Parameter Estimation for a Lithium-Ion Battery From Chassis Dynamometer Tests

Abstract

Simulation techniques are extensively used in vehicle performance evaluations, particularly for advanced vehicles such as hybrid electric vehicles ( HEVs), plug-in HEVs, and battery electric vehicles. It is necessary that the parameters used in simulation models are estimated properly so that the simulations can produce results that are close to real-world behaviors. This paper suggests methodologies for estimating parameters for a lithium-ion ( Li-ion) battery model by analyzing test data obtained from chassis dynamometer tests. A representative model based on a first-order equivalent circuit is used for the battery, and four main parameters of the model-source voltage, internal resistance, polarization resistance, and polarization capacitance-are obtained by characterizing the test results. The model is validated with the test data by applying the estimated parameters in the model, and the validation results show that the battery output voltage is calculated by the simulation model very well; 70% of simulations produce the output voltage within 1% of the root-mean-square error, as compared with the test data. Although the methodology cannot replace a conventional process that estimates the battery parameters from dedicated tests, the approach would be very feasible and save the effort and time needed to develop simulation models for Li-ion batteries.