- 113 Views
- Uploaded on
- Presentation posted in: General

Modeling of the Unsteady Separated Flow over Bilge Keels of FPSO Hulls under Heave or Roll Motions

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -

Modeling of the Unsteady Separated Flow over Bilge Keels of FPSO Hulls under Heave or Roll Motions

Yi-Hsiang Yu

09/23/04

Copies of movies/papers and today’s presentations may be downloaded from

Http://cavity.ce.utexas.edu/kinnas/fpso/

- FPSOs are tanker like floating hulls which are used for production, storage and offloading of oil.

- FPSO hulls have often been found to be subject to

excessive roll motions, and the installation of bilge keels has been widely used as an effective and economic way of mitigating the roll motions of hulls.

- The main focus of this research is to model the unsteady separated viscous flow over the bilge keels of a FPSO hull subject to roll motions and to determine its effect on hull forces.

- Numerical Formulations
- Governing Equations
- Numerical Methods
- Non-linear Term Treatment
- The Effect of the Moving Grid

- Results
- An Oscillating Flow over a Vertical Plate
- Submerged Body Undergoes Heave or Roll motions
- FPSO Hull Subjects to Roll Motions

- Conclusions and Future Work

- Governing Equation
Non-Dimensional Governing Equation (Navier-Stokes Equation & Continuity Equation)

where U represents the velocity; Q is the force term; and R indicates the viscous term. The definitions of the column matrices for the Navier-Stokes equation are given as

where the Reynolds number is define as Re = Umh/ν; and the length scale, h, is a representative length in the problem being solved.

- Cell Based Finite Volume Method
(Collocated variable, non-staggered grid arrangement)

According to the integral formulation of the Navier-Stokes equation and to the Gauss divergence theorem, a semi-discrete integral formulation of the momentum equation can be given as

where Sijis the area of the cell; and ds represents the length of each cell face. A cell center based scheme is applied where “i,j” is the center of the cell (non-staggered grid: unknown value u, v, p are located at the cell center).

when calculating the flux, the value on the cell face (at D) is needed. It can be obtained from Taylor series expansion.

- Crank-Nicolson Method for Time Marching
where f represents the summation of the convective terms, the viscous terms and the pressure terms at the present time step n and the next time step n+1.

- Pressure-correction Method
- SIMPLE method (Patankar 1980)
where p’ is the pressure correction, V’faceis the velocity correction term, әp’ /әnis the pressure correction derivative with respect to the normal direction of the cell face, V*face= (u*; v*) is the predicted velocity vector obtained from the momentum equation.

- SIMPLE method (Patankar 1980)

- Appropriate Pressure-Correction Equation
Since our 2-D Navier-Stokes solver uses non-staggered grid, the scheme has to be modified somewhat to avoid the checkerboard oscillation problem.

where aijis the coefficient of the unknown velocity in the momentum equation; "av" indicates the average value obtained from the cell center value; and "d" represents the value calculated directly at the face center.

- Non-linear Terms Treatment
The momentum equation can be rearranged as

where u* is the unknown velocity at T = n + 1; the coefficient “a” is also a function of the velocity at T = n+1 which can be obtained from the previous iteration; and dijis the coefficient of pressure in the momentum equation.

- Moving Grid

When the grid is moving,

additional terms need to

be taken into account.

where (ugrid, vgrid) is the velocity of the moving grid; and

represents the total change in the value of u with both increment in time and the corresponding change in the location of the point. When the above equation is substituted into the momentum equation

- The main focus of this research is to model the unsteady separated viscous flow over the bilge keels of a FPSO hull subject to roll motions and to determine its effect on hull forces.
- It can be simplified as three different problems:
- Oscillating flow over a vertical plate.
- Free surface effect (linear and non-linear).
- Submerged body with
or without the bilge keels.

- Consider the effect of the
bilge keels and the effect of

the free surface

- Oscillating Flow Past a 2-D
Vertical Plate

- Drag & Inertia Coefficient for a Range of
Kc=UmT/h (0.5 < Kc < 5)

- Submerged Body Motions

- Potential Flow Results of the Hull Undergoing the Heave Motion at t/T=0.25

FVM

FVM

- Potential Flow Results of the Hull Subject to the Roll Motion

The pressure distribution along the submerged hull without bilge keels

- Viscous Flow Results of the Submerged Hull with Bilge Keels Subject to the Roll Motion

- Results of the Unsteady Separated Viscous Flow over the Bilge Keels of a FPSO Hull Subject to Roll Motions

- Previous Results

- More details can be found in Kacham [2004], and Kakar [2002]

- A numerical scheme for solving the Navier-Stokes equation has been developed.
- It is well validated with experimental results in the case of an oscillating flow past a vertical plate.
- The method is also applied to the case of submerged bodies which are subject to forced heave or roll motions. The numerical results have shown good agreement with the potential solver (boundary element method).
- Then, the method is applied to the case of a FPSO hull undergoing roll motions. The effects of the bilge keels and of the free surface are also taken into account. The numerical results is improved after including the terms for a moving grid in a fixed inertial coordinate system.

- More convergence studies in time and space are necessary.
- The capability of the solver to handle the non-orthogonal grid geometry still needs to be improved (Some small oscillating behaviors exist around the bilge keels area).
- More investigations on the nonlinear free surface effect are needed.
- Extend the model in 3-D and compare with experiments and other numerical results.