Supplementary Data for

“ML-Accelerated Prediction of Radiative View Factors in Packings of Cubical Particles”

Martin Gruber, Bo Jaeger, Josef Tausendschön, Michael Mitterlindner, Martin Schiemann, Viktor Scherer, Stefan Radl

Data Statement

Due to the size of the raw data used for the ML model training, this data is stored on a  this repository. The data comprises the features (feat_*) and were generated using python routines and MCRT simulations. The data folder also contains input files for these simulations. The raw data is available online (see below).

The trained models (*.pth) can also be downloaded from a cloud server. The link also contains the hyperparameter database (*.db), predictions and targets of the models as ascii files (*.txt) and model prediction times.

Software and Data

MCRT

The Monte Carlo Ray Tracing tool used to generate view factor data is available via:

https://gitlab.tugraz.at/simSci/simSoftware/rayfactor.git

Python Routines

Python routines for pre-/post-processing of data, including machine learning:

https://gitlab.tugraz.at/simSci/simSoftware/simsci_ml.git

Raw Data

This data is available and can be downloaded from this repository:

rawData.zip

ML Models

The final trained models and scripts can be downloaded from this repository:

trainedModels_Scripts.zip