Not getting Convergence using explicit solver in fluent.

Patelp1996 · December 11, 2020, 11:34

I have been tried to simulate unsteady shock wave in the shock tube using 3D model of shock tube. For that purpose I am solving the Euler's equations (Inviscid Flow) in 3D dimensions with energy equations. The Driver gas (air) has 6.48107 Bar Pressure and Driven has atmospheric Pressure. So, I am trying to solve those equation using explicit solver with explicit transient formulation but as you can see in this picture I am not able to converge the solution. I tried changing the courant number (from 0.1 to 5) and various time step sizes (from 10^-6 to 10^-9) but getting the same graph as shown above (Straight line and then diverges). Moreover I want to know how to calculate the CFL condition exactly. Anyone Who can Help me?? One more thing, max skewness for my mesh is 0.95455 and average (skewness) is 1.7671e-002. Min orthogonal quality is 0.12922 and average (orthogonal quality) is 0.99292.
The first picture is about geometry, second one shows the elements with max skewness and third one shows the solution graph.

pakk · December 11, 2020, 23:53

Picture is missing.

Patelp1996 · December 12, 2020, 03:00

I added the Pictures. It was my mistake. I am sorry.

duri · December 12, 2020, 09:04

I could see some warning on sliding interface. Interface is not a requirement for this kind of problem. I would recommend to solve using axi symmetry mesh if it could solve the purpose, to save computational resource. Just make one domain and patch the initial condition either side of the diaphragm as like the real shock tube initial condition. Use CFL number <=1 for explicit and adjust the time step to get proper convergence, with dual time stepping the time step is user choice can be calculated based on shock speed.

LoGaL · December 12, 2020, 10:31

Ye, you must have done something stupid (it happens to me as well, it happens to everybody, i am not saying you are a bad engineer) in the setting. There shouldn't be any sliding interface in the domain. And yes, as it was suggested, it's axisymmetric, so use 2D solver.

AlexanderZ · December 14, 2020, 03:41

as it was suggested, in design modeler put all your bodies into 1 part, so you will remove interfaces.

explain which boundary conditions do you use, do you patch any part of your domain with high pressure?

Patelp1996 · December 15, 2020, 14:45

Yeah, I patched driver with higher pressure about 6 bar. Also patched the driven and ambient with 1 bar pressure.

December 14, 2020, 03:41		#6
AlexanderZ Senior Member Alexander Join Date: Apr 2013 Posts: 2,363 Rep Power: 34	as it was suggested, in design modeler put all your bodies into 1 part, so you will remove interfaces. explain which boundary conditions do you use, do you patch any part of your domain with high pressure? Patelp1996 likes this. __________________ best regards ****************************** press LIKE if this message was helpful

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Convergence	Centurion2011	FLUENT	48	June 15, 2022 00:29
Adjoint Solver Convergence	devesh.baghel	STAR-CCM+	7	October 9, 2018 08:05
CFX overwhelming Fluent in mass convergence of boundary layer separation case	Pierre1	FLUENT	7	March 26, 2015 22:43
Naca 0012 (compressible and inviscid) flow convergence problem	bipulsaha	FLUENT	1	July 6, 2011 08:51
fluent convergence problem	josip76	FLUENT	0	May 26, 2011 21:08

December 11, 2020, 23:53		#2
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	Picture is missing.

December 12, 2020, 03:00		#3
Patelp1996 New Member Punsa Join Date: Sep 2020 Posts: 14 Rep Power: 6	I added the Pictures. It was my mistake. I am sorry.

December 12, 2020, 09:04		#4
duri Senior Member duri Join Date: May 2010 Posts: 245 Rep Power: 17	I could see some warning on sliding interface. Interface is not a requirement for this kind of problem. I would recommend to solve using axi symmetry mesh if it could solve the purpose, to save computational resource. Just make one domain and patch the initial condition either side of the diaphragm as like the real shock tube initial condition. Use CFL number <=1 for explicit and adjust the time step to get proper convergence, with dual time stepping the time step is user choice can be calculated based on shock speed.

December 12, 2020, 10:31		#5
LoGaL Senior Member Lorenzo Galieti Join Date: Mar 2018 Posts: 375 Rep Power: 12	Ye, you must have done something stupid (it happens to me as well, it happens to everybody, i am not saying you are a bad engineer) in the setting. There shouldn't be any sliding interface in the domain. And yes, as it was suggested, it's axisymmetric, so use 2D solver.

December 15, 2020, 14:45		#7
Patelp1996 New Member Punsa Join Date: Sep 2020 Posts: 14 Rep Power: 6	Yeah, I patched driver with higher pressure about 6 bar. Also patched the driven and ambient with 1 bar pressure.