Research On Estimation Of SOC Based On Central Difference Kalman Filter Algorithm For Power Battery

After the20thcentury, with the global energy crisis and environmental pollutionproblems increasingly seriously. In this situation, new energy vehicles, especially elec-tric vehicles, have been highly concerned by the major automobile manufacturersaround the world. As an important part of the electric vehicles, Battery ManagementSystem (BMS) through the efficient management and control of the battery to ensurethe efficient use of battery and driving safety. But the battery management technologyis still in development stage, and is far from mature, and how to improve the batterystate of charge (SOC) estimation accuracy is one of important and difficult problems.This paper takes theLiFeO4battery as the research object, in-depth study from twoaspects which the SOC estimation method based on the model and battery model.Firstly, this paper introduces the basic principles and main parameters of lithium-ion batteries. Then through a series of experiments, this paper analyzes the basicfeatures such as relationship between the open circuit voltage (OCV) and state ofcharge (SOC), ohmic resistance and capacity. On this basis, comparison of theadvantages and disadvantages combined with common circuit model, this paperproposes an improved second-order RC equivalent circuit model which considers thetime variation of battery capacity, and obtains the initial value of model parameters byindex fitting method in Matlab software. This thesis also takes advantage of theMatlab/Simulink to update the initial parameters at each point of battery SOC. Theexperimental results show that the equivalent circuit model parameters after parame-ters correction improves the tracking accuracy of the battery voltage changes.Secondly, after analysis of the traditional SOC estimation method, this thesisestablishes theLiFeO4battery state-space equation based on second-order RC equi-valent circuit model and non-linear characteristics of theLiFeO4battery. Becauseof the accuracy limitation of the extended Kalman algorithm to estimate the nonlinearequation, proposing the Central Difference Kalman Filter (CDKF) Algorithm. Thesimulation results show that the CDKF algorithm under the same conditions on theSOC estimation, the accuracy of CDKF is better than EKF algorithm.Finally, using AVL Cruise software to build the electric vehicle models andoperate the simulation experiments in simulated urban road conditions. The simu-lation data obtained from the simulation condition of the battery are imported into Matlab software, and utilize the CDKF algorithm to estimate the SOC. The resultswhich compared with EKF clearly indicate that the CDKF algorithm has the advan-tages, such as anti-interference, astringency, and greater accuracy to the SOC estima-tion of vehicle.