Study On The Pure Electric Logistics Vehicle Battery State-of-Charge Estimation Method

In this thesis, the problem of the state of charge of the lithium battery in the key technology of the pure electric city logistics distribution vehicle was studied. Due to the chemical reaction of lithium iron phosphate battery in the electric logistics vehicles was very complex and highly nonlinear and had derivative changes, the state of charge of the lithium battery was estimated to be uncontrollable and erroneous. So the estimation method had its own difficulty and complexity. There were urgent requirements for the lithium battery state of charge estimation that the safety of pure electric logistics vehicles and economy of the lithium battery recycling and mileage information etc. So the lithium battery state of charge estimation was the key of research on pure electric vehicles. It was worthy of further research about improving estimation accuracy accuracy of charge state estimation of lithium iron phosphate battery.First, the purpose of the research was in order to improve the accuracy of estimation algorithm about the state of charge of the pure electric vehicle lithium battery. So the construction of the lithium battery model and vehicle model charge state estimation algorithm was designed. Secondly the lithium battery electrochemical reaction was with the electrochemical polarization and concentration polarization phenomenon. Considered with this and according to the impedance characteristics of the two different kinds of polarization, the thesis put forward a frequency of lithium battery equivalent circuit model characterization method according to the lithium iron phosphate battery capacity,internal resistance, open circuit voltage and temperature characteristics. And according to the actual calculation ability and precision index, the complexity of the equivalent circuit model and the selection of the model order were determined. Thirdly according to the fractional characteristics of Li cell electrochemical reaction, the equivalent circuit model was extended to fractional. And this thesis solved discrete filter approximation and solution of fractional order differential equivalent circuit model and estimated parameters of fractional differential equivalent circuit model on the basis of identification idea of recursive least squares identification. Furthermore Kalman Filter algorithm which was used to charge the theory of state estimation was studied. Based on the fractional lithium battery equivalent circuit model, this thesis compared the extended Kalman Filter to the unscented Kalman Filter on the different characteristics of the nonlinear system. Based on the idea of UT transform, this thesis proposed an algorithm of the unscented Kalman Filtering based on fractional order differential model and dynamic switching sampling,so as to avoid the error due to the extended Kalman filter Taylor series expansion and the introduction of higher order terms that were neglected and reduce the complexity of the algorithm. This algorithm which was based on the general state estimation algorithm and the memory characteristics of the fractional order calculus could improve the accuracy of the estimation algorithm. And the two kinds of SOC estimation algorithms were compared and analyzed. The experimental results were used to characterize the superiority and accuracy of the estimation algorithm.Finally, pure electric logistics vehicle model was built in the ADVISOR vehicle simulation platform, the model was based on the pure electric van truck CDW5070XXYH1 PEV as the prototype that was produced by Chengdu WangPai Motor Co., Ltd. And this model was in typical road conditions on the running process of the vehicle for simulation analysis. The results showed that the lithium iron phosphate battery charged state estimation algorithm was accurate and effective.