Abstract
For high power Li-ion batteries, an important approach to improve the accuracy of modeling and algorithm development is to consider the current dependence of internal resistance, especially for large current applications in mild/median hybrid electric vehicles (MHEV). For the first time, the work has experimentally captured the decrease of internal resistance at an increasing current of up to the C-rate of 25 and developed an equivalent circuit model (ECM) with current dependent parameters. The model is integrated to extended Kalman filter (EKF) to improve SOC estimation, which is validated by experimental data collected in dynamic stress testing (DST). Results show that EKF with current dependent parameters is capable of estimating SOC with a higher accuracy when it is compared to EKF without current dependent parameters.
Highlights
All of these findings suggest that the current dependence of direct current resistance (DCR) can no longer be neglected when the current has large variations, especially for Li-ion batteries in median Hybrid electric vehicles (HEV) (MHEV)
The resultant 2D look-up tables are presented as the contour plots in Figure 6, where the scaling factor at each current is determined by the current dependence obtained from where R1,dis, R2,dis, R1,ch, and R2,ch change with state of charge (SOC) and current
This work has verified the significance of current dependent parameters in Li-ion battery models and the advancement of its application in extended Kalman filter (EKF)-based SOC estimations
Summary
Hybrid electric vehicles (HEV) are efficient in improving fuel economy and reducing emissions. Existed electrochemical models are more sophisticated than ECMs, current dependence is still not considered in most of the work due to assumptions, such as the linearized Butler-Volmer equation As elaborated, when it comes to MHEV applications where the current has larger variations (1 C–30 C), the dependence of DCR on current becomes obvious. Only Waag et al considered the current dependence of DCR [6] They developed a first order ECM with current dependent charge transfer resistance, which is implemented it to an on-line parameter identification algorithm, a relatively simple feedback strategy. To the best of our knowledge, no work can be found in the literature that has integrated a battery model with current dependent parameters into advanced filtering algorithms, such as EKF
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
