An ANN-based framework for rapid spectral fatigue analysis of steel catenary risers

A simplified framework is presented in which an existing artificial neural network (ANN) based tool for critical stress range prediction is used in order to rapidly assess the fatigue life of a steel catenary riser (SCR). The simplified approach considers the first-order motions of the hosting floater (heave, pitch and roll motions) and irregular sea-states to assess the critical stress range within the touchdown zone (TDZ) of the SCR. Stress transfer functions are generated that approximate the SCR TDZ critical stress range due to vertical motion at the SCR hang-off point. The motion response amplitude operators (RAOs) and transfer functions are then combined to generate the SCR TDZ stress spectra and hence assess accumulated fatigue damage for all potential sea-states at the floater location. The fatigue lives of two large diameter SCRs subject to a sample irregular wave scatter diagram are calculated using the simplified framework. The results are then compared with those determined via a state of the art commercial software that uses a dynamic time-domain finite element (FE) analysis with rain-flow cycle (RFC) counting and shown to provide a good agreement. It is an important result as the time required to run the simplified analysis is an order of magnitude smaller than the more rigorous analysis (minutes versus hours). It demonstrates the usefulness of the simplified approach at the early stages of an SCR design where a large number of simulations are needed for sensitivity studies in order to select an optimized concept.