Research on Shape Reconstruction of Deep-Sea Riser Based on Neural Network

The deep-sea riser is an important equipment in marine engineering. It is responsible for transporting minerals from subsea oil wells to offshore platforms. The complex environment of the marine is easy to damage the riser. Aiming at the uncertainty of the deep-sea riser shape in the harsh service environment and the large deformation and vulnerability characteristics of the riser during the laying process, a method of riser shape reconstruction based on deep neural network was proposed. In this paper, the form of a piggyback pipeline is used to monitor the shape of the riser. An optical fiber array is laid inside the sub-tube, and the nonlinear logical relationship between the strain data measured by the optical fiber array and the deformation data of the main-tube is explored through a deep neural network model. Finally, the shape of the riser is obtained through the curve reconstruction algorithm. The results show that the predicted shape of riser obtained by this method is basically consistent with the preset shape, which can be used as a new method for long-term health monitoring of deep-sea riser.