[jacky831]

Hi, there. i want to simulate the flow on the spillway, what i concern is the high-speed flow's aeration problem. i don't know whether FLOW3D's two-fluid vof can sovle this problem. Looking forward to your advice

[JBurnham]

FLOW-3D can solve most spillway problems, including high-speed aeration. It is easier (and usually much more accurate) to use a one-fluid VOF method. If it is desirable to have a clearly-defined free surface, then 1-fluid VOF (IFVOF=4, 5, or 6) is the best choice. Usually IFVOF=4 is the place to start unless the free surface is highly curved. In these cases, once the hydraulic flow model is complete, then do a restart and activate 'air entrainment' physics. Specify the air density, surface tension coefficient (do not use full surface tension physics, just air entrainment), and a calibration coefficient for the rate of entrainment. You will probably have to calibrate the coefficient for each mesh resolution you try, using a similar known case with experimental data to calibrate the coefficient correctly. If deaeration is important in the stilling basin, then after calibrating the coefficient, activate 'drift flux' physics, and specify the air density and a drag coefficient of 0.95 (from Karamenev and Nikolov's data on bouyant spheres). Use the viscosity of water for all viscosities, to reflect the fact that the air bubbles are always separated by at least a thin film of fluid, and turn on 'escape at free surface' to allow the air to exit the flow when it rises to the surface. Good luck!

[jacky831]

Originally Posted by JBurnham

FLOW-3D can solve most spillway problems, including high-speed aeration. It is easier (and usually much more accurate) to use a one-fluid VOF method. If it is desirable to have a clearly-defined free surface, then 1-fluid VOF (IFVOF=4, 5, or 6) is the best choice. Usually IFVOF=4 is the place to start unless the free surface is highly curved. In these cases, once the hydraulic flow model is complete, then do a restart and activate 'air entrainment' physics. Specify the air density, surface tension coefficient (do not use full surface tension physics, just air entrainment), and a calibration coefficient for the rate of entrainment. You will probably have to calibrate the coefficient for each mesh resolution you try, using a similar known case with experimental data to calibrate the coefficient correctly. If deaeration is important in the stilling basin, then after calibrating the coefficient, activate 'drift flux' physics, and specify the air density and a drag coefficient of 0.95 (from Karamenev and Nikolov's data on bouyant spheres). Use the viscosity of water for all viscosities, to reflect the fact that the air bubbles are always separated by at least a thin film of fluid, and turn on 'escape at free surface' to allow the air to exit the flow when it rises to the surface. Good luck!

thank you very much.i lerarn a lot from your reply! you must be a professional.

[JBurnham]

Thanks. There's a tech note here that gives more details of the air entrainment model: http://www.flow3d.com/pdfs/tn/FloSci-TN61.pdf. Good luck; I probably won't have time to respond to further questions on this topic. - Jeff

[Jing_min]

Originally Posted by JBurnham

FLOW-3D can solve most spillway problems, including high-speed aeration. It is easier (and usually much more accurate) to use a one-fluid VOF method. If it is desirable to have a clearly-defined free surface, then 1-fluid VOF (IFVOF=4, 5, or 6) is the best choice. Usually IFVOF=4 is the place to start unless the free surface is highly curved. In these cases, once the hydraulic flow model is complete, then do a restart and activate 'air entrainment' physics. Specify the air density, surface tension coefficient (do not use full surface tension physics, just air entrainment), and a calibration coefficient for the rate of entrainment. You will probably have to calibrate the coefficient for each mesh resolution you try, using a similar known case with experimental data to calibrate the coefficient correctly. If deaeration is important in the stilling basin, then after calibrating the coefficient, activate 'drift flux' physics, and specify the air density and a drag coefficient of 0.95 (from Karamenev and Nikolov's data on bouyant spheres). Use the viscosity of water for all viscosities, to reflect the fact that the air bubbles are always separated by at least a thin film of fluid, and turn on 'escape at free surface' to allow the air to exit the flow when it rises to the surface. Good luck!

Hi
First off all, TOO MUCH Thanks for JBurnha.
I have some questions:
1- Another way to simulate these problem, is using 2pahse model. Is it better than air entrainment?
2- What and Why do you mean by "do not use full surface tension physics"?
3- Are the coefficients in the air ent. model functions of mesh size? And what are their approximation? Who do they change with the mesh size?
4-What are the differences and applications between IFVOF=4, 5, or 6 ?

Thanks in advanced!

[mf_emp]

To add some points on JBurnham's comment:
Based on the one-fluid, variable density model, can handle the high air concentration flow. It treats the flow as a variable mixture density fluid; the variation of the fluid density depends on the volume of entrained air.
In tjis option, the model of the drift flux and buoyancy are also included. The air is not treated as a passive scalar variable, and the velocity of the air bubble can be different to the surrounding (water) flow. In the computation, the relative velocity should be calculated for the transportation of the aerated flow. For closure, the quadratic drift flux model equation is applied. FLOW3D obtains the relative velocity by the input of 1) the radius of the air bubble (R = 2mm), 2) Richardson-Zaki coefficient = 2.39 when Reynolds number is larger than 500 (FlowScience, 2007), and 3) the drift coefficient. The drift coefficient is (2pwater*R2)/9ywater = 0.8845 where R is the air bubble radius, ywater and pwater are the dynamic viscosity and density of water respectively.