自组织多项式网络算法在油气产能定量预测中应用  

THE USE OF SELF ORGANIZING POLYNOMIAL NETWORK ALGORITHM IN RESERVOIR QUANTITATIVE PREDICTION OF RESERVOIR YIELD

作　　者：胡文艳[1] 汪徐焱[2] 舒雅琴[2] 曾锦光[2] 

机构地区：[1]四川石油管理局地质勘探研究院 [2]成都理工学院应用数学系

出　　处：《成都理工大学学报（自然科学版）》1998年第S1期59-64,共6页Journal of Chengdu University of Technology: Science & Technology Edition

摘　　要：油气定量预测中复杂函数的拟合与逼近一直是人们关注的热点之一。文章利用神经网络方法结合自组织理论设计了用于非线性模型描述的自组织多项式网络算法，该算法特征多项式的完全描述性和自组织功能，使其在用于陕甘宁中部气田油气产能定量预测中取得了较为理想的效果。The fitting and approximation of complex function in the reservoir quantitative prediction is always one of the hotpoints which people pay attention to. This paper designs a self organizing polynomial network algorithm which can apply to the nonlinear model description by using the neural network approach connected with the self organizing method. With its complete description of characteristic polynomial and the function of self organizing, the algorithm has been applied to the reservoir yield quantitative prediction of the center field in the Shaanxi Gansu Ningxia Border Region and got better results.

关 键 词：油气产能 K-G多项式 GMDH算法 神经网络 定量预测 

分 类 号：P618.130[天文地球—矿床学] O233[天文地球—地质学]