- Home
- Documents
*Numerical Simulation of Vortex-Induced Vibration of a ... Numerical Simulation of Vortex-Induced...*

prev

next

out of 52

View

1Download

0

Embed Size (px)

Numerical Simulation of Vortex-Induced Vibration of a Vertical Riser in Uniform and Linearly Sheared Currents

Enhao Wang and Qing Xiao

NOTICE: this is the author’s version of a work that was accepted for

publication in the Journal of Ocean Engineering. Changes resulting from the

publishing process, such as final peer review, editing, corrections, structural

formatting, and other quality control mechanisms may not be reflected in this

document. Changes may have been made to this work since it was submitted

for publication. This manuscript was accepted for publishing on 02 June

2016.

1

Numerical Simulation of Vortex-Induced Vibration of a Vertical Riser

in Uniform and Linearly Sheared Currents

Enhao Wang and Qing Xiao*

Department of Naval Architecture, Ocean and Marine Engineering, University of Strathclyde, Henry Dyer

Building, Glasgow G4 0LZ, Scotland, UK

Abstract

This paper presents a numerical study on vortex-induced vibration (VIV) of a vertical riser

subject to uniform and linearly sheared currents. The model vertical riser tested at the

MARINTEK by ExxonMobil is considered. The predicted numerical results are in good

agreement with the experimental data. It is found that the dominant mode numbers, the

maximum root mean square amplitudes, the dominant frequencies and the fatigue damage

indices increase with the flow velocity. A standing wave response is observed for the single-

mode in-line (IL) and cross-flow (CF) vibrations. Dual resonance is found to occur at most of

the locations along the riser. At some locations along the riser, a third harmonic frequency

component is observed in the CF response and a frequency component at the CF response

frequency is found in the IL response apart from the frequency component at twice the CF

response frequency. The majority of the vortex shedding shows a clear 2S pattern, whereas a

2P mode is observed near the position where the maximum vibration amplitude appears. The

higher IL fatigue damage in the present study emphasises the importance of the IL fatigue

damage especially in the design of low flow velocity or low mode number applications.

Keywords: Vortex-induced vibration (VIV); Riser; Fluid-structure interaction (FSI);

Computational fluid dynamics (CFD);

* Corresponding author. Tel: +44 1415484779.

Email address: qing.xiao@strath.ac.uk (Q. Xiao)

2

Nomenclature

Axrms/D, Ayrms/D Dimensionless in-line and cross-flow root mean square

amplitudes

max /xrmsA D , max /yrmsA D Dimensionless maximum root mean square amplitudes

c Structural damping

D, Do Riser outer diameter

E Young’s modulus

fn Natural frequency of the oscillating mode

fn, beam n th eigenfrequency for a nontensioned beam

fn, string n th eigenfrequency for a tensioned string

fox, foy In-line and cross-flow oscillation frequencies

fz,ε Zero-crossing frequency of the bending strain

I Moment of inertia of the beam

L Length of the riser

m Mass per unit length of the riser

m* = m/(ρπD2/4) Mass ratio

n Mode number

Re = VD/ν Reynolds number

T Top tension

t Instant time

tw Riser wall thickness

V Uniform flow velocity

Vmax, Vmin Maximum and minimum velocity

Vprofile Velocity profile

x In-line displacement

xmean Mean in-line displacement

y Cross-flow displacement

ε Root mean square strain

ρ Fluid density

ν Kinematic viscosity of the fluid

3

1. Introduction

Vortex-induced vibration (VIV) often causes the fatigue of offshore slender structures, such

as risers, mooring lines and pipelines. Detailed understanding of this fluid-structure

interaction (FSI) phenomenon and an efficient prediction of such self-excited and self-

sustained oscillations are required for the reliable estimation of the fatigue damage and the

development of VIV suppression techniques (Bourguet et al., 2011a, 2013).

Over the past few decades, VIV has been extensively studied. One may refer to the

comprehensive reviews by Sarpkaya (1979), Bearman (1984), Williamson and Govardhan

(2004), Gabbai and Benaroya (2005), Bearman (2011) and more recently by Wu et al. (2012).

