The Flow Around FPSO s in Steep Regular Waves.

Transcription

1 The Flow Around FPSO s in Steep Regular Waves. R.H.M.Huijsmans, J. P. Borleteau Maritime Research Institute Netherlands, Wageningen, Netherlands, r.h.m.huijsmans@marin.nl SIREHNA, Nantes, France. March 17, 2003 Abstract This paper investigates the hydrodynamical aspects of a fixed FPSO in regular waves. The emphasis is geared to the validation of state of the art CFD techniques, using particle image velocimetry measurements on the water velocities around a captive model of an FPSO in regular waves. The main focus is on the following issues: Determination of the water velocities around the vessel using PIV techniques and pressures on a girth below and just above the mean waterline of the vessel using pressure gauges. Comparison of the RAO s of the water velocities, pressures and relative wave heights. The resulting pressure profiles along the girth, especially near the waterline, showed a significant non linear effect in steep beam wave conditions. The velocities as measured near the bilges of the vessel indicated a strong separation at the bilge s of the vessel. This leads to an overestimation of the water velocities near the bilge s of the vessel as calculated using a linear diffraction programme. Key words:linear Hydrodynamics, PIV, CFD 1 Introduction This paper presents the results of the experimental programme on the FPSO, ie-captive model tests for PIV measurements, extinction tests in waves for an EU project with acronym EXPRO-CFD. The experiments aim to perform measurements (flow data and hydromechanical data) on a FPSO and to deliver data for validation of state of the art CFD tools as e.g. reported by Graham et al [12], Visonneau et al [2]. The application of PIV has been studied e.g. with liquid sloshing in moving tanks over a long time. Historically, sloshing studies started with impact problems in LNG carrier tanks. Then, the developed numerical and experimental tools and methodologies have been applied to other industrial domains : aeronautics (Airbus A300 wing tank), aerospace (Ariane V oxygen tank), automotive transportation (tanker trucks, car tanks), food industry (milk tank). In the design of mooring systems and protection barriers of FPSO s in deep water often use is made of linearised (or weakly non-linear) potential flow methods. However in steep deep water waves non-linear effect like excessive rolling or green water become predominantly non-linear. Therefore model tests are very often required even in the early design phases. In an attempt to address these non-linear effects a consortium, called EXPRO CFD, funded by the European Commission, was formed. This consortium studied the usability and applicability of advanced CFD tools in the early design stages. Also in this consortium extensive model tests on FPSO s were performed to evaluate and validate the use of these advanced CFD techniques. In an approach to gather water velocity data in a spatial area around the FPSO in steep water waves use is made of 2-D PIV techniques. In this paper the flow near the bilge s of the FPSO are studied. Earlier attempts to use PIV techniques for offshore applications have been reported by Tukker et al. [11]. Applications of detail spatial velocity measurements in civil engineering and offshore engineering applications have been reported by Gray et al ([6], [5],[9]), Earnshaw et al ([3],[4]) and Jensen et al ( [8],[1]). The linear diffraction computations used in this study are based on the work presented by van Oortmerssen [10] and Huijsmans [7]. 2 General Presentation The model of the FPSO scale (1:80) is tested in the Shallow water Basin of MARIN. The measurements included: global loads, incoming waves, by means of arrays of wave gauges, the wave profiles and by means of optical techniques, the velocity field (PIV technique) near the FPSO. All measurements will Paper No: ISOPE 2003-PF-06 R.H.M.Huijsmans Page: 1 of 8

2 Designation Sym [] FPSO Length perpendiculars L pp m 285. Breadth B m Draft 70% fully loaded draft T m Displacement volume m Longitudinal Center of gravity LCG m (with respect to station 10) Center of gravity above keel VCG m 10. Metacentric height transverse GM t m 7.33 Pitch radius of gyration in air k yy m Roll radius of gyration in air k xx m Natural roll period T φ s 12.4 be characterized by fully unsteady hydrodynamic phenomena. 3 Model of FPSO For the captive PIV tests, the model of the FPSO will be fixed with a 6 component force transducer in order to measure the wave loads. The particulars of the tanker are displayed in the table above. The frames of the vessel are presented in figure(1) Figure 2: Location of Wave probes and pressure transducers Figure 1: Hull frames of FPSO 4 Instrumentation Global loads The FPSO is be kept fixed to a 6 degrees of freedom force transducer attached to the main carriage of the basin. This system will be constructed to measure forces (Fx, Fy and Fz) and over turning moments (Mx, My and Mz), Relative wave probes: 10 relative wave probes are attached to the FPSO on the wind ward side of the vessel as indicated in figure(2). Pressure gauges on a girth on the FPSO at station 10 / 15 : An array of 5 pressure gauges is placed in the wind ward side of the vessel at station 10 and 15 as indicated in figure(2). Flow velocity near FPSO: A laser sheet coincides with a vertical plane normal to the FPSO; for this purpose the optical head will be placed fixed to the basin floor; Two cameras placed outside the FPSO with the optical axis normal to the laser sheet; each camera will have a field of view of approximately 280x200 (mm) in the laser sheet; one camera is viewing near the free surface and one camera is viewing near bilges. The cameras and laser sheet are kept fixed to the basin floor. 5 Experimental Arrangement General Arrangement for Captive PIV Measurements The experimental set up was developed by SIREHNA on the basis on its previous experience in PIV measurements in large scale hydrodynamic facilities. The general arrangement of the experimental set up, illustrated in figure(3) was composed with the following compo- Paper No: ISOPE 2003-PF-06 R.H.M.Huijsmans Page: 2 of 8

