Menu

Discussion Forum

New Discussion
Filtra

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Problem Velocity Profile 2D

Posted 18 mar 2020, 12:57 GMT-4 Fluid & Heat, Computational Fluid Dynamics (CFD), Results & Visualization Version 5.5 6 Replies

Jhoan Guerrero
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hello

I'm doing a 2D model of a transient incompressible air flow. I have 3 rectangular regions. Actually the central rectangle is intended to represent a porous medium, but I will activate this option when I'll be sure I'm obtaining the right profiles.

So, I put normal inflow velocity as inlet boundary condition and atmospheric pressure asoutlet boundary condition. Other boundaries are defined as walls. Initial pressure was defined as atmospheric pressure and initial velocity as zero.

I've tried several meshes (two of them are shown below). Element quality of all elements in higher than 0.4. I've done studies using differents final times(from 0 to 100 s) , either using user controlled tolerance (0.001; 0.005, etc) and physics controlled tolerance.

My magnitude velocity profile at any constant y for my final time (20 s for example) should be parabolic, because, I mean, it is just a gas flow through a pipe, so normally there is no mistery on it, but I found always the same profile as shown in the picture. This profile shows a higher velocity in some regions near to the wall than in the middle of the rectangle which is obviouslly ilogical.

I'm stuck since several days ago on this problem.

Do you guys have an idea why my profile is always like this??? Any suggestion is welcome.

Thanks in advance, Jhoan.

6 Replies Last Post 26 mar 2020, 19:37 GMT-4
Jhoan Guerrero
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago 18 mar 2020, 13:28 GMT-4

Actually I've tried also using a fully developed flow as inlet boundary condition in which I had to define an average velocity. This way I obtained a symmetric parabolic profile :) after a time of 15 seconds, but then when I actived porous medium option, I obtained a profile even worse than the one shown in the last message (you can see the new profile below). BTW when I use high final time ( 100 seconds or bigger), simulation doesn't converge (Maximum number of Newton interations reached).

Any suggestion of how to solve those problem is sincerely apreciated.

Thanks, Jhoan

Actually I've tried also using a *fully developed flow* as **inlet boundary condition** in which I had to define an *average velocity.* This way I obtained a symmetric parabolic profile :) after a time of 15 seconds, but then when I actived porous medium option, I obtained a profile even worse than the one shown in the last message (you can see the new profile below). BTW when I use high final time ( 100 seconds or bigger), simulation doesn't converge (Maximum number of Newton interations reached). Any suggestion of how to solve those problem is sincerely apreciated. Thanks, Jhoan
Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 5 years ago 18 mar 2020, 14:32 GMT-4

What is the Reynolds number for your flow? I am thinking it may not be laminar.

Jeff

-------------------
Jeff Hiller
What is the Reynolds number for your flow? I am thinking it may not be laminar. Jeff
Jhoan Guerrero
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago 19 mar 2020, 06:20 GMT-4

Well, in fact this is intended to be a 2D representation on a more complex 3D problem. From experimental data and for ambient conditions, I have Re= 9160 at inlet and Re=53600 for outlet (since I have cross section change). Then yeah, I was using laminar flow interface but this is actually a turbulent flow, so I'll have to choose a turbulence model. Do you have any suggestion of turbulence model for this Reynolds number interval?

Anyway, even when using turbulent flow module, velocity profile shouldn't be as shown on previous pictures, do you agree?

BTW thanks so much for your answer,

Jhoan G.

Well, in fact this is intended to be a 2D representation on a more complex 3D problem. From experimental data and for ambient conditions, I have Re= 9160 at inlet and Re=53600 for outlet (since I have cross section change). Then yeah, I was using **laminar flow interface** but this is actually a *turbulent flow*, so I'll have to choose a turbulence model. Do you have any suggestion of turbulence model for this Reynolds number interval? Anyway, even when using turbulent flow module, velocity profile shouldn't be as shown on previous pictures, do you agree? BTW thanks so much for your answer, Jhoan G.
Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 5 years ago 19 mar 2020, 08:56 GMT-4
Updated: 5 years ago 19 mar 2020, 08:57 GMT-4

Hello Jhoan,

This blog explains hot to choose the right turbulence model. In that blog post you'll also find a sketch that shows what you can expect the velocity profile to look like, namely rising from zero quickly then plateauing, as you move away from the wall.

