Research On The Realization Of Battery State Of Charge Estimation Based On FPGA

Different from the research direction of traditional automobiles,the three-electric system(battery,motor,and electronic control)has become the key technology for the development of electric vehicles,and the battery technology for the safety of electric vehicles has become the most concerned object in the three-electricity.As the core of electric vehicles,the battery management system(BMS)is responsible for power output and safety control of the whole vehicle.It is powerful and complex.Among them,the state of charge(SOC)of the battery is an important parameter in the battery management system,which has guiding significance for the vehicle mileage and charge-discharge control.An accurate SOC value is helpful for the driver to make a correct driving judgement.Since the SOC of the battery is the state variable of the battery and cannot be directly measured,it is necessary to estimate the SOC.The fast and accurate estimation of the SOC value can give feedback to the battery management system in time,and guide the BMS to make corresponding control quickly.This thesis takes the battery SOC estimation as the main research content,and completes the SOC estimation implementation based on the FPGA platform.Firstly,this paper takes the NCR18650 BD lithium-ion battery as the research object,combined with the characteristics of the battery and the characteristics of terminal voltage variation,establishes the equivalent circuit model of the second-order RC network to provide the basis for the algorithm implementation.For the needs of battery model parameter identification,the recursive least squares method with a forgetting factor of 0.97 is used to identify the resistance-capacitance parameters.Combined with the model and parameter identification results,the feasibility of the SOC estimation method is analyzed and verified in Simulink.According to the verified algorithm and the parameter values calculated by the model,an FPGA-based UKF algorithm implementation scheme is designed to estimate the battery SOC.The whole design is built with Verilog hardware description language,and the state machine is designed to control the calculation process according to the calculation steps of the UKF algorithm.The algorithm in the design involves many matrix operations,so the systolic array structure is used for some algorithms,the Cholesky triangular matrix decomposition is used to simplify the calculation process,and an error matching mechanism is introduced in the calculation process to optimize the estimation effect.After the overall design of UKF is realized,simulation tests are carried out on each functional module to ensure that the required functions are implemented correctly.The overall design adopts a high-speed parallel computing method,which can improve the estimation rate of battery SOC.A time-division multiplexing processing unit and a systolic array structure are used for a large number of matrix multiplication and addition operations,and the calculation method can save a lot of calculation time and reduce memory access.The Cholesky triangular matrix decomposition method can simplify the process of matrix inversion and square root calculation,and achieve the purpose of quickly solving complex matrix operations.The implemented SOC estimation scheme is verified on the FPGA platform.Compared with the conventional computing method and the fast parallel computing method,the UKF algorithm has a significant speedup under parallel computing,and compared with the embedded computing scheme,the FPGA full hardware implementation method combined with the Cholesky triangular matrix Decomposition can accelerate matrix inversion and square root calculations by 20 to 30 times,and the calculation speed of the entire UKF algorithm is increased by 25%.