A novel approach for motion predictions of a semi-submersible platform with neural network

A neural-network-based prediction of motion responses of a semi-submersible is presented here. The fully connected neural networks and the long–short-term memory networks were employed to establish the neural networks for motion predictions. The effects of network architectures and time steps were investigated in depth. The predicted results were compared with the measurements and the predictions based on other conventional methods. The results demonstrate that the neural-network-based approach could offer a fast and accurate prediction of heave, roll, and pitch responses of a semi-submersible. This provides a promising alternative for evaluations of hydrodynamic performances of new designs and monitoring of the dynamic behavior of in-service floating structures.