Hi friends, How can we incorporate the Liquid head due to gravity in CFX? Can it be added as volumetric source term (density*g*h) in particular direction (say z direction) or we have to give CEL for pressure? I just want to solve simple flow in tank problem.

Thanks in advance Regards Tuks

Hi,

Are you solving a two-phase (liquid+air) free-surface problem, or just a one-phase problem?

If it is a two-phase free-surface problem, you'll have to define the buoyancy refence density (Rho_ref) and the gravity vector (g). This will add a momentum source SM=(Rho-Rho_ref)*g . Rho_ref is usually set to the air density.

If it is just a one-phase flow, you may define the 3 components of the momentum source as:SMx=Rho*gx ; SMy=Rho*gy ; SMz=Rho*gz However, you will obtain theoretically the same results (if your fluid properties, including density of course, are independent of pressure) if in Post you add the the gravity contribution to pressure, P'=P+Pg, Pg=Rho*g*h . And this is probably what you should do, as in this way CFX solves the pressure field caused only by the flow itself.

Hi Rui, Thank you for your advice. I am solving the single phase problem and tried with SMx=Rho*gx, but the solution is diverging. Anyway i got the rid of problem, it was fairely simple, i tried with buoyancy for single phase (gx only) and it worked. No need to define as source. But still Question remains that why didn't it worked when defined as source term?.

Thank again you for kind advice. Regards Tuks

Hi,

When you defined Buoyancy, did you define a Buoyancy Reference Density or a Buoyancy Reference Temperature?As I think for a single phase, if the fluid density is constant CFX only allows you to employ the Bussinesq buoyancy model which requires a Buoyancy Reference Temperature.

When you tried SMx=Rho*gx, have you defined the g vector direction correctly (downwards)? Have you specified the initial pressure taking into account the gravity effect (P=Rho*g*h)?

Hi Rui,

Thank you for your advice.

Regarding the buoyancy, i had defined the Ref Temp. Bussineqs aprrox doesnot matter for me because the conditions are isothermal.

About the SMx=Rho*gx, i had defined the correct direction. The pressure was also intialised using P=Rho*g*h, with Pref=101325[Pa]. But still it didnot work and i had to go for Buoyancy approach.

Regards

Tuks

Hi,

If you see the CFX-5.7.1 documentation, Solver Theory, Basic Solver Theory, Buoyancy pages 24-25, when using the Bussinesq Model the momentum source term is: SM = -Rho*Beta*(T-T_ref)*g , where Beta is Thermal Expansivity of your fluid (defined in Material Properties). Thus, this won't give the pressure head due to gravity.If you have performed any simulation with this model, you should be able to check this in CFX-Post.

Pref doesn't have any influence on the calculations (if all the fluid properties are independent of pressure), it's just a value that is added to the pressure actually calculated by the Solver. How have you defined the boundary conditions?When does the simulation diverges? After a few iterations?

Regards,

Rui