Method, system and apparatus for real-time reservoir model updating using ensemble kalman filter

A program storage device readable by a machine, tangibly embodying a program of executable instructions when executed by the machine perform method steps for updating an ensemble of reservoir models, conducting a reservoir simulation, said method steps comprising:

• (a) generating an ensemble of reservoir models, the ensemble being represented, at a first time step, by an initial state vector of initial static and dynamic state variables;
• (b) from the first time step to a second time step, forecasting forward the initital state vector to compute a forecasted state vector which includes forecast static and dynamic state variables and forecast production data;
• (c) computing a gain based upon the forecasted state vector and measured production data;
• (d) computing an updated state vector including updated static variables utilizing the computed gain of step (c);
• (e) creating a conforming state vector including the updated static state variables of step (d) and the initial dynamic state variables of step (a);
• and
• (f) from the first time step to the second time step, forecasting forward the conforming state vector to compute a forecasted conforming state vector which includes new forecast static and dynamic state variables that are in conformance with one another;
• whereby the ensemble of reservoir models is updated with the forecasted conforming state vector which includes the new forecast static and dynamic state variables that are in conformance with one another.

A comnputer-implemented method for updating an ensemble of reservoir models to reflect the most entreat measured production data and to accurately predict reservoir performance, the method includes the steps of generating an ensemble of reservoir models such that the ensemble it represented at a first time step by an initial state vector which includes initial state and dynamic state variables, forecasting from the first time step to a second time step the initial state vector to compute a forecasted state vector which includes forecast static and dynamic state variables and forecast production data, computing a Kalman gain based upon measured production data and the forecast production data, and utilizing the Kalman gain to compute an updated state vector including updated static state variables, wherein the improvement comprises additionally performing the steps of:

• (a) creating a conforming state vector including the updated static state variables and the initial dynamic state variables;
• and
• (b) forecasting from the first time to the second time step the conforming state vector to compute a forecasted conforming state vector which includes new forecast static and dynamic state variables that are in conformance with one another;
• whereby the ensemble ot reservoir models is updated on a computer with the forecasted conforming state vector winch includes the new forecast static and dynamic state variables that are in conformance with one another.