Enhanced State and Parameter Estimation within Reconfigurable Battery Systems for Electric Vehicles

Today's electric vehicles commonly comprise battery management systems in which model-based functionalities are applied. To keep the incorporated models up to date state and parameter estimation is crucial, but only with sufficient excitation adequate estimates are achievable. In recent years, reconfigurable battery systems became more and more prominent in research, whereby switches allow to change the power path reversibly during operation. Within this work, the impact of switching actions of reconfigurable battery systems on state and parameter estimation is investigated. For this purpose, suitable switching patterns and strategies are introduced and their benefit on estimation accuracy is evaluated based on validation experiments. A reconfigurable battery pack consisting of twelve lithium-ion cells in mixed topology is excited by different current profiles whereby varying switching actions are applied. Results show that switching operation significantly improves the estimation accuracy, particularly in cases where the native current excitation is of low dynamics. Furthermore, a suitable threshold for situation-based switching is determined which uses the Goertzel algorithm to assess the current excitation.