电源学报

基于正交参数优化LSTM神经网络的SOFC性能预测

PDF下载

张妙妙中国电源学会学生会员 ¹ , 吕佳 ¹ , 吴剑 ¹^,² , 胡强 ² , 冯海林 ¹

电源学报 | 新能源发电与储能 2026,24(7): 137-146

收起

电源学报 |新能源发电与储能 2026 , 24 (7) : 137 -146

基于正交参数优化LSTM神经网络的SOFC性能预测

全屏

张妙妙中国电源学会学生会员¹, 吕佳¹, 吴剑¹^,², 胡强², 冯海林¹

作者信息

^1.浙江农林大学数学与计算机科学学院，杭州　311300

^2.浙江臻泰能源科技有限公司，丽水　321400

通讯作者:

吕佳（1989—　），女，通信作者，博士。研究方向：能源互联网。E-mail：jialv@zafu.edu.cn。

作者简介:

张妙妙（1998—　），女，中国电源学会学生会员，硕士研究生。研究方向：低碳能源。E-mail：2022611012073@stu.zafu.edu.cn。

吴剑（1985—　），男，博士。研究方向：氢能。E-mail：wujian@zafu.edu.cn。

胡强（1979—　）男，博士，研究员。研究方向：能源转换。E-mail：qihu@zetech.cn。

冯海林（1985—　），男，博士。研究方向：人工智能智慧农林业。E-mail：hlfeng@zafu.edu.cn。

Optimization of LSTM for SOFC Performance Prediction Using Orthogonal Method

Miaomiao ZHANG Student Member, CPSS¹, Jia Lü¹, Jian WU¹^,², Qiang HU², Hailin FENG¹

Affiliations

^1.College of Mathematics and Computer Science, Zhejiang A&F University, Hangzhou 311300, China

^2.Zhejiang Zhentai Energy Technology Co., Ltd., Lishui 321400, China

出版时间: 2026-07-30 doi: 10.13234/j.issn.2095-2805.2026.7.137

文章导航

摘要

收起

固体氧化物燃料电池SOFC（solid oxide fuel cell）是环保、高效、具有广泛前景的新型发电装置，近年来受到越来越多的关注。对SOFC系统运行状态的预测和评估是发电系统能够长期稳定运行的关键。针对长短期记忆LSTM（long short-term memory）神经网络模型参数选值困难导致的电池寿命预测精度不足的问题，正交试验法被用于系统性地优化基于LSTM的SOFC预测模型参数。结果表明，批大小对LSTM模型预测精度和计算时间的影响最为显著，且所提优化模型相较于循环神经网络RNN（recurrent neural network）、门控循环单元GRU（gated recurrent units）神经网络、鲸鱼优化算WOA（whale optimization algorithm）法优化LSTM（WOA-LSTM）和灰狼优化GWO（grey wolf optimization）算法优化LSTM（WOA-LSTM）的均方根误差RMSE（root mean square error）分别降低了72.36%、18.18%、35.05%和32.97%。经正交法对模型参数进行系统性调优后，LSTM模型表现出较高的预测精度，预测准确性提高。未来可将其与控制策略相结合，有效减缓燃料电池的衰退。

关键词

固体氧化物燃料电池 / 长短期记忆神经网络 / 正交试验法 / 参数优化 / 性能预测

Abstract

收起

Solid oxide fuel cell (SOFC) has attracted increasing attention due to its environmental friendliness, high efficiency and broad prospects. The life prediction of SOFC system is critical to the long-term operation of power generation system. However, the accuracy in predicting the life of SOFC using long short-term memory (LSTM) neural network model with numerous parameters is still insufficient. To address this problem, the orthogonal experiment method was employed to systematically optimize the parameters for LSTM. The results indicate that batch size has the greatest influence on the prediction accuracy and calculation time of LSTM. And the root mean square error (RMSE) of optimized model is reduced by 72.36%、18.18%、35.05% and 32.97% when compared with that of recurrent neural network (RNN), gated recurrent unit (GRU), whale optimization algorithm (WOA) optimizates LSTM (WOA-LSTM) and grey wolf optimization (GWO) optimizates LSTM (GWO-LSTM), respectively. The systematically optimized LSTM model exhibits improved performance, which may mitigate fuel cell degradation with other control strategies.

Key words

Solid oxide fuel cell (SOFC) / long short-term memory (LSTM) neural network / orthogonal experimental method / parameter optimization / performance prediction

引用本文

张妙妙中国电源学会学生会员, 吕佳, 吴剑, 胡强, 冯海林. 基于正交参数优化LSTM神经网络的SOFC性能预测. 电源学报, 2026 , 24 (7) : 137 -146 . DOI: 10.13234/j.issn.2095-2805.2026.7.137

Miaomiao ZHANG, Jia Lü, Jian WU, Qiang HU, Hailin FENG. Optimization of LSTM for SOFC Performance Prediction Using Orthogonal Method[J]. Journal of Power Supply, 2026 , 24 (7) : 137 -146 . DOI: 10.13234/j.issn.2095-2805.2026.7.137

基金

收起

国家自然科学基金青年基金资助项目(22209149)

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

[1]

Yousaf M, Lu Yuzheng, Akbar M, et al. Advances in solid oxide fuel cell technologies: Lowering the operating temperatures through material innovations[J]. Materials Today Energy, 2024, 44: 101633.

