rulsurv

Code for: "RULSurv: A probabilistic survival-based method for early censoring-aware prediction of remaining useful life in ball bearings (2025)"

Preprint: https://arxiv.org/abs/2405.01614.

Requirements

To run the models, please refer to Requirements.txt. The code was tested in a virtual environment with Python 3.9.

Training and prediction

• Download the XJTU-SY dataset from MEGA
• Unzip the data inside the main folder to create /data
• Make directories mkdir results and mkdir plots
• Make the timeseries dataset by running python src/make_dataset.py
• Run cross-validation by running python src/run_cross_validation.py
• Predict ISDs by running python src/predict_isd_curves.py
• Refer to config.py for model configuration and paths.

Citation

If you find this paper useful in your work, please consider citing it:

@article{lillelund2025rulsurv,
  title={{RULSurv: A probabilistic survival-based method for early censoring-aware prediction of remaining useful life in ball bearings}},
  author={Lillelund, Christian Marius and Pannullo, Fernando and Jakobsen, Morten Opprud and Morante, Manuel and Pedersen, Christian Fischer},
  journal={arXiv preprint arXiv:2405.01614},
  year={2025}
}