An Empirical Aging Model for Lithium-Ion Battery and Validation Using Real-Life Driving Scenarios

Lithium-ion batteries (LIBs) have been widely used as the energy storage system in plug-in hybrid electric vehicles (PHEVs) and battery electric vehicles (BEVs) due to their high power and energy density and long cycle life compared to other chemistries. However, LIBs are sensitive to operating conditions, including temperature, current demand and surface pressure of the cell. One very well understood phenomenon of lithium-ion battery is the reduction in charge capacity over time due to cycling and storage commonly known as capacity fade. Considering the need for predicting the behavior of an aged cell and the need for estimating battery useful life for warranty purpose, it is crucial to predict the capacity fade with reasonable accuracy. To accommodate this need, a novel cell level empirical aging model is built based on storage tests and cycle tests. The storage test captures the calendar aging of the lithium-ion cell while the cycle test estimates the cycle aging of the cell. In the proposed model the calendar aging is represented as a function of time, storage temperature and state-of-charge. On the other hand, the cycle aging is represented as a function of energy throughput, cell temperature and charge/discharge rate. The cycle and calendar aging models are then combined to formulate an empirical cell-level aging model. The performance of the cell-level empirical aging model is validated by comparing with the aging test data for different driving scenarios and driver habits in different geographical locations of the US. The aging prediction from the empirical aging model shows very good agreement with the test data with a maximum normalized root mean squared error (RMSE) of 0.6%.