测井评价方法研究.(E-mail:zhangzhaohui_123@163.com)。
