低阻油气层识别方法研究

在油气勘探开发中,地球物理测井资料解释的最基本任务之一是在钻孔剖面上准确地识别油(气)层、油水同层、水层、干层等。 新疆塔北地区三叠系的特殊油气层电阻率低于或接近水层电阻率，在电性上直接区分特殊油气层与水层很不现实;另外，该区油气层的特征和油气层的影响因素明显与其它油田存在差异，故解决问题方法和思路不同于其它油田。文章利用油气层、油水同层、水层和干层的测井曲线和储层参数，建立识别储层流体属性的判别模型，采用灰关联分析聚类法、BP人工神经网络等模式识别法，对实际测井资料进行了解释，识别结果与试油结果对比表明BP人工神经网络、奇异值分解等方法识别结果与实际结果基本一致，没有漏掉油气层，取得了好的解释效果。

A Study of Identification Method of Low-resistivity Reservoir

In petroleum exploration and development, one of the most fundamental tasks of geophysical well logging data interpretation is to accurately identify oil (gas) layers, transitional oil-water layers, water layers and dry layers on the borehole sections. The resistivities of the specific Tertiary oil and gas layers in Tabei area, Xinjiang, are below or near that of the water layers. It is unrealistic to directly distinguish the specific oil and gas layers from water layers according to the electrical properties. In addition, the characteristics and influential factors of the oil and gas layers in the study area are different with that in the other oilfields. Therefore, different methods and ideas should be used to solve the specific issues in the study area. A discrimination model for identifying the properties of reservoir fluids is built by using the borehole log and reservoir parameters of oil (gas) layers, transitional oil-water layers, water layers, and dry layers. Pattern recognition methods, such as grey correlation cluster analysis and BP artificial neural network, are used to interpret the real well logging data. The results, achieved with BP artificial neural network and singular value decomposition, are basically in accord with the formation testing results, and no oil and gas layers are missed.