[parth_k]

Hi all,
I'm working on a simple 3D Transient study involving sinusoidal laminar pulsating fluid flow (UDF-velocity inlet) in a rectangular geometry channel (closed system). I have the outlet modeled as an opening B.C and tried as a Pressure outlet too.
Two small questions to ask here due to some strange errors in CFX solver:

1) When I use an Pressure outlet B.C (0 Pa as static Pressure), I get warning message between every iterative loops saying wall detected consider switching to opening B.C.. any recommendation on this? (While the inlet is way far from the outlet in the channel 500mm apart)

2) When I switch to opening B.C at the outlet (0 Pa as Opening Pressure) I see from monitors that either my mass flow at the inlet is all negative sinusoidal with positive sinusoidal at the outlet, while the transverse velocity of flow at sample location is as desired... or all of it VICE-VERSA - when I use (0 Pa - Entrainment) on the opening B.C.

Any solutions could you suggest as to how it can be avoided. Thanks for replying

[Gert-Jan]

https://www.cfd-online.com/Wiki/Ansy...f_an_OUTLET.22

[parth_k]

Thanks Gert-Jan for your reference, I have followed the steps as listed. The backflow warning in my case stays persistent and causes the CFX solution to abruptly terminate in a few itr.
However using an opening B.C (as my 2nd point states from post - 0 Pa opening Pressure)I face the issue of negative mass-flow at the inlet and positive at the outlet section. Could you or anyone suggest how this can be corrected?

[ghorrocks]

If your fluid is incompressible, and domain is closed except for the inlet you are driving and an outlet, then when the outlet is entirely backflow the solver will crash. You can't have incompressible flow entering a domain but nowhere for it to go. The only response is to crash.

For the opening case - have a look at the flow directions on these boundaries. You need to work out which direction (in or out of the domain) is defined as positive flow in both cases.

[parth_k]

Thanks Glenn for your valuable suggestions. I see the model respond well with an opening condition. I have a small query though, I'm able to set the inlet mass flow as positive working with the flow directions and the outlet mass flow as expected to be opposite, but the velocity profile at a sample location appears to start from negative instead. (Referring from my monitors during runs)

Any recommendations on this behavior?

[ghorrocks]

Please show an image describing what you mean, and show what the points you are graphing are on the geometry.

[parth_k]

Hi Glen, Sure herewith attached my model snip.. The yellow crosslines are my sampling locations for velocity being in the mid section of my fluid zone. Inlet to the left face, Outlet to the Right face.

I have also attached sample transient inlet mass flow, outlet mass flow profiles from the two faces, with velocity profile at the sampling point for reference simulated at a sample small timestep. So the inflow and outflow profiles are desirable but the velocity profile is strange as it begins from negative.

Also the sinusoidal velocity profile is defined at 0.1 Hz frequency at a steady vel of 0.1 m/s and 0.2 m/s as pulsation amplitude.

Kindly recommend me how the nature of velocity profile can be rectified to positive scale using an opening B.C?

[ghorrocks]

Repeating post #4: "For the opening case - have a look at the flow directions on these boundaries. You need to work out which direction (in or out of the domain) is defined as positive flow in both cases."