I am working on a flame affected by external magnetic field using MHD module with fluent 14.5 . I am using compiled UDF to produce non-uniform profile for boundary condition. My settings are:
Model settings: 2D, k-w SST ,Non-premixed combustion (with 8 species), Magnetic Induction model with under relaxation of 0.6
Boundary conditions: axis,velocity inlet (for fuel), the non uniform magnetic field applied in x-direction for pressure inlet boundary
Solver: Pressure Coupled
But I got the following error:
# Divergence detected in AMG solver: pressure coupled -> Decreasing coarsening group size!
# Divergence detected in AMG solver: pressure coupled -> Increasing relaxation sweeps!
# Divergence detected in AMG solver: pressure coupled -> Changing to W-cycle!
After the detected error, I used Pseudo Trandient method, but nothing was changed.
I decreased “coarsen by” and increased the pre-sweep for scalar and coupled parameters in advanced solution controls, but the error persist. I also reduced the URF of the pressure,energy,density, fmean, and B_y to 0.1 step by step and I couldn’t fix the problem.
When I used BCGSTAB stabilization method for flow I encountered AMG Divergence for k Equation. As I changed the Stabilization method for k, AMG divergence detected in the next equation. Finally I changed all of the equation to BCGSTAB stabilization method, But immediately all the residual changed to 1.#QNBe+00 and the solution is converged!
I should mention that the combustion with MHD equation was solved without Lorenz Force and Joule Heating, but as I turned on Lorenz force the solution process stopped working (with b_0 Scale factor of 0.1).
When I applied magnetic field for cold flow (combustion model is off), the problem converged without any error. And when I turned off magnetic problem the combustion is converged. In other word, but I cannot combine both of them.
Thank you in advance for your help.
Dear Javad,
Please see the following papers:
www.ansys-blog.com/multiple-steps-in-mechanical/
http://www.cfd-online.com
Good luck
either there is a problem on problem defining or the mesh is not good please cross check both .
I agree with Shirish.
First, you should check your mesh. If the mesh is bad or ratio is too high, the solution will be divergent and you will get AMG error.
Have you used the FMG initializer at all to get a good approximate starting solution for your case? The FLUENT documentation describes how to use it.
You should check the smallest grid size, and try to calculate the time step size using CFL condition( the courant number- by default it is 200. Try to reduce it to 50-100). You don't need to set the time step equal that value but cant not set to large to compare with that value (implicit method) and you must set time step equal or below that value if you use explicit method.
Also reduce URF to 0.1 or 0.2 for all parameters
Hope this will help u.
# Divergence detected in AMG solver
This error crops due to
1. Check scaled geometric model with units
2. Mesh quality desire by fluent
3. Wrong setting the parameter for simulation-related to step size or species
4. Ensure UDF is doing its intended function
First check with no magnetic field applied, is the simulation working?
If working then apply magnetic field and simulated.
Thank you dear Shirish for your response, My mesh is uniform with the maximum aspect ratio of 1.41 .I also used mesh adaption,althogh mesh quality is good enough. But I cannot undrestand what you mean with problem defining,Could you explain more in details which things may affect the result?
I am going to attach the physical domain for you.
Thank you dear Dinesh for your consideration.I have used Non premixed implicit solver and fluent says I cannot use FMG on the current setting.Anyway I changed my solver to SIMPLE explicit solver with dendity based type to use FMG and I turned on verbosity to see the residuals, but 1.#QNAN is appeared for normalized residual. and finally says "Error: Divergence detected in AMG solver: uds-0".
I used even the courent number less than 1,but it did not help me.
And finally,I already reduced the URF to 0.1, but it did not work either.
And thank you Mahesh. I already get converged result without mgnetic field.But as I turn on Magnetic field AMG Divergence pops on.It also converged for cold flow (without combustion) with magnetic field.But when I turn on the combustion model, the soltion is diverged again. In fact, they do not work when I use the combination of magnetic field and combustion. When I used cold flow I checked the UDF and it works properly.
Dear Javad, could you find out why the simulation does not converge when both MHD and combustion models are solved simultaneously?
Can I know if you could find out why the simulation does not converge when both MHD and combustion models are solved simultaneously in fluent?
As I'm working on a similar problem and your answer could be very helpful.
Dear Raju. Unfortunately not. So I had to write small UDF codes for Lorentz force and Joule Heating and I added as source terms to the Momentum and energy equations.
Can I request you to share the UDF codes If you still happen to have them?
Dear Javad,
Please see the following papers:
www.ansys-blog.com/multiple-steps-in-mechanical/
http://www.cfd-online.com
Good luck
Hello Mr. Javad, you have mentioned that "I already get converged result without mgnetic field".
(Q1), Is this converged result for combustion case?
Theoretically, the grid resolution is one of big problems to get the converged result as mentioned by previous professors or researcher or Fluent users. BUT, it is clear from your attached image that your configuration is very simple. Under this situation, Gird issue may not be a big problem!! So,
(Q2). Could you please let me know that is it possible for you to establish a flame on the burner without magnetic field?
(Q3) What is diameter of your burner?
According to your image, it is also clear that your calculation is axisymmetric . If this is the case, then
(Q4) Have you maintained the proper convention of Fluent in defining the symmetric axis?
hello sir,to slove your problem please sheck the boundary condition and your mesh..
Hi Javad Hosseinpour , have you managed to get rid of the error? how? can you please share your experience? Regards. Oula
- Badges
- Science topic