利用SOFM对鸡西盆地城子河组储层综合评价  

Self-organizing feature map (SOFM) neural network method in reservoir quality synthetic evaluation and its application in Jixi basin

作　　者：郄瑞卿[1,2] 薛林福[2] 孙晶[1] 张琳[3] 

机构地区：[1]吉林农业大学土地管理系,长春130118 [2]吉林大学地球科学学院,长春130026 [3]吉林省国土资源信息中心,长春130042

出　　处：《测绘科学》2010年第3期89-92,共4页Science of Surveying and Mapping

基　　金：国家油气重大专项(XQ-2004-07)

摘　　要：通过对已有的储层评价方法优势与不足的分析,提出了在空间数据库基础上应用自组织特征映射神经网络进行油气储层评价。通过对鸡西盆地城子河组表征储层性能参数的分析,建立储层评价参数标准,并按照建立空间数据库-网格化-文件转换-文件合成-神经网络评价-类别评价-图形绘制的评价流程,生成鸡西盆地城子河组储层综合评价图。评价结果显示:Ⅰ级储集层主要发育于鸡东坳陷中央即鸡D6井东侧呈东西向条带状分布和鸡D6井西侧不规则分布;梨树镇坳陷西部:梨树镇周缘即鸡D2、鸡1、鸡3井周缘呈块状分布;Ⅱ级储集层区块主要发育于鸡东坳陷中部;Ⅲ级储集层区块最大;Ⅳ级在鸡西盆地零星分布。By analyzing advantages and disadvantages of existing method in reservoir quality synthetic evaluation, a SOFM （self- organizing feature map） neural network method was proposed based on spatial database. The parameter standard of reservoir appraisal was established by analysis of Chengzihe group reservoir. According to the flowchart of establishment of spatial database- grid- file conversion- file synthesis- neural network appraisal- category appraisal - graph plan, the synthesis appraisal map in Chengzihe group reservoir of Jixi Basin was generated. The results indicated that the first class reservoir mainly developed in Jidong central-depression - that is, the eastern margin of Ji-D6 well was zonal distribution and the western margin of Ji-D6 well was irregular distribution. The first class reservoir also distributed in the west of Pear town depression. At the same time, the peripheral area of Ji D 2 well, Ji 1 well, Ji 3 well in the central of Pear town depression was massive distribution. The second class reservoir, the main block was developed in the central of Jidong depression. The third class reservoir accounted for the biggest section in Jixi basin. While the forth class reservoir were sporadic small block distribution.

关 键 词：自组织特征映射神经网络 储层评价 鸡西盆地 

分 类 号：X83[环境科学与工程—环境工程]