Archive | 2019
In-Line VIV Based on Forced-Vibration Tests
Abstract
Excitation and added mass functions determined from forced vibration tests of a rigid cylinder undergoing harmonic motion in the flow are used in the semi-empirical software VIVANA to predict the VIV response of pipelines. An advantage of this approach, as opposed to the morecommonly-used response function approach, is that it can account for changing conditions along the length of the pipe, like changing current velocity, seabed proximity, and/or pipe diameter. This makes it useful for pipelines as well as for risers when such changes occur. Further, for pipelines, travelling wave effects play less of a role than for risers, so the VIVANA approach can be simplified by assuming the phase angle of the harmonic response is constant along the span. The interactions between cross-flow and in-line response that complicate the prediction of cross-flow VIV by the excitation function approach, do not arise for pure in∗Address all correspondence to this author. line VIV. For the latter case, using the pure in-line forced vibration test data of Aronsen (2007), it is found that both VIVANA approach and simplified ‘SIVANA’ approach thereof predict VIV amplitudes consistent with experiments on flexible pipe (Ormen Lange umbilical VIV tests), and the DNVGL-RP-F105 response function for a range of structural and soil damping values. In a companion paper, this approach is applied partially strake-covered pipeline spans, to show that a relatively small fraction of well-placed strake coverage is enough to suppress in-line VIV. NOMENCLATURE CF Cross-flow. FSP Free span pipeline. IL In-line. RP Recommended Practice. In the present paper, it refers to DNV-RP-F105 [1]. SIVANA Simplified version of VIVANA, in which the un1 Copyright © 2019 by ASME damped mode shape for the bare pipe is used in a Rayleigh-Ritz approach, leading to one equation for the natural frequency (based on the Rayleigh Quotient) and another for the amplitude of vibration (based on energy balance). UMB Umbilical. VIV Vortex-induced vibrations. VIVANA A semi-empirical VIV prediction software. Ce Excitation coefficient. Ca Added-mass coefficient. Fe Excitation force. D Diameter of the model. DH Hydrodynamic diameter of the model. EA Axial stiffness. EI Bending stiffness. L Length. TEFF Effective tension force. U Current speed. Ur Ur =U/ fnD Reduced velocity. fn Natural frequency. fosc,IL IL oscillating frequency. k Spring stiffness. m Mass per unit length. m∗ Mass ratio. nCF CF response mode. nIL IL response mode. w Weight in water. ζ Damping ratio. INTRODUCTION Pipelines lying on the seabed usually have free spans due to uneven topography of sea bottom, stiffness and weight of the pipeline, and the residual tension force. Free spanning pipelines may experience vortex-induced vibrations (VIV) due to ocean currents. VIV can result in fast accumulation of fatigue damage and severe structure failure ultimately. Predicting VIV accurately is hence important for the design of pipelines, and structural integrity of the pipeline throughout the entire service life. Several deepwater oil and gas fields have uneven seabed. One example is the Ormen Lange natural gas field on the Norwegian continental shelf. The water depth of the Ormen Lange field varies from 800 m to over 1100 m. Along the 120 km long pipeline from the subsea templates to the land terminals, there is a large number of free spans even with the optimized route. Extensive model tests of Ormen Lange gas pipeline were carried out in the Ocean Basin of MARINTEK (now SINTEF Ocean) during the period between 2000 to 2003 [2]. Full-scale prototype section of the Ormen Lange umbilical was tested in a later phase [3]. The objectives of these tests FIGURE 1. PURE IL EXCITATION COEFFICIENT CONTOUR FOR A BARE CYLINDER, OBTAINED FROM RIGID CYLINDER FORCED MOTION TESTS [8]. THE THICK LINE REPRESENTS RESPONSE AMPLITUDE WHEN EXCITATION COEFFICIENT EQUALS TO ZERO A/D|Ce=0. were to study the VIV response of free spans of the Ormen Lange pipelines and umbilicals, and assess the suppression effectiveness of helical strakes. Empirical models to calculate VIV of free span pipelines have been developed, such as VIVANA [4–7]. Rigid pipe model forced motion VIV test was performed to find the hydrodynamic coefficients such as excitation coefficient (Ce) and added mass coefficient (Ca). Pure IL response of flexible beams will take place at the primary mode at low reduced velocities simply because the frequency of IL forces is twice the CF frequency. Hence, coefficients for pure IL response are of interest, which was the motivation of Aronsen [8]. He did pure IL VIV forced motion tests on a rigid pipe at Reynolds number of 24000 in the Marine Cybernetic Laboratory (MCLab). The key results from such experiments are the hydrodynamic coefficients excitation coefficients and added mass coefficients. Figure 1 presents the excitation coefficients extracted from the forced motion experiments by Aronsen. It has been observed that CF motions will change the IL response significantly as compared to pure IL response. The IL amplitude will increase, and the CF and IL response frequencies will be decided by an adjustment of their added mass. Using the same rig, Aglen [9] performed forced motion tests with the measured orbits (combined IL and CF) from the Ormen Lange free span pipeline VIV model tests. Time domain approaches have been developed to tackle the non-linearities and complex flow [4,10]. Helical strakes are usually installed on the free spans to suppress VIV. To study the fatigue damage caused by IL 2 Copyright © 2019 by ASME 5 Ormen Lange model tests PIPE PRETENSION MEASUREMENT AXIAL STIFFNESS REGULATOR PRETENSION REGULATOR L = 4.7m, 7.0m , 9.0m, 11.4m LEFF/D = 100, 150, 200, 350 FIGURE 2. ORMEN LANGE MODEL TEST SETUP [16]. VIV on a free span pipeline partially covered with strakes, Shell carried out a series of numerical and experimental studies together with SINTEF Ocean [11–13]. A methodology to assess the VIV response for free spans partially covered with strakes SIVANA [14] is applied. The SIVANA approach is a simplified version of VIVANA proposed by Ralf Peek whereby the undamped and not-excited modes for the bare pipe are used in a Rayleigh-Ritz approximation. This readily yields the response function for bare or straked pipe, which is then used to estimate fatigue following DNVGL-RP-F105 [1]. This SIVANA approximation greatly simplifies the calculation at little cost in accuracy when applied to pipeline spans where only the lowest modes are excited and propagating wave effects are negligible. The overall objective of this paper and a companion paper [15] is to use the empirical VIV prediction software VIVANA and SIVANA, to predict IL VIV of free spanning pipelines partially covered by helical strakes. The present paper validates VIVANA as an effective tool to calculate VIV of free spanning pipelines without strakes. Selected cases from Ormen Lange free span pipeline VIV tests and Ormen Lange umbilical VIV tests are analyzed and compared with VIVANA’s results. ORMEN LANGE MODEL TESTS The Ormen Lange test setup is described in [16], see Fig. 2. A 12.0 m truss beam serves as the support structure for the pipe model. Universal joints are attached to both ends of the pipe model. At one end, the universal joint is fitted to a force transducer to measure the axial force. At the other end there is a mechanism for adjustment of axial stiffness and pretension of the pipe model. PROJECT NO. 302001981 REPORT NO. MT2016 F‐122 VERSION