ANN and wavelet network meta-models for the coupled analysis of floating production systems

• Efficient ANN and WN meta-models for the coupled analysis of FPS are presented.
• Models provide simultaneously platform motions and structural response of lines.
• Wave elevation time series are taken as exogenous inputs.
• Performance evaluated by application to a FPS for an actual deep water scenario.

This work presents the development of efficient meta-models based on Artificial Neural Networks (ANNs) and wavelet networks (WNs). The models can be associated to a coupled methodology for the analysis and design of floating production systems (FPS), providing fairly accurate results for the platform motions and the parameters of the structural response of the lines, with remarkable reductions in processing time.To accurately take into account the influence of the risers on the platform motions, previous approaches employing meta-models required the execution of coupled motion analyses, with the mooring lines and risers represented by Finite Element models; the resulting motions would subsequently be input to a meta-model for the uncoupled simulation of the line. Now, the model presented in this work will obtain simultaneously both the platform motions and the desired parameters of the response of the lines (i.e. any result that could be provided by a coupled FE simulation). This way, not only the FE structural analyses of individual lines can be avoided, but also the expensive coupled motion simulations, therefore comprising an even more rapid and effective computational tool.The performance of the ANN and WN meta-models are evaluated, in terms of accuracy and computational time, by their application to a FPS representative of those currently considered for ultra-deep water scenarios in Southeastern Brazil pre-salt fields.