ABSTRACT
The Proton Exchange Membrane Fuel Cell (PEMFC), a form of environmentally friendly energy conversion technology, has garnered significant interest and widespread adoption. Nevertheless, the durability of PEMFC has persistently posed a major hurdle to its practical implementation. This study aimed to compare the effect of neural networks with different layers on the lifetime prediction of proton exchange membrane fuel cells (PEMFC). This paper uses the Levenberg-Marquardt algorithm and Bayesian regularization to carry out experiments on 4, 10, and 15-layer neural networks. The basic working principle of PEMFC is introduced. In terms of content, this paper draws some important conclusions by analyzing the performance of different algorithms in predicting the life of PEMFC. The results show that the 15-layer neural network combined with the Levenberg-Marquardt algorithm achieves the best prediction performance. This study has important implications for optimizing PEMFC design and increasing its lifetime, guiding future fuel cell technology development.
- Gao Weitao, Lei Yijie, Zhang Xun, Hu Xiaobo, Song Pingping, Zhao Qing, Wang Cheng, Mao Zongqiang.2022. Recent progress in proton exchange membrane fuel cells, Chemical Industry and Engineering Progress, 41(3), 1539-1555.Google Scholar
- Du Zhenzhen, Wang Jun, Wang Jing, Yu Fan, Li Jiongli, Wang Xudong. 2022. Proton exchange membrane fuel cells for transportation and stationary applications, Journal of Materials Engineering, 50(12), 35-50.Google Scholar
- Chen GuangYao, Peng PeiXi, Tian YongHong. 2022. Robustness analysis for compact neural networks. Scientia Sinica (Technologica), 52(5), 689-703.Google ScholarCross Ref
- Zhao Linna, Lu Shu, Qi Dan, Xu Dongbei, Ying Shuang. 2019. Daily Maximum Air Temperature Forecast Based on Fully Connected Neural Network, Journal of Applied Meteorological Science, 33(3), 256-269.Google Scholar
- Bhavana Burramukku; Oğuzhan Ceylan; Mehdi Neshat. 2021. Power Output Prediction of Wave Farms Using Fully Connected Networks, 2021 56th International Universities Power Engineering Conference (UPEC), Middlesbrough, United Kingdom, 1-6, https://doi.org/10.1109/UPEC50034.2021.9548274.Google ScholarCross Ref
- Rubio. Jose de Jesus. 2021. Stability Analysis of the Modified Levenberg-Marquardt Algorithm for the Artificial Neural Network Training. IEEE Transactions on neural networks and learning systems, 32(8), 3510-3524. https://doi.org/10.1109/TNNLS.2020.3015200.Google ScholarCross Ref
- Ahmadi, Shervin Parvini Hansson, Anders Pakazad and Sina Khoshfetrat.2021. Distributed localization using the Levenberg-Marquardt algorithm. Eurasip journal on advances in signal processing, 1, 1-26. doi: 10.1186/s13634-021-00768-wGoogle ScholarCross Ref
- Feng, Wei Li, Qiaofeng Lu, Qiuhai Li, Chen Wang and Bo .2021. Element-wise Bayesian regularization for fast and adaptive force reconstruction. Journal of sound and vibration, 490. https://doi.org/10.1016/j.jsv.2020.115713Google ScholarCross Ref
- Zhu Xiaozhou. 2018. Design and Experimental Study of Self-humidifying Cathode Open PEMFC Control System, Southwest Jiaotong University.Google Scholar
- Akira Taniguchi, Tomoki Akita, Kazuaki Yasuda, Yoshinori Miyazaki. 2008. Analysis of degradation in PEMFC caused by cell reversal during air starvation, International Journal of Hydrogen Energy, 33(9), 2323-2329. https://doi.org/10.1016/j.ijhydene.2008.02.049Google ScholarCross Ref
- FCLAB Research.2014. IEEE PHM 2014 Data Challenge; FC Lab: Belfort, France.Google Scholar
- Ian T. Jolliffe, Jorge Cadima.2016. Principal component analysis: a review and recent developments, Philos Trans A Math Phys Eng Sci. 374(2065): 20150202. https://doi.org/10.1098/rsta.2015.0202Google ScholarCross Ref
Recommendations
Application of Neural Network approach for Proton Exchange Membrane fuel cell systems
Artificial Intelligence (AI) techniques, particularly the Neural Networks (NNs), are recently having significant impact on power electronics. In a Proton Exchange Membrane (PEM) fuel cell system, there is a strong relationship between the available ...
IRNN-Based Modeling and Simulation of Electrical Characteristics of Proton Exchange Membrane Fuel Cells
AICI '09: Proceedings of the 2009 International Conference on Artificial Intelligence and Computational Intelligence - Volume 01The proton exchange membrane fuel cell (PEMFC) is a rising research field, and PEMFC modeling is a key of PEMFC research and development. However, PEMFC mechanism models were too complicated to be suitable for PEMFC practical system control at present. ...
Modeling a Specific Commercial Single Proton Exchange Membrane Fuel Cell
Hybrid Artificial Intelligent SystemsAbstractWithin the different modeling possibilities to describe the behavior of complex dynamic systems, Artificial Neural Networks (ANN) are one of the most outstanding approaches. In this paper authors have used that approach to model the electrical ...
Comments