Bayesian optimised route and SOH estimation effect for Li-ion battery management system of electric vehicles based on LSTM

Lithium-ion batteries are widely used in electric vehicles, and accurate state of health (SOH) estimation is crucial for driving safety. This study applies a long short-term memory (LSTM) algorithm to model SOH based on health features correlated with standardised capacity. Since manual parameter tuning is inefficient and training is time-consuming with large datasets, a domain space design inspired by manual adjustment is combined with Bayesian optimisation for hyperparameter configuration. Experimental results show that the optimised LSTM improves estimation accuracy by 0.0235%. Compared with grid and random search, Bayesian optimisation reduces relative error by 50.63% on average and requires the least time, demonstrating both higher optimisation efficiency and near-optimal parameter selection.