主成分分析与模糊识别在岩性识别中的应用

发布时间：2018-02-17 02:46

本文关键词： 模糊识别主成分分析岩性识别测井参数苏里格气田　出处：《岩性油气藏》2017年05期 　论文类型：期刊论文

【摘要】：表征复杂岩性的测井曲线之间的相似度较高,在岩性识别过程中存在着大量信息冗余,造成测井曲线具有一定的模糊性与相关性,干扰识别过程,导致识别效果不理想。以苏里格气田苏东41-33区块马家沟组五段碳酸盐岩测井数据为例,采用一种基于主成分分析与模糊识别相结合的方法解决了这一难题。该方法首先提取声波时差(AC)、自然伽马(GR)、光电吸收截面指数(P_e)、补偿中子(CNL)、补偿密度(DEN)、深侧向电阻率(R_(LLD))等对岩性变化反映比较敏感的测井曲线参数,通过主成分分析构建出3个综合变量Y_1,Y_2和Y_3,再采用模糊识别方法对研究区的岩性进行识别。与传统识别方法相比,主成分分析与模糊识别相结合的岩性识别方法能有效消除特征曲线间的模糊性与相关性,并使岩性识别的正确率达到86%,是一种实用且有效的识别复杂岩性的方法,具有一定的推广和应用价值。
[Abstract]:The similarity between log curves representing complex lithology is high, and there is a large amount of information redundancy in lithology recognition, which results in the log curves having certain fuzziness and correlation, which interferes with the recognition process. The result is that the recognition effect is not satisfactory. Taking the carbonate logging data of the fifth member of Majiagou formation in Suodong 41-33 block of Sulige gas field as an example, A method based on principal component analysis (PCA) and fuzzy recognition is used to solve this problem. The method firstly extracts acoustic moveout (AC), natural gamma ray (RG), optoelectronic absorption cross section index (PSP), compensation neutron (CNL), compensation density (DEN), deep lateral electricity (DEC). The resistance rate of RX / LLDX and other parameters of logging curves are sensitive to lithologic changes. Through principal component analysis, three synthetic variables Y1C / Yak _ 2 and Ys _ 3 were constructed, and then the lithology of the study area was identified by fuzzy recognition method. Compared with the traditional identification method, The lithology recognition method combined with principal component analysis and fuzzy recognition can effectively eliminate the ambiguity and correlation between characteristic curves, and make the accuracy of lithology recognition reach 86. It is a practical and effective method to identify complex lithology. Has certain promotion and application value.
【作者单位】：中国地质大学(北京)数理学院;北京中地润德石油科技有限公司;
【基金】：国家自然科学基金项目“基于正交匹配追踪算法的相敏关联层析研究”(编号:29932016006)与“异常高压致密砂砾岩油藏衰竭式开采渗流规律研究”(编号:51674227)联合资助
【分类号】：P618.13;P631.81

【相似文献】