Optimization Tool for the Characterization of Electric Vehicle Battery Packs

Abstract

The use of vehicle scale data for the parameter characterization of electric vehicle battery packs is a challenging topic. This paper describes the implementation of a design tool that carries out both simulated annealing and genetic optimization of model parameters for a modified Nernst Open Circuit Voltage Battery model (K0, K1, K2, K3), concurrently with the dynamic transient model parameters (R0,R1,RP,C1,CP) and pack level parameters including the initial state of charge and capacity (SOCINIT and AH). This paper describes the model for the battery pack implemented in the Saber simulator and the optimization tool (written in TCL-TK) also integrated with the Saber simulator. Results were collected from rolling road tests of of a BMW i8 to validate the fidelity of the model.