Prediction of Mooring Line Top-Tensions Incorporated with Identification of Lost Clump Weights

Monitoring the top-tension of mooring lines is essential for ensuring the safe operation of floating units. This study aims to propose an innovative hybrid method combining measured motions of Floating Production Storage and Offloading Vessels (FPSO) with numerical models to estimate mooring line top-tensions using artificial neural networks. The inconsistency in numerical results of FPSO motions, attributed to the loss of clump weights on mooring lines, necessitates the development of an inversion method employing genetic algorithms. This method identifies the loss of clump weights in the mooring system by utilizing field-measured FPSO motion data to update the numerical model. The results demonstrate that, after detecting clump weight loss, the relative error of the maximum horizontal displacement between the simulated submerged turret production and the measured values is reduced to less than 5%. With the updated model in place, numerical simulations are conducted to map measurable motions and unmeasurable mooring line top-tensions. To achieve precise real-time predictions of mooring line top-tensions, a Long Short-Term Memory (LSTM) neural network is deployed, using the measured FPSO motions as input data. This approach enables highly accurate predictions of mooring line top-tensions.