A method and system for jointly estimating the health status and remaining service life of lithium-ion batteries under conditions of missing ship data.

Abstract

This invention discloses a joint estimation method for the health status and remaining service life of lithium-ion batteries under conditions of missing ship data. First, multi-source online operational data is time-aligned, anomaly cleaned, and explicitly labeled with missing information to construct a model input with a unified structure and missing information. Physically reasonable sequences are reconstructed from the missing segments, and the stability and engineering usability of the completed results are improved through trend consistency and temporal continuity checks. Then, based on the completed operational sequence, a combination of multi-scale convolutional feature extraction, degradation-sensitive segment enhancement, and long-range state-space dependency modeling is used to uniformly characterize the battery's short-term fluctuation characteristics, multi-scale degradation patterns, and long-term degradation trends. Finally, through the joint output and consistency constraints of health status and remaining service life, a stable, continuous, and physically reasonable joint estimation result of battery life status is obtained under conditions of missing data and complex ship operating conditions.

Images