联合神经网络在储层参数预测中的研究与应用论文.pdfVIP

Vol No W EL L L O G GIN G T EC H N O L O GY Apr BP RBF BP RBF P T P T E A Doi issn j Research on Committee Neural Network Model for Reservoir Physical Parameter Prediction DU A N Youxiang LI Gentian Colleg e of Co m put er and Co m m unicatio n Engineering China U niver sity of Pet roleum East China Qing dao Shando ng China Abstract Geological reservoir phy sical parameters are crucial f or building the three dimensional geological model and reservoir phy sical parameters are of ten obtained f rom logging data In conventional logging interpretation reservoir phy sical parameters are calculated by empirical f ormula or simplified geological conditions T he development of new technology has brought a new w ay f or the prediction of reservoir phy sical parameters T his paper present s committee neural netw ork CN N a new neural netw ork model w hich is based on BP neural netw ork RBF neural netw ork support vector reg ression and single lay er perception T his model could adjust netw ork structure automatically and get the optimal predicted value w hich avoids the def ect s of individual neural netw ork in parameters prediction and improves the accuracy of the prediction T he model is used and tested in three w ells logging in the same a