The value of streamlines based localization in the assimilation of 4D seismic data with the Ensemble Kalman filter
conference paper
The Ensemble Kalman Filter is a powerful data assimilation algorithm which allows a sequential update of the reservoir states from measurement data. However, results performed on a synthetic 3D model, show that too frequent 4D seismic data may cause an excessive contraction of the ensemble covariance and the filter divergence. A solution to this problem is the conditioning of the cross-covariance matrix between the ensemble variables and the predicted measurements, using information from streamlines trajectories; in this way the influence of unrelated, distant and spurious observations is eliminated. Example from a 2D synthetic case show that, even after the frequent assimilation of 4D seismic data, if localization is applied, the ensemble retains a reasonable spread and the divergence of the filter is avoided; furthermore, streamlines based localization of the EnKF applied on seismic data, with respect to the traditional EnKF, seems to provide a better quality in the estimation of the permeability field and for the reservoir production, both in the history and in the forecast period.
TNO Identifier
290661
ISBN
9781615672363
Source title
71st European Association of Geoscientists and Engineers Conference and Exhibition 2009, 8 June 2009 through 11 June 2009, CY - Amsterdam
Pages
1767-1771
Files
To receive the publication files, please send an e-mail request to TNO Repository.