A0469
Title: Bayesian learning lithium-ion open circuit voltage curve via state-space model
Authors: Tomas Iesmantas - Kaunas University of Technology (Lithuania) [presenting]
Robertas Alzbutas - Kaunas University of Technology, Lithuanian Energy Institute (Lithuania)
Abstract: The state of charge (SoC) is an indicator of the remaining battery charge and must be continuously monitored by the battery management system. However, SoC cannot be directly measured while the battery is in use. Estimating the SoC, particularly the open-circuit voltage curve, which defines the relationship between SoC and open-circuit voltage, remains a challenge. Most current techniques rely on laboratory measurements and the OCV curve derived from those measurements. However, removing the battery (e.g., from an electric vehicle) for testing is impractical for most applications. A novel application of a state-space model is presented for estimating the OCV curve based solely on current and voltage data measured during typical battery use, eliminating the need for laboratory testing. To the authors' knowledge, this is the first demonstration of estimating the OCV curve using only voltage and current measurements obtained within the context of real-world battery usage. To achieve this, the battery is modeled using an equivalent circuit model, where the SoC-OCV curve is represented by a parametric nonlinear function. The unknown parameters of the model are estimated via a Bayesian inference framework implemented using the particle MCMC algorithm. Using datasets from real batteries operating under varying workloads, the approach and its accuracy in estimating is demonstrated not only the SoC-OCV curve but also other parameters within the equivalent circuit model.
