Time domain simulation of riser VIV in current and irregular waves

Abstract

A semi-empirical time domain force model for combined cross-flow and in-line vortex-induced vibrations (VIV) is proposed, based on a series of earlier publications. The new feature is a term which represents the effect of vortex shedding in the flow direction, referred to as the in-line vortex shedding load. The latter is added to Morison's equation and a cross-flow vortex shedding force. The fundamental idea of how to model the effect of vortex shedding is the same as in previous works. An algorithm for synchronization is applied between the vortex shedding loading terms themselves and the structural response, which excites vibrations for a pre-determined frequency interval. The originality of the present study is rather how the VIV-terms are integrated as part of Morison's equation, which is there to provide a description of inertia and drag forces. All the loading terms combined enables simultaneous simulation of VIV and other response phenomena, such as wave induced motion and static drag displacement.The performance of the time domain model is verified against measurements of a vertical riser subjected to two different external flow cases. The first one is simply steady uniform current whereas the second flow case combines the latter with irregular waves. To simulate the experiments, a linear finite element model of the riser is made, and the proposed hydrodynamic force model is applied to the translation degrees of freedom along the structure. It is to be noted that the same empirical coefficients are used to simulate both experiments. In uniform flow, the results are good. Dominating frequency and vibration amplitude agree well in both cross-flow and in-line directions. The simulated time series show more regular/less tendency of amplitude modulations than the experiments, but the overall agreement is still acceptable for engineering purposes. For the second flow case, where the riser was towed in irregular waves, the model provides a highly realistic representation of the riser motion. Both when the total response (containing wave and VIV frequencies) and the filtered signal (including only VIV frequencies) are analysed, the predictions follow the measurements closely. From before, the cross-flow part of the VIV model has been tested in oscillating flow and regular waves, and the performance of the former was experimentally proven. It is hence concluded that the proposed prediction tool is applicable to a variety of non-stationary flow conditions, implying that the synchronization model captures parts of the underlying physics, despite its simple form.

Flexible cylinder flow-induced vibration

“…Based on the new phenomena observed in the experiment of the flexible cylinder in the unsteady flow, several corresponding prediction methods have later been developed. [97][98][99] Standing wave vs Traveling wave: field observations, laboratory experiments, and numerical simulations revealed that when long, tensioned cylinders were exposed to spatially non-uniform flow, they could have a standing wave response, traveling wave response, or a mixture in both the IL and the CF directions. 19,85,100,101 Traveling waves usually occur when there is a nonuniform distribution of damping (including both structural and hydrodynamic damping) along the cylinder.…”

A Modified Wake Oscillator Model for the Cross-Flow Vortex-Induced Vibration of Rigid Cylinders with Low Mass and Damping Ratios

“…CFD-based numerical prediction methods are limited in terms of calculation speed and cost for these large-scale structures [3]. Thus, empirical models are widely used in the current software (such as OrcaFlex) for the prediction of vortex-induced vibrations [4]. Due to the strong nonlinearity of vortex-induced vibration, it is difficult to accurately predict its response under all test conditions by using an empirical model.…”

Application of wake oscillators to two-dimensional vortex-induced vibrations of circular cylinders in oscillatory flows

“…Some empirical coefficients were proposed and limited to the available low-amplitude data range. This model has recently been extended to account for in-line VIV in oscillatory flows by Ulveseter et al (2018).…”