Fast multilayer temperature distribution estimation for lithium-ion battery pack

Abstract

Fast and accurate temperature estimation is crucial for ensuring battery packs' safety and operation performance. However, the full-scale online temperature estimation is still challenging. In this work, a novel reduced-order multi-physics model for a lithium-ion battery pack is proposed for real-time temperature distribution estimation, containing the distributed equivalent circuit model, the three-heat-source thermal model, and the flow resistance network model. The proposed model is conducted on a direct contact liquid-cooled battery pack composed of three modules connected in series. An online parameterization methodology with a closed loop observer is designed, and the key parameters can be automatically identified and corrected. The validation results suggest that the multilayer temperature distribution of cell, module, and pack levels can be commendably described under both steady and transient conditions, where the maximum error can be controlled within 2.8 °C. Besides, the temperature variation of the coolant can be estimated during the operation. The proposed model shows excellent potential in onboard temperature estimation with tens of milliseconds for each temperature update.