Hi

I am trying to model an accelerated flow over a wing. I want to accelerate at 5 m/s^2 and use the time step as my change in time. How would I write a CEL expression to model this?

thanks

Hi,

What do you mean accelerated flow? Do you mean the wing accelerates in velocity? In that case you just write the inlet boundary condition velocity as a function of time and calculate the velocity as a CEL expression.

To help you write CEL expressions have a look as the tutorial expamples, a number of them use CEL.

Glenn Horrocks

If the wing is accelerating and you want to keep the domain constant (i.e. accelerate the entire domain), you'll need to change the inlet boundary condition to reflect the change in velocity. The expression is therefore the velocity as a function of time:

Vo = <initial velocity> [m/s] dvdt = 5 [m/s^2] Vin = Vo + dvdt*t

...and insert 'Vin' as the expression for your inlet velocity. You could also write this as a single expression in the boundary condition (i.e. <initial velocity> + 5 [m/s^2]*t), but breaking the variables out in this manner makes it easier to change variables later. You can also define one of these variables as a DX parameter to do a series of runs automatically.

If the fluid is compressible, you'll also need to add an acceleration term to your domain. To do so just put 'dvdt' in for the appropriate component of the acceleration due to gravity in the Domain form.

-CycLone

Hi Cyclone

Are you CFD professor? I and my friend think you are. And how old are you if you dont mind us asking.

Hi Sana,

I'm not a professor, but I'll take that as a compliment. I'm old enough to know and young enough to want to know more ;-)

-CycLone

The equations should have read:

Vo = _initial velocity_ [m/s] dvdt = 5 [m/s^2] Vin = Vo + dvdt*t

You are very smart.

-Sana

This will only work for incompressible flows. Otherwise the inlet velocity will not be 'felt' by the entire domain at each timestep, only those cells which are within c*dt (where c is the speed of sound) will feel the inlet change in V. Unless you set dt to something big, in which case your transient is wrong....

I agree, but that is why I recommended adding an acceleration to the domain for compressible flow.

-CycLone

[cpath]

Hi HekLer

I recently came across this thread.


Does this mean that, in the case of incompressible flows, any changes made at the inlet V are felt across the domain instantaneously?


I'm not still sold on why this happens in incompressible flows, can you please elaborate on this?

(or point in the right direction of literature)


Either of them works.


Thanks!

[ghorrocks]

changes made at the inlet V are felt across the domain instantaneously?

Yes.

Because the speed of sound is infinite in incompressible flow, so any change to the flow instantly changes the pressure throughout the entire flow field. In a compressible fluid a pressure wave propagates out at a finite speed of sound.

The speed of sound is infinite as acoustic velocity = sqrt(d(Pressure)/d(Density)), and d(Density) = 0 as it is incompressible so acoustic velocity is infinite.