微型直接甲醇燃料电池多级退化参量融合预测

董杰; 苏雨临; 张大骋

doi:10.13700/j.bh.1001-5965.2023.0517

微型直接甲醇燃料电池多级退化参量融合预测

doi: 10.13700/j.bh.1001-5965.2023.0517

董杰¹,
苏雨临²,
张大骋^{1, 3, ,}

1.
昆明理工大学信息工程与自动化学院，昆明 650500
2.
北京麻豆精品秘国产传媒软件学院，北京 100191
3.
云南省绿色能源与数字电力量测及控保重点实验室，昆明 650500

基金项目:

国家自然科学基金(62103174)；云南省基础研究计划(202202AD080006)

详细信息

通讯作者:
E-mail：dacheng.zhang@kust.edu.cn

中图分类号: TM911.4
计量
- 文章访问数: 159
- HTML全文浏览量: 58
- PDF下载量: 1
- 被引次数: 0
出版历程
- 收稿日期: 2023-08-08
- 录用日期: 2023-09-28
- 网络出版日期: 2023-10-20
- 整期出版日期: 2025-10-31

Ensemble-based prediction using multi-level degradation parameters for micro direct methanol fuel cells

DONG Jie¹,
SU Yulin²,
ZHANG Dacheng^{1, 3
, ,}

1.
Faculty of Information Engineering and Automation，Kunming University of Science and Technology，Kunming 650500，China
2.
School of Software，Beihang University，Beijing 100191，China
3.
Yunnan Key Laboratory of Green Energy，Electric Power Measurement Digitalization，Control and Protection，Kunming 650500，China

Funds:

National Natural Science Foundation of China (62103174); Yunnan Fundamental Research Projects (202202AD080006)

More Information

Corresponding author: E-mail：dacheng.zhang@kust.edu.cn

摘要

摘要:
微型直接甲醇燃料电池(µDMFC)膜电极性能会随运行时间和工况衰退，影响其效率和使用寿命。准确的健康状态(SOH)估计与剩余使用寿命(RUL)预测可以确保µDMFC安全运行。输出电压的退化趋势受运行工况影响而波动，传统基于退化趋势回归的预测方法无法准确捕获这种随机变化。在数据驱动和机理模型相结合的思路下，提出基于输出电压和等效电路模型(ECM)的RUL融合预测方法。针对动态工况，引入负载电流这一退化协变量，利用随机过程重构了未来时刻的负载变化，结合ECM参量的退化趋势，实现对输出电压的准确估计和RUL预测，并在中国轻型汽车测试工况(CLTC)下进行验证。实验结果表明：基于多级退化参量的融合预测方法能更好地适应变工况，RUL预测精度、准确度分别为88.18%和85.71%，高于单一退化模型最优值3.27%和14.28%。
- 微型直接甲醇燃料电池 /
- 动态工况 /
- 剩余使用寿命 /
- 多级退化参量 /
- 融合预测
Abstract:
The deterioration of the membrane electrode assembly in micro direct methanol fuel cells (µDMFC) limits cell efficiency and lifespan. Accurate prediction of the state of health (SOH) and remaining useful life (RUL) is essential for ensuring the safe and reliable operation of µDMFCs in industrial applications. The degradation trend of the output voltage fluctuates depending on the operating conditions. Traditional trend regression methods, however, are inadequate for capturing such stochastic fluctuations. Therefore, an RUL ensemble-based prediction method based on the output voltage and equivalent circuit model (ECM) was proposed under a combination of data-driven and mechanistic models. The degradation covariate of load current was introduced to account for dynamic operating conditions. The load changes in the future were reconstructed using a random process, and combined with the degradation trend of ECM parameters, the accurate estimation of the output voltage and RUL prediction were achieved. The proposed approach was validated using accelerated aging tests under the China light vehicle test cycle (CLTC). Experimental results show that the ensemble-based method can adapt to varying operating conditions. RUL prediction accuracy and precision are 88.18% and 85.71%, respectively, outperforming the best individual model by 3.27% and 14.28%.
- micro direct methanol fuel cell /
- dynamic operating conditions /
- remaining useful life /
- multi-level degradation parameter /
- ensemble-based prediction