[2]

Lu Jia, Hu Yubo, Zhang Miaomiao, et al. Optimization of Pt infiltrated Sr₂Fe_2-xMo_xO_6-δ-Ce_0.8Sm_0.2O_1.9 oxygen electrode for reversible solid oxide cell and CH₄-assisted electrolysis process[J]. International Journal of Hydrogen Energy, 2024, 55: 786-795.

[3]

郭喜峰, 黄裕海, 单丹, 等. 改进NGO优化MVMD-BiLSTM的锂电池剩余使用寿命预测[J]. 电源学报, 2026, 24(4): 158-168. Guo Xifeng, Huang Yuhai, Shan Dan, et al. Improved NGO optimized MVMD-BiLSTM for remaining useful life prediction of lithium batteries[J]. Journal of Power Supply, 2026, 24(4): 158-168. (in Chinese)

[4]

汪建锋, 王荣杰, 林安辉, 等. 质子交换膜燃料电池退化预测方法[J]. 电工技术学报, 2024, 39(11): 3367-3378. Wang Jianfeng, Wang Rongjie, Lin Anhui, et al. Degradation prediction method of proton exchange membrane fuel cell[J]. Transactions of China Electrote-chnical Society, 2024, 39(11): 3367-3378. (in Chinese)

[5]

敖文杰, 陈家伟, 陈杰, 等. 燃料电池-锂电池混合供电系统的无源控制策略及参数设计方法[J]. 电工技术学报, 2024, 39(2): 580-594. Ao Wenjie, Chen Jiawei, Chen Jie, et al. Passivity-based control strategy and parameter design method for fuel cell-lithium battery hybrid power supply system[J]. Transactions of China Electrotechnical Society, 2024, 39(2): 580-594. (in Chinese)

[6]

Oh H S, Kim J M, Chang C Y. A study on LSTM-based lithium battery SoH estimation in urban railway vehicle operating environments[J]. Journal of Electrical Engineering & Technology, 2024, 19(4): 2817-2829.

[7]

Wang Xiaohua, Ni Nanbing, Hu Min, et al. A MDA-LSTM network for remaining useful life estimation of lithium batteries[J]. Signal, Image and Video Processing, 2024, 18(1): 129-140.

[8]

Peng Jingxuan, Huang Jian, Jiang Chang, et al. Generalized spatial-temporal fault location method for solid oxide fuel cells using LSTM and causal inference[J]. IEEE Transactions on Transportation Electrification, 2022, 8(4): 4583-4594.

[9]

Sheng Chuang, Zheng Yi, Tian Rui, et al. A comparative study of the Kalman filter and the LSTM network for the remaining useful life prediction of SOFC[J]. Energies, 2023, 16(9): 3628.

[10]

Zuo Bin, Cheng Junsheng, Zhang Zehui. Degradation prediction model for proton exchange membrane fuel cells based on long short-term memory neural network and Savitzky-Golay filter[J]. International Journal of Hydrogen Energy, 2021, 46(29): 15928-15937.

[11]

Xia Zetao, Wang Yining, Ma Longhua, et al. A hybrid prognostic method for proton-exchange-membrane fuel cell with decomposition forecasting framework based on AEKF and LSTM[J]. Sensors, 2023, 23(1): 166.

[12]

高金武, 贾志桓, 王向阳, 等. 基于PSO-LSTM的质子交换膜燃料电池退化趋势预测[J]. 吉林大学学报(工学版), 2022, 52(9): 2192-2202. Gao Jinwu, Jia Zhihuan, Wang Xiangyang, et al. Degradation trend prediction of proton exchange membrane fuel cell based on PSO-LSTM[J]. Journal of Jilin University (Engineering and Technology Edition), 2022, 52(9): 2192-2202. (in Chinese)

[13]

Zhang Mengru, AnLisha, Wang Enhua, et al. Effects of sintering parameters on the low-temperature densification of GDC electrolyte based on an orthogonal experiment[J]. Catalysts, 2022, 12(8): 831.

[14]

Zhou Yaqin, Mao Jingwen, Wang Enhua, et al. Parameters optimization for electrophoretic deposition of Mn_1.5Co_1.5O₄ on ferritic stainless steel based on multi-physical simulation[J]. Journal of the Electrochemical Society, 2024, 171(6): 062509.

[15]

Lu Jia, Liu Wei, Wu Jian, et al. Electrochemical characterization and degradation analysis of a flat-chip solid oxide fuel cell fed by ambient temperature inlet gas[J]. Electrochimica Acta, 2023, 470: 143231.

