Research On Prediction Of State-of-charge Oriented Li-ion Battery

The energy crisis and the economic crisis have led to the rapiddevelopment of electric vehicles. As a source of energy of electricvehicles, accurate estimation of the state-of-charge (SOC) of Lithiumbatteries is always important for battery management system (BMS) tothe full utilization for energy, the better user experience use of electricvehicle as well as the accurate estimation of car’s Mileage.Start from the characteristics of the lithium battery, this paperanalysis the main factors affecting the SOC. Then this paper comparedseveral estimation algorithms, and proposed an improved estimationalgorithm based on the UKF estimation algorithm, which integratedampere hour counting method as well as the open circuit voltage method,it also take full account of the impact from discharge current and ambienttemperature, in order to reduce the noise from drift current and hysteresiseffect, a better solution was proposed, the results of simulation showsthat the improved algorithm has better effect.The system was designed on the basis of accurate estimation of theSOC. This system includes the lithium battery data acquisition, CAN busdata communication, data storage, data processing, GPS automaticallyfinding, with these modules to complete the battery voltage, current,temperature, and other real-time data acquisition, and SOC estimation ofthe real-time calculation, At the same time, this system successfullyintroduced UCOS-II Real Time Operating System, greatly improved thereal-time and reliability of the system. As for the phenomenon that thereare not enough charging stations at this stage, proposed an innovativesolution, which introduced by the GPS/GPRS module, the GPSautomatically searches for the nearby charging stations when the batteryis low, and to establish a communications link with the most convenientcharging station, enabling online booking for charging service, whichgreatly reducing the waiting time of the drivers.