自组织特征映射网络在储层识别中的应用

Application of the SOM Network on Identification of the Unconventional Reservoir

针对泌阳坳陷安棚油田下第三系致密砂岩储层低孔、低渗、高含水饱和度所造成储产层与非储产层在测井曲线上不易识别的特点，选用了自组织特征映射网络模型（ Ｓ Ｏ Ｍ） ，利用测井曲线参数，建立了安棚油田非常规储产层的预测识别模型．经验证，已知样本吻合率达８２ ．２ ％

The paper presents character of unconventional reservoir that the productive reservoirs and nonproductive formations are difficultly identificated in well logging. The unconventional reservoir is dense sandrock with lower pore, lower permeability and higher water saturation in Anpeng oil field, Biyang Depression. The identification and predication system of network model is set up with well logging datum. The coincidense rate of the known samples reaches 82.2 percent.