Method for history matching and uncertainty quantification assisted by global optimization techniques utilizing proxies
AssigneeChevron U.S.A. Inc.
InventorsBurak Yeten, Alexandre Castellini
Life of this patent
Estimating the term…
Abstract
A method for forecasting production from a hydrocarbon producing reservoir, the method includes defining an objective function and characteristics of a history-matched model of a reservoir and acceptable error E. At least one geological realization of the reservoir is created representing a probable geological setting. For each geological realization, a global optimization technique is used to perform history matching in a series of iterative steps to obtain acceptable models. Production of the reservoir is forecasted based upon simulation runs of the respective models.
Claims20
A method for forecasting production from a hydrocarbon producing reservoir, the method, comprising:
- (a) defining an objective function and characteristics of a history-matched model of a reservoir and acceptable error E;
- (b) creating at least one geological realization of the reservoir representing a probable geological setting;
- (c) for each geological realization, use a global optimization technique to perform history matching in a series of iterative steps to obtain acceptable models;
- and
- (d) forecasting production of the reservoir based upon simulation runs of the respective models.
A method for forecasting production from a hydrocarbon producing reservoir, the method, comprising:
- (a) defining an objective function and characteristics of a history-matched model of a reservoir and acceptable error E;
- (b) creating at least one geological realization of the reservoir representing a probable geological setting;
- (c) for each geological realization, use a global optimization technique to perform history matching in a series of iterative steps to obtain acceptable models, the optimization technique employing:
- (i) creating an initial population of parent models having discrete sets of parameters;
- (ii) running simulations on, and calculating errors E for, the parent models;
- (iii) creating a plurality of proxy surfaces based on the discrete set of parameters and errors E associated with the parent models;
- (iv) utilizing the proxy surface to at least one of selecting parent models from minimums of the proxy surfaces and utilizing the proxy surface as filters for selecting offspring models prepared from the parent models;
- (v) selecting acceptable reservoir models from the offspring models;
- and
- (d) forecasting production of the reservoir based upon simulation runs of the respective models.
A method for forecasting production from a hydrocarbon producing reservoir, the method, comprising:
- (a) defining an objective function and characteristics of a history-matched model of a reservoir and acceptable error E;
- (b) creating at least one geological realization of the reservoir representing a probable geological setting;
- (c) for each geological realization, use a global optimization technique to perform history matching in a series of iterative steps to obtain acceptable models, the optimization technique employing:
- (i) creating an initial population of parent models having discrete sets of parameters;
- (ii) running simulations on, and calculating errors E for, the parent models;
- and
- (iii) determining whether the parent models have converged to a predetermined acceptability criteria, accepting the parent models as acceptable models when the parents have converged to meet the predetermined acceptability criteria and proceeding to step (d), when the parents have not converged to meet the predetermined acceptability criteria proceeding to step (iv);
- (iv) creating a proxy surface based upon the parameters and errors E associated with the parent models;
- (v) preparing a set of offspring models from the set of parent models;
- and
- (vi) determining whether the proxy surface passes a test for proxy goodness according to predetermined criteria;
- and
- (d) forecasting production of the reservoir based upon simulation runs of the respective models.
Dependencies are inferred from the claim text.