人工神经网络模型在石油资源预测中的应用

英文篇名：Application of neural network model in petroleum resource predication
中文刊名：资源调查与环境
英文刊名：Volcanology & Mineral Resources
作者：朱庆杰 ; 谷延斌 ; 卢时林 ; 岳文珍
英文作者：ZHU Qing-jie1 ; GU Yan-bin2 ; LU Shi-lin2 ; YUE Wen-zhen2 ; (1 Earthquake Engineering Research Center of Hebei Province ; Tangshan 063000 ; China)(2 Daqing and Tangshan company of China National Petroleum Group ;  Daqing 163000 ; Tangshan 063000 ; China)
中文关键词：储层预测 ; 神经网络模型 ; 地震资料 ; 大凌河油层 ; 油气勘探
英文关键词：reservoir predication ; neural network model ; seismic data ; Dalinghe oil layer ; petroleum exploration
出版日期：2002-12-26
机构：河北省地震工程研究中心 , 中国石油集团大庆和唐山公司 , 中国石油集团大庆和唐山公司 , 中国石油集团大庆和唐山公司河北唐山066000 ,大庆163000,河北唐山063000 ,大庆163000,河北唐山063000 ,大庆163000,河北唐山063000
年：2002
期：04
出版单位：资源调查与环境

摘要

锦270井区大凌河油层是盆地大幅度沉降条件下的产物,以深陷湖环境为主,发育湖底扇浊积岩。准确预测大凌河油层的砂体分布和泥质含量,是勘探目标选择的关键。由于钻井资料少,不能反映储层横向变化。为了准确预测储层的横向变化,综合钻井地质资料和地震勘探资料,并为了反映地震资料的多个参数与储层横向变化间的非线性相关关系,采用人工神经网络模型进行预测,计算了大凌河油层砂层厚度和泥质含量的平面分布。依据计算结果,分析了有利油气区的分布,并给出了几点结论。
Dalinghe oil layer is the result of basin's large extent subsiding, with environments of deep rift lake, and turbidite of sublacustrine fan were deposited. Prediction of sand thickness distribution and shale content exactly is the key for exploration targets selection. For the lack of drilling data, it is difficult to predict the lateral variety of reservoir. In order to predict the lateral variety of reservoir exactly, both geological data and seismic data are used, and neural network model is used to analyze the nonlinear relations between the factors of seismic data and the lateral variety of reservoir. The distributions of sand thickness and shale content are calculated. According to the calculating results, the distributions of favorable area for petroleum exploration are analyzed, and some conclusions are given.

引文

[1]　朱庆杰,王波.轮南奥陶系裂缝研究的综合定量方法[J].石油地球物理勘探,1999,34(2):180-189
[2]　何明一.神经计算[M].西安:西安电子科技大学出版社,1992,35-46682