3 the tests, the scanning frequency was set to the image frame rate, i.e. 60 Hz. Figure 4: RAO s Influence of position of Camera housing on Figure 3: Set up of PIV measurement system nents : One seeding pipe, One underwater laser sheet to light a vertical plane parallel to the wave propagation direction, Two cameras, placed inside a camera housing, to record images of the measurement area illuminated by the laser sheet. The camera housing positioning should have been such that no hydrodynamic interference with the flow at the midship position would be present. The influence of the camera position on the results is displayed in figure (4). Here the RAO s of the relative wave height at station 6 on the windward side is calculated with a linearised diffraction code for two situations. One situation with the camera housing present in the calculations and one without the presence of the camera housing. From the comparison one must conclude that for a proper CFD validation also the position of the camera housing has to be taken into account. Laser Sheet The laser sheet was created in a vertical plane, parallel to the direction of the wave propagation, thanks to an underwater optical head (diameter 60mm, length 140mm) and a fiber optic linked to a continuous Argon laser source see figure (5). The thickness of the laser sheet was on the order of 3 to 4 mm. For Figure 5: Set up of Laser Sheet on the Basin Camera Images of the flow visualization were recorded by means of the two cameras. Progressive scan cameras allowing full frame rates of 60 Hz are used (644 pixels x 480 pixels). The two cameras were installed inside a camera housing placed at a fixed position in the basin with a distance of 1.45 m (in the basin) to the laser sheet. Each camera has a Field Of View (FOV) of approximately 280 mm x200 mm in the laser sheet. One FOV was set-up to observe the bottom of the model and the bilge. The other FOV was tangent to the side of the model with the mean free surface place in the middle of the FOV. Paper No: ISOPE 2003-PF-06 R.H.M.Huijsmans Page: 3 of 8

4 Image Recording For the present tests, the image acquisition rate and the number of recordable images were set to 900 images at 60 Hz for each computer (i.e. camera) which represent a time duration of 15 seconds in the basin or 135 second in full scale Seeding The water was seeded by means of a seeding pipe placed 6m upstream the laser sheet. The seeding pipe is presented on figure (3) which shows the general set up for PIV measurements. Summary of the Tests The program for the tests was based on the following parameters : Incoming waves monochromatic waves among 10 waves defined by 2 amplitudes and 5 periods, wave amplitudes of 2.0 and 5.0 m and wave periods from 9 to 16 seconds Incidence of the model with respect to the waves : Beam seas, Bow quartering seas Position of the PIV measurements area along the model Station 17.5 Station 10 (amidships), Station 6 6 Discussion of Results PIV Images near bilge s of FPSO model The results of the PIV measurements consist of a time sequence of frames where the water velocity are displayed. A sample is displayed in figure (10) to (13). For two points the RAO of the y and z water velocity are determined. The two points, indicated as A and B in figure(1), are near the bilge of the FPSO vessel. From linear diffraction analysis the y-z water velocity RAO are determined. They are plotted together with the results from the PIV analysis in figure (6) to figure (9). As can be seen from the calculated and measured water velocity RAO s an definite over estimation from the linear computations is observed as expected. The calculated vertical water velocities at point A, just below the keel of the vessel, showed no significant amplitude. This was also observed from the measured vertical water velocities at point A. However the measured data did produce a significant amount of scatter in the RAO s. The difference between the two measured RAO s, ie for the 2.0m and 5.0m wave amplitude case, did not lead to a definite conclusion with respect to the non-linearity of the observed watervelocities. From the displayed water velocity vector plots ( see figure(10), (11) and (12) one observes a marked separation region of the flow near the bilge s of the vessel. Frame of Reference The optical measurements are provided in a right-hand orthogonal frame of reference (O,x,y,z) which is fixed with respect to the basin and defined as follows: x-axis is parallel to the wave propagation, positive from upstream to downstream, z-axis is vertical, pointing upwards, y-axis is perpendicular to the x-axis and the z-axis. The origin, O, of the frame is located at the intersection of the plane of the mean free surface and the vertical line on the side of the model illuminated by the laser sheet. Figure 6: Y- Water velocity RAO s compared with PIV measurements at A PIV Analysis The analyses of the images recorded during this test campaign was composed of the two following main steps: Global cross correlation, Adaptive cross correlation. Each step was completed by a global filtering, a local filtering, an interpolation and a smoothing process. The figure (10) presents an example of velocity map determined from the two consecutive PIV images. Global Wave Load Response on FPSO The measured global wave load response in sway heave and roll are displayed in figure (14) to figure (15), together with the calculated linear response RAO s. Here the non linearity effects are observed from the two wave amplitudes, 2.0m and 5.0m that have been used in the experiments. One observes that the calculated RAO s of the wave loads fairly well agree with the RAO taken for the wave amplitude of 2.0m tests. Paper No: ISOPE 2003-PF-06 R.H.M.Huijsmans Page: 4 of 8

