Model-Based Parameter Identification of Healthy and Aged Li-ion Batteries for Electric Vehicle Applications

Electric vehicles are receiving considerable attention because they offer a more efficient and sustainable transportation alternative compared to conventional fossil-fuel powered vehicles. Since the battery pack represents the primary energy storage component in an electric vehicle powertrain, it requires accurate monitoring and control. In order to effectively estimate the battery pack critical parameters such as the battery state of charge (SOC), state of health (SOH), and remaining capacity, a high-fidelity battery model is needed as part of a robust SOC estimation strategy. As the battery degrades, model parameters significantly change, and this model needs to account for all operating conditions throughout the battery's lifespan. For effective battery management system design, it is critical that the physical model adapts to parameter changes due to aging. In this paper, we present an effective method for offline battery model parameter estimation at various battery states of health. An equivalent circuit with one voltage source, one resistance in series, and several RC pairs modeled the battery charging and discharging dynamics throughout the lifespan of the battery. Accelerated aging tests using real-world driving cycles simulated battery usage. Three lithium nickel-manganese-cobalt oxide (LiNiMnCoO2) cells were tested at temperatures between 35 degrees C and 40 degrees C, with interruptions at every 5% capacity degradation to run reference performance tests for tracking changes in the battery model parameters. The equivalent circuit-based model was validated using real-world driving cycles. The parameter estimation procedure resulted in an efficient model that keeps track of the battery evolution as it ages.