Method and system for performing reservoir simulations using look-ahead models

1 . A method, comprising:

obtaining grid model data regarding a geological region of interest;

obtaining well data regarding one or more wells in the geological region of interest;

obtaining a grid model for the geological region of interest based on the grid model data and the well data;

obtaining a time selection for a first look-ahead simulation among a first plurality of look-ahead simulations;

determining, by a reservoir simulator comprising a central processing unit, a first look-ahead model for the first look-ahead simulation based on the grid model data, the well data, and a first coarsening function,

wherein the first look-ahead model simulates the geological region of interest at a faster rate than the grid model;

performing, by the reservoir simulator, the first look-ahead simulation using the first look-ahead model, a first parallel processor among a plurality of parallel processors, and the time selection;

determining, by the reservoir simulator, look-ahead data based on the first look-ahead simulation;

determining, by the reservoir simulator, a composite reservoir parameter based on the well data and the look-ahead data,

wherein the composite reservoir parameter comprises a weighted value based on a number of the first plurality of look-ahead simulations; and

performing, by the reservoir simulator, a first reservoir simulation of the geological region of interest using the grid model, a second parallel processor among the plurality of parallel processors, and the first look-ahead simulation,

wherein the first reservoir simulation and the first look-ahead simulation are performed concurrently using the first parallel processor and the second parallel processor.

2 . The method of claim 1 , further comprising:

determining, at a first time step in an iterative process, a first well potential for a plurality of wells in the geological region of interest;

determining a second well potential for the plurality of wells using the first look-ahead simulation,

wherein the first well potential describes a simulated production rate for the plurality of wells for the first time step, and

wherein the second well potential describes a simulated production rate for the plurality of wells over the time selection; and

determining a composite well potential based on the first well potential and the second well potential,

wherein the first reservoir simulation is performed using the composite well potential.

3 . The method of claim 1 , further comprising:

determining a plurality of look-ahead models for a plurality of wells in the geological region of interest,

wherein a respective look-ahead model among the plurality of look-ahead models corresponds to a respective well among the plurality of wells,

wherein the respective look-ahead model comprises an area of interest around the respective well, and

wherein the respective look-ahead model comprises a first subset of cells outside the area of interest with greater coarsening than a second subset of cells within the area of interest; and

performing a second reservoir simulation based on a second plurality of look-ahead simulations using the plurality of look-ahead models.

4 . The method of claim 1 , further comprising:

simulating the geological region of interest using a plurality of reservoir simulations comprising the first reservoir simulation,

wherein simulating the geological region of interest corresponds to a predetermined time period, and

wherein the time selection is a subset of the predetermined time period that is less than the predetermined time period.

5 . The method of claim 1 ,

wherein the time selection comprises a target time,

wherein the target time corresponds to a predetermined time step in an iterative process where the first look-ahead simulation ends, and

wherein the first reservoir simulation continues in the iterative process past the predetermined time step.

6 . The method of claim 1 ,

wherein the first look-ahead simulation is performed before beginning the first reservoir simulation.

7 . The method of claim 1 , further comprising:

determining a second look-ahead model for a second look-ahead simulation based on the grid model data, the well data, and a second coarsening function; and

performing the second look-ahead simulation in parallel with the first look-ahead simulation using the second look-ahead model and the time selection,

wherein the first coarsening function and second coarsening functions are different functions.

8 . The method of claim 1 ,

wherein the first coarsening function is selected from a group consisting of a fluid characterization coarsening, a streamline coarsening, or a grid coarsening.

9 . The method of claim 1 ,

wherein the first look-ahead model is a proxy model comprising a plurality of well constraints that are the same as the grid model, and

wherein the proxy model has a first plurality of grid cells that are less than a second plurality of grid cells in the grid model.

10 . A system, comprising:

a network comprising a plurality of parallel processors, wherein the plurality of parallel processors comprise a first parallel processor and a second parallel processor, and

a reservoir simulator comprising a central processing unit (CPU), wherein the reservoir simulator is coupled to the network, the reservoir simulator being configured to perform a method comprising:

obtaining grid model data regarding a geological region of interest;

obtaining well data regarding one or more wells in the geological region of interest;

obtaining a grid model for the geological region of interest based on the grid model data and the well data;

obtaining a time selection for a look-ahead simulation among a first plurality of look-ahead simulations;

determining a look-ahead model for the look-ahead simulation based on the grid model data, the well data, and a first coarsening function,

wherein the look-ahead model simulates the geological region of interest at a faster rate than the grid model;

performing the look-ahead simulation using the look-ahead model, the first parallel processor, and the time selection;

determining, by the reservoir simulator, look-ahead data based on the look-ahead simulation;

determining, by the reservoir simulator, a composite reservoir parameter based on the well data and the look-ahead data,

wherein the composite reservoir parameter comprises a weighted value based on a number of the first plurality of look-ahead simulations; and

performing a first reservoir simulation of the geological region of interest using the grid model, the second parallel processor, and the look-ahead simulation,

wherein the first reservoir simulation and the look-ahead simulation are performed concurrently using the first parallel processor and the second parallel processor.

11 . The system of claim 10 , wherein the method further comprises:

determining, at a time step in an iterative process, a first well potential for a plurality of wells in the geological region of interest;

determining a second well potential for the plurality of wells using the look-ahead simulation,

wherein the first well potential describes a simulated production rate for the plurality of wells for the time step, and

wherein the second well potential describes a simulated production rate for the plurality of wells over the time selection; and

determining a composite well potential based on the first well potential and the second well potential,

wherein the first reservoir simulation is performed using the composite well potential.

12 . The system of claim 10 , wherein the method further comprises:

determining a plurality of look-ahead models for a plurality of wells in the geological region of interest,

wherein a respective look-ahead model among the plurality of look-ahead models corresponds to a respective well among the plurality of wells,

wherein the respective look-ahead model comprises an area of interest around the respective well, and

wherein the respective look-ahead model comprises a first subset of cells outside the area of interest with greater coarsening than a second subset of cells within the area of interest; and

performing a second reservoir simulation based on a second plurality of look-ahead simulations using the plurality of look-ahead models.

13 . The system of claim 10 ,

wherein the time selection comprises a target time,

wherein the target time corresponds to a predetermined time step in an iterative process where the look-ahead simulation ends, and

wherein the first reservoir simulation continues in the iterative process past the predetermined time step.

14 . The system of claim 10 ,

wherein the network is a parallel cluster comprising a plurality of graphical processing units (GPUs).

15 . A non-transitory computer readable medium storing instructions executable by a computer processor, the instructions being configured to perform a method comprising:

obtaining grid model data regarding a geological region of interest;

obtaining well data regarding one or more wells in the geological region of interest;

obtaining a grid model for the geological region of interest based on the grid model data and the well data;

obtaining a time selection for a look-ahead simulation among a plurality of look-ahead simulations;

determining a look-ahead model for the look-ahead simulation based on the grid model data, the well data, and a coarsening function,

wherein the look-ahead model simulates the geological region of interest at a faster rate than the grid model;

performing the look-ahead simulation using the look-ahead model, a first parallel processor among a plurality of parallel processors, and the time selection;

determining look-ahead data based on the first look-ahead simulation;

determining a composite reservoir parameter based on the well data and the look-ahead data,

wherein the composite reservoir parameter comprises a weighted value based on a number of the first plurality of look-ahead simulations; and

performing a first reservoir simulation of the geological region of interest using the grid model, a second parallel processor among the plurality of parallel processors, and the look-ahead simulation,

wherein the first reservoir simulation and the first look-ahead simulation are performed concurrently using the first parallel processor and the second parallel processor.