Optimal parameter estimation of battery model for pivotal automotive battery management system

The battery management system (BMS) is an integral part of a battery electric vehicle (BEV). To ensure the optimal performance of the battery, BMS should measure estimation battery parameters and battery capacity over battery life accurately. The traditional procedure for evaluation of battery parameters are time-consuming, high-priced and demand high computational power. A methodology based on heuristic optimization techniques has been implemented to overcome this problem. The proposed model was tested using six different state-of-the-art heuristic optimization approaches. Evaluation of parameters has been performed by optimally resembling predicted voltage curve's from model to curve acquired from manufacturers catalog. The practicability of particular optimization approaches is assessed by the precision of estimated model and convergence rate of prediction. Investigation showed that Differential Evolution (DE) and Teaching Learning Based Optimization (TLBO) algorithms demonstrate sufficient accuracy amongst heuristic optimizers for parameter evaluation of battery. Further analysis showed that DE algorithm produced the most consistent results with high convergence rate for accurate estimation of the parameter.