Computational Fluid Dynamics (CFD) offers a powerful tool to simulate the fully three-dimensional Vortex Induced Vibration (VIV) around flexible bodies such as marine risers. In this paper two riser geometries with L/D = 1400 from Norwegian Deepwater Program and L/D = 4200 from Miami experiments in Gulf of Mexico are studied. The former case includes the VIV fatigue prediction of a horizontal riser in uniform currents and shear currents. The latter case has the VIV simulation of a vertical riser in non-uniform current. The results from both studies are validated using field experimental data. The motion of the riser is solved using a tensioned beam motion equation discretized with a finite difference scheme. The flow field around the riser is obtained by unsteady Navier-Stokes equation along with Large Eddy Simulation (LES) method. The flow field solver is coupled with the riser motion solver using a feedback loop to get the Fluid Structure Interaction. An overset (chimera) grid system is applied to decompose the structure and flow field in blocks of overlapping grids. Both inline and crossflow motion are resolved and processed to obtain the corresponding fatigue damage. The instantaneous displacements obtained at different axial locations on the risers are converted to bending stress time series which are used in rainflow counting algorithm with Palmgren-Miner’s rule to obtain fatigue damage.
As the oil and gas industry is moving in ultra-deep waters the need to accurately predict the motion of long flexible marine risers has become of prime importance. The constant high currents occurring at the production site causes the excitation of slender structures due to periodic shedding of vortices leading to Vortex-Induced Vibration (VIV) fatigue. Recent years have seen a tremendous surge of research activities to understand and model the VIV phenomenon. Field experiments by Norwegian Deepwater Program (NDP) has led to significant understanding of the VIV fatigue response of bare and straked riser and the results were published by Trim et al (2005). Experiments conducted by British Petroleum and Exxon Mobil have also shed light on the prediction of VIV motion in long marine risers. In 2006, Deepstar JIP conducted experiments in Gulf Stream using a 500ft riser (L/D = 4200). Jhingran et al. (2007) and Vandiver et al. (2006) used these results to understand the occurrence of high mode numbers during VIV. Although experimental investigations of the VIV have been very promising, they have been increasingly limited due to constraints on the experimental facilities.
Computational Fluid Dynamics (CFD) offers a viable alternative to estimate the VIV fatigue with reasonable accuracy. CFD has been a useful tool in recent years to simulate the VIV motion and many favorable results have been documented by Chaplin et al. (2005) and Holmes et al. (2006). The Finite-Analytic Navier-Stokes (FANS) numerical method has been developed in Pontaza, Chen and Chen (2004, 2005a, 2005b), Pontaza, Chen and Reddy (2005) and Pontaza and Chen (2007). This method along with the overset (chimera) grid technique can reasonably predict the vortex induced vibrations of marine risers as shown in Huang, Chen and Chen (2007, 2008, 2010, 2011, 2012). These studies employ the FANS algorithm to predict the flow field around flexible bodies.
In this paper, we present numerical simulations of free span pipelines under vortex-induced vibrations (VIV) and pipe-soil interactions. Pipeline is simplified as a tensioned beam with uniformly distributed tension. The flow field around the pipeline is computed by numerically solving the unsteady Navier-Stokes equations. Fluid domain is discretized using an overset grid system including embedding, overlapping and matching grids. Simulation results are compared to experiments for validation in two cases: one for an isolated pipeline in uniform current without boundary effect, and the other for a free span pipeline lying on soil bottom. General agreements are observed.
The vortex-induced motion (VIM) of semi-submersible platforms becomes an important issue with the recent development of deep draft semi-submersible platforms. As a result of the increased draft, the semi-submersibles are susceptible to coherent vortex shedding, and the platform VIM increases significantly. The VIM of semi-submersibles is more complex than those of spars and mono-column hulls due to the wake interaction of vortices shed from multiple columns. In general, the vortex-induced motion of deep draft semi-submersible platform is characterized mainly by three degree-of-freedom motions with surge (in-line), sway (transverse), and yaw motions. In the present study, numerical simulations are performed for a semi-submersible with four square columns subjected to a current at a 45 degree incidence angle. Calculations were performed using the Finite-Analytic Navier-Stokes (FANS) code in conjunction with a moving overset grid approach to accommodate the relative motions between the semi-submersible hull, wake, and background grid blocks. Simulations are performed both for the full scale and the 1:70 model platforms to check the validity of the Froude scaling law. Various current speeds corresponding to different reduced velocities are simulated. Motion responses and the flow fields for both the model and full scale platforms are studied. Comparisons are made with experimental data to demonstrate the capability of the present CFD approach.
Slamming is a common phenomenon as a ship navigates in rough seas, and it can cause severe structural damage to the hull structure. Full-domain Computational Fluid Dynamics (CFD) simulation of random wave and structure interaction is considered impractical by many researchers. Simplified approaches are usually adopted to alleviate the expensive CFD random wave simulation. In this paper, we present a rigorous methodology that solves the Navier-Stokes equations entirely without any need of matching. In our simulation scenario, a container ship cruising at a constant speed is allowed to heave and pitch in random waves. Both head sea bow slamming at 6 knots speed, and following sea stern slamming at 0 and 5 knots are studied. Irregular waves based on the Bretschneider spectrum for 25-year return sea states are used to simulate a realistic seaway environment. A very effective procedure is developed to capture the desired waves at a specific part of the sailing vessel. Our multi-block overset grid code is fully parallelized and greatly reduces the computation time to make the simulation practical. A level-set function is employed to capture the violent free surface and to simulate the interaction of the random wave and the ship. This rigorous Navier-Stokes numerical approach is able to capture complex mechanisms and show results that are possible only with CFD simulations, thus provides useful guidance for ship designs.
