Bi-subgroup optimization algorithm for parameter estimation of a PEMFC model. (15th June 2022)

• Record Type: Journal Article
• Authors: Chen, Yang; Pi, Dechang; Wang, Bi; Chen, Junfu; Xu, Yue
• Is Part Of: Expert systems with applications. Volume 196(2022)
• Publication Date: 2022-06-15
• Subjects: Parameter estimation -- Swarm intelligence -- Convergence analysis -- Bi-subgroup optimization -- PEMFC model; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2022.116646
• Languages: English
• ISSNs: 0957-4174

Abstract

Abstract: Proton exchange membrane fuel cell (PEMFC) has the advantages of cleanliness, environmental protection, and high stability, and is widely used in the fields of astronautics, military, and so on. How to extract the PEMFC model parameters more precisely is a key problem, which can be transformed into a highly nonlinear optimization problem. Therefore, an efficient optimization algorithm is needed. Swarm intelligence algorithms are used for solving various complex optimization problems. In this paper, a bi-subgroup optimization algorithm (BSOA) is proposed. The BSOA divides the population into two different sub-populations, one of which is to develop more optimal solutions and the other is to mine more optimal solutions within the population. Hence, it can overcomes the drawback that a single population search tends to fall into local optimal solutions and enhances the diversity of the population. By constructing and analyzing the Markov model, it is proved that the BSOA is a global convergence algorithm. Finally, the BSOA is applied to the estimation of the unknown parameters of the PEMFC modules and compared with the state-of-the-art algorithms. The simulation results show that BSOA is an effective algorithm for estimating the unknown parameters of the PEMFC model. Highlights: Development of novel bi-subgroup algorithm named BSOA. BSOA uses two hierarchical subpopulations to perform collaborative search. Constructing the Markov model of BSOA to prove its convergence. Validating the effectiveness of BSOA through the popular PEMFC models.