A PSO-Based Fuzzy-Controlled Searching for the Optimal Charge Pattern of Li-Ion Batteries

Abstract

This paper proposes a searching strategy based on particle swarm optimization, combined with a fuzzy-deduced fitness evaluator (FDFE), to find the optimal multistage charging pattern that delivers the most discharged capacity within the shortest charging time (CT). The objective function of the optimization problem is to maximize the cost effectiveness for the applied charging pattern based on the CT and normalized discharged capacity (NDC). Therefore, this paper proposes an FDFE to combine CT and NDC into a unified cost function to properly evaluate the multiple performance characteristics index in the charge problem. The experimental results show that the obtained pattern is capable of charging the batteries to over 88% capacity within 51 min. Compared with the conventional constant current–constant voltage method, the CT, the obtained life cycle, and the charging efficiency of the lithium-ion (Li-ion) battery for the devised approach are improved by approximately 56.8%, 21%, and 0.4%, respectively. The presented charging approach is suitable for the increasingly applications, in which the batteries are “sealed” inside the products to extend the life span.