Parameter identification of PEMFC via feedforward neural network-pelican optimization algorithm

Yang, Bo, Liang, Boxiao, Qian, Yucun, Zheng, Ruyi, Su, Shi, Guo, Zhengxun and Jiang, Lin ORCID: 0000-0001-6531-2791 (2024) Parameter identification of PEMFC via feedforward neural network-pelican optimization algorithm. Applied Energy, 361. p. 122857.

Text
Parameter identification of PEMFC via feedforward neural network-pelican optimization algorithm copy.pdf - Author Accepted Manuscript
Available under License Creative Commons Attribution.
Download (4MB) | Preview

Official URL: http://dx.doi.org/10.1016/j.apenergy.2024.122857

Abstract

Parameter identification is a critical task in the research of proton exchange membrane fuel cells (PEMFC), which provides the basis for establishing an accurate and reliable PEMFC model. However, the nonlinear characteristics of PEMFC model as well as inevitable noise data and insufficient measurement data often overwhelm traditional optimization techniques. In particular, noise data and inadequate measurement data can introduce bias or lead to data loss. To address this problem, a novel hybrid optimization strategy is proposed. Firstly, a feedforward neural network (FNN) is employed to preprocess the measured data (i.e., reducing noise data and enriching measurement data). Furthermore, Gaussian noise and Rayleigh noise with three signal-to-noise ratio levels are introduced to simulate various disturbances of noise. Then, the pelican optimization algorithm (POA) is used to identify the parameters of PEMFC based on preprocessed data. Lastly, the effectiveness of the proposed strategy named FNN-POA is verified by comparing it with seven advanced competitive algorithms. Simulation results demonstrate that FNN-POA has higher robustness and optimization quality by comparing original data and preprocessed data. For instance, the root-mean-square error obtained by FNN-POA is reduced by 99.44% under medium temperature and medium pressure through noise reduction.

Item Type:	Article
Divisions:	Faculty of Science and Engineering > School of Electrical Engineering, Electronics and Computer Science
Depositing User:	Symplectic Admin
Date Deposited:	07 Mar 2024 15:59
Last Modified:	01 May 2024 15:14
DOI:	10.1016/j.apenergy.2024.122857
Related URLs:	Publisher
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3179220