Main content area

Degradation model and cycle life prediction for lithium-ion battery used in hybrid energy storage system

Liu, Chang, Wang, Yujie, Chen, Zonghai
Energy 2019 v.166 pp. 796-806
algorithms, data collection, electrochemistry, energy, lithium batteries, models, prediction
Lithium-ion battery/ultracapacitor hybrid energy storage system is capable of extending the cycle life and power capability of battery, which has attracted growing attention. To fulfill the goal of long cycle life, accurate assessment for degradation of lithium-ion battery is necessary in hybrid energy management. This paper proposes an improved degradation model of lithium-ion battery based on the electrochemical mechanism of capacity fade, in which the influence of cycling current is taken into consideration. Moreover, genetic algorithm is applied to identify the initial values of model parameters. A particle filter based data driven framework is also designed to track the variation of model parameters and states during the cycling process. Short-term and long-term degradation prediction methods are then developed, to forecast the essential indicators of battery health. Datasets of battery cycling tests under both constant cycling current and dynamic cycling current are used for verification. The root mean square error results for state of health prediction (less than 18 cycles) and remaining useful life prediction are below 8% and 40 cycles, respectively, showing the accuracy and applicability of the proposed methods. Additionally, the baseline degradation prediction for case with cycling current information unknown is also illustrated.