Non-Linear Time Domain Drilling Riser VIV Modelling

Abstract

A new time-domain Vortex Induces Vibration (VIV) model is here proposed and applied to modelling of drilling riser and wellhead fatigue. The model challenges the industry best practices for VIV predictions which are known to be conservative. The paper is divided in two main parts. The first part is devoted to a validation process considering a rigorous benchmark with the ExxonMobil database. The second part treats the application of the model to a study of a drilling riser configuration. In both parts the proposed time domain model is compared to the industry best practice representated by SHEAR7 v4.10. The model development has been driven by project requirements. It carefully balances CPU consumption and targeted / required accuracy. The model is based on a hybrid approach integrating concepts from both frequential model and existing time domain wake oscillators. It consists of a hydrodynamic solver capable of balancing both Vortex Induced Excitation and Hydrodynamic Damping at each predetermined node along the mechanical model. The solver is then coupled with a Finite Element (FE) model (or equivalent) to assess the resulting riser dynamics. The different nonlinearities of the riser model (e.g. soil, riser flexible joint) can consequently be accounted for in the usual manner for the global riser analysis. The proposed model is first benchmarked using a reference database. The model produces riser responses that match well for both uniform and shear current conditions. Results obtained from the proposed model are also compared to results obtained from SHEAR7 with different choices of parameters (including the recommended values). Although mode shapes are in good agreement, the conservatism of the latter is demonstrated and challenged by the proposed model which appears to be a promising candidate for VIV analysis. A nonlinear model of a 400m long drilling riser is considered as the use case. The VIV analysis is first performed using SHEAR7 and a linearized version of the riser model. The results are compared to those obtained for the proposed time-domain model with the same linearized drilling riser model. While the two approaches again produce the same mode shapes, the proposed model produces much less conservative results (approx 50% reduction of the vibration amplitude). Finally, the different nonlinearities of the riser model (wellhead, riser flexible joint) are considered to assess their possible impact on global riser behavior and local impact (i.e. at well head). Depending on the nonlinearities to be considered, the local impact of the model is clearly demonstrated.