Deriving representative load profiles for battery simulation: The use case of system imbalances induced by cell-to-cell variations

Abstract

Intrinsic battery cell parameter variations and their impact on the overall system behavior is topic of many experimental and simulative research papers. Usually, studies use short operational profiles to measure the extent of behavior inhomogeneities. However, load profile choices highly influence the extent of differences between each cell’s behavior. Simultaneously, actual vehicle usage history is high-dimensional. Until now, studies lack the framework to derive representative load profiles to reduce complexity for simulations measuring cell-to-cell variation effects. We propose the usage of micro-trip representations to synthesize suitable load profiles. By augmenting State-of-the-Art methods with a battery-specific autoregressive generation algorithm, we can compress high-dimensional vehicle history information into significantly shorter load profiles. During synthetization of trip sequences, signal consistency and representativeness are ensured. The obtained framework enables realistic battery design choices.