A reduced-order thermal model of battery thermal management system for online applications based on proper orthogonal decomposition and Galerkin projection method

The temperature of lithium-ion batteries (LIBs) manifests significant hysteresis effects that substantially impede precise temperature regulation within battery thermal management systems (BTMS). The application of model predictive control (MPC) has been identified as a potentially effective strategy for mitigating thermal hysteresis. However, the existing thermal models for LIBs lack the requisite accuracy and computational efficiency for effective implementation in online MPC frameworks. In this paper, a reduced-order thermal model (ROTM) of BTMS is established based on the proper orthogonal decomposition (POD) and Galerkin projection. Firstly, a finite element model (FEM) of three parallel and eight series air-cooling battery module is constructed in aim of generating original data. The basis vectors of the flow and temperature fields of the battery module are subsequently extracted from the original data by the POD method. Finally, the Navier-Stokes equation and the Fourier's law of heat conduction are projected on the basis vectors previously described. The ROTM can thus be obtained. In comparison with the FEM, the ROTM exhibits a significantly reduced computational time and maintains adequate accuracy. The computational time for ROTM is merely one ten-thousandth of that required by FEM, whilst under 1.25 C constant-current conditions the maximum error between the two methods is less than 0.2°C.