用遗传人工神经网络进行精细油藏描述  

Fine Reservoir Redescription with ANN Technique Based on Genetic Algorithms.

作　　者：赵彦超[1] 汪立君[1] 李丽[1] 王希明[2] 熊敏[2] 

机构地区：[1]中国地质大学石油系 [2]胜利石油有限公司临盘采油厂

出　　处：《测井技术》2001年第6期473-476,共4页Well Logging Technology

摘　　要：商河油田商二区沙二下油藏单元是一个经过多年开发的老的、断层控制的油藏单元。为了提高油藏的采收率 ,必须开展精细油藏描述 ,建立油藏地质模型。该油藏单元具有岩性细 ,泥质含量高 ,油层低电阻的特征。采用人工神经网络测井识别技术识别油层并进行储层物性及含油性的解释。在此基础上 ,对油层参数进行了解释并编绘了反映油层参数分布的平面图及油藏剖面图 ,建立了油藏模型。经生产动态资料检验 。Shanger reservoir unit is a fault controlled reservoir which has been developed for many years. In order to enhance oil recovery of the reservoir, fine reservoir description has to be made by building up new reservoir geologic model. However, the reservoir is characterized by high lithology, high clay content and low resistivity. Therefore, it is very difficult to distinguish oil with water layer and to evaluate oil layer by using logging data. For this reason, an efficient ANN (artificial neural network) technique based on genetic algorithms is used to identify oil layers and interprete reservoir property and oil bearing capability, based on which, the oil bed parameters have been calculated, the planar figures as well as reservoir profiles indicating oil bed parameters worked out, and further the detailed reservoir description model set up. The dynamic production data proved that the new established model is creditable.

关 键 词：精细油藏描述 油水层识别 遗传人工神经网络 地质模型 商河油田 

分 类 号：TE321[石油与天然气工程—油气田开发工程] TE319