Mooring Line Tensions Prediction Based on MLP for Soft Yoke Mooring System

The soft yoke mooring system is widely used for the floating production storage and offloading (FPSO) in shallow water. However, there were accidents of mooring system being pulled down by the FPSO in field. It is hence necessary to conduct realtime prediction of mooring line tensions. Under complex environmental loads, the coupled effects of the mooring system and floaters lead to nonlinear characteristics of the mooring line tensions. In recent years, scholars explore the application of neural networks to predict mooring line tensions due to the strong nonlinear fitting capacity of neural networks. The coupled time-domain model of FPSO and soft yoke mooring system is established and sensitivity analyses are conducted to identify input parameters. Accordingly, a series of numerical analyses are carried out to extract the datasets for prediction. A multilayer perceptron (MLP) neural network model is established to predict the maximum mooring line tensions of the soft yoke mooring system in Bohai Sea in this work. The results demonstrate that the model is applicable for predicting the mooring tensions in short term sea states.