Research And Application Of SOC Monitoring Technology For Lithium Battery Based On Artificial Intelligence

With the development of power batteries,lithium batteries have been widely applied because of its advantages of no pollution and long life.However,because of its nonlinear characteristics in the process of system discharge,it affects the estimation accuracy of the SOC(State Of Charge,charge state),so that the lithium battery encounters a "bottleneck" in its own development.For this reason,a SOC prediction method based on neural network is proposed in this paper,and on this basis,the research of residual energy detection of power battery is carried out.By integrating the system on the LaBVIEW platform,the on-line dynamic monitoring function of battery parameters is realized.The main research work is as follows:Firstly,the characteristics and influenc ing factors of the battery are analyzed.A SOC prediction model based on BP neural network is established.In view of the shortcomings of slow convergence rate and large error in BP network,this paper proposes a combination of additional momentum method and LM algorithm to improve the BP neural network,the experimental results show that the method has high prediction accuracy and fast convergence speed.Based on this,in view of the working state of the battery,the improved BP network is used to train the battery discharge parameters,so that the estimated value is gradually approximated to the actual value.The experimental results show that the prediction error of the improved BP algorithm is kept within 2%,and a higher accuracy prediction is achieved.Secondly,with the wide use of electric vehic les and other products,a large number of power battery losses will occur.In order to maximize the use of battery resources to enable it to continue to be used in energy storage batteries such as solar energy and wind power generation,this paper has carried out research on the method of detection of residual energy.By intercepting parameters of the discharge characteristic curve is sent to the neural network model established above to predict the residual energy.The test result shows that: the prediction error of residual energy is within 5%,which make up for the shortage of the national residual energy detection standard and basically realized the non-destructive testing of batteries,which has a good application prospect.Finally,based on the LaBVIEW platform,the dynamic monitoring system of lithium battery parameters is designed,which can realize the function of collecting,displaying and saving the battery parameters in real time through the serial port and EBC-A10 communication.Through repeated calls to the MATLAB program,the on-line dynamic prediction of the lithium battery SOC is realized.The prediction results can change follow the nonlinear change of the battery,and the prediction accuracy of the battery in the process of use is improved.To sum up,the neural network model established in this paper improves the prediction precision of battery SOC,and it can detect the residual energy of the battery,the result is good and the application prospect is broad.