On-board diagnostics of Li-Ion battery packs for electric vehicles using a combination of spectroscopy and identification methods

An accurate parameter estimation becomes one of most important goals of modern battery management systems in electrical vehicles and smart grids. Actually, only state of the charge is used to analyse the battery conditions. However, information about internal resistance and capacity as well as open circuit voltage allows optimizing the battery operation mode either while charging or discharging and therefore increase the lifetime of the battery. Hence the battery parameters like capacity or the internal resistance are not constant and are changing due to environmental influences like temperature or cell ageing, the need for online identification methods becomes more and more important. The key idea of the proposed paper lies in the application of impedance spectroscopy methods without injection of an additional current and using the transient signals during operation of an electric vehicle. The accuracy of estimation with three typical dynamic load profiles is analysed. It is shown that the open circuit voltage impacts the results of an impedance spectroscopy in ultra-low frequency range. To compensate this impact and to enhance the estimation accuracy the classical impedance spectroscopy is combined with an open circuit voltage estimation algorithm in time domain. The simulation results as well as experiments with a real battery pack show the correctness of a proposed approach.