[16]

Wu J M, Tsai M H, Huang Yongzhi, et al. Applying an ensemble convolutional neural network with Savitzky-Golay filter to construct a phonocardiogram prediction model[J]. Applied Soft Computing, 2019, 78: 29-40.

[17]

MaRui, YangTao, BreazE, et al. Data-driven proton exchange membrane fuel cell degradation predication through deep learning method[J]. Applied Energy, 2018, 231: 102-115.

[18]

Greff K, Srivastava R K, Koutník J, et al. LSTM: A search space odyssey[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28(10): 2222-2232.

[19]

Rao Mumin, Wang Li, Chen Chuangting, et al. Data-driven state prediction and analysis of SOFC system based on deep learning method[J]. Energies, 2022, 15(9): 3099.

[20]

Mondal P, Bhavsar D, Mittal K, et al. Estimating state-of-charge in lithium-ion batteries through deep learning techniques: A comparative evaluation[J]. IEEE Access, 2024, 12: 78773-78786.

[21]

Qu Yaxin, Zhao Xue. Application of LSTM neural network in forecasting foreign exchange price[J]. Journal of Physics: Conference Series, 2019, 1237(4): 042036.

[22]

Song Shaohui, Xiong Xingyu, Wu Xin, et al. Modeling the SOFC by BP neural network algorithm[J]. International Journal of Hydrogen Energy, 2021, 46(38): 20065-20077.

[23]

Oyedotun O K, Papadopoulos K, Aouada D. A new perspective for understanding generalization gap of deep neural networks trained with large batch sizes[J]. Applied Intelligence, 2023, 53(12): 15621-15637.

[24]

Li Mingfei, Wu Jiajian, Chen Zhengpeng, et al. Data-driven voltage prognostic for solid oxide fuel cell system based on deep learning[J]. Energies, 2022, 15(17): 6294.

[25]

袁建华, 谈顺, 刘闯. 基于改进灰狼算法优化LSTM的光伏发电功率短期预测[J]. 电力学报, 2024, 39(2): 111-118. Yuan Jianhua, Tan Shun, Liu Chuang. Short term forecasting of photovoltaic power generation based on LSTM optimized by improved grey wolf algorithm[J]. Journal of Electric Power, 2024, 39(2): 111-118. (in Chinese)

[26]

Wang Tongying, Huang Haozhong, Li Xuan, et al. Optimal estimation of proton exchange membrane fuel cell model parameters based on an improved chicken swarm optimization algorithm[J]. International Journal of Green Energy, 2023, 20(9): 946-965.

[27]

徐泽明, 李凯昂, 顾玉琦, 等. 基于鲸鱼算法优化CNN对叉车电池SOH进行预测[J]. 电源学报, 2026, 24(4): 196-204, 270. Xu Zeming, Li Kaiang, Gu Yuqi, et al. Predicting SOH of forklift batteries based on CNN optimized by whale optimization algorithm[J]. Journal of Power Supply, 2026, 24(4): 196-204, 270. (in Chinese)

2026年第24卷第7期

PDF下载

引用本文

BibTeX

文章信息

doi: 10.13234/j.issn.2095-2805.2026.7.137

接收时间：2024-06-13
首发时间：2026-10-29
出版时间：2026-07-30

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2024-06-13
修回日期：2024-08-13
录用日期：2024-09-20

基金

National Natural Science Foundation of China Young Scientists(22209149)

国家自然科学基金青年基金资助项目(22209149)

作者信息

^1.浙江农林大学数学与计算机科学学院，杭州　311300

^2.浙江臻泰能源科技有限公司，丽水　321400

通讯作者:

吕佳（1989—　），女，通信作者，博士。研究方向：能源互联网。E-mail：jialv@zafu.edu.cn。

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/dyxb/CN/10.13234/j.issn.2095-2805.2026.7.137

分享至

全文二维码

扫描看全文

引用本文

BibTeX

本文的引用情况

2种不同金属材料的力学参数

科 Family	属数 Number of genus	种数 Number of species	占总种数比例 Percentage of total species (%)	属 Genus	种数 Number of species	占总种数比例 Percentage of total species (%)
鹅膏菌科Amanitaceae	2	11	5.26	鹅膏菌属 Amanita	10	4.78
小菇科 Mycenaceae	2	12	5.74	丝盖伞属 Inocybe	5	2.39
多孔菌科 Polyporaceae	8	14	6.70	蜡蘑属 Laccaria	5	2.39
红菇科 Russulaceae	3	23	11.00	小皮伞属 Marasmius	6	2.87
				小菇属 Mycena	11	5.26
				光柄菇属 Pluteus	5	2.39
				红菇属 Russula	17	8.13
				栓菌属 Trametes	5	2.39

关闭全屏

BibTeX
EndNote
RefWorks
TxT