State-of-Charge modelling and State-of-Health determination of batteries for all-electric vehicles

The aim of this EPSRC funded project is to develop accurate battery modelling strategies so that the State-of-Charge (SoC) and State-of-Health (SoH) of a battery system for an all-electric vehicle can be monitored allowing optimised charging scenarios to be implemented, thereby extending battery lifetime.

With the increasing drive towards the use of X-by-wire control systems, and the desire for passenger comfort (i.e. air-conditioning/in-vehicle entertainment systems) the large transient power demands that are characteristic of such all-electric vehicles are becoming of utmost importance to the operational safety and reliability of the primary battery source, thereby necessitating the use of advanced battery State-of-Function (SoF) prediction techniques. Battery circuit models such as the Randles´ model below are traditionally used with static circuit parameters to estimate the voltage across the main charge-store (Cb) which is linked to SoC.

Randles’ battery model and its mapped equivalent circuit

For improved accuracy, a new battery circuit model has been proposed that is `mapped´ directly from the Randles´ model, allowing improved state estimation techniques. Moreover, the application of Subspace parameter estimation algorithms has been shown to provide excellent estimates of the dynamic change in circuit parameter values, which, when combined with state-observer techniques, allows excellent estimation of the variation in the voltage associated with SoC. Using this information, estimation of the variation in the magnitude of the capacitance associated with the main charge store has been achieved, and a relationship between this parameter and the total available capacity of the battery (i.e. SoH) has been found.

Estimation of model parameters linked with SoC and SoH for a 48Ahr VRLA battery