5 Figure 7: Z- Water velocity RAO s compared with PIV measurements at A Figure 9: Z- Water velocity RAO s compared with PIV measurements at B Figure 8: Y- Water velocity RAO s compared with PIV measurements at B Figure 10: Vector Plots of Watervelocity obtained from two Consecutive PIV Images Pressure Gauges and Relative Wave heights At the midship (station 10) and station 15 a set of 5 pressure gauges are mounted onto the FPSO model. From the captive tests the time series of the pressures are obtained. A sample from these time series is displayed in figure (16) for a beam wave condition of 11.3 seconds wave period and a wave amplitude of 5.0 m. In order to evaluate the effective non linearity of the pressure and/or relative wave height a plot is presented where the pressures are plotted against the relative wave height. These plots are displayed in figure (17), figure (18) and figure (19), corresponding to the time serie as displayed in figure (16). The pressure maps of the gauges just above the waterline, as indicated in figure(17) as P10top, shows a hysterysis type of behavior. When no non-linear effects would be present, then a straight line is expected. This last behavior is visible in the pressure maps below the mean waterline as presented in figures (18) and (19). Conclusions From the PIV experiments one can conclude that the quality of the velocity map data provides rich enough information for the validation of advanced CFD codes. Clear separation bubbles are observed and quantified from the PIV experiments. The initial validation of the water velocity RAO s are done with standard linear diffraction theory. As can be expected in the regions not close to the bilge of the vessel one sees a fair comparison of the theory and measurements. The main deviations, especially below the keel of the vessel in the vertical velocity RAO s are to be related to low values of the vertical Paper No: ISOPE 2003-PF-06 R.H.M.Huijsmans Page: 5 of 8

6 Figure 11: Vector Plots of Watervelocity obtained from two Consecutive PIV Images Figure 13: Vector Plots of Watervelocity obtained from two Consecutive PIV Images Figure 12: Vector Plots of Watervelocity obtained from two Consecutive PIV Images water velocity, which hampered the accuracy of the measured data. Also separation of the fluid flow near the bilges will give strong deviation from linear theory. As far as the pressure measurements are concerned, the pressures measured showed a strong asymmetry, due to wet-dry regions on the vessel. References Figure 14: RAO of Vertical Wave force [1] Clamond D. and Grue J. and Richon J-B. and Gray C. Jensen, J. and Sveen. Accelerations in water waves by extended particle image velocimetry. Experiments in Fluids, [2] G.B. Deng and J. Piquet and X. Vasseur and M. Visonneau. A fully coupled method for computing incompressible turbulent flows. In Proceedings of 16th Inter- Paper No: ISOPE 2003-PF-06 national Conference on Numerical Methods in Fluid Dynamics, Arcachon France, [3] Greated C. and Easson W.J. Earnshaw, H.C. and Bruce. Piv measurements of oscillatory flow over a rippled bed. In Proceedings of 24th Int. Conference on Coastal Eng. (ASCE), R.H.M.Huijsmans Page: 6 of 8

7 Figure 16: Time series of a set of pressure gauges and relative wave heights Figure 15: RAO of Roll Moment [4] Greated C. and Easson W.J. Earnshaw, H.C. and Bruce. Low-cost particle image velocimetry: System and application. In Proceedings of 1998 Int. Conference Offshore and Polar Engineering, [5] Greated C. Gray, C. and Skyner. The measurement of breaking waves using particle image velocimetry. In Proceedings of 1988 Int. Conference ICALEO, [6] T. Gray, C. and Bruce. The application of particle image velocimetry (piv) to offshore engineering. In Proceedings of 1995 Int. Symposium on Offshore and Polar Engineering, [7] R.H.M. Huijsmans. Mathematical modeling of the mean wave drift force in current: A numerical and experimental study. PhD thesis, technical University Delft, [8] Huseby M. Jensen, A. and Sveen. Measurements of velocities and accelerations in steep irregular water waves. Report of Department of Mathematics University of Oslo, Norway, [9] McCluskey D.R. Gray, C. and Greated. An analysis of the scanning beam piv illumination system. Meas. Sci. Technol., 2, [10] Oortmerssen G. van. The motions of a moored ship in waves. PhD thesis, technical University Delft, Figure 17: Pressure map for the gauge just above the mean waterline [11] Tukker J. and Kuiper G. and Huijsmans R.H.M. Wake flow measurements in towing tanks with piv. In Proceedings of 9th Int. Symposium on Flow Visualization, Paper No: ISOPE 2003-PF-06 R.H.M.Huijsmans Page: 7 of 8

8 Figure 18: Pressure map for the gauge just below the mean waterline [12] D.Xiu, G.E.Karniadakis. A semi-lagrangian high-order method for navier-stokes equations. Journal of Computational Physics, 172, Figure 19: Pressure map for the gauge just above the bilge of the Vessel Paper No: ISOPE 2003-PF-06 R.H.M.Huijsmans Page: 8 of 8