针对质子交换膜燃料电池（PEMFC）在动态工况下的电压多步提前预测精度较低的问题,提出一种基于离散小波变换-长短时记忆神经网络（DWT-LSTM）混合驱动的燃料电池多步提前老化预测方法。首先为消除动态工况下操作参数波动对电压预测的影响,对操作参数与电压进行相关性研究,使用LSTM对相关性强的操作参数和DWT分解电压分别构建预测模型,模型超参数和融合权重由粒子群优化算法（PSO）给出。同时为防止模型多步提前预测误差累计过快,搭建物理老化模型对多步提前预测结果进行修正。最后基于动态工况下老化的电堆耐久性数据进行测试,其中短期预测结果均方根误差（RMSE）和决定系数（R2）达到0.00254 V和0.9926;多步提前预测结果显示该方法能有效预测电压的老化趋势及局部波动,窗口值选择为4 h时均方根误差和决定系数达到0.0224 V和0.8941,程序消融实验也显示了多参数输入以及引入物理模型有利于电压多步提前预测的精度提升。
Abstract
Aiming at the problem of low accuracy of multi-step advance prediction of voltage of proton exchange membrane fuel cell （PEMFC） under dynamic conditions, a fuel cell multi-step advance prediction method based on DWT-LSTM hybrid drive is proposed. Firstly, in order to eliminate the influence of operating parameter fluctuations on voltage prediction under dynamic conditions, the correlation between operating parameters and voltage is studied, and the prediction models are constructed for the operating parameters with strong correlation and discrete wavelet transform （DWT） decomposition voltage respectively using long short-term memory neural network （LSTM）. The model hyperparameters and fusion weights are given by particle swarm optimization algorithm （PSO）. At the same time, in order to prevent the multi-step advance prediction error of the model from accumulating too quickly, a physical aging model is built to correct the multi-step advance prediction results. Finally, tests were carried out based on the durability data of the aged battery stack under dynamic conditions, in which the root mean square error （RMSE） and determination coefficient （R2） of the short-term prediction results reached 0.00254 and 0.9926; the multi-step advance prediction results show that this method can effectively predict the aging trend and local fluctuation of voltage. When the window value is selected as 4 hours, the root mean square error and determination coefficient reach 0.0224 and 0.8941. The program ablation experiment also shows that multi-parameter input and the introduction of physical models are conducive to improving the accuracy of multi-step advance prediction of voltage.
关键词
质子交换膜燃料电池 /
离散小波变换 /
长短时记忆 /
老化预测
Key words
proton exchange membrane fuel cell /
discrete wavelet transform /
long short-term memory /
aging prediction
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] HUA Z G, ZHENG Z X, PAHON E, et al.A review on lifetime prediction of proton exchange membrane fuel cells system[J]. Journal of power sources, 2022, 529: 231256.
[2] TIAN Q C, CHEN H T, DING S, et al.Remaining useful life prediction method of PEM fuel cells based on a hybrid model[J]. Electronics, 2023, 12(18): 3883.
[3] WANG Y P, WU K C, ZHAO H H, et al.Degradation prediction of proton exchange membrane fuel cell stack using semi-empirical and data-driven methods[J]. Energy and AI, 2023, 11: 100205.
[4] OU M Y, ZHANG R F, SHAO Z F, et al.A novel approach based on semi-empirical model for degradation prediction of fuel cells[J]. Journal of power sources, 2021, 488: 229435.
[5] CHU T K, ZHANG R F, WANG Y B, et al.Performance degradation and process engineering of the 10 kW proton exchange membrane fuel cell stack[J]. Energy, 2021, 219: 119623.
[6] AO Y J, LAGHROUCHE S, DEPERNET D, et al.Proton exchange membrane fuel cell prognosis based on frequency-domain Kalman filter[J]. IEEE transactions on transportation electrification, 2021, 7(4): 2332-2343.
[7] CHEN H, ZHAN Z G, JIANG P X, et al.Whole life cycle performance degradation test and RUL prediction research of fuel cell MEA[J]. Applied energy, 2022, 310: 118556.
[8] ZHANG R F, CHEN T, XIAO F, et al.Bi-directional gated recurrent unit recurrent neural networks for failure prognosis of proton exchange membrane fuel cells[J]. International journal of hydrogen energy, 2022, 47(77): 33027-33038.
[9] LI H L, CHEN Q H, ZHANG L Y, et al.Degradation prediction of proton exchange membrane fuel cell based on the multi-inputs bi-directional long short-term memory[J]. Applied energy, 2023, 344: 121294.
[10] ZHANG F, WANG B W, GONG Z C, et al.Short-term performance degradation prediction of proton exchange membrane fuel cell based on discrete wavelet transform and Gaussian process regression[J]. Next energy, 2023, 1(3): 100052.
[11] 陈景文, 杨淇, 兰天一, 等. 基于集成经验模态分解和极限学习机的质子交换膜燃料电池寿命预测[J]. 太阳能学报, 2025, 46(2): 135-141.
CHEN J W, YANG Q, LAN T Y, et al.Lifetime prediction of PEMFC based on ensemble empirical mode decomposition and extreme learning machine[J]. Acta energiae solaris sinica, 2025, 46(2): 135-141.
[12] HUA Z G, ZHENG Z X, PAHON E, et al.Multi-timescale lifespan prediction for PEMFC systems under dynamic operating conditions[J]. IEEE transactions on transportation electrification, 2022, 8(1): 345-355.
[13] 王艳琴, 谢卓峰, 韩国鹏, 等. 基于极端梯度提升的PEMFC长短期老化趋势预测[J]. 太阳能学报, 2024, 45(7): 232-239.
WANG Y Q, XIE Z F, HAN G P, et al.Short-and long-term degradation prediction for proton exchange membrane fuel cell based on extreme gradient boosting[J]. Acta energiae solaris sinica, 2024, 45(7): 232-239.
[14] MA R, XIE R Y, XU L C, et al.A hybrid prognostic method for PEMFC with aging parameter prediction[J]. IEEE transactions on transportation electrification, 2021, 7(4): 2318-2331.
[15] HONG J C, YANG H X, LIANG F W, et al.State of health prediction for proton exchange membrane fuel cells combining semi-empirical model and machine learning[J]. Energy, 2024, 291: 130364.
[16] YANG X T, ZOU A K, CAO J D, et al.Systemic risk prediction based on Savitzky-Golay smoothing and temporal convolutional networks[J]. Electronic research archive, 2023, 31(5): 2667-2688.
[17] ZHAO Y, LUO M J, YANG J W, et al.Numerical analysis of PEMFC stack performance degradation using an empirical approach[J]. International journal of hydrogen energy, 2024, 56: 147-163.
[18] FCLAB Research Federation.2014 IEEE PHM data challenge[EB/OL].[2018-01-02]. http://eng.fclab.fr/ieee-phm-2014-%20data-challenge/.
[19] ZHANG Z D, WANG Y X, HE H W, et al.A short-and long-term prognostic associating with remaining useful life estimation for proton exchange membrane fuel cell[J]. Applied energy, 2021, 304: 117841.
[20] BENAGGOUNE K, YUE M L, JEMEI S, et al.A data-driven method for multi-step-ahead prediction and long-term prognostics of proton exchange membrane fuel cell[J]. Applied energy, 2022, 313: 118835.
PDF(2194 KB)
