Research On Li-battery SOC Estimation Based On Double Adaptive Unscented Kalman Filter

With the increasingly severe situation of energy and environment,the new energy electric vehicle has a broad prospect for industrialization.As the core driving energy source of electric vehicles,the lithium battery faces some limits for large-scale application that its state cann’t been estimated accurately and the dynamic cycle life is not high enough.As a result,it becomes the key technology to estimate the SOC value of the lithium battery accurately and efficiently for applying the electric vehicles.In this study,the LiFePO4 battery is regarded as the research object,and an estimation algorithm of SOC based on Adaptive Unscented Kalman Filter is proposed by building a test bench of battery.The main research contents are as follows:Firstly,the types of lithium battery are introduced in detail and its working principle is analyzed,especially the performance index of lithium-ion battery.A battery charging and discharging test bench is established based on the ITS5300 instrument including four modules,that is the programmable linear power supply ITECH,the LiFePO4 battery,the programmable electric load and the industrial PC.The relevant tests are designed to confirm the relationship between the open-circuit voltage and its SOC value based on the test bench.Secondly,several commonly used models are analyzed in detail,and the two-order RC model is chosen to estimate the SOC of LiFePO4 battery.Besides,the stability of model is analyzed,and the online identification and off-line identification methods for the model parameters are introduced,especially the online identification algorithm based on the traditional Kalman Filter is analyzed in detail.It can have the better estimation precision to estimate the SOC value using the Unscented Kalman Filter(UKF)algorithm based on the UT transformation,but the application condition of the algorithm is that the statistical properties of the system process noise and observation noise are known clearly.And then,an Adaptive Unscented Kalman Filter(AUKF)algorithm and its implementation procedure are described.By combining the parameter online identification method based on the traditional Kalman filter and the SOC estimation method based on the AUKF algorithm,the Double Adaptive Unscented Kalman Filter(DAUKF)is introduced in this study,and the implementation steps of the algorithm are designed in detail.Lastly,the dynamic tests of LiFePO4 battery under the condition of the constant current discharging,the constant current charging and discharging,and the UDDS dynamic cycle are performed based the test bench respectively.The experimental results show that: the estimation precision can be achieved less than 2% to apply the proposed DAUKF algorithm with a better estimation precision and adaptive tracking ability compared to the traditional DUKF algorithm.