Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
errors when solving a Laval nozzle with hmnf module
Posted Jun 19, 2012, 3:37 a.m. EDT Fluid & Heat, Studies & Solvers Version 4.2a, Version 4.3 1 Reply
Please login with a confirmed email address before reporting spam
I want to simulate the velocity field and the temperature of an ideal gas in a laval nozzle. I use the High Mach Number Flow module. The problem is, that I always get errors like the one in the attached file. Do you know what I can do that these kind of errors don't appear? Do I have to change some parameters of the solver?
Thanks
"Failed to find a solution for the initial parameter.
Maximum number of segregated iterations reached.
Returned solution is not converged."
Attachments:
Hello Tilmann John
Your Discussion has gone 30 days without a reply. If you still need help with COMSOL and have an on-subscription license, please visit our Support Center for help.
If you do not hold an on-subscription license, you may find an answer in another Discussion or in the Knowledge Base.
Please login with a confirmed email address before reporting spam
you should define your inlet and outlet-boundaries. The selection is empty. I see, that you have chosen the hmnf-diffusor example. I did that too at first place. But the example is very special. You should chose a Material for the liquid. Look out that you take the right properties of the material, because there are two types (basic and ideal gas).
I had some help from the support to get my Laval-nozzle to work:
"
1. CFL-number:
You find this by clicking your "High Mach Number Flow 1" physics node and
then go to the "Advanced Settings". I have changed the default settings for
the second "ramp up" to "if(niterCMP>=260,40*1.3^min(niterCMP-260,9),0)"
and divided the entire expression for CFL by 10.
The CFL-number controls the "pseudo-timestepping", that is a iterative
development of the final solution using the iteration steps as
"pseudo-timesteps". The CFL-number is the step-width. And it is a local
variable, as it is dependent on the local mesh size. So at parts of the
mesh with small elements, "pseudo-time" will run slower - which is a good
thing, because at those parts you are probably expecting high gradients.
Nevertheless you should ramp this CFL-number up to higher values at some
points during your solution process to get a final solution in decent time.
The default settings ramp the CFL-number by using if-operators. But you
could enter any function using the variable "niterCMP", which is the
iteration number.
2. Finer Mesh:
I used a swept mesh of blocks for the rectangular part in the middle of
the geometry, which is a efficient way of meshing for those "structured"
regimes.
3. Isotropic Diffusion:
You find this in the "High Mach Number Flow 1" physics node in
"Inconsistent Stabilization". This is really inconsistent - you typically
get a mass loss due to this additional diffusion term. But your solution
will be very smooth and therefore convergence better. After getting a
solution with isotropic diffusion you can use that solution as initial
values for another study. In that study you should lower the isotropic
diffusion value to zero, then it will be disabled.
In general: To get your model to convergence you should lower the
CFL-number, ramp it up at later iteration point and refine the mesh at
parts of high gradients. During solving you should watch the "plot while
solving plot" for the velocity, which you activate at "Study>Solver
Configurations>Solver>Stationary Solver>Segregated Solver" -> "Results
while Solving".
"