Obtenção de metamodelo para o problema do resfriamento de petróleo estagnado no interior de tubulações do tipo pipe-in-pipe

Abstract

The formation of solid deposits within pipelines transporting hydrocarbons is a major issue in maintaining fluid flow, as it can result in blockages and substantial financial losses. Low temperatures in the seabed facilitate such deposits, particularly in the case of paraffins and hydrates. To address this challenge, various mitigation strategies are employed, including the use of a pipe-in-pipe system combined with active electric heating and computational simulation tools. However, these simulations can be computationally intensive. This study proposes the utilization of metamodels based on Echo State Neural Networks for the problem of oil cooling in underwater pipein-pipe pipelines that use direct active electric heating. The results show that on average, the metamodel is thirty-one times more computationally efficient than simulation using the Finite Volume Method.