DeepField-app

Web application for simulation and visualization of reservoir models.

Lightweight. Modern. Open source.

Features

Use the application to

• read reservoir models given in ECLIPSE file format;
• simulate models using JutulDarcy simulator;
• view and explore static and dynamic data in 3D, 2D, and 1D;
• export simulated data in ECLIPSE and CSV formats;
• write and execute custom python scripts for reservoir transformations and calculations;
• work locally or on a remote server.

Main page of the application:

Simulated reservoir dynamics in 3D:

Filtering of grid cells and indication of well status (producing, injecting, inactive):

Selection of cells along well trajectories:

2D slice view:

Construction of a multiline 1D plot to compare dynamic properties:

Visualization of PVT and relative permeability tables:

Description of the reservoir model:

Script writing:

...and the results of its execution:

Performance

Loading time and memory usage for benchmark models in the benchmarks directory measured on a PC with Intel Core Ultra 7, 3.9GHz, 64Gb CPU:

Number of cells	Loading time	Memory usage
1M	23s	0.5Gb
10M	3m 49s	2.6Gb
50M	19m 26s	10.3Gb

Installation as a package

We recommend creating a new virtual environment with python 3.11 to install the project dependencies:

conda create -n app python=3.11

Activate the new environment:

conda activate app

To install the project dependencies, run in the terminal:

pip install "git+https://github.com/deepfield-team/DeepField-app.git"

After installation, run in the terminal:

deepfield-app

This should open a new tab in your default browser to http://localhost:8080/ with the application's home page.

You can add a few optional parameters to the application start command:

• --server - use to prevent a new tab from opening in the browser
• --app - use to launch the application in a separate window rather than in the browser
• --port 1234 - to change the default port 8080 to, e.g., 1234

When the application is running, you can click on the help icon in the upper right corner to read a brief description of the page. Hover over buttons and icons to see a tooltip with textual information about them.

Note

The installation of JutulDarcy simulator will be done automatically at the very first launch of the application. It may take some time.

Installation from source code

Another option to run the application is to clone the entire repository:

git clone https://github.com/deepfield-team/DeepField-app.git

Addionally, you will need to clone the repository DeepField into the same directory as the DeepField-app:

git clone https://github.com/deepfield-team/DeepField.git

Install dependencties in both repositories using

pip install -r requirements.txt

Then navigate to the directory DeepField-app and run in the terminal

python -m deepfield_app.app

to start the application.

Note

The installation of JutulDarcy simulator will be done automatically at the very first launch of the application. It may take some time.

Rendering options

By default, rendering is performed locally in the user's browser using vtk webassembly functionality. To enable vtk remote rendering, use the -vr or --vtk_remote option when starting the application. Note that the functionality of the application is slightly different between local and remote rendering.

Open-source reservoir models

An example reservoir model with dynamics simulation can be found in the open_data directory in the DeepField repository https://github.com/deepfield-team/DeepField, as well as links to a number of other open source models.

Script writing

The application allows you to write and execute python scripts for reservoir model transformations and calculations. The script should be based on the DeepField framework https://github.com/deepfield-team/DeepField. Read the documentation and see examples in the DeepField repository to prepare a script.

Next releases

The project is developing. We are preparing new releases with new features. Your suggestions and issues reports will help to make the application even better.

What's inside

We use

• trame to build the web application
• DeepField to read and process reservoir models
• JutulDarcy for reservoir simulation

Citing

We hope that this project will help you in your research and you will decide to cite it as

DeepField web application (2024). GitHub repository, https://github.com/deepfield-team/DeepField-app.