Research On Aging Performances And Cycle-life Predictions Of Li-ion Battery

New energy vehicles with low emissions, no pollution, and the energy can be used indirectly to achieve the purpose of energy saving. The power battery as the power source of new energy vehicles, is the core part of the electric vehicle power system. The lithium ion battery because of its high voltage, high energy density, good circulation performance, small self discharge, no memory effect and other advantages, has become a relatively better choice of power battery. Useful Life Remaining (RUL) is the key of the battery monitoring and maintenance, which has become an important part of the battery technology. Accurate estimation of battery RUL, predict to lithium ion battery residual life to make, helps to improve the efficiency in the use of the battery to ensure the safety and reliability of the battery in use, is of great significance for improving the control performance of the whole vehicle and predicting the driving range. The main contents of this paper are the analysis of lithium ion battery cycle life experiment and aging characteristics, establish the equivalent circuit model of the lithium ion battery, and cycle life prediction based on data driven. The details are as follows:Firstly, based on the internal electrochemical reaction and working principle of the lithium ion battery, the mechanism of the performance degradation of the battery was analyzed. Cycle life experiment platform to build, design cycle experiment and collection of data, including maximum available capacity test, different discharge test, composite pulse HPPC test, cycle life testing. Then focus on the experimental analysis of the power battery aging characteristics:discharge rate, charge discharge efficiency, discharge voltage, ohmic resistance and other factors, to select the capacity as an index prediction and the cycle life of the battery laid the foundation.Secondly, the equivalent circuit model of Li ion battery was established. After analysis and comparison of various battery model selection both accuracy and calculation of second-order RC model as the main research object, and select open circuit voltage function was fitted, the nonlinear relationship between open circuit voltage and SOC is solved. According to the current circuit model can not reflect battery capacity decline and parameters with cell aging changes, respectively to different batteries in the state of life of fitting parameters and contrast battery using initial, middle, end of a three-stage battery model parameters, it is concluded that the general law of change of battery parameters.Finally, the choice of two kinds of prediction method based on data driven life: ANN and SVM is used to forecast the battery life, and the comparison of two kinds of methods for the performance of the pros and cons, ultimately determine the support vector machine method is of high accuracy, less training samples, to overcome the shortcomings of battery life estimation using neural network method.