...< ISBN 0-96261044-5 FLAREBOOM VIBRATIONS - VORTEX SHEDDING AND DAMPING OF WELDED INDIVIDUAL STEEL MEMBERS Arne KVltl ud and Lei! J Dalsgam d NorwegIan Petroleum Duectorate Stavanger, Norway ABSTRACT The Norwegian Petroleum Directorate recommend the structural damping ratio of welded individual members to be taken as 0.15% of critical damping when evaluating vortex induced vibrations. This paper descibres the background for the recommendation which mainly is based on the measurements of the vibrations on the Heimdal flareboom. Severe vortex induced vibrations in Norwegian flarebooms_ have occured twice. The first occurence was in 1978 on the flareboom at Statfjord A. The second incident was at the Heimdal flareboom in 1984/85. Both incidents gave valuable information about vortex shedding problems. KEY WORDS Vibrations, damping, vortex shedding, flareboom, wind loading, critical wind velocities STATFJORD-A VIBRATIONS At Statfjord A the vibration was an overall vibration of the tie, see figure 1. Calculations of the natural frequency of the individual panel sections of the main chord showed that it was close to the recorded frequency and to the vortex shedding frequency for the chord. The calculation of this natural frequency had not been inCluded in the design. The problem was solved by spirally wrapping the main chords of the ties with 25mm diameter ropes. Three ropes was wrapped on each chord at a phase difference of 1200 , each rope spiral had a pitch of one metre. The spirals on diagonally opposite chords of the tie had the same rotational direction (clockwise) and the other pair was wound in the opposite direction. 172 Afterwards no more vibrations have been reported (Mobil, 1978). A large measurement program was planed, but was cancelled after the successful use of the ropes. Measurements were later performed to study the overall dynamic effects in the flare boom. It was conCluded that the design gave a slight under estimation, but it was compensated for by conservative load calculations (Aas- Jakobsen, 1979). No measurements were performed to actually obtain values for the damping in the vibrating parts of the structure. HEIMDAL VIBRATIONS The Heimdal platform was installed during the summer 1984. A general drawing of the platform is showed on figure 2. A study requested by NPD was carried out by Bergen Engineering (1984). The study revealed that the dynamic amplification was left out in the design. The study concluded that several elements were critical to vortex shedding based on fatigue and reduced velocity calculations. The later observed performance was also in a good agreement with the predictions from Bergen Engineering. Based on an evaluation of the consequences of possible vortex vibration, the NPD gave concent for installation. However, NPD required observations to be incorporated in the normal structural inspection program for Heimdal. These observations lead to the discovery of the vibration problems November 1st 1984. The vibrating members are shown on figures 3 and 4. All nodes connected to critical members were inspected with Non Destructive Examination methods. Magnetic particle inspection and dye penetrant of brace to chord welds (all welds at node) was used in order to determine whether cracks developed. Two cracks were observed and later on repaired. Material from one of the cracks was removed for laboratory testing. It was later established that the crack was caused by fatigue. In order to gain further information on the dynamic characteristics of the Heimdal flare boom, a measurement programme was conducted by Veritec (1985). This programme comprised of exciting six members of the flare boom with a rubber hammer and measuring the resulting vibrations with 1-3 transducers mounted on the member. The output signals were recorded on tape for subsequent analysis. A spectral analysis was performed on the free vibration signals using the Fast Fourier Transform (FFT) method. This produced information relating to the natural frequencies on the flareboom members. The damping values were calculated from the decay in the vibration amplitude over time, the signals first being filtered at the resonance frequencies. A comparison between the measured and calculated frequencies for the members in the top chord was made. The calculated frequencies was selected to correspond to vibration of the specific members only. The calculated frequencies was relatively close to the actual measurements. However, the measurements showed significant responses at more than one frequency. These measured frequencies did not appear to be simple multiples of each other. It was concluded that the multiple frequencies was not caused by different vibration modes of the single members, but could be a result of interaction with surrounding members. In December 1984 and January 1985 a total of 19 members of the flare boom was wounded with ropes. Rope diameter was approximat1y 10-12 % of diameter. One loop of the rope was covering a length of 5 times the diameter. Three ropes were used. The ropes were equally spaced around the perimeter. On roped members a damping ratio of approximately 1% of critical was measured (Veritec, 1985b). After winding small vibrations observed in 5 of the members. were In January and February 1985 a total of 4 nodes and 2 members were braced by fitting a wire to the structure and then pretensioning the wire. In total these methods were effective. Nevertheless small vibrations were observed, but since the observed amplitude was approximatly 10 mm, the vibrations were not considered critical. 173 Different methods to permanently stop the vibrations were evaluated by DnV (1985), Technip Geoproduction (1985) and Brown and Root (1985). Based on these evaluations additonal wind tunnel tests were carried out by Laboratorie Eiffel, Paris. The shroud solution gave the best results. Shrouds were used for the final modification. All members being critical according to the final design criteria, were modified. No vibration has been reported after the modifications with shouds. Later measurements on the flare boom shows satisfactory behavior of the structure. A description of the measurements is given by Doucet and Nordhus (1987). VORTEX SHEDDING IN DESIGN The design against cross-flow vibrations on Heimda1 was performed according to DnV (1976 and 1977) stating that cross-flow vibrations may occur where: VR > 3.5 reduced velOCity) and KS where VR KS using V f 0 ME 0 g DAMPING < 16 stability number) V I ( f * 0 ) (2 * ME * 6 ) I ( g * 02 particle vind velocity fundamental frequency of the member member diameter member mass pr unit lenght logaritmic decrement mass density of fluid A damping of 0.5% was used...

