Research On SOC And Battery Capacity Estimation Of Lithium-ion Battery For Electric Vehicle

Nowadays,energy resource crisis and environmental pollution have become the main concerns of human world.Electrified transportation,as an alternative energy solution,can effectively reduce hazardous emission and mitigate dependence on fossil fuel.Electric vehicles and hybrid electric vehicles are currently the promising candidates,in which the electrical energy storage system is equipped to store and supply the complete propelling power,or just assist extra power for internal combustion engines.Lithium-ion batteries have been widely adopted for their numerous advantages,such as wide environmental temperature operation capability,high energy density,long lifespan,and large charge/ discharge current.For lithium-ion batteries,state of charge and available capacity,usually provided by battery management systems,are crucial parameters for battery electrical performance evaluation as well as vehicle control.Accurate estimation for batteries state of charge and available capacity is the necessity to guarantee proper operation of on-vehicle battery management system and also provides indispensable information to drivers.Taken a ternary lithium-ion battery as the research object,this paper carries out a series of researches on the battery state of charge estimation and capacity estimation.Firstly,considering the basic working characteristics of lithium-ion batteries,a second-order resistance capacity equivalent circuit model is selected to model the battery on the basis of comprehensive analysis of model complexity and accuracy.An adaptive forgetting factor-recursive least square method is adopted to complete the on-line identification of battery model parameters.The adaptive forgetting factor-recursive least square is an improved method derived from forgetting factor-recursive least square by adapting the forgetting factor,contributing to enhance the identification capability for dynamic parameter.Then,the state of charge estimation algorithm based on square root cubature Kalman filter for lithium-ion battery is proposed.In this algorithm,the discrete expressions of state equation and measurement equation for state of charge estimation are constructed in combination with the established battery model and the identified parameters.The performance of the proposed algorithm is evaluated and analyzed by the simulation results which is based on the experimental data under the Urban Dynamometer Driving Schedule profile and the Dynamic Stress Test profile.The experimental results reveal that compared with traditional extended Kalman filter,unscented Kalman filter and cubature Kalman filter,the proposed algorithm has the advantages of shortest execution time,the smallest error and the highest precision.Furthermore,the proposed algorithm shows strong convergence,robustness and adaptability under different initial deviations of state of charge and different operating conditions.Finally,for degradation process of battery capacity in real-world application,a model-based adaptive joint estimation algorithm of state of charge and capacity during entire lifespan is also proposed in this paper.Based on the open-circuit voltage and SOC relationship curves at different aging stages,a three-dimensional response surface with respect to the capacity,SOC and open-circuit voltage is constructed.The available battery capacity and model parameters are identified and updated by genetic algorithm.After updating the battery capacity and model parameters,combined with the square root cubature Kalman filter,the joint estimation of battery capacity and state of charge in their entire lifespan is completed.The performance of the co-estimation method is verified by experiments under different degradation states and different experimental conditions.The experimental results show that the proposed the co-estimation method is capable of estimating the battery capacity adaptively even in cases of aged batteries,which improves the precision of state of charge estimation in the whole life cycle.The proposed algorithm can effectively actualize joint estimation of state of charge and battery capacity for lithium-ion batteries during the entire lifespan with high estimation accuracy,which offers a reference for performance optimization and applications of batteries,and has valuable significance promoting the development of the key technology for power lithium-ion batteries management system.