Fast and robust state-of-power estimation for lithium-ion batteries using pegasus algorithm

The State of Power (SOP) of a battery is the maximum power it can safely deliver or absorb over a given period without violating electrochemical limits. Respecting these limits is essential to prevent thermal runaway. Accurate SOP estimation is therefore a key Battery Management System function for ensuring safety. SOP can be estimated by solving an optimization problem, but doing so on embedded systems with limited resources is challenging, mainly because it requires simulating differential equations over a time horizon. After reformulating SOP estimation as a current-constrained problem, this paper proposes an efficient optimization method based on the Pegasus algorithm. It offers a better trade-off between computation time and accuracy than methods such as bisection, Model Predictive Control (MPC), or Multi-Layer Perceptron (MLP) prediction. The predictive SOP calculation is combined with a closed-loop approach using measurements to ensure constraints are never exceeded, protecting the battery from thermal runaway. Hardware-in-the-loop tests show an average computation time of 456 µs per step and RAM usage of 673 bytes. The method outperforms other reported solutions, is robust, and is suitable for embedded applications.