Abstract:

This paper proposes a collaborative monitoring and evaluation framework for the operation status of lithium-ion battery energy storage power plants, which integrates machine learning and deep learning, to address the difficulties in multi-source data fusion and the inability of traditional methods to effectively capture nonlinear degradation features. A multi-level feature extraction and fusion mechanism has been systematically designed to solve the problems of temporal and spatial misalignment of multimodal data such as voltage current temperature, and insufficient extraction of implicit degradation features, providing a new solution for the monitoring and evaluation of the operation status of lithium-ion energy storage power plants.

Key words: lithium-ion battery, energy storage system, energy storage power station, status monitoring