Objective In marine engineering, the reliable operation of marine engines is critical to ensuring maritime safety and operational efficiency. However, the complex and harsh marine environment exposes marine engines to a range of issues, including corrosion of key components, clogged pipelines, and inadequate lubrication. These problems not only degrade engine performance but also increase maintenance costs and pose a threat to the safety of ships at sea. Therefore, developing an effective fault diagnosis method, especially one capable of diagnosing faults under cross-operating conditions, is crucial. The primary objective of this study is to enhance the diagnostic performance of marine engine fault diagnosis models under cross-operating conditions. By addressing this issue, we aim to deliver a more accurate and reliable fault diagnosis solution for marine engines, thereby enhancing the overall safety and reliability of ships.

Method To achieve this goal, a novel intelligent fault diagnosis framework based on model fine-tuning technology is proposed. This framework employs ResNet-18, a deep learning model, as the feature extractor for identifying engine faults. First, a pre-trained model is constructed using data from the source domain. The pre-training process minimizes the discrepancy between the model's predictions and ground truth labels using a suitable loss function. The Adam optimizer is employed to perform gradient descent and parameter optimization. A 1D-CNN serves as the core computational module, comprising convolutional layers, batch normalization, activation functions, and pooling layers. Subsequently, a fine-tuning strategy is applied. Specifically, the feature extraction layers are frozen while the classifier layers are fine-tuned using data from the target domain. This approach helps align the feature distributions of the source and target domains, thereby enhancing the model's performance under cross-operating conditions. The effectiveness of the proposed framework is verified using fault data from the lubrication and intake/exhaust systems of marine engines.

Results In fault diagnosis tasks under cross-operating conditions, the proposed framework achieves diagnostic accuracies exceeding 94% for the engine lubrication system and 98% for the intake/exhaust system respectively. Compared with the VGG-11 model used as a feature extractor, the ResNet-18-based framework outperforms it significantly in terms of diagnostic accuracy. The ResNet-18 model shows superior stability and robustness in different transfer tasks, exhibiting lower variance in classification accuracy.

Conclusion In conclusion, the proposed intelligent fault diagnosis framework based on model fine-tuning technology significantly enhances diagnostic performance under cross-operating conditions for marine engines, improving the model's generalization and adaptability to diverse operating scenarios. However, this study has certain limitations. Future work may focus on integrating agent technology to improve the model's interactive and text understanding capabilities, enhancing the model's ability to identify and diagnose unknown fault types, and conducting experiments on cross-device transfer diagnosis for diesel engines. This will further improve the practical application value of the fault diagnosis technology for marine engines.