As riser pipes often possess a length-to-diameter ratio (L/D) of the order of 103 (Chaplin et al.,

2005), many experiments have been carried out on deepwater risers with large L/D

(Tognarelli et al., 2004; Chaplin et al., 2005; Trim et al., 2005; Lie and Kaasen, 2006;

Vandiver et al., 2006; Tognarelli et al., 2008; Vandiver et al., 2009; Huang et al., 2011b; Gu

et al., 2013; Gao et al., 2015). These experiments investigated flexible riser VIV responses

under different flow conditions and some also assessed the effectiveness of VIV suppression

techniques, such as using helical strakes. Better insights into some important VIV aspects

(i.e., response amplitude, dominant mode, dominant frequency and fatigue damage etc.) were

obtained from these experiments, and thus provided some good benchmarks for verifying

numerical prediction models.

Apart from the various experimental investigations, there have been a number of

computational fluid dynamics (CFD) studies on VIV of flexible cylinders.

Willden and Graham (2001) used a quasi-three-dimensional (Q3D) method to simulate the

transverse vibration of an L/D = 100 cylinder subject to a sheared inflow at low Reynolds

numbers. A high tension was applied to the cylinder so that the fundamental mode would be

excited. A maximum amplitude of 0.36D was found at L/D = 44 which was slightly below

the midpoint of the cylinder span. The results also showed that the majority of the shedding

frequencies along the cylinder were modified towards the natural frequency and a significant

spanwise correlation was observed.

4

Meneghini et al. (2004) and Yamamoto et al. (2004) presented the numerical simulations of

long marine risers with L/D up to 4600 with Q3D discrete vortex method (DVM). In their

simulations, the riser tended to select a vibration mode which could keep the reduced velocity

Vr = V/fnD in the range of 4 ≤ Vr ≤ 7 where the energy was transferred from the fluid to the

structure. Visualisations of the wake indicated a hybrid mode of vortex shedding along the

span with a 2S mode being found in regions of small amplitudes, changing to a 2P mode in

regions of larger amplitudes.

The simulations described above were based on Q3D method with several two-dimensional

(2D) strips over the length of the riser. However, Q3D simulations have many shortcomings,

e.g., three-dimensional (3D) vortex structures cannot be treated correctly and straked risers

and variations in the angle of attack cannot be studied directly. Therefore, a series of fully 3D

numerical simulations emerged.

Newman and Karniadakis (1997) simulated VIV of an infinitely long flexible cable at Re =

100 and Re = 200 with a spectral/hp element method. Both the standing wave and travelling

wave responses were realized. It was found that an interwoven pattern of vorticity was

associated with a standing wave cable response while oblique vortex shedding was produced

by a travelling wave cable response. A mixed standing wave/travelling wave response

together with chevron-like vortex shedding was found to be related to a sheared inflow.

Evangelinos and Karniadakis (1999) studied VIV of an infinitely long flexible cylinder at Re

= 1000. The structure’s bending stiffness was varied to obtain different responses. The

authors found that the modulated travelling wave motion of a free-free beam or cable led to a

mixed response consisting of oblique and parallel shedding. In the case of structures with

pinned endpoints a standing wave response was obtained with lace-like flow structures.

Holmes et al. (2006) and Menter et al. (2006) investigated riser VIV with fully 3D finite

element method (FEM) and finite volume method (FVM), respectively. Both of the

simulations used relatively coarse meshes with high element aspect ratios and the results

were in good agreement with the experimental data by Trim et al. (2005) and Chaplin et al.

(2005), respectively.

5

Constantinides and Oakley (2008) compared their CFD results with the data obtained in the

field experiments by Jhingran and Vandiver (2007). The results were able to match the

experimental data. Both the first and third harmonic components were well captured. The

authors emphasised the importance of the third harmonic component in fatigue damage

analysis due to the fact that it produced strains of the same order of magnitude as the first

harmonic component and had a frequency of three times the first harmonic component, which

returned roughly three times more fatigue damage.

Huang et al. (2009, 2011a) performed finite-analytic Navier-Stokes (FANS) simulations on