# US 12,055,673 B2 · System for improved reservoir exploration and production

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,055,673 B2
- Document ID: US12055673B2
- Status: Granted
- Kind: B2
- Application number: 17/880,298
- Filed: 2022-08-03
- Published: 2022-12-01
- Granted: 2024-08-06
- Earliest claimed: 2017-10-11
- Assignees: Beyond Limits, Inc.
- Applicants: Beyond Limits, Inc.
- Inventors: Shahram Farhadi Nia; Zackary H Nolan; Azarang Golmohammadizangabad
- Web page: https://patentgrep.com/patent/US12055673B2

## CPC

- G01V 20/00: Geomodelling in general
- G06F 30/20: Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30)
- G06N 3/043: based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
- G06N 3/0499: Feedforward networks
- G06N 3/088: Non-supervised learning, e.g. competitive learning
- G06N 3/09: Supervised learning
- G01V: GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS (means for indicating the location of accidentally buried, e.g. snow-buried, persons A63B29/02)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An architecture for predicting and modeling geological characteristics of a reservoir includes one or more neural networks, a static modeling module, a dynamic modeling module, and a fuzzy inference engine to provide recommendations for drilling a wellbore. The neural networks receive log data for coordinates along a well trajectory, and determine a geophysical relationship for a property of a subterranean formation as a function of distance vectors between the coordinates along the well trajectory and one or more sets of randomly generated coordinates. The static modeling module generates three-dimensional static models of a volume of interest based on predicted properties of formations residing therein from the neural networks. The dynamic modeling module determines connectivity values between clusters of formations based on nodal connectivity of neighboring clusters, assigns pressure values across the volume of interest, and generates a three-dimensional dynamic model for the volume of interest based on the pressure values.

## Claims

### Claim 1

A method for modeling geological characteristics of a volume, the method comprising:
generating, using a static modeling module that leverages one or more trained neural networks, a three-dimensional static model for a volume of interest based on geophysical characteristics for one or more sets of randomly generated coordinates of a volume of interest surrounding a well trajectory, wherein the geophysical characteristics for the one or more sets of randomly generated coordinates are a function of distance vectors between the one or more sets of randomly generated coordinates and measured geophysical characteristics measured at known coordinates along the well trajectory;
assigning portions of the volume of interest to one or more clusters based on the geophysical characteristics;
determining connectivity values between the one or more clusters based on a nodal connectivity of neighboring clusters;
assigning saturation values across the volume of interest of the three-dimensional static model based on the connectivity values; and
generating a three-dimensional dynamic model for the volume of interest based on the three-dimensional static model and the saturation values.

### Claim 2 (depends on 1)

The method of claim 1, further comprising predicting the geophysical characteristics for the one or more sets of randomly generated coordinates based on one or more neural networks.

### Claim 3 (depends on 1)

The method of claim 1, further comprising adjusting the three-dimensional dynamic model based on predicted changes in the saturation values over time.

### Claim 4 (depends on 1)

The method of claim 1, further comprising:
evaluating the three-dimensional static model and the three-dimensional dynamic model based on a set of rules executed by a fuzzy inference engine; and
providing a recommendation for drilling a wellbore in the volume of interest.

### Claim 5 (depends on 1)

The method of claim 1, wherein the one or more trained neural networks is trained to identify a geophysical relationship for a property of a subterranean formation as a function of distance vectors using the distance vectors as an input and log data for coordinates along the well trajectory as an output.

### Claim 6 (depends on 1)

The method of claim 1, further comprising:
generating a plurality of static model realizations using one or more neural networks based on a plurality of sets of randomly generated coordinates; and
quantifying an uncertainty value using the one or more neural networks based on the plurality of the static model realizations.

### Claim 7 (depends on 1)

The method of claim 1, further comprising:
assigning pressure values across the volume of interest of the three-dimensional static model based on the connectivity values; and
adjusting the three-dimensional dynamic model based on predicted changes in the pressure values over time.

### Claim 8 (depends on 1)

The method of claim 1, wherein the geophysical characteristics include at least one of a location, a porosity, a permeability, or an initial saturation.

### Claim 9 (depends on 1)

The method of claim 1, further comprising defining a fault plane in the volume of interest for at least one of the three-dimensional static model or the three-dimensional dynamic model, wherein the at least one fault plane acts as a node with zero or reduced permeability.

