This application discloses a method for the operation and maintenance analysis of
power equipment for
new energy ships. It generates a high-quality coupled data sample set by collecting multi-
physics field operation,
microstructure, and historical fault data, followed by time-series deviation correction and
standardization. By constructing a physical constraint layer embedding
physical field coupling equations and a fault mechanism
knowledge graph, and injecting a model training
loss function, the resulting multi-
physics condition calculation model can accurately characterize multi-field
coupling characteristics and fault evolution patterns, improving adaptability to complex operating conditions. By extracting multi-dimensional features and weighted fusion for
dimensionality reduction, a core fault
feature vector is generated, accurately capturing the essence of the fault. A large-scale
fault analysis model is obtained by fine-tuning the pre-trained model using a meta-learning
algorithm, improving the analysis accuracy and generalization ability for
small sample scenarios. This allows for clear tracing of fault evolution paths, definition of
coupling critical conditions, and establishment of micro-
macro mapping relationships. Real-ship feedback enables closed-loop updates of the model and
knowledge graph, ensuring long-term stability of operation and maintenance accuracy.