Online Lithium-ion Battery State of Health Prognostics (LiBSoHP)

To ensure the battery safety and reliability in operation, it is of great importance to be able to online estimate state of health (SoH) and predict its remaining useful life (RUL) of lithium-ion batteries. Our research team gets access to a fine-grained and large-sample battery cycling data, which is, as far as we are concerned, is the largest publicly available for nominally identical commercial lithium-ion batteries cycled under controlled conditions. Based on the unique data, this proposal aims to address two emerging challenges for online lithium-ion battery SoH prognostics.  First, we will develop and train a deep learning model that is able to learn the relationship between battery capacity and partial charge/discharge. It is to decipher the complex nonlinear dependencies of the battery capacity degradation on the current, voltage, and temperature based on the thousands of battery degradation trajectories at different working conditions. Second, a deep learning model will be developed to learn the long-term internal dependencies of capacity degradation trajectories for RUL predictions as per a small set of early-cycle data. The contents perfectly fit the themes of the call by utilising real data in machine learning with the integration of knowledge- and data-driven techniques.

The total budget for this project is 300,000 SEK. 

Partner organizations

  • Volvo Group (Private, Sweden)
Start date 01/09/2020
End date 28/02/2021

Published: Fri 24 Jul 2020.