On-line Estimation Of State Of Charge Of Energy Storage Battery

The application of battery energy storage technology improves the operation efficiency and power supply quality of the power system.Accurate estimation of battery State Of Charge(SOC)is an important basis for monitoring and managing energy storage systems.In this paper,on-line SOC estimation research is carried out for lithium iron phosphate batteries suitable for power grid energy storage.The main work is as follows:(1)Design and simulate the actual energy storage battery working condition to obtain the data needed for SOC estimation.Based on the investigation and analysis of the characteristics of energy storage battery,energy storage conditions and SOC estimation requirements,the corresponding experimental process and scheme are worked out,the relevant experimental equipment is selected,and the experimental platform is built.According to the self-defined energy storage conditions,the corresponding charge and discharge experiments were carried out.Through the recording and cleaning of experimental data,the data foundation is laid for battery model parameter identification and SOC estimation verification.(2)Taking the second-order RC equivalent circuit model as the object,the on-line updating of model parameters is realized by using recursive least square method with forgetting factor.On the basis of experimental data and battery characteristic analysis,the second-order RC equivalent circuit is selected as the battery circuit model through comparison.Recursive least square method with forgetting factor can adjust the proportion of new and old data,and has the ability to quickly respond to data changes and converge to true values.The experimental verification of self-defined energy storage conditions shows that the method can realize online updating of model parameters and has good effect.(3)Through analysis and comparison,the adaptive unscented Kalman filter(AUKF)is selected as the SOC estimation algorithm of the energy storage battery.Based on the second-order RC circuit model and the custom charging and discharging conditions of power grid energy storage,the effects of extended Kalman filter(EKF),unscented Kalman filter(UKF)and adaptive unscented Kalman filter(AUKF)on SOC estimation are analyzed and compared.The results show that the AUKF algorithm has real-time estimation correction for the noise characteristics of the system due to the introduction of adaptive covariance matching in UKF.Compared with the other two algorithms,the AUKF algorithm has higher robustness and estimation accuracy,and can meet the accuracy requirements of SOC estimation for energy storage batteries.(4)On the basis of SOC estimation of single battery,an SOC estimation rule of battery pack based on logic tree analysis is proposed.According to different definitions of series and parallel battery packs,the structure of battery packs is analyzed by using logic tree analysis method,and a SOC estimation rule based on logic tree analysis method is proposed.The SOC estimation of battery packs in simplified energy storage power station is realized by using the algorithm in MATLAB.On the premise of ensuring safe and stable operation of battery packs,the SOC of battery packs can be estimated by using the SOC of single battery cells.