We present a coupled level-set (LS) and volume-of-fluid (VOF) method for time-domain simulation of violent free surface flows in an overset grid system. The advection of the VOF function is performed by using a mixed second-order Eulerian and Lagrangian scheme with a piecewise linear interface calculation (PLIC). A volume correction scheme is implemented to compensate for the mass change and to maintain a divergence-free velocity field. The LS function is solved by using a fifth-order Weighted Essentially Non-Oscillatory (WENO) scheme. The coupled level-set and volume-of-fluid (CLSVOF) interface-capturing method is employed in conjunction with the Finite-Analytic Navier-Stokes (FANS) method for violent free surface flow problems. Moreover, a chimera domain decomposition approach is implemented by using an overset grid system, including embedding and overlapping grids for accurate resolution of flow around structures. The simulation results demonstrate the capability of the CLSVOF method in an overset grid system for violent free surface flow problems.
Computational fluid dynamics (CFD) simulations of vortex-induced vibrations (VIV) and wake-induced vibrations (WIV) have been performed for two vertical risers in tandem and side-by-side arrangements. Both risers have the same outer diameter of 0.016 m, and the same total length of 1.5 m (L/D = 93.75). However, only the lower 40% of the risers is submerged in water. Two different cases were considered for dual-riser VIV in side-by-side arrangements with center-to-center distances of 2 and 4 diameters. Simulations were also performed for the same two risers in tandem arrangement with different combinations of top-tension, flow velocity, and center-to-center spacing. For completeness, calculations were also performed for an isolated riser to provide a more detailed comparison of the riser motion with available experimental data. In all simulations, the flow field was solved using an unsteady Reynolds-Averaged Navier-Stokes (RANS) numerical method in conjunction with a chimera domain decomposition approach with overset grids. The riser inline and cross-flow motion responses were calculated using a tensioned beam motion equation. The external force terms were obtained by integrating viscous and pressure loads on the riser surface. The computed vortex patterns and dynamic responses of the risers are in good agreement with the experimental data.
A coupled level-set (LS) and volume-of-fluid (VOF) method has been developed for time-domain simulation of violent free surface flows around two- and three-dimensional structures. The advection of the VOF function is performed using a mixed second-order Eulerian and Lagrangian scheme with a piecewise linear interface calculation (PLIC). A mass correction scheme is implemented to compensate for the mass change and maintain a divergence-free velocity field. The level-set function is solved using a fifth-order Weighted Essentially Non-Oscillatory (WENO) scheme. The coupled level-set and volume of fluid (CLSVOF) interface-capturing method is employed in conjunction with the Finite-Analytic Navier-Stokes (FANS) method for time-domain simulation of violent free surface flow problems. Moreover, a chimera domain decomposition approach is implemented using an overset grid system including embedding and overlapping grids for accurate resolution of flow around structures. The simulations results demonstrated the capability of the CLSVOF method in maintaining mass conservation for violent free surface flow problems.
Greenwater on jack-up structures occurs when storm waves significantly exceed the air gap and slam into the jack-up hull and its topside equipment. In the present study, a Navier-Stokes code was employed in conjunction with the interface-capturing level-set method for the prediction of wave impact on a jack-up structure under hurricane wave conditions. An overset grid system is employed to facilitate the simulation of complex flow around a generic jack-up structure with 3 supporting legs and simplified topside equipment. Time-domain imulation of greenwater and wave impact loads were performed for random 3D short-crested waves based on the directional wave spectra of Hurricane Katrina. The simulation results successfully captured the evolution of the wave crest as it propagates through the jack-up structure while exerting impact loads on the topside equipment in its path. A detailed evaluation of 6 different extreme wave events indicates that the greenwater impact loads are highly localized due to the short crestedness of the random 3D waves. Nevertheless, the maximum wave impact load exerting on the entire jack-up structure is strongly correlated with the crest height of the incident waves. The simulation results provide useful guidance for estimating the peak impact loads produced by directional short-crested waves.
Huang, Kevin (Ocean Engineering Program, Department of Civil Engineering, Texas A&M University) | Chen, Hamn-Ching (Ocean Engineering Program, Department of Civil Engineering, Texas A&M University) | Chen, Chia-Rong (Department of Mathematics, Texas A&M University)
This paper studied a vertical riser VIV under sheared current using numerical simulation and presented the results and their comparisons to published experimental data. The riser was made of a 9.63-m brass pipe with an OD of 0.02 m (L/D D 482) and mass ratio of 1.75. In the experiment the riser was positioned inclined with top tension of 817N and pinned at its 2 ends to the test rig. Rotating the rig in the wave tank would simulate the sheared current. In our numerical simulation we pinned the riser’s ends to the ground and imposed a linearly sheared far field incoming current. The riser and its surrounding fluid were discretized using 1.5 million elements. The flow field was solved using an unsteady Reynolds- Averaged Navier-Stokes (RANS) numerical method in conjunction with a chimera domain decomposition approach with overset grids. The riser was also discretized into 250 segments. Its motion was predicted through a tensioned beam motion equation with structural damping. The external force terms were obtained by integrating viscous and pressure loads on the riser surface. We then processed the critical parameters including riser VIV a/D, vorticity contours, response histories and spectra, and VIV-induced fatigue. Finally, comparisons were made to the experimental data and conclusions drawn. In general the VIV simulation results agree well with the experimental data. It is concluded that the present CFD approach is able to simulate the vertical riser VIV in sheared current. In addition, it can also predict the VIV-induced fatigue damage.
Riser VIV numerical simulation has been an area of interest for many years. Many of these investigations are limited to a 2D or 3D rigid riser. A good summary of the publications on VIV simulation of a stationary and oscillating cylinder was provided by Dong and Karniadakis (2005).