# NEW STUDIES ON META-MODELING FOR LAZY-WAVE STEEL CATENARY RISERS

## Palavras-chave:

Artificial Neural Networks, Optimization, PSO, Offshore Systems, Riser Lazy-wave## Resumo

Offshore production systems are very complex structures involving safety and economic aspects. For

these reasons, several research is done to reduce these costs. At design phase, riser layout and its configuration are

a key factor as this is a high-cost component and suffers from the action of environmental loads and dynamic

movements of the floating unit where it is connected. This component is mathematically modeled, and calculations

are performed using the Finite Element method to assess the structural feasibility and optimization algorithms can

be used to determine the viable configuration with the lowest associated cost. The problem with this methodology

is that the analysis of different configurations in an optimization process demands a very large computational

power, consequently a very high processing time. Previous works have demonstrated that meta-models such as

ANN - Artificial Neural Networks can be used to replace the finite element procedure in the evaluation of riser

configurations. In this work, novel developments were studied considering two approaches, first the optimization

process starts and after a specified number of generations, the training of the meta-model is carried out and from

that point the optimization uses the meta-model as a method of evaluation. The second approach has the first part

the same as the previous one, however, after the first training, re-trainings are carried out, incorporating new

individuals to the knowledge of the meta-model during the optimization process itself.