摘要
根据油气层综合解释的特点,提出了单井油气层综合解释的人工神经网络方法。采用的神经网络模型为包含28个输入神经元、26个隐含神经元和4个输出神经元的三层网络。选择了松辽盆地北部大庆长垣以西地区91个试油层段的资料对网络进行了训练,通过对已知样品的交叉验证,训练后的神经网络识别符合率为93.4％,证明人工神经网络是一种有效的油气层综合解释方法。
Based upon the characteristics of integrative interpretation of pay beds,a new method using ANNs for integrative pay beds interpretation is presented. The network consists of three layers,and the input, hidden and output layers have 28,26 and 4 neurons respectively. To train the network, 91 pay beds are selected from the western area to Daqing oil fields of Songliao basin. By cross validation of 91 samples,the correct ratio of the new method to interprete pay beds is 93.4% ,the results confirm that new method can be used for practical interpretation.
出处
《录井工程》
1995年第2期1-5,共5页
Mud Logging Engineering
关键词
油气层
综合解释
神经网络
模式识别
模型
应用
Pay bed Integrative interpretation Artificial neural network Pattern recogtion Model Application
