CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Gravity component and inlet velocity (

Vidya July 24, 2006 22:26

Gravity component and inlet velocity
Hi all,

I'm modeling a flow in which the flow direction is opposite to the direction of gravity. I'm aware that the direction of gravity is set by the user in Fluent. The flow in my model is from the bottom to top (it enters at the bottom of the geometry and exits at the top in the z direction).

While running the simulation in Fluent, I activated the gravity component in the Boundary Conditions, and gave a value of 9.81 m/s2 in the Z direction. Then, when defining the velocity inlet, I gave a steady velocity of -0.175 m/s at the inlet face. The negative sign was used to specify that the flow is opposing gravity. Isn't this the right way to specify my velocity?

But for some reason, Fluent doesn't seemt o accept this and the computation goes on forever........ never converges. The outlet face in my model is to be maintained at atmospheric pressure. So I use the PRESSURE OUTLET bc and retained the default values. I'm also expecting some backflow in my model. Is there any way to define backflow?

Can anyone tell me why my solution doesn't converge and what may be the cause for my simulation to backfire?


Albert Franco July 25, 2006 02:25

Re: Gravity component and inlet velocity
Hello, I don't know if you alredy know that the blackflow parametres can be imposed at the presure outlet bc window. Apart from that I suppose you've imposed the signs depending on your arrows.


Razvan July 25, 2006 05:54

Re: Gravity component and inlet velocity
The "+" or "-" signs that you apply to a velocity boundary condition have two possible interpretations ONLY:

- in case you defined velocity as "normal to boundary", "+" = fluid enters domain, "-" = fluid exits domain;

- if you defined velocity as "magnitude and direction" or "components", then "+" and "-" are strictly limited to the reference cartesian coordinate system of your grid.

So, in your case you should define velocity boundary condition:

- as "normal to boundary" and with a positive value of v=175m/s,

- or as "magnitude and direction" with an absolute value of 0.175m/s and 0,0,1 for x,y,z components.

I think this is clear enough.


P.S.: the same applies to gravity direction too...

Vidya July 25, 2006 16:26

Re: Gravity component and inlet velocity

I understood what you explained about the velocity direction. But if I want to activate the gravity component, I just have to specify the value and say, if I want it in the Z direction, I can say gravity = 9.81 m/s2 in the Z direction with X,Y values as zero. Is this right? This would mean that gravity is in the Z direction. Now since my flow is opposing gravity, and I use the magnitude and direction methos, I have to give an absolute value of 0.175 and 0,0,1 as the vector components, right?

Please correct me if I'm wrong.


Vidya July 25, 2006 22:48

Re: Gravity component and inlet velocity
Hi Razvan,

I tried to use the 'magnitude and direction' option to specify the velocity inlet bc of 0.175 m/s.

Should I use 0,0,1 or 0,0,-1 to indicate that flow is in the -Z direction? I need to run two simulations- one in which there are gravity effects and the other in which gravity can be neglected. In the case where gravity has to be considered, since the flow is opposing gravity, should the Z direction vector be +1 or -1?

And in the case where gravitational effects are not considered, how do we impose the direction of flow. In my simulation, the first case is with the geometry standing vertical, in which case gravity is considered. The other case is where the geometry is horizontal and resting on the ground. So there are no gravity effects. However, in both cases, flow enters and exits in the same direction and in the same faces. How can I specify the velocity inlet bc in these two cases?

Thanks and regards.

Razvan July 26, 2006 01:10

Re: Gravity component and inlet velocity
As I already mentioned, settings for velocity components directions have nothing to do with anything else but the reference cartesian coordinate system of your grid (it has absolutely nothing to do with the direction of the gravity force in your model). To be sure of the right direction, simply plot the grid and verify direction using the small cartesian system displayed on the left side of the window.

So, for the gravity case, if the flow is poiting into the negative z direction, then in the velocity-inlet panel set 0,0,-1 for the x,y,z components of the velocity. In the non-gravity case, the new reference cartesian coordinate system would be the one you chose when constructing the grid in Gambit, so plot the grid on your screen and identify the new direction for the flow and set the x,y,z values accordingly.


Cappy July 27, 2006 03:29

Re: Gravity component and inlet velocity

-Flow goes up (positive Z) -Inlet is at bottom, Outlet is at top -Gravity pulls down (negative Z)


set inlet velocity to +0.175 m/s (positive into inlet) OR set inlet velocity to 0,0,+0.175 m/s (positive Z)

set gravity to 0,0,-0.175 m/s (negative Z)

Vidya July 30, 2006 08:55

Re: Gravity component and inlet velocity
Hi Cappy,

How can I set gravity to 0,0, -0.175 m/s? When the gravity option is activated, the X,Y, Z values are in the units of m/s2, which is acceleration. So I gave the values of 0,0,9.81m/s2.

I also use the "magnitude and direction" method to specify the inlet velocity. I gave a positive value of 0.175 m/s at theinlet and the direction vectors as 0,0,-1.

The solution ws initialized from the inlet zone. Still, for some reason, the computation goes on forever and the solution never converges. The residuals too keep on increasing. I don't understand why this happens. I tried with both segregated and coupled solvers, but both gave same results........ i.e. solution never converged.

Do you have any thoughts on this?


Cappy July 31, 2006 08:28

Re: Gravity component and inlet velocity
My mistake, I meant 9.81 m/s^2 for gravity.

For your convergence problems, I would make sure the initial condition for velocity (under solution-->initialization) is the same as the velocity inlet.

If changing your initial conditions doesn't work, look at your grid in Fluent and make sure that there are no unintended wall boundaries that block flow.

All times are GMT -4. The time now is 19:45.