An improved lithium-ion battery model with temperature prediction considering entropy

Abstract

This paper presents a system-level simulation model of a commercially available lithium-ion battery considering temperature prediction. For the 18650 battery modeling, four equations are derived based on a typical battery equivalent circuit to characterize the discharge behaviors. Current- and temperature- dependencies, as well as the internal resistance variations are taken into account to improve the model accuracy. In our extensive simulations, the calculated battery terminal voltage as a function of the state of discharge matches well with the manufacturer's data for the 18650 battery. The original model for 18650 is properly modified to characterize the Ultralife UBBL 10 battery package assembled by cylindrical 18650 batteries. The model uses a simplified equivalent circuit and modified correction factors to improve the precision. Compared with our measurement data on UBBL 10, the simulated results using our proposed model shows small deviations in both charging and discharging characteristics. Moreover, due to the considerations of the internal resistance changes and entropic heat in the thermal dynamic description, the model is able to predict the battery temperature to circumvent the safety concerns on undesirable temperature rise.