SUMMARY In this computational study we combine 4D seismic data with production data to continuously update the reservoir model during production. The inversion is based on the ensemble Kalman filter (EnKF ), and the forward method is a combination of a fast seismic modeling tool and a reservoir simulator. The EnKF method is a Monte Carlo approach, and state variables, as fluid saturations and pres sure, and model input parameters as the porosity and permeability, are updated in the reservoir model at each assimilation step. The assimilat ed measurements are seismic waveforms data and production data, which are measured gas-oil ratio, water cut and flowing bottomhole press ure. The updated models allow for improved estimation of the parameters and the state variables during the production, because additional well data and repeated seismic surveys are sequentially included in the inversion. The method is applied to a synthetic 2D reservoir model, and it is shown that introduction of production and seismic waveforms data gives a fairly good estimation of the porosity and permeability fields.
[1]
Tor Arne Johansen,et al.
Incorporating 4D Seismic Data in Reservoir Simulation Models Using Ensemble Kalman Filter
,
2005
.
[2]
G. Evensen,et al.
An ensemble Kalman smoother for nonlinear dynamics
,
2000
.
[3]
Håvar Gjøystdal,et al.
Simulated Prestack Local Imaging: a robust and efficient interpretation tool to control illumination, resolution, and time-lapse properties of reservoirs.
,
2003
.
[4]
Albert Tarantola,et al.
Inverse problem theory - and methods for model parameter estimation
,
2004
.
[5]
S. Aanonsen,et al.
History Matching Using Time-lapse Seismic (HUTS)
,
2003
.
[6]
Geir Evensen,et al.
The Ensemble Kalman Filter: theoretical formulation and practical implementation
,
2003
.
[7]
A. Buland,et al.
Bayesian linearized AVO inversion
,
2003
.