Towards a methodology to estimate environmental loadings from time history motions of offshore platform by using Artificial Neural Networks

Resumo

Floating production systems (FPS) for offshore oil exploration are subject to environmental loads such
as waves, wind and current in different directions of incidence and varying intensities that result in dynamic
movements of this same system. Nowadays, FPS has several sensors, in this particular case, its position is
monitored by GPS and accelerometers. On the other hand, it is hard to monitor environmental loadings in a deep
water that depends on oceanographic buoys. Therefore, this paper presents the first steps to estimate the parameters
of wave loading from a time history motions of an offshore platform by using Artificial Neural Networks (ANN).
From this, it may be possible to verify oceanographic forecast models and to know the environmental conditions
at the moment of an event, such as, line break or equipment breakdown. In addition to this, we can estimate real
environmental data for the generation of digital twins, which is a digital replica of the real system. In the case
study, ANN training process is performed from data of a rigorous Finite Element (FE) analysis. From the results,
we can observe that ANN presents a high level of accuracy in this kind of application, which allows to move
forward with research in this area.