Abstract

Abstract—The estimation of temperature inside battery cells requires accurate information about the cooling conditions even when the temperature of the battery surface is measured. This paper presents a novel approach of estimating temperature dis-tribution inside cylindrical batteries under unknown convective cooling conditions. A computationally efficient thermal model is first developed using a polynomial approximation of the temperature profile inside the battery cell. The Dual Extended Kalman Filter (DEKF) is then applied for the identification of the convection coefficient and the estimation of temperature inside the battery. In the proposed modeling approach, the thermal properties are represented by volume averaged lumped values with uniformly distributed heat generation. The model is parameterized and validated using experimental data from a 2.3 Ah 26650 Lithium-Iron-Phosphate (LFP) battery cell with a forced-air convective cooling during hybrid electric vehicle (HEV) drive cycles. Experimental results show that the proposed DEKF-based estimation method can provide an accurate prediction of core temperature under unknown cooling condition by measuring the cell current, voltage, and surface and ambient temperature. The accuracy is such that the scheme cam also be used for fault detection of a cooling system malfunction.