HTML全文

图 1 直接甲醇燃料电池单芯结构

Figure 1. Structure of µDMFC single cell

下载: 全尺寸图片幻灯片

图 2 直接甲醇燃料电池加速老化测试平台

Figure 2. µDMFC accelerated aging test platform

下载: 全尺寸图片幻灯片

图 3 不同甲醇溶液浓度的输出性能比较

Figure 3. Output performance comparison with different methanol concentrations

下载: 全尺寸图片幻灯片

图 4 µDMFC一次放电循环

Figure 4. A discharge cycle of µDMFC

下载: 全尺寸图片幻灯片

图 5 µDMFC输出电压

Figure 5. µDMFC output voltage

下载: 全尺寸图片幻灯片

图 6 不同时刻EIS测量示例

Figure 6. EIS measurement examples at different time

下载: 全尺寸图片幻灯片

图 7 LSTM神经元结构

Figure 7. LSTM neuron structure

下载: 全尺寸图片幻灯片

图 8 µDMFC的典型电化学阻抗谱

Figure 8. Typical EIS of µDMFC

下载: 全尺寸图片幻灯片

图 9 等效电路模型

Figure 9. Equivalent circuit model

下载: 全尺寸图片幻灯片

图 10 负载电流重构的马尔可夫链

Figure 10. Markov chain-based load current reconstruction

下载: 全尺寸图片幻灯片

图 11 t_p= 600 h时刻的重构电流

Figure 11. Current reconstruction at t_p=600 h

下载: 全尺寸图片幻灯片

图 12 RUL 预测原理

Figure 12. RUL prediction illustration

下载: 全尺寸图片幻灯片

图 13 t_p=600 h时刻基于外部电压的RUL预测

Figure 13. DTM-based RUL prediction at t_p=600 h

下载: 全尺寸图片幻灯片

图 14 电化学阻抗谱拟合示例

Figure 14. EIS fitting example

下载: 全尺寸图片幻灯片

图 15 t_p=600 h时刻基于电化学阻抗谱的参量估计与预测

Figure 15. EIS-based covariate estimation at t_p=600 h

下载: 全尺寸图片幻灯片

图 16 t_p=600 h时刻基于等效电路模型的RUL预测

Figure 16. ECM-based RUL prediction at t_p=600 h

下载: 全尺寸图片幻灯片

图 17 基于模型融合预测的技术路线

Figure 17. Model ensemble-based prediction framework

下载: 全尺寸图片幻灯片

图 18 t_p=600 h时刻模型1与模型2融合的RUL预测

Figure 18. Ensemble-based RUL prediction (Model 1 and Model 2) at t_p=600 h

下载: 全尺寸图片幻灯片

图 19 不同模型的RUL预测结果

Figure 19. RUL prediction results by different models

下载: 全尺寸图片幻灯片

图 20 不同模型预测时长比较

Figure 20. Time complexity comparison of different models

下载: 全尺寸图片幻灯片

表 1 基于CLTC工况的负载电流设置

Table 1. Load current settings based on CLTC operating conditions

速度区间	运行状态	占比/％	平均车速/(km·h⁻¹)	实验负载电流/A
低速	减速	23.67	11.85	0.018
	匀速	21.36	24.91	0.038
	加速	22.18	35.02	0.053
	怠速	32.79	0	0
中速	减速	26.67	44.52	0.068
	匀速	28.29	55.53	0.089
	加速	27.32	62.49	0.096
	怠速	17.72	0	0
高速	减速	19.56	75.13	0.114
	匀速	57.11	87.71	0.133
	加速	19.78	90.92	0.138
	怠速	3.56	0	0

