Accurate state of temperature estimation for Lithium-Ion batteries based on square root cubature Kalman filter

Accurate state of temperature estimation for lithium -ion batteries is essential to ensure its safe and efficient operations. In this study, a state of temperature estimation framework for cylindrical lithium -ion batteries is proposed based on square root cubature Kalman filter algorithm, wherein the internal and surface temperatures of power battery are accurately estimated. Firstly, the electro-thermal coupling characteristics of lithium -ion battery is characterized based on an equivalent circuit model, Bernardi battery heat generation model and two-state lumped thermal model. On this basis, the discrete time-domain characteristics of internal and external temperature of the battery are addressed. Then, the genetic algorithm and forgetting factor recursive least squares algorithm are respectively employed to identify the circuit parameters and thermal physical parameters, and the internal and surface temperatures prediction is achieved based on square root cubature Kalman filter algorithm. The feasibility of the proposed temperature estimation method is verified at different charge and discharge process under wide range ambient temperature of -20 degrees C to 60 degrees C. The experimental results manifest that the proposed method can accurately online estimate the internal temperature and surface temperature with the mean absolute error of less than 1.16 degrees C.