FluentCFD協(xié)助NATCO對(duì)海上浮動(dòng)分離器的設(shè)計(jì)

1、1cfd simulation assists natco with offshore floating separator designby chang-ming lee, ph.d.kenneth warren, ph.d. ted frankiewicz, ph.d.21. semi-sub, tlp, and fpso: (floating production storage & offloading) -crude oil exploration and production companiesfrequently deal with fluid flows in proc

2、essing and storage equipment under different sea states. (weather conditions ranging from sometimes “calm” to frequently “stormy”).1. problem solutions: - installing internal baffles to suppress the liquid sloshing inside storage or processing vessels.3 eight cases have completed: - case 1: empty ve

3、ssel at normal sea state (only has tpvane section in gas phase). - case 2: empty vessel at storm sea state (only has tpvane section in gas phase). - case 3: vessel w/ pre-designed internals at normal sea state (including tpvane, performax, perforated plates). - case 4: vessel w/ pre-designed interna

4、ls at storm sea state (including tpvane, performax, perforated plates).4 eight cases have completed: - case 5: vessel w/ modified internals at storm sea state (opening at bottom was reduced from xx to xx, and height was raised to xx above liquid level). - case 6: vessel w/ all perforated plates at s

5、torm sea state (both performax sections have been replaced). - case 7: vessel w/ horizontal baffle at storm sea state (horizontal perforated plates was added at vessel tail). - case 8: internals were same as case #7 w/ flows into vessel (gas = xx mmscf/day; liquid = xxxxxx bbl/day).561. surge: y2. s

6、way: x3. heave: z4. pitch: zy5. roll: zx6. yaw: xyzxy543216789one-directionalmovementtwo-directionaldiagonally movement1011table deck size:7 ft. x 10 ft. reference elevation: 8 ft. 8 inch. displacement: +/- 12 inch. angular motions: +/- 7.5 degree. motion periods: 3 to 15 seconds. maximum load: 4000

7、 lb. location: indoors (tulsa r&d)1213geometry build upphysical modelmathematical modelmeshing domainb.c. & case set upcomputing (iterations)results (visualization)validation141. cfd model assumptions - 3-d geometry is based on the approximate-dimensions of the fwko separator (10 x 40 exclud

8、ing heads). - revolution inlet device is simplified as two solid blocks; sand pan, sampling wells and vortex breaker are neglected.- volume of fluid (vof) with “geometric-reconstruction” scheme (multiphase flows model) was applied for two phases only (oil and gas).3-d cfd simulations approach153-d c

9、fd simulations information applied software packages - 3-d geometry and volume meshing generation:gambit 1.3.0 (fluent inc., lebanon, nh)“hex” & “tet/hybrid” volume elements. - cfd simulation and post-processing:fluent 5.4.8 and fluent 5.5.5 3d implicit segregated solver (unsteady)with standard

10、“k-e e” turbulence model. - 3d wave motion user defined function (udf).161. material properties: - gas phase: density = xxxx kg/m3; viscosity = 0.015 cp - oil phase: density = xxxx kg/m3; viscosity = 15.0 cp(note: operating condition: t = 125 f , p = 260 psiggas flow rate = xxxx mmscf/day , andcombi

11、ned liquid flow rate = xxxxxx bbl/day)1. boundary conditions: - tpvane:porous zone - perforated plates: porous-jump - performax: porous zone3-d cfd simulation case set up172-d drawing of separator from autocad18layout of separator on offshore platformne (bow)pitchrollcenter of rotation “storm” sea s

12、tatecenter of rotation “normal” sea statefwkokeel plate deck surge 19pmax1 (porous zone)pmax2 (porous zone)weirplate (wall)perplatehead (porous-jump)perplatetail (porous-jump)inletdevice(solid zone)tpvane (porous zone)vessel tailvessel headliquid leveloriginally designed internals of the vessel20coa

13、rse mesh“tet/hybrid”39,739 cellsfine mesh“tet/hybrid”125,879 cellsvolume meshing comparison (coarse vs. fine)21case 1 & 2:“normal & storm”“empty” - vessel with tpvane only 22drag coefficient on vessel tail (normal empty)23contour plot of oil volume fraction (empty vessel)19.60 sec20.80 sec24

14、contour plot of oil volume fraction (empty vessel)25.20 sec26.60 sec25case 3 &4:“normal & storm”“porous” - vessel w/ all internals(tpvane, performax, perforated plates) 26drag coefficient on vessel tail (storm porous)27contour plot of oil volume fraction (w/ internals)19.646 sec20.80 sec28co

15、ntour plot of oil volume fraction (empty vessel)25.246 sec26.60 sec29entire domain is completely meshedwith “hexahedral” elementvolume meshing generation from “gambit”total grid number = 28,77030modified internals of the fwko separatorperformax1 (porous zone)performax2 (porous zone)weirplate (solid

16、wall)perplatehead (porous-jump)perplatetail (porous-jump)liquid inlet(velocity-inlet)tpvane (porous zone)vesseltailvesselheadliquid levelhorizontal plate (porous-jump)gas outletliquidoutletgas inlet(velocity-inlet)31drag coefficient on vessel tail (case #7)32oil-gas interface inside the separator (c

17、ase #8)time = 66.0 sec33animation of oil-gas interface (case #8)34contour plot of velocity magnitude (case #7)23.70 sec35contour plot of turbulent kinetic energy (case #7)23.70 sec36velocity vectors colored by velocity magnitude (case #4)time = 21.0 sec37velocity vectors colored by velocity magnitud

18、e (case #4)time = 27.0 sec38 consideration of sea state (“normal” vs. “storm”) - due to the variation of “center of rotation”, the sloshingmagnitudes inside the separator vessel were not much different for 1-year or 10-year storm sea state. mitigation of wave motion (“empty” vs. “porous”) - the adde

19、d internal baffles can very effectively suppress the liquid sloshing inside the vessel due to wave motion. comparison of internals (“pmax” vs. “perplate”) - replacing “performax” section with perforated plates has illustrated a positive effect on suppressing wave motionbased on the decreasing of dra

20、g coefficient on vessel wall. 39comparison of liquid sloshing (“storm” sea state)26.60 sec2d contour plots of oil volume fraction26.60 sec40comparison of drag coefficients (“storm” sea state)41comparison of drag coefficients (case #5 & #6)42 effect of addition of “horizontal baffle” - adding a h

21、orizontal perforated plates at the vessel tail (oil well section) did not show any apparent benefit. effect of “flows” into the vessel - there is no adverse result was observed with both gas andoil flow into the vessel for simulation time up to 85.60 sec. consideration of “roll” motion - according to the 3-d animated plots of interface tracking,the vessel i