ParSweeper: Mitigating Model Crashes and Streamlining Parameter Sweeping for Improved Physics-Based Modelling of Lithium-Ion Batteries

Authors:
Preprint
Publisher website
Google Scholar
Abstract
The physics-based model of lithium-ion batteries (LIBs) is a powerful tool in designing LIBs. However, the model needs to be validated by experimental data. This can be done by estimating the model parameters using multiple experimental techniques, then fine tuning the parameters. Many papers in the literature have used this approach, but the model fitting process is not normally described in the papers. This seems to be easy but there can be all kinds of errors in solving the PDEs. Running ageing simulations are even more difficult, as the model can crash or get stuck. All kinds of errors hinder us from effectively parametrising the ageing model. To cope with these difficulties, we have created ParSweeper, a Python-based script aiming to carry out efficient parameter sweeping for the physics-based model of lithium-ion batteries. This script can deal with lots of errors that can occur during modelling. Moreover, some strategies are used to accelerate the simulation, which enable us to generate lots of modelling results and therefore fit the experimental results more efficiently.

This publication has 22 references indexed in Scilit: