Improved multi-scale cosine control whale optimization–error feedforward double Kalman filtering for the online state of charge and state of health co-estimation of lithium-ion batteries

Lithium-ion batteries play a pivotal role in diverse applications, necessitating the precise estimation of their parameters for safe and efficient energy utilization. In the context of ternary lithium-ion battery research, this study introduces the cosine control whale optimization (CCWO) method to real-time optimization of the forgetting factor. Concurrently, an innovative multi-scale error feedforward extended double Kalman filter (EFDKF) algorithm, featuring model error feedforward, is presented for dynamic optimization of the state noise covariance matrix. This approach diminishes reliance on the model while jointly estimating the state of charge (SOC) and state of health (SOH). The algorithm’s feasibility is validated under varied operational conditions, encompassing both normal and low temperatures. The maximum absolute errors of SOC and SOH estimation are 1.09% and 3.16% at normal temperature, and 3.46% and 4.96% at low temperature, respectively. The results affirm the algorithm’s enhanced precision in joint SOC and SOH estimation, heightened robustness, and superior convergence. This contribution introduces a novel methodology for lithium-ion battery state estimation.