ΔLogR技术与BP神经网络在复杂岩性致密层有机质评价中的应用

发布时间：2018-04-12 18:28

本文选题：复杂岩性 + 致密层　；参考：《地球物理学进展》2017年03期

【摘要】：本文系统分析ΔLogR技术应用于复杂岩性致密层有机质评价中存在两方面的局限性:参数选取方面,测井曲线选取过于单一,无法有效削弱致密层段复杂岩性和孔隙度等因素对计算有机碳含量的影响;构建模型方面,人为剔除异常点存在随机性与偶然性误差,影响建模准确性.针对上述问题,本文建立了BP神经网络模型,并将其应用于柳河盆地柳参1井下桦皮甸子组烃源岩有机质评价.研究结果表明,在不剔除异常点情况下,BP神经网络模型计算TOC值和实测116组TOC值相关性达到0.886,显示建模效果良好.分别应用BP神经网络和ΔLogR模型,计算研究区致密层纵向上连续的TOC曲线,BP神经网络模型的计算TOC曲线与实测TOC数据基本吻合,而ΔLogR模型的计算TOC曲线吻合度较差.因此在测井资料完善的情况下,本文建议使用BP神经网络评价复杂岩性的致密层有机质.
[Abstract]:In this paper, we systematically analyze the limitations of 螖 LogR technique in evaluating organic matter in complex lithologic dense beds. The parameters are chosen, the logging curve is too single.The influence of complex lithology and porosity on the calculation of organic carbon content can not be effectively weakened. In the construction of model there are randomness and fortuitous errors in the artificial elimination of abnormal points which affect the accuracy of modeling.In view of the above problems, a BP neural network model is established and applied to the evaluation of organic matter in source rock of Huapianzi formation in Liushen 1 downhole in Liuhe Basin.The results show that the correlation between the calculated TOC value of BP neural network model and the measured TOC value of 116 groups is 0.886, which shows that the modeling effect is good.BP neural network and 螖 LogR model are used to calculate the longitudinal continuous TOC curves in the study area. The calculated TOC curves are in good agreement with the measured TOC data, but the calculated TOC curves of the 螖 LogR model are less consistent.Therefore, under the condition of perfect logging data, it is suggested that BP neural network be used to evaluate the organic matter of dense layers with complex lithology.
【作者单位】：中国石油大学(华东)非常规油气与新能源研究院;中国石油大学(华东)地球科学与技术学院;
【基金】：国家自然科学基金资助项目(41330313,41402122) 中央高校基本科研业务费专项资金资助(13CX05013A) 博士后科学基金面上项目(2014M561980)联合资助
【分类号】：P631.81

【相似文献】