Research And Implementation Of The SOC Estimation Of Lithium Battery In Storage

With the increasing awareness of environmental protection and the gradual increase of power demand,the development of new energy has attracted more and more attention.In order to improve the stability of new energy supply and guarantee the quality of power,the storage technology of large capacity plays an important role.Lithium batteries because of small volume?high voltage?high specific capacity?good environmental protection and without polluting characteristics,so it is more and more widely applied in the field of energy storage,became the focus of research.However,how to improve the performance and effici ency of the battery is a difficult problem during the use of lithium batteries.In this paper,the state of charge(SOC)of lithium battery is accurately estimated,which can be used more effectively and scientifically to improve battery life and improve energy storage efficiency.The SOC of lithium battery cannot be directly measured,but indirect estimation can be carried out by other parameters and methods.Based on this characteristic,this topic research first with single-chip microcomputer and the LabVIEW software platform developed a real-time online monitoring system for lithium battery's parameter information,and then lithium battery monitoring system acquires the lithium battery's voltage?the discharge current and temperature parameters,which get the lithium battery's remaining capacity by calculating.We can display and storage these parameters in LabVIEW user interface.Next analyzes all kinds of commonly used lithium battery SOC estimation method.According to the characteristics of lithium battery and actual working condition,the open circuit voltage method ?ampere-hour integral method and back propagation neural network method are used to forecast the battery's SOC in the battery's different working stages.Discharge experiments are divided into constant current status and variable flow status of two kinds of discharge situation.System to record the temperature of the battery voltage and current parameters and calculated the residual capacity as the sample data of BP neural network.Because of the shortcomings of BP neural network,it needs to beimproved to improve the effectiveness of network training and to make the estimation results more accurate.Therefore,the lm-bp algorithm is selected in the constant current state and the genetic-BP algorithm is selected in the variable flow state,to establish the BP neural network on MATLAB was trained,then with the trained network to test the samples of SOC estimation,which achieved good results.Finally,the network that was trained on MATLAB platform was called to LabVIEW,and the SOC of the real-time prediction battery was realized,and the reliability of the system was verified.At the same time,the prediction method of this paper realizes the high nonlinear mapping between the input and output of the battery,which has high accuracy and feasibility.