... of the form x + 2~WO~ + wax - F(t) The input-output relations in the time domain are: -x(t) = J h(T)F(t-T)dT (1) a where hIT) is the impulse-response function: hIT) = fIJ~l eXPflJoT) sin (flJdT) ; 1/;or T 2: 0 and zero elsewhere and flJd = Wo (1 - ~) is the damped natural frequency. In what follows...

... upstream oil &amp; gas modulation correspond amplitude modulation vehicle frequency ship velocity cable experimental run operation drag coefficient drag force vibration triantafyllou configuration experiment flow-induced vibration ship tow cable Abstract Drag...

... bi-frequency domain sum frequency frequency monochromatic wave mooring system equation upstream oil &amp; gas surge response experiment dual wave subsea system coefficient drift force drift motion wave drift amplitude surge drift motion prediction excitation slow drift motion...

... ABSTRACT The mooring line dynamics is modelled using the finite element method, and its non-Gaussian response property is investigated using the time domain method. By comparison to the time domain (TD) approach, the application of the hybrid time-frequency domain (HTFD) fatigue analysis...

... to be conservative at wave excitation close to the system natural frequencies. A non-linear analysis was undertaken to confirm this hypothesis. The introduction of representative damping through viscous drag forces was expected to reduce the conservatism. The results from a non-linear time history analysis...

...PlOceedlngs o/the First (1990) Ewopean OjfshOle Mechamcs SymposIUm T! ondhelm, NOI way, 20-22 August 1990 Copyl/ght © 1990 by The InternatIOnal SO(lety o/OjfshOie and Polar Engineers ISBN 0-9626/04-4-5 LOW-FREQUENCY MOTIONS OF SEMISUBMERSIBLES Davtd Pzzer The Manne Technology Centre, Umverslty...

... certain conditions e.g. a) when the wave frequency and the natural frequency of the structure lie within close proximity of each other or b) assumption of a quasi-static response of the structure neglecting the effect of the natural vibration which is applicable only in sever sea states...

... of N regular waves with different amplitudes and frequencies, that is, N 71(0, t) = L(an coswnt + bn sinwnt) (11) n=l where the Fourier coefficients, an and bn are obtained from a direct Fast Fourier Transform (FFT) of the input time series. At any instant of time the kinematics and pressure are given...

... amplification (Triantafyllou et al, 1985) attains its maximum value within the wave frequency range. This is entirely due to the action of the nonlinear drag force, Which, by preventing the mooring line from lateral motion, forces it to respond by stretching along its tangential direction. It is therefore...

... structures. IPOS consists mainly of the following subsystems (see Figure 1): L General-purpose response analysis program for any type of off- shore structure (IPOSa) 2. Stability analysis program for any type of floating structure 3. Frequency-domain motion and structural analysis program for floating...

... frequency maximum value response calculation short crested sea 100-year value sea state crested sea fatigue life spectra independent exponent long crested sea exponent main wave direction calculation upper limit mitsuyasu spectrum response calculation svene havel statoll...

... of interaction, has to be envisaged. In this paper main parameters of concern for risk of pipe damage from dragging ship anchors are discussed, in particular the ones which give a significant contribution to the assessment of the frequency of pipe-anchor interaction and of the pipe damage. 1. INTRODUCTION...

... ABSTRACT It is demonstrated how information about the undisturbed wave at all frequencies of interest can be retrieved by constrained deconvolution of the outputs of several sensors with different known linear response characteristics, placed on or around a floating structure. A model...

... and Vignoli, 1989). As an example, the recorded data and the corresponding Fourier spectra of wind speed, wave elevation, and strain are shown in Figs. 2, 3 and 4. There is a substantial agreement between the fundamental frequencies of the waves and the structural response shown in Figs. 3 and 4 which...

... interacting with the SOI~, whIch contributes to increase the natural frequencIes and consequently to reduce the occurrence of resonant oscillations in the operating life of the pipeline (Ref. 6). Purthermore, the pipe-soil interaction introduces an energy removal mechanism in the free span. Indeed...

...Ploceedzngs afthe Fllst (1990) EUiopean OffshOie MeLharn(s Symposium TlOndhe,m, NOI way, 20-22 August 1990 CopYllght © 1990 by The IntelnatlOnal Souety of Off,hm e and Polar EngzneelS ISBN 0-9626/04-4-5 ESTIMATION OF LOW-FREQUENCY MOTION DAMPING AND RESPONSE OF A DEEP-DRAUGHT FLOATER IN WAVES R G...

... had been confirmed by many researchers. 68 Considering some of the large diameter fixed offshore struc· tures, the natural frequency of the structure falls within the wave energy frequency range of the moderate sea. Thus the problem of resonance occurs. In the vicinity of natural fre- quency...

... with sufficient accuracy. However, with regard to second-order effect~, there are variations both in excitation forces and the damping models. Normally, the second-order (low-frequency) motions are assumed to be uncoupled from the first-order mo- tion. this assumption is not generally valid, and that may...

... and bottom axial stiffness flexural stiffness n-th actual eigenfrequency (quasi-static condition) frequency of vortex shedding, calculated from Strouhal's relationShip nominal natural frequency (static condition) drag average load, per unit length lift average load, per unit length length of the model axial...