[kuplulumert]

Dear all,

I would like to ask a question about defining a specific boundary condition for my analysis. I am trying to simulate a pipeline which transfers polymer inside it. The boundary conditions are; inlet pressure 3500 psi, inlet mass flow 0.8 kg/s, outlet pressure 1000 psi (it is not open to the atmosphere, there is another pump at the exit.). Whatever I do about defining boundary conditions, I could not obtain these boundary conditions at the same time. If I give pressure inlet, I cannot obtain velocity/mass flow rate or if I give mass/flow rate, I cannot obtain the pressure value.

I attached the mass flow inlet and obtained pressure images. Even though i put the exact pressure inside the flow rate, the inlet pressure does not take that value.

I would be glad if someone has a solution about it.

[MKuhn]

Originally Posted by kuplulumert

I could not obtain these boundary conditions at the same time.

It is not possible at the same time. Overdetermined! The pressure drop (pressure inlet minus pressure outlet) depends on the mass flow through your pipe. If you have a mass flow inlet (or velocity inlet), fluent will calculate the corrosponding pressure drop. If you have an pressure inlet, fluent will calculate the corrosponding mass flow.

[kuplulumert]

Dear Mkuhn thank you for the answer. We have a DCS system that can track the values for this pipeline. I read 3500 psi inlet pressure and 0.8 kg/s mass flow inside it. The problem is when I put the pressure value it give more mass flow rate than we see from DCS and when I give mass flow the pressure value is different from the reality. That is why I tried to define both of them. Since it is a overdefinition as you said for the ANSYS, I think I will go with the pressure value for the analysis.

[MKuhn]

Than the pressure drop is higher in your test set up as in your CFD-Model. Check if you have somewhere additional pressure drop in your test set up, mainly at the in- and outlets? Are the fluid properties correct? Roughness of the pipeline? It is turbulent or laminar? Check the turbulence! 2000 psi is quite a lot, heating effects?

[kuplulumert]

Actually the test setup has much more fewer pressure drop than the ANSYS simulation. The test setup has 100 psi pressure drop in 4 meters where the ANSYS simulation has almost 3000 psi pressure drop. The flow is fully developed laminar flow and the fluid is shear thinning non-newtonian fluid. There is also a heat flux through the walls. It looks like all the parameters are well put into the ANSYS however the simulation does not obtain the real results. I am new to ANSYS so I guess there might be a mistake in the boundary conditions.

[LoGaL]

Not really in the boundary conditions, it is that your pressure drop is function of the mass flowing in your system--> you can't impose both pressure drop and mass flow rate. Either you impose the two pressures or a pressure and a mass flow. You can see it clearly in a pipe. If you impose delta P, the mass flow in your pipe is driven by this delta P, the greater delta P, the higher the mass flow.

If by imposing mass flow rate you don't get the pressure drop of the experiment, that means you are not reproducing the conditions of your experiment. Could be due to wrong models, could be due to bad mesh, could be due to stuff happening in the experiment and not in your cfd, or viceversa. The point is that if you impose mass flow and 1 pressure and your CFD setup is "correct", you should get the other pressure out of the solution. Same goes if you impose the two pressures, you should get the correct mass flow rate. No way to impose them both, this is unphysical.

[kuplulumert]

Dear LoGaL thank you for the answer. Actually I first imposed the pressure boundary condition and observed the mass flow output that the Ansys gave. After, I put the velocity inlet b.c. and observed the pressure differentiation that the simulation created. Neither of the conditions gave the real values that we obtain from the test setup. It looks like the geometry that is defined does not reflect the test setup as it has a lot of bends, mixers and valve orifices inside it. Still trying to obtain the test setup values.

[pakk]

The viscosity in your simulation might be different than the real viscosity.

[kuplulumert]

Originally Posted by pakk

The viscosity in your simulation might be different than the real viscosity.

This also another case. I know the entrance viscosity value and the final viscosity value however implementing the same viscosity build up in the ANSYS is a bottleneck of this simulation. It looks like I defined the material correct (non-newtonian power law fluid) but the result does not approve this.