Thermal and state-of-charge (SOC) imbalance is well known to cause non-uniform ageing in batteries. This paper presents the electro-thermal control of a multi-level converter (MLC) based modular battery to address this issue. The modular battery provides a large redundancy in synthesizing terminal voltage, which gives extra degrees-of-freedom in control on cell level. There are multiple tightly coupled control objectives including the simultaneous thermal and SOC balancing as well as battery terminal voltage control. The main purpose of this paper is to devise an electro-thermal control scheme for cases where full future driving information is not accessible. The control scheme is based on decomposition of controller into two orthogonal components, one for voltage control and the other for balancing control. This problem decomposition enables the application of constrained linear quadratic model predictive control scheme to solve the balancing problem elegantly. The control scheme is thoroughly evaluated through simulations of a four cell modular battery. The results show that a rather short prediction horizon is sufficient to achieve robust control performance.

No abstract

This paper investigates two alternative approaches to precompute a discharge strategy for the main commuter route of a plug-in hybrid electric vehicle. The first approach is based on the idea of computing a state of charge reference trajectory by solving a convex program; while the second approach utilizes dynamic programming to determine an optimal cost-to-go function. During real-time operation the torque split is decided by an equivalent consumption minimization strategy where the main difference between the two approaches is how the equivalence factor is determined. With the first approach it is adapted to track the state of charge reference trajectory and in the second approach it is given by the partial derivative of the cost-to-go function with respect to state of charge. To evaluate the two approaches a simulation study is performed in the dynamic vehicle modelling software Autonomie using logged commuter driving data. The simulation results indicate no clear difference between the two approaches in terms of fuel economy and battery usage. Both approaches are, however, significantly better than a charge depleting charge sustaining discharge strategy.

No abstract