Best regards,

Jeff

-------------------
Jeff Hiller
Hello Jhoan, [This blog](https://www.comsol.com/blogs/which-turbulence-model-should-choose-cfd-application/) explains hot to choose the right turbulence model. In that blog post you'll also find a sketch that shows what you can expect the velocity profile to look like, namely rising from zero quickly then plateauing, as you move away from the wall. Best regards, Jeff
Jhoan Guerrero
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago 22 mar 2020, 14:58 GMT-4

Thank you so much for your help.

Thank you so much for your help.
Jhoan Guerrero
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago 26 mar 2020, 19:37 GMT-4

Hello, it's me again.

So, in order to understand my problem better, I decided to plot velocity profiles, for the same geometry and mesh, so I could obtain whether laminar or turbulent flow just by changing inlet velocity (and COMSOL physics). I understand that a minimum tube length is needed to obtain a fully developed flow.

On laminar flow simulation for a long length I obtain a good profile (as shown in the picture 1), but there is a problem at wall section near to the inlet. 2D visualization shows a region where velocity seems to be bigger than in the centre of the figure for no reason (FIGURE 2).

I've read this could happen because I didn't have an extended domain just for inlet boundary condition (as suggested in this blog ),so, then, I stablished symmetry conditions in a part of the wall near to inlet boundary to get over this problem. Even by doing this, I keep having samekind of 2D visualization (FIGURE 3). Maybe anyone has an idea on why is it happening?

On turbulent flow simulation, Reynolds number is 6300, I'm using k-epsilon turbulence model since it's most well-known model. Main issue is velocity magnitude at wall is not zero (about 0.17 m/s as shown ih the FIGURE 4), even when changing Re, mesh or whatsoever :(.

I've also tried with other turbulence models as SST and Low Reynolds k-epsilon but I keep having a non-zero velocity at wall ( about 0.003 m/s). Since k-epsilon is the most studied and well-known model, normally I should have the best results, isn't it? Any insights about this?

Thanks in advance and stay home as much as possible guys, Jhoan

Hello, it's me again. So, in order to understand my problem better, I decided to plot velocity profiles, for the same geometry and mesh, so I could obtain whether laminar or turbulent flow just by changing inlet velocity (and COMSOL physics). I understand that a minimum tube length is needed to obtain a fully developed flow. On **laminar** flow simulation for a long length I obtain a good profile (as shown in the picture 1), but there is a problem at **wall section near to the inlet**. 2D visualization shows a region where velocity seems to be *bigger than in the centre* of the figure for no reason (FIGURE 2). I've read this could happen because I didn't have an extended domain just for inlet boundary condition (as suggested in this [blog](https://www.comsol.com/blogs/how-to-assign-fluid-pressure-in-cfd-simulations/) ),so, then, I stablished symmetry conditions in a part of the wall near to inlet boundary to get over this problem. Even by doing this, I keep having samekind of 2D visualization (FIGURE 3). Maybe anyone has an idea on why is it happening? On **turbulent** flow simulation, Reynolds number is 6300, I'm using **k-epsilon** turbulence model since it's most well-known model. Main issue is velocity magnitude at wall is not zero (*about 0.17 m/s as shown ih the FIGURE 4*), even when changing Re, mesh or whatsoever :(. I've also tried with other turbulence models as **SST** and **Low Reynolds k-epsilon** but I keep having a non-zero velocity at wall ( *about 0.003 m/s*). Since k-epsilon is the most studied and well-known model, normally I should have the best results, isn't it? Any insights about this? Thanks in advance and stay home as much as possible guys, Jhoan

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content