# US 2021/0278935 A1 · User interface for proxy modeling of reactive transport modeling

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

- Number: US 2021/0278935 A1
- Document ID: US20210278935A1
- Status: Published application
- Kind: A1
- Application number: 16/810,115
- Filed: 2020-03-05
- Published: 2021-09-09
- Earliest claimed: 2020-03-05
- Assignees: Saudi Arabian Oil Company
- Applicants: Saudi Arabian Oil Company
- Inventors: Yupeng Li; Peng Lu
- Web page: https://patentgrep.com/patent/US20210278935A1

## CPC

- G06F 3/0482: Interaction with lists of selectable items, e.g. menus
- G01V 20/00: Geomodelling in general
- G06F 30/27: using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/09: Supervised learning
- G06N 20/00: Machine learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G01V: GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS (means for indicating the location of accidentally buried, e.g. snow-buried, persons A63B29/02)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Systems and methods include a method for providing, for presentation to a user, a graphical user interface (GUI) for defining and generating machine learning-based proxy models as surrogates for process-based reactive transport modeling (RTM). User selections of training parameters for generating training sample data are received. The training sample data is generated in response to receiving a parameter files generation indication. A training cases generation indication is received. Training sample cases are executed using the training sample data. User selections of proxy models training parameters are received. A set of parameter-specific proxy models represented by a neural network are trained. Each parameter-specific proxy model corresponds to a specific RTM parameter from a set of RTM parameters. Blind tests are performed using the set of parameter-specific proxy models. Each blind test tests a specific one of the parameter-specific proxy models. Predictions are generated using the set of parameter-specific proxy models.

## Claims

### Claim 1

A computer-implemented method comprising:
providing, for presentation to a user, a graphical user interface (GUI) for defining and generating machine learning-based proxy models as surrogates for process-based reactive transport modeling (RTM);
receiving user selections of training parameters for generating training sample data, and generating the training sample data in response to receiving a parameter files generation indication from the user;
receiving a training cases generation indication from the user, and executing training sample cases using the training sample data;
receiving user selections of proxy models training parameters, and in response to receiving training selection indications, training a set of parameter-specific proxy models represented by a neural network, wherein each parameter-specific proxy model corresponds to a specific RTM parameter from a set of RTM parameters;
receiving user selections of blind tests to be run, and performing blind tests using the set of parameter-specific proxy models, wherein each blind test tests a specific one of the parameter-specific proxy models; and
receiving user selections for generating predictions, and generating predictions using the set of parameter-specific proxy models.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, wherein generating the training sample data includes generating parameter files based on an initial number of training cases and a number of target analysis factors.

### Claim 3 (depends on 1)

The computer-implemented method of claim 1, wherein the set of RTM parameters comprises one or more of dolomite, porosity, permeability, pH, magnesium ions (Mg2+), and calcium ions (Ca2+).

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, wherein training the set of parameter-specific proxy models includes defining training parameters for the training.

### Claim 5 (depends on 4)

The computer-implemented method of claim 4, wherein the training parameters include a training data percentage, a validate data percentage, a test data percentage, a neuron number in a first hidden layer, and a neuron number in a second hidden layer.

### Claim 6 (depends on 1)

The computer-implemented method of claim 1, further comprising receiving, using the GUI, user inputs for controlling steps for training the machine learning-based proxy models as surrogates for the process-based RTM.

### Claim 7 (depends on 1)

The computer-implemented method of claim 1, wherein at least one parameter-specific proxy model depends on an output of one or more previously-executed parameter-specific proxy models.

