As the scale of ship power systems expands, it becomes increasingly necessary to quickly and accurately assess the transient stability of the system to ensure its safe operation. In recent years, data-driven artificial intelligence (AI) methods have gained attention for their speed, accuracy, and strong generalization capabilities. Given the current limitations in utilizing existing specialized software to perform precise transient calculations for ship power systems, this study establishes a comprehensive mathematical model of the ship power system. Time-domain simulations under various disturbances are conducted to build a dataset for training neural networks. A neural network classification model combining a denoising autoencoder (DAE) and a long short-term memory (LSTM) network is developed to achieve real-time transient stability assessment of ship power systems. To address the decline in model performance under different operating conditions and the time-consuming process of re-collecting data and retraining models, transfer learning (TL) is employed for the transient stability assessment of ship power systems.
Research on Data-Driven Transient Stability Assessment Method for Ship Power Systems
2024-10-10
483454 byte
Conference paper
Electronic Resource
English
Data-driven submarine cable protection ship anchor damage risk assessment method
European Patent Office | 2025
|Ship stability - A research strategy
Tema Archive | 1982
|