/Domain-Adaptive-Remaining-Useful-Life-Prediction-with-Transformer

Pytorch implementation for Domain Adaptive Remaining Useful Life Prediction with Transformer

Primary LanguagePython

Update extracted & processed N-CMAPSS dataset, 6/10/2023

To promote relevant researches on domain adaptive RUL prediction in the community, we provide the N-CMAPSS dataset used in this paper, with 7 domains readily available. Download here.

Domain Adaptive Remaining Useful Life Prediction with Transformer

Pytorch implementation for Domain Adaptive Remaining Useful Life Prediction with Transformer, https://doi.org/10.1109/TIM.2022.3200667

  • Prognostic health management (PHM) has become a crucial part in building highly automated systems, whose primary task is to precisely predict the remaining useful life (RUL) of the system. In this paper, we leverage domain adaptation for RUL prediction and propose a novel method by aligning distributions at both the feature level and the semantic level. The proposed method facilitates a large improvement of model performance as well as faster convergence. Besides, we propose to use Transformer as backbone, which can capture long-term dependency more efficiently. We test our model on CMAPSS dataset and its newly published variant N-CMAPSS provided by NASA, achieving state-of-the-art results on both source-only RUL prediction and domain adaptive RUL prediction tasks. 1677138917679

Environment

This is a strict constraint!

  • pytorch==1.10
  • torchvision==0.11.0

Dataset

  • Unzip CMAPSS.zip into CMAPSS folder, which contains processed CMAPSS dataset used in this code.

Usage

  • train on CMAPSS
    python train_cmapss.py --source $S --target $T
    where $S is source domain, $T is target domain. Domains include "FD001,FD002,FD003,FD004". Trained models are saved to /online.
  • evaluate on CMAPSS
    You can evaluate our best performing models saved in folder save/final by running:
    python validation_cmapss.py --source $S --target $T
    where $S is source domain, $T is target domain. Domains include "FD001,FD002,FD003,FD004".

Results

image

Citation

@article{li2022domain,
  title={Domain Adaptive Remaining Useful Life Prediction With Transformer},
  author={Li, Xinyao and Li, Jingjing and Zuo, Lin and Zhu, Lei and Shen, Heng Tao},
  journal={IEEE Transactions on Instrumentation and Measurement},
  volume={71},
  pages={1--13},
  year={2022},
  publisher={IEEE}
}