.png){kind=icon}
.png){kind=icon}
Return
High-performance parameter extraction for solid oxide fuel cells under dynamic conditions using the griffon vulture optimization algorithm
.png){kind=icon}
OA
AI
M
R
P
A
G
DOI：10.1007/s40243-026-00384-4.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Solid oxide fuel cells (SOFCs) offer high energy conversion efficiency alongside excellent fuel flexibility. However, their complex underlying electrochemical processes require successful parameter identification so that modeling and control can be achieved. Traditional optimization algorithms on the other hand often face limitations in this problem, such as premature convergence to suboptimal solutions, or an inability to explore the entire solution space. In this manuscript a new metaheuristic based on vulture foraging is derived i.e., Griffon Vulture Optimization Algorithm (GVOA). It is superior in terms of accuracy, stability and computational speed in case of complicated optimization, such as SOFC parameter identification. GVOA’s guided convergence and controlled diversity provide a theoretically sound and practical solution. Because of its efficiency, it can be used for real-time modeling and adaptive control. The GVOA algorithm was successfully used to estimate seven nonlinear model parameters of a simplified SOFC model in ten different operating conditions (varying temperature and pressure), based on simulated V-I data. In comparison with 9 well-known metaheuristic algorithms, such as BKA, AO, LSO, and PEOA, GVOA always gave the lowest minimum mean squared error (MSE) results under any temperature and pressure conditions. Furthermore, it occupied the highest rank in the Friedman test in all situations and had the quickest convergence behaviour. These results validate GVOA’s superior accuracy, stability and speed with respect to the challenging task of identification of SOFC parameters. Theoretically, this is evidence in support of the efficacy of its combination of guided convergence and controlled diversity, and, practically, is a reliable method for modeling and adaptive control of fuel cell systems. Upcoming research will be spread to investigational validation and application to other electrochemical systems.
Keywords:
Parameter extraction
Machine learning
SOFC
Fuel cell
Operating temperature
Operating pressure
Journal
IF:
5.5
Papers:
223
Citations:
955
Organization
Cited Papers