Comsol convergence on Venturi valve - Covid help

tabsquare · April 11, 2020, 08:59

Hi all, I happened to restart some simulation on COMSOL. We are working to supply some hospitals here in Italy with emergency Venturi valves used ventilation of covid patients. No profit then. Here you can see a few of those valves https://encrypted-tbn0.gstatic.com/i...np57I&usqp=CAU

You have two inlets, one for oxygen and the other for air. High pressure oxygen is injected and it sucks air from the surroundings. The mix is then injected in the patient's facial mask. Oxygen inlet flow scale should be around 10 L/min. That converts to an inlet velocity of around 5 m/s. I have used an outlet boundary condition of p=0 and an atmospheric box with p=0 boundary. Mesh is decently fine on the valve features whilst I had to leave it quite coarse on the surroundings.

I am having problems on making it converge on a stationary solution. I managed to make it with k-eps model but I had to use fine parametric sweep on inlet velocity and that is too much computation time. I noticed there are a few parameters in the physics advanced options - velocity scale, length scale factor. The first was at 1 m/s, I now changed it to 10m/s and running the computation so I am waiting to see if it converges. I left the length scale factor at 0.035.

I am merely interested in calculating stationary mass fluxes. What approach would you suggest to have fairly acceptable results in small times? I am using a laptop so small computation time is mandatory.

Thank you for help

tabsquare · April 11, 2020, 11:26

I struggled for days and now that I posted a question the simulation converged

, and in an acceptable time. The velocity scale was apparently the main problem. It took around 2 hours and a half.
But I am open to ways to ameliorate the output accuracy and computation time as I am not too expert.

Thank you

habibjalili1995 · April 15, 2020, 19:44

Hi,

I have read your comment carefully and I have followed the news about the COVID-19 on all over the world so I know you are being in a tough situation in Italy but we are too in Iran. If you still have a problem with this simulation I can help you. We can assume the oxygen and air as a single-phase and lets first simulate it as a laminar flow. Are you worried about the mixing of oxygen and air? if Yes, we have to add mass equations.

My Whats app number is +989370712643
Have a nice day!

April 11, 2020, 08:59	Comsol convergence on Venturi valve - Covid help	#1
tabsquare New Member Riccardo Join Date: May 2014 Posts: 12 Rep Power: 11	Hi all, I happened to restart some simulation on COMSOL. We are working to supply some hospitals here in Italy with emergency Venturi valves used ventilation of covid patients. No profit then. Here you can see a few of those valves https://encrypted-tbn0.gstatic.com/i...np57I&usqp=CAU You have two inlets, one for oxygen and the other for air. High pressure oxygen is injected and it sucks air from the surroundings. The mix is then injected in the patient's facial mask. Oxygen inlet flow scale should be around 10 L/min. That converts to an inlet velocity of around 5 m/s. I have used an outlet boundary condition of p=0 and an atmospheric box with p=0 boundary. Mesh is decently fine on the valve features whilst I had to leave it quite coarse on the surroundings. I am having problems on making it converge on a stationary solution. I managed to make it with k-eps model but I had to use fine parametric sweep on inlet velocity and that is too much computation time. I noticed there are a few parameters in the physics advanced options - velocity scale, length scale factor. The first was at 1 m/s, I now changed it to 10m/s and running the computation so I am waiting to see if it converges. I left the length scale factor at 0.035. I am merely interested in calculating stationary mass fluxes. What approach would you suggest to have fairly acceptable results in small times? I am using a laptop so small computation time is mandatory. Thank you for help

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
IC engine valve dynamic mesh modelling	soogeng	FLUENT	0	April 14, 2018 10:27
Problems with Simulation of a Valve with Cavitation	WilbertS	CFX	4	April 4, 2018 09:22
Best solver for hydrulic valve cfd with Comsol?	Mr Matti Karvonen	COMSOL	0	July 3, 2008 05:56
convergence problem in comsol	anderson	COMSOL	0	February 24, 2008 16:54
Defect correction and convergence	ganesh	Main CFD Forum	4	June 30, 2006 14:20

April 11, 2020, 11:26		#2
tabsquare New Member Riccardo Join Date: May 2014 Posts: 12 Rep Power: 11	I struggled for days and now that I posted a question the simulation converged , and in an acceptable time. The velocity scale was apparently the main problem. It took around 2 hours and a half. But I am open to ways to ameliorate the output accuracy and computation time as I am not too expert. Thank you

April 15, 2020, 19:44		#3
habibjalili1995 New Member habibjalili Join Date: Aug 2018 Posts: 15 Rep Power: 7	Hi, I have read your comment carefully and I have followed the news about the COVID-19 on all over the world so I know you are being in a tough situation in Italy but we are too in Iran. If you still have a problem with this simulation I can help you. We can assume the oxygen and air as a single-phase and lets first simulate it as a laminar flow. Are you worried about the mixing of oxygen and air? if Yes, we have to add mass equations. My Whats app number is +989370712643 Have a nice day!