# US 11,852,778 B2 · Static engine and neural network for a cognitive reservoir system

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

- Number: US 11,852,778 B2
- Document ID: US11852778B2
- Status: Granted
- Kind: B2
- Application number: 17/497,477
- Filed: 2021-10-08
- Published: 2022-01-27
- Granted: 2023-12-26
- Earliest claimed: 2017-10-11
- Assignees: Beyond Limits, Inc.
- Applicants: Beyond Limits, Inc.
- Inventors: Azarang Golmohammadizangabad; Shahram Farhadi Nia
- Web page: https://patentgrep.com/patent/US11852778B2

## 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

Implementations described and claimed herein provide systems and methods for developing a reservoir. In one implementation, observed data points in a volume along a well trajectory and well logs corresponding to observed data points are received at a neural network. Feature vectors are generated using the neural network. The feature vectors are defined based on a distance between each of the observed data points and randomly generated points in the volume. A 3D populated log is generated by propagating well log values of the feature vectors across the volume. Uncertainty is quantified by generating a plurality of realizations including the 3D populated log. Each of the realizations is different and equally probable. core values are generated from the realizations, and a static model of the reservoir is generated by clustering the volume into one or more clusters of rock types based on the core values.

## Claims

### Claim 1

A system for developing a reservoir, the system comprising:
one or more processors; and
at least one non-transitory computer readable medium having stored therein instructions executed by the one or more processors to:
generate, by a neural network, a set of feature vectors defined based on a distance between each observed data point of a set of observed data points in a volume along a well trajectory and a set of randomly generated points in the volume, the set of feature vectors corresponding to well log values captured using one or more measuring tools, the neural network generating a three-dimensional populated log by propagating the well log values across the volume, the neural network quantifying uncertainty by generating a plurality of realizations including the three-dimensional populated log;
generating, by a static modeler in communication with the neural network, a static model of the reservoir by clustering the volume into one or more clusters of rock types based on core values generated from the plurality of realizations;
generating, by a dynamic modeler in communication with the static modeler, a dynamic model of the reservoir using the static model and dynamic data by constructing a reservoir graph representing the clusters as graph vertices of the static model; and
generating, by a reasoner in communication with the dynamic modeler, a ranking of sub-volumes of the reservoir for drilling based on the dynamic model.

### Claim 2 (depends on 1)

The system according to claim 1, wherein the dynamic data includes pressure data or fluid saturation data.

### Claim 3 (depends on 2)

The system according to claim 2, wherein vertex properties of the reservoir graph includes at least one of location, porosity, pore volume, permeability, or initial oil saturation, and wherein each vertex is defined to contain a spatially continuous voxel set and the dynamic modeler defines graph connectivity through nodal connectivity of neighboring clusters.

### Claim 4 (depends on 3)

The system according to claim 3, wherein the dynamic modeler determines pressure data based on the nodal connectivity of neighboring clusters.

### Claim 5 (depends on 4)

The system according to claim 4, wherein the dynamic modeler is operable to determine fluid saturation data across the reservoir based on 4D seismic data.

### Claim 6 (depends on 1)

The system according to claim 1, wherein the reasoner is further operable to generate supporting information in relative and absolute terms detailing how the ranking was produced.

### Claim 7 (depends on 1)

The system according to claim 1, wherein each of the plurality of realizations is different and equally probable.

### Claim 8 (depends on 7)

The system according to claim 7, wherein the set of randomly generated points is changed at the static modeler such that each of the plurality of realizations is different and equally probable.

### Claim 9 (depends on 1)

The system according to claim 1, wherein the core values include at least one of porosity, permeability, or initial saturation.

### Claim 10

A method for developing a reservoir, the method comprising:
receiving static data including a set of observed data points in a volume along a well trajectory and well logs corresponding to the set of observed data points;
generating a set of feature vectors, the set of feature vectors defined based on a distance between each of the observed data points and a set of randomly generated points in the volume;
generating a three-dimensional populated log by propagating well log values of the set of feature vectors across the volume;
quantifying uncertainty by generating a plurality of realizations including the three-dimensional populated log;
generating core values from the plurality of realizations;
generating a static model of a reservoir by clustering the volume into one or more clusters of rock types based on the core values;
generating a dynamic model of the reservoir using the static model of the reservoir and dynamic data by constructing a reservoir graph representing the clusters as graph vertices of the static model; and
generating a ranking of sub-volumes of the reservoir for drilling based on the dynamic model.

### Claim 11 (depends on 10)

The method according to claim 10, wherein the dynamic data includes pressure data or fluid saturation data.

### Claim 12 (depends on 10)

The method according to claim 10, wherein vertex properties of a reservoir graph includes at least one of location, porosity, pore volume, permeability, or initial oil saturation, and wherein each vertex is defined to contain a spatially continuous voxel set and further comprising defining graph connectivity through nodal connectivity of neighboring clusters based on the static model.

### Claim 13 (depends on 12)

The method according to claim 12, further comprising determining pressure data based on the nodal connectivity of neighboring clusters.

### Claim 14 (depends on 13)

The method according to claim 13, further comprising determining fluid saturation data across the reservoir based on 4D seismic data.

### Claim 15 (depends on 14)

The method according to claim 14, further comprising generating supporting information in relative and absolute terms detailing how the ranking was produced.

### Claim 16 (depends on 10)

The method according to claim 10, wherein each of the plurality of realizations is different and equally probable.

### Claim 17 (depends on 10)

The method according to claim 10, wherein the core values include at least one of porosity, permeability, or initial saturation.

### Claim 18 (depends on 10)

The method according to claim 10, wherein the core values are generated using a k-nearest neighbor algorithm.

### Claim 19 (depends on 10)

The method according to claim 10, wherein the one or more clusters of the static model are obtained through k-means clustering.

### Claim 20 (depends on 19)

The method according to claim 19, wherein the static model includes one ore more fault planes fitted to the one or more clusters.
