Real-time prediction of mooring tension for semi-submersible platforms

Abstract

The harsh operating environment of deepwater semi-submersible production platforms poses severe challenges to the safe operation and maintenance of their production processes. Using artificial intelligence technology to improve the operation and maintenance level of large-scale offshore engineering equipment has become an inevitable trend. Accurate mooring tension prediction for semi-submersible production platforms can effectively identify potential damage and ensure the platform's safe and efficient production operation. Compared with traditional cyclic neural networks, transformer neural networks have faster calculation speeds, better generalization abilities and scalability, and better abilities to capture long-distance dependencies. In this study, the first deepwater semi-submersible production platform in the South China Sea, “DeepSea One,” was selected as the research object, and an improved novel transformer neural network architecture was proposed for the mooring tension prediction of semi-submersible production platforms using actual long-term mooring tension data obtained during the platform's production process. The proposed method was compared with the commonly used long short-term memory neural network model, and the comparison proved the effectiveness of the proposed method in predicting the dynamic tension of semi-submersible production platform mooring. In addition, the proposed method is also applicable for the prediction of other time series data in the field of shipbuilding and offshore engineering, which is of certain value for generalization.