SOC Estimation And Balancing Control For Serially Connected High Capacity Lithium-ion Battery

With the advantages of high energy density, safety and environmental protection, easy to combine, lithium battery is widely used in the field of electric car, micro grid energy storage, aeronautics and astronautics. When high capacity batteries are in series, because of the aging and the processing technology of battery, the SOC(State of Charge) of battery will show discreteness, which reduces the effective capacity of series batteries, and this will short the service life of battery pack. So, the study of large capacity battery modeling, SOC estimation and energy management has a significant meaning to the application of lithium battery.First of all, an improved first-order Thevenin model is proposed. In the model parameters, there is a function relation between the electromotive force and the SOC. In order to obtain the model parameters of battery, the detail derivation of least square method and RLS(recursive least square) is shown in the paper. What’s more, the HPPC(Hybrid Pulse Characterization Test) is improved to collect the voltage and current signals of working battery pack, which can build a database. Finally, Use MATLAB to analyze and calculate the database and get the parameters of lithium battery model.Second of all, charge-discharge efficiency, temperature, self-discharge and aging impacts the SOC estimation of lithium battery, which is fully analyzed. It also brings correction coefficient when estimating the SOC. A kind of SOC estimation which combines improved Open Circuit Voltage and Ampere-hour method is adopted and working voltage is amended by battery model. Therefore, it can realize on-line estimation and increase the Ampere-hour method’s correction ability to the initial value.Third of all, active equalization and passive equalization are combined together and the active equalization is accomplished by multi-winding flyback converter whose working principle is analyzed in detail. When the SOC discreteness is caused by different reasons, different control strategies have different impacts on balancing speed. Standard deviation which represents the degree of discreteness and coefficient which represents the reasons of discreteness are brought into the paper to choose different control strategies according to the discrete state of SOC. Simulation model is established in MATLAB and it verifies the feasibility of equalization and control strategy.At last, key points are introduced in detail during hardware designing and they are as follows, key parameters, choice of key device and creation of experiment platform. Moreover, software method to realize SOC estimation and control strategy is presented in the paper. The goal of experiment is to estimate SOC and balance energy for series connection batteries and the feasibility of equalizer and control strategy is verified.