[ooo]

For a particular purpose, I need to know the density in my flow simulation. But the only parameter can be set, is the kinematic viscosity.
For example consider a simulation with the nu=0.01.
If i don't have the dynamic viscosity, how can i find the density?!

[alexeym]

Hi,

I guess you are doing your incompressible flow simulation. So density is constant everywhere. So if you take random simulation with random value of kinematic viscosity there's no way to get density out of it.

[mturcios777]

As long as you have set your kinematic viscocity correctly (in transportProperties) for the fluid you are simulating, the density is whatever you need it to be. Just remember that the pressure for incompressible cases is the kinematic pressure (p/rho_0). So if you have the kinematic viscocity of say 7.4E-7 m^2/s, then you could either have methanol at 20 C or Isobutane at around -60 C (just a table I have in front of me at the moment). So depending on the fluid you want to be simulating, all the relevant properties should be scaled by the proper density. In the cases I said earlier, densitites are 788.4 and around 650 kg/m^3 respectively, and you would have to multiply your kinematic viscocity, kinematic pressure and any similarly scaled property accordingly.

[ooo]

Originally Posted by mturcios777

As long as you have set your kinematic viscocity correctly (in transportProperties) for the fluid you are simulating, the density is whatever you need it to be. Just remember that the pressure for incompressible cases is the kinematic pressure (p/rho_0). So if you have the kinematic viscocity of say 7.4E-7 m^2/s, then you could either have methanol at 20 C or Isobutane at around -60 C (just a table I have in front of me at the moment). So depending on the fluid you want to be simulating, all the relevant properties should be scaled by the proper density. In the cases I said earlier, densitites are 788.4 and around 650 kg/m^3 respectively, and you would have to multiply your kinematic viscocity, kinematic pressure and any similarly scaled property accordingly.

Thanks for your answers.
I've implemented a method to simulate the flow around a 2d cylinder without having the geometry of the cylinder.But i have forces on the coordinates of the cylinder.
So i would like to use these forces to compute drag coefficient!
The formula is : Cd = Fd / (0.5 * rho * u^2 * D ) . the points is that the dimension of Fd in my algorithm is m/s^2 and its not divided by rho !
I would appreciate if you have some idea about this.

[mturcios777]

How are you calculating Fd?

[ooo]

Originally Posted by mturcios777

How are you calculating Fd?

The velocity is interpolated to a lagrangian grid, then a force in lagrangian grid computes by : F=U/dt (then this force is distributed to the eulerian grid and ....)
But the points is that the dimension of the force is m/s2 ...

[mturcios777]

Well that makes sense, as Force is actually d(mU)/dt:

http://en.wikipedia.org/wiki/Newton'...27s_second_law

The units are consistent with an incompressible fluid, as rho is not changing, and m=rho*V (where V is volume).

Because your rho is already been divided out in the incompressible formulation, Cd = Fd/(1/2 u^2 D)