Real-World Aging Prediction of a Lithium-Ion Battery Using a Simulation-Driven Approach

<div class="section abstract"><div class="htmlview paragraph">A large increase in GHG emissions has led to a substantial increase in EV adoption. Due to its complexity, predicting the states of LIB remains to be a roadblock for mass adoption. Furthermore, the ability to predict the performance of an EV through its lifetime continues to be a difficult task. The following work provides how a detailed electro-thermal P2D battery model, GT-AutoLion1D, can be implemented along with a 1D vehicle model to predict how the system will age over 40 weeks of operation. The battery is calibrated using experimental data and is capable of predicting performance and aging. It considers aging mechanisms like solid electrolyte interphase (SEI) layer growth, active material isolation (AMI), and SEI cracking. It is also coupled with a lumped thermal model. The 1D vehicle model considers aerodynamic, rolling resistance, driveline inefficiency, motor-inverter losses, battery resistive losses and auxiliaries. The results showed that simulation is over 30000 times faster than real time and the capacity decreased over 7% assuming a recurrent weekly routine and charging pattern.</div></div>