Methods and systems for machine—learning based simulation of flow

• US 10,087,721 B2
• Filed: 05/19/2011
• Issued: 10/02/2018
• Est. Priority Date: 07/29/2010
• Status: Active Grant

• Current Assignee: Young Living Essential Oils LC
• Original Assignee: Exxonmobil Upstream Research Company (Exxon Mobil Corporation)
• Inventors: Usadi, Adam, Li, Dachang, Parashkevov, Rossen, Terekhov, Sergey A., Wu, Xiaohui, Yang, Yahan
• Primary Examiner(s): Chad, Aniss
• Assistant Examiner(s): Guill, Russ
• Application Number:
  US13/805,647
• Publication Number:
  US 20130096898A1
• Time in Patent Office: 2,693 Days
• CPC Class Codes: E21B 43/00 Methods or apparatus for obtaining oil, gas, water, soluble or meltable materials or a slurry of minerals from wells applicable only to water E03B
• CPC Class Codes: G01V 1/282 Application of seismic models, synthetic seismograms
• CPC Class Codes: G01V 20/00 Geomodelling in general
• CPC Class Codes: G01V 2210/66 Subsurface modeling
• CPC Class Codes: G06F 2111/10 Numerical modelling
• CPC Class Codes: G06F 30/23 using finite element methods [FEM] or finite difference methods [FDM]
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/042 Knowledge-based neural networks; Logical representations of neural networks

Abstract

There is provided a method for modeling a hydrocarbon reservoir that includes generating a reservoir model that has a plurality of sub regions. A solution surrogate is obtained for a sub region by searching a database of existing solution surrogates to obtain an approximate solution surrogate based on a comparison of physical, geometrical, or numerical parameters of the sub region with physical, geometrical, or numerical parameters associated with the existing surrogate solutions in the database. If an approximate solution surrogate does not exist in the database, the sub region is simulated using a training simulation to obtain a set of training parameters comprising state variables and boundary conditions of the sub region. A machine learning algorithm is used to obtain a new solution surrogate based on the set of training parameters. The hydrocarbon reservoir can be simulated using the solution surrogate obtained for the at least one sub region.

Claims

1. A method for modeling a hydrocarbon reservoir, comprising:

generating a reservoir model comprising a plurality of sub regions, wherein the reservoir model is segmented into the plurality of sub regions that represent different portions of the reservoir model and the reservoir model is utilized to model fluid flow within the hydrocarbon reservoir;

obtaining a solution surrogate for at least one sub region of the plurality of sub regions by;

searching a database of existing solution surrogates to determine if an approximate solution surrogate exists based on a comparison of physical, geometrical, or numerical parameters of the at least one sub region with physical, geometrical, or numerical parameters associated with the existing solution surrogates in the database;

if the approximate solution surrogate exists, then associating one of the existing solution surrogates with the at least one sub region as the solution surrogate;

if the approximate solution surrogate does not exist in the database;

simulating the sub region using a training simulation of the sub region to obtain a set of training parameters comprising state variables and boundary conditions of the sub region, wherein the training simulation is a reservoir simulation having resolution to represent predetermined physical processes wherein the simulating the sub region comprises generating a set of inputs, a set of desired outputs, and weighting values, w;

using a machine learning algorithm to obtain a new solution surrogate based on the set of training parameters wherein the machine learning algorithm calculates a set of computed outputs, compares the set of computed outputs to the set of desired outputs to calculate a value for an objective function between the set of computed outputs and the set of desired outputs, and uses the value of the objective function to alter the weighting values, w, in at least one subsequent iteration of calculating a value for the new solution surrogate;

storing the new solution surrogate to the database; and

associating the new solution surrogate with the at least one sub region as the solution surrogate; and

simulating fluid flow in the hydrocarbon reservoir using the reservoir model comprising the plurality of sub regions and the solution surrogate obtained for the at least one sub region;

generating a data representation of a physical hydrocarbon reservoir in a non-transitory, computer-readable, medium based at least in part on the results of the simulation; and

causing hydrocarbon to be produced from the hydrocarbon reservoir based, at least in part, upon the results of the simulation.

2. The method of claim 1, wherein using the machine learning algorithm to obtain the new solution surrogate comprises approximating an inverse operator of a matrix equation that provides a solution to the fluid flow through a porous media.

3. The method of claim 1, wherein the machine learning algorithm is used to obtain a plurality of effective phase permeability values for a coarse grid, wherein the plurality of effective phase permeability values is used to calculate fluid flow through a porous media.

4. The method of claim 1, wherein using the machine learning algorithm to obtain the new solution surrogate comprises training a neural net using the training parameters, wherein the boundary conditions are used as input to the neural net and the state variables are used as a desired output.

5. The method of claim 1, wherein simulating the sub region using the training simulation of the sub region comprises specifying a set of boundary condition types for each boundary of the sub region and generating a fine-gridded computational mesh of the sub region.

6. The method of claim 1, wherein storing the new solution surrogate to the database comprises associating the solution surrogate with a set of physical, geometrical, or numerical parameters used to generate the solution surrogate and storing the parameters to the database.

7. The method of claim 6, wherein the new solution surrogate is re-used in subsequent hydrocarbon reservoir simulations based on a comparison of the set of physical, geometrical, or numerical parameters stored to the database with the solution surrogate and a new set of physical, geometrical, or numerical parameters that characterize a new sub region.

8. The method of claim 7, wherein the physical parameters comprise distribution of permeability, porosity, grid geometry of the training simulation of the sub region, or some combination thereof.

9. A method for producing a hydrocarbon from a hydrocarbon reservoir, comprising:

generating a reservoir model comprising a plurality of sub regions, wherein the reservoir model is segmented into the plurality of sub regions that represent different portions of the reservoir model and the reservoir model is utilized to model fluid flow within the hydrocarbon reservoir;

obtaining a solution surrogate for at least one sub region of the plurality of sub regions by;

searching a database of existing solution surrogates to determine if an approximate solution surrogate exists based on a comparison of physical, geometrical, or numerical parameters of the at least one sub region with physical, geometrical, or numerical parameters associated with the existing solution surrogates in the database;

if the approximate solution surrogate exists, then associating one of the existing solution surrogates with the at least one sub region as the solution surrogate; and

if the approximate solution surrogate does not exist in the database, simulating the sub region using a training simulation of the sub region to obtain a set of training parameters comprising state variables and boundary conditions of the sub region, wherein the simulating the sub region comprises generating a set of inputs, a set of desired outputs, and weighting values, w, using a machine learning algorithm to obtain a new solution surrogate based on the set of training parameters, wherein the machine learning algorithm calculates a set of computed outputs, compares the set of computed outputs to the set of desired outputs to calculate a value for an objective function between the set of computed outputs and the set of desired outputs, and uses the value of the objective function to alter the weighting values, w, in at least one subsequent iteration of calculating a value for the new solution surrogate, storing the new solution surrogate to the database, and associating the new solution surrogate with the at least one sub region as the solution surrogate, wherein the training simulation is a reservoir simulation having resolution to represent predetermined physical processes;

simulating fluid flow in the hydrocarbon reservoir using the reservoir model comprising the plurality of sub regions and the solution surrogate obtained for the at least one sub region; and

producing a hydrocarbon from the hydrocarbon reservoir based, at least in part, upon the results of the simulation.

10. The method of claim 9, wherein producing the hydrocarbon comprises:

drilling one or more wells to the hydrocarbon reservoir, wherein the wells comprise production wells, injection wells, or both;

setting production rates from the hydrocarbon reservoir;

or

any combinations thereof.