Research On Voltage Equalization Of Multi-module Serial Connected Supercapacitors And Their State Estimation

In high-power energy storage systems,supercapacitors are becoming more and more important as a new type of energy storage component.Its rapid power absorption and release capabilities make it more suitable than other battery-based energy storage components in areas where fast power conversion is required.In order to ensure the reliability and service life of the super capacitor energy storage system,it needs to be effectively managed.This paper studies the supercapacitor management system.It mainly includes two functions: supercapacitor voltage equalization and supercapacitor state estimation.In the dissertation,the multi-cell series structure is considered firstly.The difference between the cells causes a gradual voltage difference during the charging and discharging process of the energy storage system.The accumulation of this voltage difference will seriously affect its normal operation.To solve this problem the voltage equalization method is studied.In this paper,the multi-winding transformer voltage equalization topology is modeled and analyzed to predict the equalizing speed within a module.At the same time,this topology has the disadvantage of the limited number of equalized cells.For this reason,a novel voltage equalization method for cells between modules is proposed based on the fact that voltage equalization is achieved using a multi-winding transformer based topology in each module so as to increase the number of equalized cells that can be balanced,and because the equalization principle is the same as the multi-winding transformer voltage equalization topology,its advantages can be maintained.Both of these topologies are validated by Simulink simulations and experiments.The supercapacitor management system needs to provide the real-time power indicator of the supercapacitor while realizing the voltage equalization,and it should be capable of detecting the performance decay of the supercapacitor during use,so as to perform fault diagnosis,so it is also required to have online state estimation capabilities(including state of charge SOC and state of health SOH)of supercapacitors.First of all,in this paper,the Kalman filter and the sliding mode observer are used to estimate the SOC of the classic triple RC branch model of supercapacitors.To estimate the SOC,model parameters are required by both observers.For this reason,this article uses Simulink’s integrated offline parameter identification function to identify triple RC branch model parameters,and proposes that the model can be simplified as two RC branches,thereby reducing the number of parameters to be determined.Secondly,the supercapacitor SOH estimation is equivalent to the estimation of equivalent serial resistance or capacitance.So this paper proposes to use Linear Least Squares and EKF algorithm to identify its nonlinear capacitance,and proposes to use EKF to estimate the SOC of the two-RCbranch model and identify the parameters simultaneously.The validity of the above algorithm is verified by Simulink simulation.In the research process of the thesis,combined with the "City Supercapacitor Line Energy Storage System" project of China Railways,the principle research,simulation analysis and experimental research of supercapacitor management system were carried out.The research results can provide reference for the further application of supercapacitors.