Reversed Flow Problem

HusnaMotin · April 23, 2022, 13:21

Hi everyone!

This is my first post here and I'm somewhat new to Star.

I have taken on the work of a past post-grad student and I am simulating a problem within a box that is meant to protect doctors from harmful aerosols during intubations. I have access to the post-doc students sims.

I have moved the suction pipe- ran as a pressure outlet- around the z axis from the side of the box (see streamlines scene) to the front of the box (see geometry scene). However my simulation output window says there is reversed flow on my domain inlet (see highlighted pink region). The domain inlet is being treated as a stagnation inlet and is identical in both sims.

I have checked my mesh, boundary conditions, physics etc. but I'm confused as to why I have reversed flow despite my scenes indicating good results and my residuals are ok (see attached).

Thank you for taking the time to read this and any help is much appreciated

fluid23 · April 29, 2022, 10:22

Reversed flow at a boundary may or may not be a problem, it really depends on what you are modeling. Unfortunately, the description you offer is very confusing so I am not totally sure what you are trying to accomplish.

I would guess that your domain inlet is the wrong shape or size and/or at the wrong location which is leading to reversed flow at the inlet.

If your model is converging and your results are 'good' then I really wouldn't worry too much about it. However, I don't think your residuals indicate a suitable mesh and/or physics setup.

General guidance is 5 orders of magnitude (<0.00001) for turbulence parameters and 3 orders for everything else (<0.001). However, residuals aren't the arbiter of accuracy that some would have you believe, either.

I suspect that if you provide a better domain and boundary conditions that your residuals will improve.

HusnaMotin · April 29, 2022, 15:49

Hi there!

Thank you for replying back. I understand that the description may have not been brilliant; is there any thing you'd like me to clarify?

Essentially my domain inlet allows for air to enter the box; this allows the person inside the box to breathe whilst suction occurs (this removes any harmful aerosols produced during intubations). I have spoken with my supervisor who has inspected the simulations and he believes mesh quality, physics selections etc. are good.

However I will try to increase the domain inlet size

Nikpap · May 1, 2022, 07:58

The convergence of your turbulent quantities does not seem great.

You can try different under relaxation factors and see if there is any improvement.

You could also make a vector seen close to your inlet and observe your mesh there to ensure if it is or not the source of your problems.

This is an isothermal problem, what have you selected in your physics tree?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Intuition for why flow follows convex surfaces	lopp	Main CFD Forum	47	February 1, 2022 13:14
SU2-7.0.1 on ubuntu 18.04	hyunko	SU2 Installation	7	March 16, 2020 04:37
Reversed flow problem	VasCFD	ANSYS	6	December 15, 2015 11:04
transient, impregnating flow problem	fgommer	FLUENT	0	February 29, 2012 16:10
Reversed flow problem & null point error	narendra	FLUENT	0	May 20, 2010 12:43

April 29, 2022, 10:22		#2
fluid23 Senior Member Matt Join Date: Aug 2014 Posts: 947 Rep Power: 17	Reversed flow at a boundary may or may not be a problem, it really depends on what you are modeling. Unfortunately, the description you offer is very confusing so I am not totally sure what you are trying to accomplish. I would guess that your domain inlet is the wrong shape or size and/or at the wrong location which is leading to reversed flow at the inlet. If your model is converging and your results are 'good' then I really wouldn't worry too much about it. However, I don't think your residuals indicate a suitable mesh and/or physics setup. General guidance is 5 orders of magnitude (<0.00001) for turbulence parameters and 3 orders for everything else (<0.001). However, residuals aren't the arbiter of accuracy that some would have you believe, either. I suspect that if you provide a better domain and boundary conditions that your residuals will improve.

April 29, 2022, 15:49		#3
HusnaMotin New Member Husna Join Date: Apr 2022 Posts: 2 Rep Power: 0	Hi there! Thank you for replying back. I understand that the description may have not been brilliant; is there any thing you'd like me to clarify? Essentially my domain inlet allows for air to enter the box; this allows the person inside the box to breathe whilst suction occurs (this removes any harmful aerosols produced during intubations). I have spoken with my supervisor who has inspected the simulations and he believes mesh quality, physics selections etc. are good. However I will try to increase the domain inlet size

May 1, 2022, 07:58		#4
Nikpap New Member Nikolaos Papafilippou Join Date: Apr 2019 Posts: 25 Rep Power: 7	The convergence of your turbulent quantities does not seem great. You can try different under relaxation factors and see if there is any improvement. You could also make a vector seen close to your inlet and observe your mesh there to ensure if it is or not the source of your problems. This is an isothermal problem, what have you selected in your physics tree?