下载: 导出CSV

表 2 不同时间的 ECM 模型参数识别结果示例

Table 2. ECM parameters identification results at different time

时间/h	E_r/V	R₁/Ω	R₂/Ω	R₃/Ω	R₄/Ω	R₅/Ω
0	0.724	0.531	0.280	4.530	1.940	1.290
90	0.717	0.590	1.220	5.250	2.720	0.551
190	0.705	0.615	1.010	5.560	2.660	1.460
250	0.678	0.635	0.719	5.920	3.000	1.680
340	0.639	0.620	1.550	6.490	3.340	1.770

下载: 导出CSV

表 3 各模型 RUL 预测性能评价结果

Table 3. Prognostic performance evaluation for different models

模型	A_cc/%	$\alpha_ {A_{\mathrm{c}} } $/%	平均预测时长/s
模型1	77.87	33.33	50.20
模型2	84.91	71.43	143.91
模型1与模型2融合	88.18	85.71	146.11

下载: 导出CSV

参考文献(24)

[1]	JIAO K, XUAN J, DU Q, et al. Designing the next generation of proton-exchange membrane fuel cells[J]. Nature, 2021, 595(7867): 361-369. doi: 10.1038/s41586-021-03482-7
[2]	MARAPPAN M, PALANISWAMY K, VELUMANI T, et al. Performance studies of proton exchange membrane fuel cells with different flow field designs–review[J]. The Chemical Record, 2021, 21(4): 663-714. doi: 10.1002/tcr.202000138
[3]	ESMAEILI N, GRAY E M, WEBB C J. Non‐fluorinated polymer composite proton exchange membranes for fuel cell applications: a review[J]. ChemPhysChem, 2019, 20(16): 2016-2053. doi: 10.1002/cphc.201900191
[4]	LEE J, WU F J, ZHAO W Y, et al. Prognostics and health management design for rotary machinery systems: reviews, methodology and applications[J]. Mechanical Systems and Signal Processing, 2014, 42(1-2): 314-334. doi: 10.1016/j.ymssp.2013.06.004
[5]	LI X, ZHANG W, DING Q. Deep learning-based remaining useful life estimation of bearings using multi-scale feature extraction[J]. Reliability Engineering & System Safety, 2019, 182: 208-218.
[6]	刘浩. 质子交换膜燃料电池的寿命预测研究[D]. 杭州: 浙江大学, 2019: 1-14. LIU H. Study on life prediction of proton exchange membrane fuel cell[D]. Hangzhou: Zhejiang University, 2019: 1-14 (in Chinese).
[7]	ZHANG D, CADET C, BÉRENGUER C, et al. Some improvements of particle filtering based prognosis for PEM fuel cells[J]. IFAC-PapersOnLine, 2016, 49(28): 162-167. doi: 10.1016/j.ifacol.2016.11.028
[8]	FUTTER G A, LATZ A, JAHNKE T. Physical modeling of chemical membrane degradation in polymer electrolyte membrane fuel cells: influence of pressure, relative humidity and cell voltage[J]. Journal of Power Sources, 2019, 410: 78-90.
[9]	JOUIN M, GOURIVEAU R, HISSEL D, et al. Prognostics of PEM fuel cell in a particle filtering framework[J]. International Journal of Hydrogen Energy, 2014, 39(1): 481-494. doi: 10.1016/j.ijhydene.2013.10.054
[10]	VICHARD L, HAREL F, RAVEY A, et al. Degradation prediction of PEM fuel cell based on artificial intelligence[J]. International Journal of Hydrogen Energy, 2020, 45(29): 14953-14963. doi: 10.1016/j.ijhydene.2020.03.209
[11]	DONG H C, JIN X N, LOU Y B, et al. Lithium-ion battery state of health monitoring and remaining useful life prediction based on support vector regression-particle filter[J]. Journal of Power Sources, 2014, 271: 114-123. doi: 10.1016/j.jpowsour.2014.07.176
