(A) study on novel battery management system using nonlinear battery cell model and consensus algorithm

Abstract

In this dissertation, a battery pack management system which consists of a robust state of charge estimation method and a full-state-based active balancing control method is proposed. In the use of a battery pack comprising a lot of battery cells connected in series and parallel, various imbalances between battery cells are inevitable factors. If the internal imbalances are not sufficiently compensated, this not only reduces the usable capacity of the battery pack in the short time period, but also causes premature failure of battery system by accelerating the degree of aging distribution. Therefore, in this paper, first, the nonlinear state space model for the single battery cell is established, and the nonlinear-model-based robust state observer is proposed. Compared with the KALMAN-filter-based optimal state estimation method, the proposed state of charge estimating method shows accurate state of charge estimating performance under the case that there exist parameter variations and model uncertainties. Second, the full-state-based active balancing control command generating method is proposed. By means of the proposed nonlinear battery cell model and distributed consensus protocol, it is possible to simultaneously generate balancing control command for all series connected battery cells. Based on the capacitance and internal resistance relationship with respect to the aging level, this can suppress the further charge imbalances resulted from the engaged current in advance. In this paper, the effectiveness of the proposed method are verified through simulations and experiments of various current cycles. Since the proposed method can be applied regardless of the type and usage of the battery cells, it can be applied to any fields requiring accurate battery state estimation and charge balancing control, in addition to the electric vehicles and mid-to-large-sized energy storage devices.