### Claim 10

A system for modeling geological characteristics of a volume, the system comprising:
one or more processors; and
one or more tangible, non-transitory computer-readable storage media that store instructions executable by the one or more processors to:
generate, by a static modeling module leveraging one or more trained neural networks, a three-dimensional static model for a volume of interest based on geophysical characteristics for one or more sets of randomly generated coordinates of the volume of interest surrounding a well trajectory, wherein the geophysical characteristics for the one or more sets of randomly generated coordinates are a function of distance vectors between the one or more sets of randomly generated coordinates and measured geophysical characteristics measured at known coordinates along the well trajectory;
assign, by a dynamic modeling module in communication with the static modeling module, portions of the volume of interest to one or more clusters based on the geophysical characteristics;
determine, by the dynamic modeling module, connectivity values between clusters based on a nodal connectivity of neighboring clusters;
assign, by the dynamic modeling module, saturation values across the volume of interest of the three-dimensional static model based on the connectivity values; and
generate, by the dynamic modeling module, a three-dimensional dynamic model for the volume of interest based on the three-dimensional static model and the saturation values.

### Claim 11 (depends on 10)

The system of claim 10, wherein the one or more trained neural networks predict the geophysical characteristics for the one or more sets of randomly generated coordinates.

### Claim 12 (depends on 10)

The system of claim 10, further comprising a fuzzy inference engine in communication with the static modeling module and the dynamic modeling module, the fuzzy inference engine is configured to:
evaluate the three-dimensional static model and the three-dimensional dynamic model based on a set of rules executed by the fuzzy inference engine; and
provide a recommendation for drilling a wellbore in the volume of interest based on a predetermined policy ruleset.

### Claim 13 (depends on 10)

The system of claim 10, wherein the one or more trained neural networks are trained to determine a geophysical relationship for a property of a subterranean formation as a function of distance vectors using the distance vectors as an input and log data for coordinates along the well trajectory as an output.

### Claim 14 (depends on 10)

The system of claim 10, wherein the one or more trained neural networks are further configured to:
generate a plurality of static model realizations based on a plurality of sets of randomly generated coordinates; and
quantify an uncertainty value based on the plurality the static model realizations.

### Claim 15 (depends on 10)

The system of claim 10, wherein the dynamic modeling module is further configured to:
assign pressure values across the volume of interest of the three-dimensional static model based on the connectivity values; and
adjust the three-dimensional dynamic model based on predicted changes in the pressure values over time.

### Claim 16 (depends on 10)

The system of claim 10, wherein the dynamic modeling module is further configured to adjust the three-dimensional dynamic model based on predicted changes in the saturation values over time.

### Claim 17 (depends on 10)

The system of claim 10, wherein the geophysical characteristics include at least one of a location, a porosity, a permeability, or an initial saturation.

### Claim 18

A tangible, non-transitory, computer-readable storage media having instructions encoded thereon, the instructions executable by a processor to perform a method for modeling geological characteristics of a volume, the method comprising:
generating, using a static modeling module that leverages one or more trained neural networks, a three-dimensional static model for a volume of interest based on geophysical characteristics for one or more sets of randomly generated coordinates of a volume of interest surrounding a well trajectory, wherein the geophysical characteristics for the one or more sets of randomly generated coordinates are a function of distance vectors between the one or more sets of randomly generated coordinates and measured geophysical characteristics measured at known coordinates along the well trajectory;
assigning portions of the volume of interest to one or more clusters based on the geophysical characteristics;
determining connectivity values between clusters based on a nodal connectivity of neighboring clusters;
assigning saturation values across the volume of interest of the three-dimensional static model based on the connectivity values; and
generating a three-dimensional dynamic model for the volume of interest based on the three-dimensional static model and the saturation values.

### Claim 19 (depends on 18)

The tangible, non-transitory, computer-readable storage media of claim 18, further comprising instructions executable the processor to adjust the three-dimensional dynamic model based on predicted changes in the saturation values over time.

### Claim 20 (depends on 18)

The tangible, non-transitory, computer-readable storage media of claim 18, further comprising instructions executable by the processor to:
evaluate the three-dimensional static model and the three-dimensional dynamic model based on a set of rules executed by a fuzzy inference engine; and
provide a recommendation for drilling a wellbore in the volume of interest.
