Study Of Lithium-ion Battery Modeling And Prognostics Method

Lithium-ion batteries, as a kind of widely-used energy storage power sources, have been taken more seriously, especially for its safety and stability. Thus a battery management system (BMS) should be established to monitor the capacity, assess its health state and predict the remaining useful life of a battery or battery pack. Before a BMS is established, it is necessary to analyze and design the functionality of the three parts in details.The main function of the first part "monitoring battery capacity" is to predict the state of charge (SOC) of a battery. In other words, it is to predict the battery capacity with the measurements of voltage and current, where a battery model should be built. A first-order RC equivalent circuit model is established and the parameters are identified, which are used to predict the SOC of battery with the open circuit and extended Kalman filtering (EKF) method. The second part "battery health assessment" is to predict the state of health (SOH) of a battery, that is "to predict the largest capacity a battery can discharge", which is presented with a percentage-the ratio of the largest dischargeable capacity and its nominal capacity. Voltage curve fitting method is employed to analyze the relationship between SOH and voltage change based on existing battery data. Then SOH can be estimated by the change of voltage.The prediction of battery remaining useful life (RUL), which is the third part, is mainly about how to predict the RUL. Cycle number is used to express the RUL. Based on battery capacity fading model, particle filtering (PF) method is employed to predict the RUL.Comparisons indicate that the model built is of high precision and the EKF method improves the accuracy of prediction; voltage curve fitting method contribute to estimating the battery SOH; the PF method provides the probability expression of RUL prediction, which can be accurately performed with this method. Wholly, the theoretic analysis and design of the main three parts of a BMS are implemented here, and will guide other researchers to design the battery system in some way.