Battery Modeling for Multi-cell Battery Pack in Electric VehiclesUsing Bias Correction Method
Abstract
The inconsistency between lithium-ion polymer battery (LiPB) cells impacts the power, durability and safety of the battery pack used in electric vehicles. Therefore, it is necessary to estimate the battery performance with good accuracy. In order to get an average cell model to represent the performance for every single cell in battery pack, a method for battery modeling using bias correction technique was proposed. In this method, the equivalent circuit model (ECM) was used as the basic model, and a novel model bias function considering the polarization effect of battery was proposed to correct the basic model. Furthermore, a Back-Propagation neural network based uncertainty quantification algorithm was proposed for constructing response surface approximate model (RSAM) of model bias function. Finally, 6 single cells in a small LiPB pack were used to verify and evaluate the proposed method. The results indicated that the proposed average cell model could be applied to every single cell in battery pack and has realized accurate terminal voltage prediction.
Keywords
Battery modeling, uncertainty, model bias correction, BP neural network
DOI
10.12783/dteees/iceee2018/27858
10.12783/dteees/iceee2018/27858
Refbacks
- There are currently no refbacks.