Skip to content

A collection of datasets for RUL estimation as Lightning Data Modules.

Notifications You must be signed in to change notification settings

tilman151/rul-datasets

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

RUL Datasets

Master Release Code style: black

This library contains a collection of common benchmark datasets for remaining useful lifetime (RUL) estimation. They are provided as LightningDataModules to be readily used in PyTorch Lightning.

Currently, five datasets are supported:

  • C-MAPSS Turbofan Degradation Dataset
  • FEMTO (PRONOSTIA) Bearing Dataset
  • XJTU-SY Bearing Dataset
  • N-C-MAPSS New Turbofan Degradation Dataset
  • Dummy A tiny, simple dataset for debugging

All datasets share the same API, so they can be used as drop-in replacements for each other. That means, if an experiment can be run with one of the datasets, it can be run with all of them. No code changes needed.

Aside from the basic ones, this library contains data modules for advanced experiments concerning transfer learning, unsupervised domain adaption and semi-supervised learning. These data modules are designed as higher-order data modules. This means they take one or more of the basic data modules as inputs and adjust them to the desired use case.

Installation

The library is pip-installable. Simply type:

pip install rul-datasets

Contribution

Contributions are always welcome. Whether you want to fix a bug, add a feature or a new dataset, just open an issue and a PR.