Research On Charging Strategy Of LiFePO4 Battery Based On SOC

Electric vehicles have the advantage of zero emissions during their usage phase,and have received a lot of research and application.However,range anxiety has always been a major obstacle in the promotion of electric vehicles,and it can be addressed primarily through fast charging strategies.Based on the analysis of relevant literature on power battery charging strategies,the current research trend focuses on charging strategies that predict the optimal charging current using either an equivalent circuit model or an electrochemical model.The equivalent circuit model is computationally simple but cannot accurately determine the safe charging range of the battery,while the electrochemical model accurately characterizes the internal reaction mechanism of the battery but has higher order and is not suitable for iterative optimization calculations.In this thesis,we conducted research on the charging strategy of power batteries for pure electric vehicles,improved the equivalent circuit model,and proposed a phased optimization charging strategy based on state of charge（SOC）for LiFePO₄ batteries throughout the entire charging process.The main achievements of this paper are as follows:（1）To address the voltage hysteresis phenomenon observed in the constant current charge-discharge process of LiFePO₄ batteries,a second-order RC equivalent circuit model based on the single-state hysteresis model was proposed.The parameter identification of the second-order RC equivalent circuit model was conducted using experimental data from the Hybrid Pulse Power Characteristic（HPPC）test,and the hysteresis voltage was employed to describe the difference between the second-order RC equivalent circuit model and the measured results,thereby improving the estimation accuracy of the model.Experimental results under multiple-step constant current charging conditions showed that the deviation between the simulation results and the measured results was within 1.8%,effectively simulating the voltage hysteresis phenomenon of the battery.The pseudo-two-dimensional electrochemical model（P2D model）was adopted to determine the relationship between the charging current and charging time during the lithium battery charging process when there is no graphite anode side reaction rate.A P2D model was built in COMSOL to verify its effectiveness.Under the condition of 0.5C constant current charging,the results showed that the deviation between the simulation results and the measured results was within 2.1%,providing a good reflection of the battery’s charging characteristics.（2）To mitigate the impact of environmental temperature and capacity degradation on battery state estimation and improve state of charge（SOC）estimation accuracy,a temperature and capacity-corrected Unscented Kalman Filter algorithm（TQ-UKF algorithm）was proposed.The effectiveness of the TQ-UKF algorithm was evaluated using a second-order RC equivalent circuit model based on the single-state hysteresis model.Simulation results under a 0.5C constant current charging condition with an initial SOC of 80%demonstrated that the proposed algorithm outperformed the Extended Kalman Filter（EKF）algorithm in terms of faster convergence,higher accuracy,and better correction effects under the same initial conditions.Simulation results under FUDS and DST conditions indicated that the TQ-UKF algorithm exhibited superior estimation accuracy and computational time compared to the EKF algorithm.（3）A staged optimization charging strategy based on the state of charge（SOC）of LiFePO₄ batteries throughout the entire charging process is proposed.This charging strategy consists of four stages,retaining the pre-charge stage of the conventional threestage charging strategy.Within the SOC range of 5%-70%,a traditional positive pulse charging method is employed to improve charging efficiency.Within the SOC range of 70%-90%,the safe charging current is determined based on the P2D model with the condition that no side reactions occur on the graphite anode surface.In the equivalent circuit model,a genetic algorithm is used for iterative optimization,targeting charging loss power and charging time,to determine the charging current at 1%SOC intervals.This variable current positive pulse charging approach reduces the average charging current and eliminates the occurrence of side reactions on the graphite anode.Depending on the optimization objective,the charging strategy can be classified as a comprehensive low-consumption charging strategy or a comprehensive fast-charging strategy.The fourth stage still employs a constant voltage trickle charging method.A charging test platform was established to assess the effectiveness of the proposed charging strategy.The experimental results demonstrate that both the comprehensive low-consumption optimization charging strategy and the comprehensive fast-charging strategy have shorter charging times and higher energy input compared to the constant current-constant voltage（CCCV）charging strategy.They also yield higher energy input than the positive pulse charging strategy with slightly longer durations and lower maximum temperature rise.