Eventos Anais de eventos
COBEM 2023
27th International Congress of Mechanical Engineering
EVALUATING THE IMPACTS OF A NEURAL NETWORK TOPOLOGY ON ANCHOR RADIUS’ ESTIMATION IN MOORING LINES
Submission Author:
Johne Trindade , SC
Co-Authors:
Johne Trindade, Thiago Pontin Tancredi, BERNARDO LUIS RODRIGUES DE ANDRADE
Presenter: Johne Trindade
doi://10.26678/ABCM.COBEM2023.COB2023-0931
Abstract
The anchoring system connecting the ocean bottom to the hull of offshore platforms is crucial for limiting platform offset. This study evaluates the impact of neural network topology on estimating anchor radius in mooring lines. A dataset was generated using Exmoor software, incorporating line length, wet weight, top traction, and water line depth, while calculating top angle, horizontal force, and anchor radius. The dataset analysis guided the selection of inputs and outputs for the neural network models. Fifty-four different topologies were constructed and trained separately, with mean squared error (MSE) used to evaluate accuracy. Each model was trained five times to determine the standard deviation of the MSE, totaling 270 training processes. The results revealed high precision in predicting anchor radius. Notably, the number of hidden layers and neurons did not always directly correlate with accuracy. The optimal balance was found within a middle range, exhibiting the lowest MSE. These findings demonstrate the potential of neural networks for anchor radius estimation in mooring lines. They underscore the importance of selecting appropriate network topology and achieving a balance in hidden layers and neurons for accurate predictions among with lower computational cost, contributing to the development of reliable models for enhancing the safety and efficiency of offshore operations.
Keywords
Applied AI, Mooring lines, offshore platform, Optmization, design methodology
