How do you identify the predicted SOE for lithium-ion batteries based on EV operating conditions?

Liu Weilong, Wang Lifang, Wang Liye, Power Electronics and Power Transmission Key Laboratory of Chinese Academy of Sciences, researchers at the University of Chinese Academy of Sciences, pointed out in 2018 the first phase of the proceedings of electro technics that energy state (SOE) is the electric car power battery important status indicators, which directly affects the range of electric cars, affected by the working condition of the electric car significantly.

To estimate, based on the working condition of electric cars, SOE to SOE estimation method, mode recognition algorithm, launches the research condition of prediction algorithm, based on the model battery residual energy state (SOR) estimation method, based on the recognition algorithm of running condition of information entropy theory, the application of the theory of Markov chain prediction algorithm, build the driving electric vehicle system model is established, the simulation for electric vehicle driving modes corresponding battery prediction, implementation based on the working condition of electric vehicle recognition and prediction of SOE is estimated. The simulation results demonstrate the effectiveness of the proposed method.

At present, the electric car has become a hot research and automobile industry, but the "range anxiety" problem of flood limit its development [1], the charged State (State - of - Charge, SOC) as indicators of power battery remaining power of the parameters is widely used in electric cars, has to alert users to timely charging function.

But as a result of power battery under the working condition of discharge voltage of a declining trend, makes the power battery SOC in larger range of energy (W * h) supply capacity is reduced, thus in the process of the electric car run SOC index showed accelerate downward trend. SOC as charging indicator parameters, easy to cause the charging time of misjudgment, caused inconvenience to the electric car users.

Power battery energy state (State of Energy, SOE) as the electric car W x h unit scale parameter, the proportion of the residual energy is a direct description of battery power supply capacity, as the electric car user charging indicator parameter.

At present, the battery SOE estimation method for the application of SOC estimation algorithm usually extend, battery is estimated based on the mapping relationship of the SOC and SOE SOE [2, 3], is obtained by SOC conversion of SOE is not accurate, however, this is due to the power battery in practice, with the change of the vehicle traffic conditions will lead to different degree of energy loss, make the change in the actual supply energy to generate power battery. Therefore, based on the prediction of future vehicle driving cycle power battery remaining available energy states (SOE) has more practical significance.

History and future traffic condition prediction is the premise of traffic condition recognition, this paper carried out based on the working condition of electric vehicle recognition lithium ion batteries SOE estimation research.

Battery model is the basis of estimate the battery status, can be divided into main electrochemical mechanism model (4, 5), experience model [6, 7] and [8, 9] three kinds of equivalent circuit model. Among them, because of the equivalent circuit model is easy to analysis application, good commonality of advantages, has achieved a wide range of applications.

SOE estimation algorithm based on the battery equivalent circuit model, based on the model of the battery residual energy state (State of Residual Energy, SOR) estimation algorithm, based on the recognition algorithm of running condition of information entropy theory and prediction algorithm based on the theory of Markov chain driving cycle; Build the model of electric vehicle, get battery predict conditions, to estimate based on the working condition of identification and prediction of SOE; The results of simulation analysis, verify the accuracy of the algorithm.


Figure 1 lithium ion battery equivalent circuit model structure

Figure 2 electric vehicle system model

Figure 3 a battery SOE estimation algorithm

Conclusion

This paper puts forward a kind of based on the working condition of electric vehicle recognition and prediction of lithium ion battery SOE estimation algorithm.

First of all, based on the theory of electrode impedance spectrum, different order battery equivalent circuit model is constructed, and the GA algorithm is applied to implement the model parameter identification;

Secondly, based on AUKF, puts forward the SOR estimation algorithm based on different order battery model, the simulation results show that the algorithm can achieve rapid convergence;

Again, FCMIE algorithm, realizes the identification of electric vehicle driving cycle, and the condition of prediction algorithm based on Markov chain theory, the electric car forecast running condition, the simulation results show that the algorithm to predict conditions effect is good;

Finally, the integration of different order battery equivalent circuit model in electric vehicle system model, the simulation to obtain the corresponding battery electric vehicles to predict condition of working condition of prediction, the battery of battery energy loss in the cases of calculation, and then to estimate the battery of SOE.

SOE estimates party error within 2.45%, the proposed battery SOE estimation algorithm has a good effect.

The page contains the contents of the machine translation.