search

yellowshippo / penn-neurips2022

PENN code for NeurIPS 2022
https://openreview.net/forum?id=B3TOg-YCtzo
Apache License 2.0
37 stars 3 forks source link

OpenFOAM files and preprocessing settings #1

Closed martenlienen closed 1 year ago

martenlienen commented 1 year ago

Hey, do you also have the OpenFOAM files and the YAML files available that you used to generate and preprocess the data? That would be of great help to understand your work in detail.

Best, Marten

yellowshippo commented 1 year ago

Hi,

Thank you for the issue! Let me consider it. Once I get a conclusion, I will come back here.

Regards, Masanobu

yellowshippo commented 1 year ago

Hi Marten,

I have uploaded OpenFOAM data. Please find the updated README.md for more details: https://github.com/yellowshippo/penn-neurips2022#raw-data-for-the-incompressible-flow-dataset

The corresponding YAML file for preprocessing can be found here: https://github.com/yellowshippo/penn-neurips2022/blob/main/data/fluid/data.yml

Thank you very much, Masanobu

martenlienen commented 1 year ago

Thanks!

martenlienen commented 1 year ago

It looks like you adapted your case from the motor bike example, though there are also files that are not included with the motor bike example. Do you remember which adaptations you made and which parts of the case files are actually relevant for the dataset?

yellowshippo commented 1 year ago

We added FOsurfaceDistance and FOwriteCellVolumes in the system directory, which enabled us to compute the distance from the wall and cell volumes. These are used for machine learning inputs.

In the constant/triSurface directory, we have many .obj files, which are necessary to define the shape for analysis.