奇异值分解法在建立气水层识别模型中的应用  

Application of SVD Method to the Identifying Model of Establishing Gas-Water-Bearing Formation

作　　者：邸维振[1] 潘和平[1] 韩浩东[1] 卢波[1] 

机构地区：[1]中国地质大学地球物理与空间信息学院,武汉430074

出　　处：《工程地球物理学报》2009年第2期229-234,共6页Chinese Journal of Engineering Geophysics

摘　　要：由于气层、水层层段上物性的复杂性,导致测井曲线特征复杂,利用测井曲线直接识别油气层的效果不好,有可能造成误判或错判,为此,提取一些特征参数,建立识别气层、水层的判别函数,来提高识别的准确率。本文对气层、水层的测井响应进行分析,提取气层、水层的测井曲线特征参数,利用模糊模式识别渗透层,奇异值分解方法建立判别函数,最后利用判别函数来识别油(气)层、气水同层、水层和干层,实际解释结果表明,取得了符合率高的效果。As the complexity of the physical properties of the gas - bearing formation and water-bearing formation lead to the complicated characteristics of the logging curve, it is not good to directly identify oil - gas bearing formation using logging curve and it is likely to make wrong judgment or false judgment. Therefore, some characteristic parameters are extracted to increase the accuracy of identification. This paper analyzes the logging response on both of the gas - bearing formation and water - bearing formation. It also adopts fuzzy model to identify permeable formation and SVD to establish discriminate function. Fi- nally, the discriminate function is used to identify oil （gas）-bearing formation, gas- water - bearing formation, water - bearing formation and dry - bearing formation. The result of the actual interpreting is very good.

关 键 词：模糊模式识别 奇异值分解法 判别函数 油气识别 

分 类 号：P631.8[天文地球—地质矿产勘探]