Tsai I.-H.Yu K.-H.Tseng A.Yen J.-Y.Fu T.-T.Huang E.JIA-YUSH YEN2019-10-222019-10-22201802533839https://www.scopus.com/inward/record.uri?eid=2-s2.0-85052100852&doi=10.1080%2f02533839.2018.1490203&partnerID=40&md5=bb975cf8da59266b38564c4d599d76a0https://scholars.lib.ntu.edu.tw/handle/123456789/427110Battery models are vital to the development of electric vehicles. Different models have been proposed over the years to describe the battery dynamics in various degree of detail. More detail comes at the cost of more computation. This paper proposes using a combination of the Kinetic Battery Model (KiBaM) and the dual capacity network model to capture both the nonlinear state of charge variation and linear transient response. After capture, we then derived a recursive formula for the online implementation of the proposed model. In the first part of this study, MATLAB® was used to build a battery model of a battery cell. The battery model simulates how the magnitude of the discharge current and state of charge influence the parameters of the model, with the results showing that it can predict the voltage response within a voltage error of 4% under dynamic loading. In the second part of the study, a model-based Kalman filter was adopted for estimating the state of charge. This algorithm was compatible with the recursive formula and could be used in conjunction with the online batter model. © 2018, © 2018 The Chinese Institute of Engineers.battery model; Kalman filter; Lithium-Ion battery; parameter estimation; SOC[SDGs]SDG7[SDGs]SDG11Battery management systems; Charging (batteries); Dynamic loads; Kalman filters; Parameter estimation; System-on-chip; Transient analysis; Battery modeling; Discharge currents; Dynamic loadings; Network modeling; On-line estimation; Online implementation; Recursive formula; Voltage response; Lithium-ion batteriesjournal article10.1080/02533839.2018.14902032-s2.0-85052100852WOS:000443901100006