Accurately predicting state of health(SOH)of batteries is crucial for ensuring their long-term safe and effective operation.Aiming at the challenges of limitation of health feature extraction

randomness of attention allocation

and scientificity of model hyperparameter setting in the prediction study of SOH of lithium battery

a hybrid neural network based on squeeze-and-excitation attention enhancement based on quantum particle swarm optimization(QPSO)is proposed

and applied to the SOH estimation of lithium-ion battery.Firstly

multi-dimensional health features

are extracted and the principal components obtained through dimensionality reduction are input into a convolutional neural network(CNN)with squeezed excitation attention enhancement to fully extract features.Then

a long short-term memory(LSTM)neural network is used to fully extract long-term dependencies.Finally

quantum particle swarm optimization is used for hyperparameter optimization to achieve global optimization.Extensive experimental verification and comparative studies show that the proposed method has better predictive performance

with prediction error root mean square error(RMSE)decrease of 0.29% and coefficient of determination(

R

2

)increase of 0.61%

compared to the optimal baseline model.Therefore

the proposed method is expected to become an effective information guide for battery health management.