Figures of NASA
Figures of CALCE
Due to the length of the paper, the two parameters of dropout and noise_level are not discussed. By setting these two parameters, better results can be obtained than in the paper.
-
noise level = 0.01: Setting the value of 1% disturbance is best: too large will degrade performance, too small will have little effect.
-
dropout = 1e-4~1e-3: Set a small value for the network dropout to ensure the robustness of the model.
-
pytorch 1.8.0
-
pandas 0.24.2
-
mixture_of_experts 0.2.1 (for AttMoE, github: https://github.com/lucidrains/mixture-of-experts)
- 6/5/2024, add figures of model and prediction
- 1/3/2024, upload the open sorce of AttMoE
- 24/2/2022,Change some variable names
Dataset CALCE processing reference
https://github.com/konkon3249/BatteryLifePrediction
Please feel free to contact me: zhouxiuze@foxmail.com
-
马里兰大
学 锂电池 数 据 集 CALCE,基 于 Python的 锂电池 寿命 预测: https://snailwish.com/437/ -
NASA 锂电
池 数 据 集 ,基 于 Python的 锂电池 寿命 预测: https://snailwish.com/395/ -
NASA 锂电
池 数 据 集 ,基 于 python的 MLP 锂电池 寿命 预测: https://snailwish.com/427/ -
NASA
和 CALCE 锂电池 数 据 集 ,基 于 Pytorch的 RNN、LSTM、GRU寿命 预测: https://snailwish.com/497/ -
基 于 Pytorch的 Transformer 锂电池 寿命 预测: https://snailwish.com/555/ -
锂电
池 研究 之 七 ——基 于 Pytorch的 高 斯函数 拟合时间序列 数 据 : https://snailwish.com/576/
@article{chen2022transformer,
title={Transformer network for remaining useful life prediction of lithium-ion batteries},
author={Chen, Daoquan and Hong, Weicong and Zhou, Xiuze},
journal={Ieee Access},
volume={10},
pages={19621--19628},
year={2022},
publisher={IEEE}
}
@article{chen2024attmoe,
title={AttMoE: Attention with Mixture of Experts for remaining useful life prediction of lithium-ion batteries},
author={Chen, Daoquan and Zhou, Xiuze},
journal={Journal of Energy Storage},
volume={84},
pages={110780},
year={2024},
publisher={Elsevier}
}