Response of Pipelines Under VIV: An Experimental Investigation

Author(s):  
Prashant K. Soni ◽  
Carl M. Larsen

Pipelines laid on an uneven bottom often have free spans. For cases with long spans, one may have several modes and eigenfrequencies that can be excited by vortex shedding. Furthermore, due to the sag effect of a long free-span, the dynamic properties are different in vertical and in horizontal directions. This causes a complex response pattern in the cross-flow (CF) and in-line (IL) directions. From previous research we know that pure IL response at relatively low current velocities may significantly contribute to fatigue damage. This response type must be studied in addition to the combined IL and CF response. The objective of this paper is to present experimental results from flexible beam experiments where both response types are studied, as well as to present results from an empirical response model for the same cases. The empirical model is based on two types of experiments. The first set of experiments were conducted with a flexible pipe for both single and double span configuration. Pure IL and combined IL and CF motions were observed. In the second set of experiments, forces on a rigid cylinder were measured under forced motions in a current. The motions were found from measurements of cross section in the flexible pipe tests. Hydrodynamic coefficients such as drag, added mass, excitation and damping force coefficients were found and then applied in the empirical response model. In the present paper the results from the flexible beam experiments are presented and also compared with the results from the empirical response model. The results so far are encouraging, but further work and more data are needed in order to have a general and robust response model for combined CF and IL VIV.

Author(s):  
Decao Yin ◽  
Elizabeth Passano ◽  
Carl M. Larsen

Slender marine structures are subjected to ocean currents, which can cause vortex-induced vibrations (VIV). Accumulated damage due to VIV can shorten the fatigue life of marine structures, so it needs to be considered in the design and operation phase. Semi-empirical VIV prediction tools are based on hydrodynamic coefficients. The hydrodynamic coefficients can either be calculated from experiments on flexible beams by using inverse analysis or theoretical methods, or obtained from forced motion experiments on a circular cylinder. Most of the forced motion experiments apply harmonic motions in either in-line (IL) or crossflow (CF) direction. Combined IL and CF forced motion experiments are also reported. However, measured motions from flexible pipe VIV tests contain higher order harmonic components, which have not yet been extensively studied. This paper presents results from conventional forced motion VIV experiments, but using measured motions taken from a flexible pipe undergoing VIV. The IL excitation coefficients were used by semi-empirical VIV prediction software vivana to perform combined IL and CF VIV calculation. The key IL results are compared with Norwegian Deepwater Programme (NDP) flexible pipe model test results. By using present IL excitation coefficients, the prediction of IL responses for combined IL and CF VIV responses is improved.


Author(s):  
Prashant K. Soni ◽  
Carl M. Larsen ◽  
Chittiappa Muthanna

Pipeline laid on irregular seabed terrain may have free spans. Due to current, such spans may experiences vortex induced vibrations (VIV), which may lead to fatigue failure. The dynamic properties of free spanning pipelines cause a very complex response pattern and adjacent spans may also have some kind of dynamic interaction. A first set of experiments with flexible pipe model, see Soni & Larsen (2006), showed that the maximum response amplitudes for two interacting spans are higher than for equivalent single span cases. The interaction between IL and CF response will probably have some influence on the response level, in addition to the interaction between adjacent spans. A second set of experiments has been conducted with a motion controlled rigid cylinder in order to find the hydrodynamic coefficients for different flow conditions and also to observe how combination of IL and CF motions will influence the hydrodynamic forces. The cylinder was forced to follow an oscillatory pattern found from the first set of experiments with flexible pipe model. The Reynolds’s number and the dimensionless frequency were kept the same for both types of tests in order to ensure that the flow conditions are identical. The vortex shedding process for the motion controlled rigid cylinder has been mapped using Particle Imaging Velocimetry (PIV) under varying oscillation conditions. Improved understanding of correlation along a flexible beam and the interaction between cylinder motions and vortex shedding is hence obtained. The variation of lift coefficient along the pipe length supports the theory given by the authors; see Soni & Larsen (2005), for energy transfer between the spans. Thus, the spans interact dynamically.


Author(s):  
Ida M. Aglen ◽  
Carl M. Larsen

The importance of cross-flow (CF) response generated by vortex induced vibrations (VIV) of free spanning pipelines has long been recognised. The significance of in-line (IL) vibrations has recently been understood and hence also been subjected to research. The combined effect of CF and IL vibrations is, however, still not fully described. This paper highlights the CF-IL interaction with focus on the transition zone from pure IL to CF dominated response, giving special attention to how the added mass influences the interaction. Results from extensive flexible beam tests connected to the Ormen Lange (OL) development have been used as a basis for this study. Trajectories for cross sectional motions from the flexible beam test were identified, and then used as forced motions of a large diameter rigid cylinder exposed to uniform flow. Non-dimensional parameters like Reynolds number (Re), amplitude ratio and reduced frequency were identical for the two tests. Hence, forces found from the forced motion test could be used to find hydrodynamic coefficients valid for the flexible beam case. This paper discusses the results from the flexible beam tests with a relatively short length to diameter ratio (L/D) of 145. Modal analyses by Nielsen et al. (2002) show that the first mode dominates in both directions for these particular tests, even though the IL response frequency is twice the CF frequency. In this paper the added mass variations along the OL flexible beam is studied. Forces acting on 4 different cross sections along the beam are measured for 7 different prototype velocities. For each test the hydrodynamic coefficients are calculated, and the results show how the added mass changes along the beam for increasing velocities, and thereby creates resonance for both IL and CF response. The stability of the added mass coefficient throughout the time series is also evaluated.


