基于EnKF的油藏自动历史拟合方法对三相相对渗透率的反演  

Inversion of three-phase relative permeability using the automatic reservoir history fitting method based on ensemble Kalman filter

作　　者：王玉斗[1] 李茂辉[1] 

机构地区：[1]中国石油大学(华东)物理科学与技术学院,山东东营257061

出　　处：《西安石油大学学报（自然科学版）》2011年第4期33-37,7-8,共5页Journal of Xi’an Shiyou University（Natural Science Edition）

基　　金：山东省自然科学基金项目"河流相油藏历史拟合的迭代集合卡尔曼滤波与核方法耦合研究"(编号:ZR2010EM053);中央高校基本科研业务费专项资金项目"基于集合Kalman滤波的河流相油藏自动历史拟合及反演方法研究"(编号:10CX04042A)

摘　　要：为解决标准集合Kalman滤波在非线性油藏反演中更新模型与更新动态场的统计不一致性问题,提出了一种改进的集合Kalman滤波方法,历史拟合过程中只对模型参数进行更新,然后利用更新的模型参数从初始状态重新运行油藏数值模拟软件进行下一拟合时刻的生产预测,协方差矩阵由从时间零运行油藏数值模拟软件得到的预测数据进行计算.理论证明了该方法与集合随机最大似然法的一致性.对某油藏的相对渗透率曲线进行了估计,比较了标准集合Kalman滤波方法与改进方法的参数估计及生产预测结果.由于保证了更新模型与更新动态场的一致性,改进的集合Kalman滤波方法得到了更好的参数估计和生产预测结果.An improved ensemble Kalman filter is presented for solving the inconsistency between the updated model and the updated dynamical variables in the nonlinear reservoir inversion using standard ensemble Kalman filter.Only are the model parameters updated in this modified method in the process of reservoir history fitting,and the reservoir production at next moment can be predicted by rerunning the simulator from initial time using the updated model parameters.The covariance is calculated from the prediction data obtained by running simulator from initial time.The equivalence of the improved ensemble Kalman filter to ensemble random maximum likelihood method is theoretically proved.The relative permeability curves of a reservoir are estimated using the modified ensemble Kalman filter,and compared with those obtained using standard ensemble Kalman filter.It is shown that the estimated reservoir parameters and the predicted reservoir production data by using the modified ensemble Kalman filter are better because of the inconsistency of the updated model with the updated dynamical variables.

关 键 词：储层历史拟合 集合KALMAN滤波 储层参数估计 储层生产预测 

分 类 号：TE353[石油与天然气工程—油气田开发工程]