Model-based estimation of state of charge and state of power of a lithium ion battery pack and their effects on energy management in hybrid electric vehicles

This paper presents the effect of modeling uncertainty of a lithium ion battery pack on the accuracies of state of charge (SOC) and state of power (SOP) estimates. The battery pack SOC is derived from the SOCs of all parallel cell modules in the pack, which is computed using a sequential estimation process. SOC and SOP estimates are essential for optimizing the energy management of hybrid electric vehicles (HEV). Two model types, an average model (AM) and a hysteresis model (HM), are considered here. The HM captures the hysteresis phenomenon, prominently observed in lithium ferro-phosphate cells commonly used in HEVs. This paper first demonstrates the accuracies of SOC and SOP estimates for AM and HM for laboratory experimental data. The feasibility of on-board implementation of the scheme is demonstrated through a closed-loop vehicle-level hardware-in-the-loop simulation. It is shown that estimates obtained using the HM result in an improved overall energy economy, as well as improved operational limits for batteries, and the underlying reasons, are revealed from the improved choice of operating points enabled by the improved estimates.