pcnns 0.1.2

No project description provided

Physically Consistent Neural Networks

This repository contains a clean and efficient version of Physically Consistent Neural Networks (PCNNs), where a physics-inspired module runs in parallel of a black-box one (neural networks) to capture physical effects.

Installation

After cloning the repository on your computer, go to the pcnn folder with cd path_to_the_folder/pcnn.

The fastest way to run the code is to use poetry, which can be installed from here.
You can then run poetry install to install all the required dependencies.
Once the dependencies are installed, you can for example run jupyter-lab with poetry run jupiter lab or VS code with poetry run code ..

Alternatively, you can install requirements from requirements.txt.

Related publications

Physically Consistent Neural Networks for building thermal modeling: Theory and analysis
Loris Di Natale, Bratislav Svetozarevic, Philipp Heer, and Colin N. Jones
Applied Energy 325 (2022). https://doi.org/10.1016/j.apenergy.2022.119806

Towards Scalable Physically Consistent Neural Networks: an Application to Data-driven Multi-zone Thermal Building Models
Loris Di Natale, Bratislav Svetozarevic, Philipp Heer, and Colin N. Jones
Submitted to Applied Energy (2023). https://arxiv.org/abs/2212.12380.

Contact

For additional information, pleasae contact [e-mail]