Making fluid slightly compressible
 

Hi Everyone,

I am trying to simulate an incompressible fluid flow in a pipe. The problematic part of this simulation is that the outlet gets closed at some point during this transient simulation, and we have one inlet and no outlet. During this time, the inlet flow is supposed to inflate the walls by increasing the pressure inside the pipe. I tried to simulate this using an incompressible fluid but got convergence issues. After reading related posts here, I realized that making the fluid slightly compressible improves convergence. I did this using the following expression for the fluid density:

density=(Absolute pressure)/(R*Temperature)

Since the temperature is equal to 310K and absolute pressure should be near atmospheric pressure, to get the fluid density of 1050 kg/m3, the value of R needs to be equal to 0.3113.

Would you please let me know if this way of making the fluid slightly compressible causes any stability problems? I am asking this because my simulation can only be converged if the density and pressure URFs is less than 0.3.

Many thanks in advance!

Before you make weird changes to the fluid model - are you sure your model is well-posed? If your model is not well posed then you are never going to get useful results from it, even if it does pretend to converge. To answer this question, can you post your CCL or output file so we know what you are doing?

If you make the fluid mildly compressible be aware that you will then get compressible fluid pressure waves bouncing around everywhere, and your time step will be very small so you can resolve these waves. The speed of these pressure waves will be the acoustic velocity and you will need a CFL number of around 1 to resolve them - so this will tell you the time step size you will need.

Also - why make up a compressibility function when a real one already exists? I would recommend you use the bulk modulus for water for your compressibility function, as this is physically real. It is the cause of things like water hammer so it is very well known (https://en.wikipedia.org/wiki/Water_hammer).

Dear Glenn,

Thank you very much for your attention.

I am trying to perform this simulation using fluent. I asked this question here because I thought it is a general question that could be helpful for CFX users as well. In fluent there are just two options to model compressible fluids, the ideal gas model and the compressible-liquid option.

Since this is a Multiphysics simulation, I don't know where I should search for Fluent's output file. This was so straightforward in CFX. I'll keep you posted.

Thanks again for your help.

If you are using Fluent then we will not be able to help you with the details of the simulation. Try the Fluent forum for that.

But the second and third paragraphs of my previous post are applicable to any CFD code, so those comments should be relevant.