Full text loading...
-
Crossover Use of Ensemble Kalman Filter and Ensemble Smoother for Efficient History Matching
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 77th EAGE Conference and Exhibition 2015, Jun 2015, Volume 2015, p.1 - 5
Abstract
To make reasonable decisions, it is important to characterize reservoirs accurately using limited data available. A number of studies were introduced for a reservoir characterization by integrating static data and dynamic data together. Ensemble Kalman Filter (EnKF) and Ensemble Smoother (ES) are widely used methods for this task.
The difference between EnKF and ES is that ES computes one global update, rather than using recursive updates like EnKF. EnKF typically provides more accurate results but takes longer simulation time than ES. To offset each’s disadvantages, crossover use of EnKF and ES (CoEnKF) is suggested for efficient history matching. The proposed method uses EnKF first and then switches to ES to update all the remaining observation data together. To select proper time to switch from EnKF to ES, the authors suggest to use the change of estimation error covariance, which represents degree of update.
By applying the proposed method to synthetic reservoirs, it gives much shorter simulation time than EnKF and more reliable results than ES. Also, the proposed method characterizes low and high permeability distribution better than EnKF or ES. Therefore, the proposed method is useful for efficient history matching.