Novel Mesoscale Electrothermal Modeling for Lithium-Ion Batteries

Artikel i vetenskaplig tidskrift, 2020

This paper devises an innovative mesoscale electrothermal model for Li-ion batteries. This model manipulates the mesoscale calculation grid in finite element analysis as independent small cell sandwiches and establishes a lumped equivalent circuit model for each cell sandwich. Then, such electrical models are arranged in parallel to form a multilayer equivalent circuit to simulate electrical characteristics of a whole battery, through capturing the current and terminal voltage of each constituent cell sandwich. This modeling idea overcomes the entrenched disadvantage of heat generation models with lumped parameters, i.e., the unavailability of heat generation distribution inside a battery. Besides the current and terminal voltage, the temperature and state of charge dependent open-circuit voltage and entropy coefficient are incorporated into a Newman's heat generation model to estimate the heat generated in the calculation grid. The battery temperature distribution is eventually derived by solving the heat conduction equation with thermal conductivity as a function of the battery temperature. We leverage the developed electrothermal model to track the temperature evolution of an 18650 Li-ion battery at different ambient temperatures and discharge rates, for the first time. Experimental results demonstrate that the electrothermal model can precisely emulate the battery thermal dynamics with an average error of 0.72 °C. Moreover, a comparative study shows that the proposed model outperforms common resistance-based thermal models that do not consider the heat generation distribution and the interdependence between the battery temperature and thermal conductivity.