Hi,
I am simulating flow around a model. This model have an inlet and an outlet. Flow is incompressible. I know that value of pressure at the outlet is very much different from ambient pressure(Operating pressure). But, I don't know real value of pressure at the outlet. Now, I applied value of zero gauge pressure at the outlet (Is this OK? Why?) and this means that pressure at the outlet is equal to operating pressure while this is not. How do i find correct boundary condition for outlet?
Thank.
Ali,
try to find Fluent (c) user manual and read about boundary conditions inside it. To my mind you need somewhat like so-called "pressure far field" boundary condition. It is also called "soft" or "non-reflecting" boundary.
Dear Ali
When you do not know the real values you can use the outflow as a boundary condition for incompressible Flow .
Thanks. But some people say that when flow is incopressible and boundary condition at inlet is velocity, we can use pressure-outlet as a boundary condition at the outlet because flow is not dependent on outlet pressure (so, we can use zero gauge pressure at the outlet). Why?
Yes, you can test pressure-outlet and see that you programe converge .
The isothermal flow of an incompressible fluid is not at all influenced by the pressure, but by pressure gradients. Once the continuity equation becomes simply divergence of V=0, the pressure is not involved anymore. Therefore when you set the pressure to 0 or to 1000 has no effect, but just an arbitrary shift on the absolute pressure value. You can consider this as a reference pressure. However, setting this at the outlet of a simulation in which the domain is not sufficiently far from wakes and other flow features can be dangerous. The best is to set pressure gradient in the outlet.
I don't know what solver you are using, but when you work with an incompressible and isothermal flow you have to solve for two equations: continuity and momentum.
For an incompressible flow, the continuity reduces simply to div(V)=0, and becomes more a constraint than a real equation to solve. In this condition, the density becomes just a multiplicative constant (this is why softwares such as OpenFoam typically do not even specify the density as a variable but rather use a 'reduced pressure' of the sort p/rho). Since the equation of state linking density and pressure is no longer needed, the pressure actually never appears in your set of equations. Only its gradient is involved, in the momentum equation.
Therefore, if p(x,y,t) is the pressure field that solves your problem, then any p(x,y,t)+c, with c an arbitrary constant, is a solution to your problem. Therefore, your software does not generally know what is your 'reference' pressure and usually you have to specify it in at least one point, else it will take the defaul it is designed to get.
So, if in the output of your simulation you set the pressure to be 0 or to be 100, the result will be exacly the same, up to a global shift in the pressure. This of course, if you do not fall into another possible problem: your boundary is for example in a region where there should be a non-uniform pressure gradient but you impose it to be uniform. As a very first test, when one has to go completely blinded, it is prefearable to set the pressure gradient to be zero ( Neumann condition).
- Badges
- Science topic
- Similar topics
- Thermodynamics
- Pressure