This invention discloses a method and
system for predicting the health status of
lithium batteries under varying operating conditions. First, key operating condition features such as
energy recovery power,
energy recovery current, equivalent power, average
discharge current, and equivalent current are extracted based on historical battery data, and parameters of a decay relationship model between health status and equivalent
cycle count are fitted. Second, a
Gaussian process regression model is trained using early data to learn the mapping relationship between operating condition features and decay parameters. Then, by combining real-time monitoring and external information, future
energy recovery power, average
discharge power, and equivalent power are predicted, and corresponding current feature sequences are derived. Finally, the predicted future operating condition features are input into the trained model to obtain predicted decay parameter values, achieving accurate prediction of future health status. This invention significantly improves the accuracy, adaptability, and foresight of health status prediction under varying operating conditions by integrating battery physical models and data-driven methods and introducing future operating condition prediction.