https://github.com/rikeshamin/mlcharring
Instructions on how to deploy machine learning models to predict the charring rate of mass timber for structural calculations
https://github.com/rikeshamin/mlcharring
containerization fire fire-engineering fire-science hazelab machine-learning mlops python regression statistics structural-engineering vaqt
Last synced: 6 months ago
JSON representation
Instructions on how to deploy machine learning models to predict the charring rate of mass timber for structural calculations
- Host: GitHub
- URL: https://github.com/rikeshamin/mlcharring
- Owner: rikeshamin
- License: mit
- Created: 2024-02-20T15:44:33.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-12-26T03:07:09.000Z (9 months ago)
- Last Synced: 2025-02-13T07:36:56.375Z (8 months ago)
- Topics: containerization, fire, fire-engineering, fire-science, hazelab, machine-learning, mlops, python, regression, statistics, structural-engineering, vaqt
- Language: Python
- Homepage: https://doi.org/10.5281/zenodo.14238389
- Size: 35.2 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Predicting the Average Charring Rate of Mass Timber Using Data-Driven Methods for Structural Calculations
MLCharring is a collection of data-driven models developed to predict the charring rate of timber in fire scenarios, providing assistance for structural engineering calculations. The dataset, [VAQT](https://zenodo.org/records/14238389) , aggregates timber furnace tests conducted in an ISO 834 fire environment, offering a robust foundation for training advanced statistical and machine learning models. These models are designed to accurately estimate the average charring rate of mass timber, supporting fire-safe and efficient structural design practices in the built environment.
---
## Setting up VAQT on your local machine
Follow these steps to set up the VAQT dataset:
1. [Download VAQT](https://zenodo.org/records/14238389) from Zenodo.
2. Install Microsoft Access or migrate to open database platform (e.g., PostgreSQL, MySQL)
3. Open the `Timber Charring fe` file.
4. Navigate to the `External Data` tab and select:
- `New Data Source` -> `From Database` -> `Access`.
5. Choose `Link to data source...` and select the `Timber Charring be` file.
6. In the dialog box, press `SELECT ALL` and then click `OK`.
7. **VAQT is now linked and ready to usee**
---
## Running models locally
1. Ensure Python is installed and set up a virtual environment. Guidance can be found here:
- [Python Installation Guide](https://www.python.org/downloads/)
- [Virtual Environments Guide](https://docs.python.org/3/tutorial/venv.html)
2. Run `pip install -r requirements.txt` to install all dependencies.
3. Ensure the model `statistical_model.pkl` is in the same directory as `ml_charring_statistical.py`.
4. Execute the script using:
```bash
python ml_charring_statistical.py
```
6. Provide the input dataset to the model.
7. The terminal will output the predicted charring rates(s):
```bash
Charring Rate: [0.71605273]
```
If you encounter any problems, please create an issue directly in this repository or email [myself](rikamin95@gmail.com) or [Prof. Guillermo Rein](g.rein@imperial.ac.uk).
---
### Additional Resources:
- [Publication DOI](https://doi.org/10.1007/s10694-024-01593-x)
- [Hazelab - Imperial College London](https://www.imperial.ac.uk/hazelab)
---
### To Do (when i have some more time):
* Update VAQT compatability to open databases
* Upload deep learning models (currently on my own server)
* Implement MLOps for containerization and deployment (Docker and Kubernetes)
---
### Citation
```text
@article{Amin2024,
author = {Amin, Rikesh and Yaxin, Mo and Richter, Franz and Kurzer, Christoph and Werther, Norman and Rein, Guillermo},
title = {Predicting the Average Charring Rate of Mass Timber Using Data-Driven Methods for Structural Calculations},
journal = {Fire Technology},
ISSN = {1572-8099},
DOI = {10.1007/s10694-024-01593-x},
url = {https://doi.org/10.1007/s10694-024-01593-x},
year = {2024},
type = {Journal Article}
}
```