### Claim 8

A non-transitory, computer-readable medium storing one or more instructions executable by a computer system to perform operations comprising:
providing, for presentation to a user, a graphical user interface (GUI) for defining and generating machine learning-based proxy models as surrogates for process-based reactive transport modeling (RTM);
receiving user selections of training parameters for generating training sample data, and generating the training sample data in response to receiving a parameter files generation indication from the user;
receiving a training cases generation indication from the user, and executing training sample cases using the training sample data;
receiving user selections of proxy models training parameters, and in response to receiving training selection indications, training a set of parameter-specific proxy models represented by a neural network, wherein each parameter-specific proxy model corresponds to a specific RTM parameter from a set of RTM parameters;
receiving user selections of blind tests to be run, and performing blind tests using the set of parameter-specific proxy models, wherein each blind test tests a specific one of the parameter-specific proxy models; and
receiving user selections for generating predictions, and generating predictions using the set of parameter-specific proxy models.

### Claim 9 (depends on 8)

The non-transitory, computer-readable medium of claim 8, wherein generating the training sample data includes generating parameter files based on an initial number of training cases and a number of target analysis factors.

### Claim 10 (depends on 8)

The non-transitory, computer-readable medium of claim 8, wherein the set of RTM parameters comprises one or more of dolomite, porosity, permeability, pH, magnesium ions (Mg2+), and calcium ions (Ca2+).

### Claim 11 (depends on 8)

The non-transitory, computer-readable medium of claim 8, wherein training the set of parameter-specific proxy models includes defining training parameters for the training.

### Claim 12 (depends on 11)

The non-transitory, computer-readable medium of claim 11, wherein the training parameters include a training data percentage, a validate data percentage, a test data percentage, a neuron number in a first hidden layer, and a neuron number in a second hidden layer.

### Claim 13 (depends on 8)

The non-transitory, computer-readable medium of claim 8, the operations further comprising receiving, using the GUI, user inputs for controlling steps for training the machine learning-based proxy models as surrogates for the process-based RTM.

### Claim 14 (depends on 8)

The non-transitory, computer-readable medium of claim 8, wherein at least one parameter-specific proxy model depends on an output of one or more previously-executed parameter-specific proxy models.

### Claim 15

A computer-implemented system, comprising:
one or more processors; and
a non-transitory computer-readable storage medium coupled to the one or more processors and storing programming instructions for execution by the one or more processors, the programming instructions instructing the one or more processors to perform operations comprising:
providing, for presentation to a user, a graphical user interface (GUI) for defining and generating machine learning-based proxy models as surrogates for process-based reactive transport modeling (RTM);
receiving user selections of training parameters for generating training sample data, and generating the training sample data in response to receiving a parameter files generation indication from the user;
receiving a training cases generation indication from the user, and executing training sample cases using the training sample data;
receiving user selections of proxy models training parameters, and in response to receiving training selection indications, training a set of parameter-specific proxy models represented by a neural network, wherein each parameter-specific proxy model corresponds to a specific RTM parameter from a set of RTM parameters;
receiving user selections of blind tests to be run, and performing blind tests using the set of parameter-specific proxy models, wherein each blind test tests a specific one of the parameter-specific proxy models; and
receiving user selections for generating predictions, and generating predictions using the set of parameter-specific proxy models.

### Claim 16 (depends on 15)

The computer-implemented system of claim 15, wherein generating the training sample data includes generating parameter files based on an initial number of training cases and a number of target analysis factors.

### Claim 17 (depends on 15)

The computer-implemented system of claim 15, wherein the set of RTM parameters comprises one or more of dolomite, porosity, permeability, pH, magnesium ions (Mg2+), and calcium ions (Ca2+).

### Claim 18 (depends on 15)

The computer-implemented system of claim 15, wherein training the set of parameter-specific proxy models includes defining training parameters for the training.

### Claim 19 (depends on 18)

The computer-implemented system of claim 18, wherein the training parameters include a training data percentage, a validate data percentage, a test data percentage, a neuron number in a first hidden layer, and a neuron number in a second hidden layer.

### Claim 20 (depends on 15)

The computer-implemented system of claim 15, the operations further comprising receiving, using the GUI, user inputs for controlling steps for training the machine learning-based proxy models as surrogates for the process-based RTM.
