Study On The Soc Estimation Based On The Improved EKF

As the environmental problems and the energy crisis becoming more serious, the governments all over the world pay more attention on the Electric Vehicles, with the characteristic of energy conservation and environmental protection. And the electric vehicles industry becomes the development direction of the global traffic energy transformation. The power battery is the important part of the Electric Vehicles, its state and lifetime largely determine the total performance of the Electric Vehicles.SOC of the power battery is the important parameter that describe the remain energy. It provides a reference for the BMS to manage the battery, and it is also the key for the energy manage strategy. Estimating SOC exactly has a great meaning to extend the lifetime of the power battery and develop the performance of the Electric Vehicles. However, the power battery is strongly nonlinear with the complex electrochemical reaction inside. So it is difficult to establish the model and estimate the SOC of the power battery. In this paper, establishing the model and SOC estimation must be the main contents.Taking the LiFePO4 battery (3.2V/10Ah) as the object of study, based on a series of correlative experiments, such as charge and discharge with the different current rate, discharge under different environmental temperatures, HPPC circulation experiment and so on, this paper analyzes the charge-discharge feature and the temperature feature of the battery. Analyzing the existing models, the improved model based on the Thevenin equivalent circuit model is established, taking the charge and discharge direction and the discharge rate into account. This paper use the HPPC circulation experiment, adopt the least-square method to obtain the parameters at the different SOC. Then it makes use of the cftool workbox in Matlab to obtain all the parameters, considering the different SOC and the current direction. According to the fitting date, it establishes and simulates the improved model in Matlab. Compared the simulation model with the experimental data, the simulation results show that the improved equivalent circuit model has high precision, and it can simulate the dynamic characteristics of the LiFePO4 battery.In this paper, based on the established equivalent circuit model, the SOC of the LiFePO4 battery is calculated by the extended kalman filter (EKF). EKF can iterative calculate constantly, gradually approaching the true value, overcoming the shortcomings of the ampere-hour model that the initial SOC is difficult to determine. But EKF strongly depends on the model. In order to improve the accuracy of SOC estimation, taking the strongly nonlinear and the inevitable error of the equivalent circuit model into account, two aspects of improvements in the algorithm have been applied:correcting the observation error covariance:introducing a constant gain. Finally, verified by the dynamic conditions of the LiFePO4 battery, it compares the estimation result of the improved algorithm with it before. The results show that the improved algorithm can effectually estimate the SOC of the LiFePO4 battery, solve the difficulty in determining the initial value of the AH integration method, and improve the accuracy of SOC estimation to a certain extent.