Optimal model based design of experiments applied to high current rate battery cells

In this paper, a model based design of experiments algorithm for creating optimal test sequences for identification of battery cell models is presented. Based on an optimality criterion, which is determined using the Fisher information matrix, the test sequence is optimized by a gradient based algorithm. Real load cycle analysis defines the current ranges of interest. By adding constant charge and discharge currents, the test sequence is forced to pass the whole range of state of charge. The constant currents added to the test sequence, furthermore, take relaxation and voltage hysteresis effects into account. The results show that the optimized test sequence improves the high dynamic battery cell model significantly.