广义回归神经网络技术在新油田快速评价中的应用  

Application of Generalized Regression Neural Network in Rapid Evaluation of New Oilfields

作　　者：陈晖 胡泽根[2] 王永平 王欣然 郭秩瑛 CHEN Hui;HU Zegen;WANG Yongping;WANG Xinran;GUO Zhiying(Bohai Petroleum Research Institute,Tianjin Branch of CNOOC Limited,Tianjin 300459,China;China Oilfield Services Limited,Tianjin 300459,China)

机构地区：[1]中海油天津分公司渤海石油研究院,天津300459 [2]中海油田服务股份有限公司,天津300459

出　　处：《重庆科技学院学报（自然科学版）》2019年第1期42-45,共4页Journal of Chongqing University of Science and Technology：Natural Sciences Edition

基　　金：国家科技重大专项"渤海油田加密调整及提高采收率油藏工程技术示范"(2016ZX05058001)

摘　　要：新油田的动态和静态资料较少,油藏认识程度低,开发指标预测难度大。提出运用广义回归神经网络技术进行非线性优化回归,通过对比相似油田开发模式和开发效果,快速评价及预测新油田的开发方式和开采指标。以处于前期研究阶段的某油田为例,运用此方法进行指标预测,计算结果符合实际情况。New oilfields have few dynamic and static data,so it is difficult to predict the development index due to the low level of reservoir recognition.Nonlinear optimal regression is carried out in this paper based on generalized regression neural network technology to fast evaluate and predict the new oilfield′s development program and production index data,by comparing the development modes and effects of similar oilfields.Take an oilfield in the early stage of development as an example,the prediction results of the method fit the actual situation.

关 键 词：新油田 开发指标预测 快速评价 广义回归神经网络 

分 类 号：TE347[石油与天然气工程—油气田开发工程]