Abstract
An open-source Python package named PyBatteryID has been developed for data-driven identification of battery models using the linear parameter-varying (LPV) modelling framework. The package provides a convenient way to specify the model dependence on relevant signals affecting the battery behaviour, such as the battery state-of-charge (SOC), current magnitude, current direction, and temperature. Furthermore, multiple optimisation methods to estimate the model parameters can be employed in a sequence, for instance, the user may employ the least absolute shrinkage and selection operator (LASSO) followed by the ridge-regression method to perform the variable selection and parameter estimation steps, respectively. Moreover, the package allows the generation of suitable current and temperature profiles to obtain an informative dataset for the model identification procedure. In addition, the package provides various utilities to streamline the model identification procedure, such as (i) analysing experimental datasets, (ii) plotting desired signals, (iii) exporting the identified models in a cross-compatible format, and more. The code, documentation and examples related to the package can be found at https://github.com/tue-battery/PyBatteryID.