Estimation of Permeability Field via Ensemble Kalman Filter

History matching, as commonly referred to in petroleum industry, is an inverse parameter estimation problem which uses sparse observations of the states of the reservoir/aquifer (e.g. pressure) to tune spatially distributed parameters such as permeability and porosity. Even though significant progress is made in improving history matching algorithms, current history matching techniques are not suitable for real-time updating of the reservoir models using the frequent observations obtained form smart/intelligent fields and time-lapsed seismic. Recursive data assimilation algorithms, such as those based on the Kalman filter, can update the reservoir model (by tuning unknown parameters) without including previously used data, thereby saving computational cost. They are very convenient to apply with reservoir simulators, do not require adjoint models, and can easily incorporate 4D seismic attributes as areal measurements. Further, these algorithms readily provide an estimate of the uncertainty of the updated parameters and the states as a byproduct of the estimation process. This fits very well into the uncertainty and risk analysis aspects of the reservoir management. Our research focuses on the application of Ensemble Kalman Filter (EnKF) technique in updating reservoir models by integrating static/dynamic point and areal measurements such as frequent production and 4D seismic data. We present our preliminary results from the application of the EnKF in updating reservoir models with and without tuning heterogeneous permeability fields. A discussion of the issues and challenges encountered in our preliminary work will be provided.