Research On Secondary Battery Cycle Life Prediction Methods Under Time-varying Stress

In recent years, with the development of materials science and manufacturing processes, life of the secondary battery is becoming longer and longer. As a result, the traditional life prediction methods depending on a large number of failure data which are harder and harder to obtain cannot meet the needs that predict secondary battery life. Since degradation data contain abundant life information, life prediction methods based on performance degradation is widely adopt to predict lifetime of products with long life and high reliability. Presently, the achievements of degradation modeling are mainly in data from traditional degradation test(normal stress, constant-stress, step-stress and progressive stress degradation tests, etc), but touches little on the modeling of time-varying stress degradation data. In the cycle life test of secondary batteries, the varying temperature stress influences both the degradation data and the fading rates. The temperature varies with time is defined as Time-varying Temperature stress. This dissertation modifies the existing degradation models by taking into account the influence of different stresses on key performance parameters to solve the problem on lifetime prediction of secondary batteries at time-varying temperatures. The main achievements are as follows:(1)Traditional degradation test and lifetime prediction methods based on performance degradation are studied and time-varying stress degradation test is proposed.(2)The cycle life prediction methods of satellite Cadmium-Nickel secondary batteries are studied. Through the analysis of cycle life testing data, we choose EODV(End of Discharge Voltage) as the key performance parameter and establish its degradation model under time-varying temperature stresses with the consideration of the relation between EODV and temperature. The cycle life of Cadmium-Nickel batteries is predicted through the Pseudo life method and the random parameter vector method, respectively.(3)The cycle life prediction methods of Li-ion batteries are studied. Through the analysis of cycle life testing data, we choose usable capacity as the key performance parameter and establish its dynamic degradation model with the consideration of the effects of different levels of stresses on both degradation data and fading rates. Cycle life of Li-ion batteries at different temperatures are estimated through Pseudo life method.