Efficient computation of safe, fast charging protocols for multiphase lithium-ion batteries: A lithium iron phosphate case study

Abstract

Fast charging is a desirable feature for lithium-ion batteries. Charging at high currents, however, can damage the battery and accelerate aging processes. Fast charging protocols are typically computed by solving an optimization in which the cost function and constraints encode the conflicting requirements of safety and speed. A key element of the optimization is the choice of the dynamic model of the battery, with an inherent tradeoff between model accuracy and computational complexity. An oversimplified model may result in unreliable protocols, whereas a complex model may result in an optimization that is too computationally expensive to be suitable for real-time applications. This article describes an approach for embedding a complex battery model into charging optimization while having low computational cost. Multiphase Porous Electrode Theory is used to provide an accurate description of batteries characterized by multiphase materials, and the optimization is solved by transformation into mixed discrete-continuous simulation of a set of Differential–Algebraic Equations. The methodology is applied to an MPET model of commercially available Lithium Iron Phosphate batteries. Protocols based on a variety of operational constraints are computed to assess both the effectiveness of the approach, and the advantages and disadvantages of the charging protocols.