A comparative study of Li-ion battery models and nonlinear dual estimation strategies

Due to their high energy density, durability, low cost, and inherent safety, lithium-ion (Li-ion) batteries are quickly becoming the most popular energy storage method for electric vehicles. Difficulty arises in properly modeling these types of batteries due to a large number of parameters and different architectures. This paper looks at studying six different Li-ion battery models found in literature, and compares their relative performance based on a benchmark dataset. Kalman-based filtering strategies are also employed to estimate important battery parameters such as capacitance, hysteresis, and state of charge (SOC). In addition, the relatively new smooth variable structure filter (SVSF) is used and compared with these Kalman-based strategies.