A System Identification Approach to Estimate Lithium-Ion Battery Entropy Coefficients

Improving lithium-ion battery thermal simulations requires identification of all the relevant heat source terms. While Joule heating is always considered, reversible heat is important for mild to low C-rates, but is often ignored. Characterising reversible heat requires determining the cell's entropy coefficient as a function of the state-of-charge, which is experimentally time-consuming. In this paper, we take a dynamic view point between the cell open-circuit-voltage and cell temperature to arrive at the entropy coefficient. Called the kernel-based method, a system identification procedure in the frequency domain is developed to estimate the entropy coefficient at each state-of-charge (SoC). The methodology is validated against the potentiometric method, which is considered as the reference procedure. The entropy coefficient was estimated over 21 SoC points and compared against the potentiometric approach and showed very good agreement across the entire SoC interval with a 57% reduction in time. All the experimental data and the Matlab code for data processing and reproducing of the results is made available together with the paper.