Research on Digital Twin-Enhanced Cross-Individual Fault Diagnosis for Marine Main Engines

Objectives To address the feature drift induced by individual variability of marine main engines and the scarcity of fault data in the target domain, this study leverages digital twin and transfer learning to achieve cross-individual diagnosis. Methods Based on the incorporation of digital twin parameters and mechanisms, a digital twin-enhanced cross-individual fault diagnosis model (DT-DANN) is established. Particle Swarm Optimization (PSO) is employed to adaptively adjust the twin parameters, aiming to construct a high-fidelity individual digital twin model and generate healthy data samples under variable operating conditions. By integrating a Multi-scale 1D Convolutional Neural Network (1D-CNN) with an improved Domain Adversarial Neural Network (DANN), a healthy-state anchor strategy is proposed to achieve cross-individual fault diagnosis for marine diesel engines. Results The adaptively tuned twin parameters achieve a MAPE of 0.0520%, with the DT-DANN model reaching 100% diagnostic accuracy. Conclusions The proposed method effectively addresses the model mismatch problem caused by individual variability and data scarcity, achieving high-accuracy zero-shot cross-individual fault diagnosis for marine main engines.