[12]	ZHANG D, BARALDI P, CADET C, et al. An ensemble of models for integrating dependent sources of information for the prognosis of the remaining useful life of proton exchange membrane fuel cells[J]. Mechanical Systems and Signal Processing, 2019, 124: 479-501. doi: 10.1016/j.ymssp.2019.01.060
[13]	ZHAO Z G, ZHANG F, ZHANG Y H, et al. Performance optimization of μDMFC with foamed stainless steel cathode current collector[J]. Energies, 2021, 14(20): 6608. doi: 10.3390/en14206608
[14]	薛瑞. 微型直接甲醇燃料电池高浓度传质阻挡层技术研究[D]. 哈尔滨: 哈尔滨工业大学, 2019: 46-62. XUE R. Research on the mass transfer barrier layer of concentrated direct methanol fuel cell[D]. Harbin: Harbin Institute of Technology, 2019: 46-62(in Chinese).
[15]	LIU Y, WU Z X, ZHOU H, et al. Development of China light-duty vehicle test cycle[J]. International Journal of Automotive Technology, 2020, 21(5): 1233-1246. doi: 10.1007/s12239-020-0117-5
[16]	王鑫, 吴际, 刘超, 等. 基于LSTM循环神经网络的故障时间序列预测[J]. 北京麻豆精品秘国产传媒学报, 2018, 44(4): 772-784. WANG X, WU J, LIU C, et al. Exploring LSTM based recurrent neural network for failure time series prediction[J]. Journal of Beijing University of Aeronautics and Astronautics, 2018, 44(4): 772-784(in Chinese).
[17]	PASTOR-FERNÁNDEZ C, DHAMMIKA WIDANAGE W, MARCO J, et al. Identification and quantification of ageing mechanisms in Lithium-ion batteries using the EIS technique[C]//Proceedings of the 2016 IEEE Transportation Electrification Conference and Expo. Piscataway: IEEE Press, 2016: 1-6.
[18]	YUAN T, ZOU Z Q, CHEN M, et al. New anodic diffusive layer for passive micro-direct methanol fuel cell[J]. Journal of Power Sources, 2009, 192(2): 423-428. doi: 10.1016/j.jpowsour.2009.03.032
[19]	SØRENSEN T S, KJELSTRUP S. A simple Maxwell-Wagner-Butler-Volmer approach to the impedance of the hydrogen electrode in a Nafion fuel cell[J]. Journal of Colloid and Interface Science, 2002, 248(2): 355-375. doi: 10.1006/jcis.2001.8210
[20]	ZHU Y L, GAO L, LI J Y. A novel button-type micro direct methanol fuel cell with graphene diffusion layer[J]. Micromachines, 2019, 10(10): 658. doi: 10.3390/mi10100658
[21]	YUAN T, YANG J, WANG Y L, et al. Anodic diffusion layer with graphene-carbon nanotubes composite material for passive direct methanol fuel cell[J]. Electrochimica Acta, 2014, 147: 265-270. doi: 10.1016/j.electacta.2014.09.124
[22]	JIANWEI G, ZONGQIANG M, JINGMING X. Studies on the electrochemical behavior of polymer electrolyte membrane fuel cell (PEMFC) by AC impedance method[J]. Chemical Journal of Chinese Universities, 2003, 24(8): 1477-1481.
[23]	KIM T, KIM H, HA J, et al. A degenerated equivalent circuit model and hybrid prediction for state-of-health (SOH) of PEM fuel cell[C]//Proceedings of the 2014 International Conference on Prognostics and Health Management. Piscataway: IEEE Press, 2014: 1-7.
[24]	苏雨临, 连冠, 张大骋. 等效电路模型法预测动态工况下微型直接甲醇燃料电池剩余使用寿命[J]. 上海交通大学学报, 2024, 58(10): 1525-1532. SU Y L, LIAN G, ZHANG D C. Equivalent circuit model-based prognostics for micro direct methanol fuel cell under dynamic operating conditions[J]. Journal of Shanghai Jiao Tong University, 2024, 58(10): 1525-1532(in Chinese).