Estudio de la Aplicación de Redes Neuronales Artificiales en la Predicción de Amplificación Rápida de la Amplitud de Roll de Buques en Olas

For a vessel which is in longitudinal waves, the wave passage, together with the vertical motions (heave and pitch), makes that the submerged geometry varies periodically with time, generating a variation of the ship’s restoring properties and from a small perturbation, it can be developed a rapidly growing movement of progressive roll. To this end, we develop a system for predicting parametric resonance, starting from the analysis of the motion of the vessel, capable of detecting short-term appearance of it. The human being is endowed with complex brain neural circuitry. Thus, the idea of computationally modeling these neural connections to bring forth intelligent behavior also in machines. In this context emerges artificial neural networks (ANNs), inspired by the very nature of neural networks and biological synapse. In this work, we are trying to take advantage of ANNs to predict the behavior of ships, platforms and other floating devices, with emphasis on detection of dynamic amplification of ship’s roll in parametric resonance conditions.