Author(s):  
Carl M. Larsen ◽  
Rune Yttervik ◽  
Kristoffer Aronsen

Pure in-line (IL) vibrations will in many cases contribute significantly to fatigue damage for free spanning pipelines. This might be the case even if IL amplitudes are smaller than cross-flow (CF). While CF response has been subjected to research for a long time, little attention has so far been given to the pure IL VIV case. The hydrodynamic coefficients needed for response calculation have in fact not been available until recently, but results from forced IL oscillations have improved this situation. Data for added mass and force in IL direction has been used to establish a general response model along the same lines as for traditional CF response analysis. This has made it possible to calculate stresses from IL VIV in free spanning pipelines, and include the influence from interaction with the seafloor at the span shoulders. A brief presentation of the analysis method is given, but the main part of the paper gives results from a case study that illustrates important effects and the significance of IL response as compared to CF.


Author(s):  
Decao Yin ◽  
Carl M. Larsen

Long marine risers exposed to ocean currents will experience vortex induced vibrations (VIV), which may cause severe fatigue damage. VIV is, however, generally less understood than other load effects. The offshore oil industry has therefore supported an intensive research activity on VIV during the last two decades. High mode VIV model tests with long flexible riser models were initiated by the Norwegian Deepwater Programme (NDP). A 38 m horizontally towed instrumented riser was tested in uniform and linearly sheared current profiles with varying towing speed. A second series of experiments has been conducted with a motion-controlled rigid cylinder in order to find the hydrodynamic coefficients for realistic cross-section trajectories. The pipe was forced to follow periodic patterns found from the NDP tests with flexible beam. The Reynolds’ number and the non-dimensional frequency, as well the amplitude ratio was kept identical for both types of tests, ensuring that the flow conditions for these two experiments remain the same. The hydrodynamic coefficients calculated from natural trajectories show a general agreement with pure harmonic forced motion tests. A slight change of excitation regions was, however, found for cross-flow response. Another observation is that in-line excitation force coefficients have much higher values than found from pure in-line test.


Author(s):  
Decao Yin ◽  
Carl M. Larsen

Vortex induced vibrations (VIV) of long, slender marine structures may cause severe fatigue damage. However, VIV is still not fully understood, which calls for further research on this topic. This paper discusses results from experimental and numerical investigations of forces on rigid cylinders subjected to combined in-line (IL) and cross-flow (CF) motions, and it aims at improving the understanding of the interaction between IL and CF response components. Model tests with a long flexible beam were conducted at MARINTEK for the Norwegian Deepwater Programme (NDP). The model was 38 m long and it was towed horizontally so that both uniform and linear sheared current profiles could be obtained. Orbits for cross section motions at selected positions along the beam were identified in these tests. Forced motion experiments using these orbits were later carried out in the Marine Cybernetic Laboratory at Norwegian University of Science and Technology (NTNU). A 2 m long rigid cylinder was towed horizontally and forced to follow the measured orbits with identical amplitude ratio, non-dimensional frequency and Reynolds number as for the flexible beam tests. Parts of the results from these tests were published by Yin & Larsen in 2010. In this paper results from an investigation of trajectories for six positions along the beam in a uniform current condition will be shown. Three orbits have nearly the same CF amplitude ratio at the primary CF frequency, and the other three have similar IL amplitude ratio at the primary IL frequency, which is twice the CF frequency. Hydrodynamic coefficients have been found from experiments and numerical computations were carried out to find vortex shedding patterns for these cases. The main conclusions are that the IL motion component is a significant influence factor, and that higher order displacement components are more pronounced in IL direction than CF. Significant displacements in IL direction at 6 times the primary CF frequency were also observed, the ‘2T’ vortex pattern was captured when strong IL motion components were present. It is also seen that hydrodynamic coefficients should be found for combined CF and IL orbits and thereby improve the empirical models for prediction of VIV.


