Deriving physics-informed system-scale models for lithium-ion batteries  

Project description

Rechargeable batteries and other energy storage technologies are key elements for reaching a sustainable carbon-free energy market. The increasing usage of batteries needs to be supported by more accurate and faster mathematical models to be integrated in online control units. State-of-the-art battery models for control applications are typically Ordinary Differential Equations (ODEs), built by analogy with Equivalent Electrical Circuits. Despite their computational simplicity, these models fail to capture the physical meaning and the dependence of the parameters on the underlying chemical-physical processes. Another mostly unresolved issue is the simulation of irreversible and complex non-linear phenomena such as fast (dis)charge and degradation.

In this project, we aim to derive new efficient system-scale models, as an alternative to classical equivalent circuit models, to enable the fast, yet accurate, simulation of short and long-term behaviour of lithium-ion cells. By projecting the PDE continuum models of the porous electrode theory onto a low-dimensional manifold, we aim to derive reduced equations that can retain the interesting features of the full model (e.g., memory, non-linearities). Since a full first-principle characterisation of these models is out of question, the parameter identifiability from real data will also be investigated.

This project will see the participation of several academic and industrial partners in the UK and overseas.