Real-time reservoir model updating in thermal recovery: Application of analytical proxies and Kalman filtering

Kalman filtering has many applications in engineering and technology development and has been widely employed in applied geosciences. As an estimator, it is used to predict system behavior for forecasting analyses. It considers observations obtained over time with uncertainties and other inaccuracies to simulate system behavior. This study customizes a physics based analytical proxy with the extended Kalman filter for nonlinear systems to characterize reservoir rock properties including porosity and permeability under the thermal recovery process. The analytical proxy represents a reservoir simulator for thermally enhanced oil recovery, and accounts for flow and geomechanical processes for unconventional heavy oil reservoirs. This paper also shows how inverse theories can be utilized to rank and update reservoir models.