A systematic and low-complexity multi-state estimation framework for series-connected lithium-ion battery pack under passive balance control. (April 2022)

• Record Type: Journal Article
• Authors: Zhang, Shuzhi; Peng, Nian; Lu, Haibin; Li, Rui; Zhang, Xiongwen
• Is Part Of: Journal of energy storage. Volume 48(2022)
• Publication Date: 2022-04
• Subjects: Series-connected lithium-ion battery pack -- State-of-charge (SOC) -- State-of-health (SOH) -- Cell SOC inconsistences -- Passive balance control; Energy storage -- Periodicals; Energy storage -- Research -- Periodicals; 621.3126
• DOI: 10.1016/j.est.2022.103989
• Languages: English
• ISSNs: 2352-152X

Abstract

Highlights: A systematic, low-complexity multi-state estimation method for battery pack is proposed. Pack's states estimation is simplified to that of the "representative cell". A rapid and reliable selection method for "representative cell" is developed. A variable multi time-scale based method is used to co-estimate pack's SOC and SOH. A second-order EKF is designed to monitor cell inconsistence in macro time-scale. Abstract: To reduce computational burden and achieve accurate states estimation, this paper presents a systematic and low-complexity multi-state estimation framework for series-connected lithium-ion battery pack under passive balance control, including pack state-of-charge (SOC), state-of-health (SOH) and cell SOC inconsistences estimation. Firstly, through SOC and SOH calculation simplification of battery pack under passive balance control, "representative cell" is determined among all in-pack cells to reflect battery pack's behavior by a rapid and reliable selection method. Secondly, a variable multi time-scale based framework is further applied to co-estimate SOC and SOH of the selected "representative cell", which are also the co-estimation results of battery pack. Subsequently, with the estimated SOC of "representative cell", a second-order extended Kalman filter is designed to realize cell SOC inconsistences tracking in macro time-scale, and the non-representative cells' SOC can be further calculated. The validation results through sophisticated driving simulation show that both mean-absolute-error (MAE) and root-mean-square-error (RMSE) between battery pack's real SOC and representative cell's estimated SOC are below 1%, and the relative error band between battery pack's precise capacity and representative cell's calculated capacity is between 0 and 1.5%. Moreover, based on the monitored cell SOC inconsistences, both MAE and RMSE between the non-representative cells' estimated SOC and reference SOC can be roughly limited below 3%. Graphical abstract: Image, graphical abstract