Author(s):  
Prashant K. Soni ◽  
Carl M. Larsen

For prediction of vortex-induced vibrations (VIV) the empirical models apply hydrodynamic coefficients to represent the fluid forces on the slender structures. The coefficients are in most cases found by measuring forces on a rigid cylinder under harmonic pure in-line (IL) or pure cross-flow (CF) forced motions, and are generally presented as functions of non-dimensional motion amplitude and frequency. The objective of the present work has been to find hydrodynamic coefficients for realistic combinations of CF and IL motions. Such trajectories were found from measured VIV of a flexible beam, and then used as forced motions of a rigid cylinder in uniform flow. Hydrodynamic forces were measured and used for calculation of hydrodynamic coefficients. The diameter of the rigid cylinder was larger than for the flexible beam in order to obtain optimum conditions for both experiments. However, both Reynolds number and non-dimensional frequency were identical for the two test types. The flexible beam oscillations were not perfectly periodic, but close to. More than one periodic trajectory could hence be identified as representative for the observed response, and these were used as forced motions in order to study the variability of the hydrodynamic coefficients. Alternative harmonic loops were also constructed in order to investigate the potential for using coefficients from harmonic tests as basis for empirical models. The vortex shedding process behind the cylinder has been mapped using Particle Image Velocimetry (PIV). PIV planes can picture the difference in cylinder wake for these trajectories and thus help to understand the process. The vorticity patterns at instantaneous positions for both periodic and harmonic trajectories are obtained. The vortices were mapped and the forces were measured simultaneously. Higher order harmonic components of the force are seen for all types of trajectories, and a correlation between these components and the vortex shedding pattern is observed.


Author(s):  
Chongyao Zhou ◽  
Gang Xu ◽  
Zhiming Huang ◽  
Dagang Zhang ◽  
Naiquan Ye ◽  
...  

Subsea pipeline laid on the seabed will experience free span when the lay path is long and seabed is rugged. Hydrodynamic loads caused by the currents around the pipeline can induce oscillations in both cross-flow and in-line directions. This phenomenon is called vortex-induced vibration (VIV) which is the most common case that could induce serious fatigue problems. The pipe-soil interaction is one of the main factors that influence the vibration. In this paper, a study focusing on the effect of pipe-soil interaction on VIV for different types of free span is presented. The Milan wake oscillator is applied to calculate the dynamic response induced by VIV in Orcaflex, and the results are compared with experimental data to identify its validity. A sensitivity study is also performed to study the parameter influence of the Milan wake oscillator model. Four types of free span (including the multiple free spans) are modeled in Orcaflex and time domain VIV analysis is carried out to study the influence of pipe-soil interaction. Comparison among different types of free span is discussed. The influence of structural damping is studied for flexible pipe only because its influence on steel pipe is negligible. The influence of structural damping on flexible pipe is studied by means of a predefined moment-curvature curve. In addition, several cases are studied to investigate the influence of tension on VIV by Milan wake oscillator.


Author(s):  
Djoni E. Sidarta ◽  
Kostas F. Lambrakos ◽  
Carl M. Larsen

A methodology for analyzing risers for in-line VIV fatigue damage has been developed that is based on the code SHEAR7, and laboratory in-line VIV coefficients. The in-line VIV fatigue in many instances governs the design of the riser since in-line VIV starts at a reduced velocity of about 1 whereas the threshold reduced velocity for cross flow VIV is about 4. The methodology can treat sheared currents on the basis of the cross flow VIV modeling in SHEAR7. Through the SHEAR7 modeling, the methodology removes conservatism implicit in the present ad hoc procedures for calculating riser in-line VIV response on the basis of the DNV-RP-F105 code. The reduction in conservatism is due to accounting properly for the power-in region in the VIV excitation, the hydrodynamic damping, and competing modal excitation (multiple mode response). The inline VIV coefficients have been derived from laboratory tests at the Norwegian University of Science and Technology (NTNU). The paper presents the in-line VIV coefficients, and examples to demonstrate the methodology for riser in sheared currents. The coefficients derived from the NTNU tests are functions of both the in-line VIV response amplitude and the reduced velocity. The coefficients presented in the paper are scaled test coefficients. The scaling of the NTNU coefficients assures that the methodology calculates in-line VIV amplitudes that are consistent with the response amplitudes in DNV-RP-F105 for the case of a simply supported riser in uniform current. This DNV code, although written for pipelines, has been extended to risers in sheared currents on the basis of conservative approaches.


Author(s):  
Jamison L. Szwalek ◽  
Carl M. Larsen

In-line vibrations have been noted to have an important contribution to the fatigue of free spanning pipelines. Still, in-line contributions are not usually accounted for in current VIV prediction models. The present study seeks to broaden the current knowledge regarding in-line vibrations by expanding the work of Aronsen (2007) to include possible Reynolds number effects on pure in-line forced, sinusoidal oscillations for four Reynolds numbers ranging from 9,000 to 36,200. Similar tests were performed for pure cross-flow forced motion as well, mostly to confirm findings from previous research. When experimental uncertainties are accounted for, there appears to be little dependence on Reynolds number for all three hydrodynamic coefficients considered: the force in phase with velocity, the force in phase with acceleration, and the mean drag coefficient. However, trends can still be observed for the in-line added mass in the first instability region and for the transition between the two instability regions for in-line oscillations, and also between the low and high cross-flow added mass regimes. For Re = 9,000, the hydrodynamic coefficients do not appear to be stable and can be regarded as highly Reynolds number dependent.


Sign in / Sign up

Export Citation Format

Share Document