divergence detected in AMG solver: momentum
 

Hello Friends, i am working on shell and tube heat exchanger using seggregated solver and laminar,and alsoheat exchanger models My simulation today give me an error message:"Divergence detected in AMG solver:x momentum Error object:()

I have no idea what it means and how to get around it.

I like to be helped by members of the forum.

With advanced thanks

sudhakar

Re: divergence detected in AMG solver: momentum
 

I have same problem.

Re: divergence detected in AMG solver: momentum
 

Tighten your under relaxation factors and then run. For instnace check on momemtum under relaxation factors now say its 0.5 tighten it to 0.1. Then this error should go away. However doing this increase your tun time. KP

Re: divergence detected in AMG solver: momentum
 

I tried from urf from .7 to .001 .but i could not get. in my problem no of cells are more than 8 lakh. my problem is shell and tube heat exchanger in 3 dimensional using seggregated solver.previous i solved same problem with out headers .but now i inserted headers now i got above problem divergence detected in AMG solver: momentum

please help me to solve this problem

Re: divergence detected in AMG solver: momentum
 

Hi,

I used coupled solver instead of seggregated one. Then reduced urf. The case is working.

divergance
 

hi everybody...
i have same problem,i changed urf from 0.5 to 0.05 but again error message showed up!
dear wang i did not understand ur method please explain more and help me out.
thank u in advance for any help provided:)

So by coupled I think he means under the solutions methods tab and under Pressure-Velocity coupling Scheme to change it to coupled. And then urf means Under relaxation Factors in the solution controls tab.

However I am having the same divergence in x-momentum error and this isnt working either!

Hello. i am doing a multiphase simulation in a packed bed using the mixture model. the volume fraction used was 0.5316.i keep getting the error message "divergence detected in AMG solver: x-momentum" for all the different solvers used and i even changed the under relaxation factors. i really have no idea how to proceed with this simulation!plzz help!thank u.

mine too
 

Dear friends,

I'm having same problem , as i ve checked the meshing was ok. but, x-axis still not 0. how to state the axis as needed plus how you re sure your boundary condition is correct?

This x-momentum problem make me stuck forever.

Help me

to check whether your boundary conditions are properly defined just have a look at your values obtained from the velocity vector profiles.you should generally have an idea where should be the inflow or outflow this can also help you to check your suitable conditions.

which version of fluent are you using?

i got this proble because i was using an old version of fluent.maybe you should update yours.

my problem on the x-momentum was never solved.

try the different relaxation factors and your conditions setings.

goodluck!

Solution for Error : Divergence detected in AMG solver: x-momentum
 

This error is due to large mesh element size. If the mesh was created in units other than metre the mesh has to be scaled. Go to problem setup/general/mesh/scale and check on the specify scaling factors.

Hi all,
i have this error...

# Divergence detected in AMG solver: x-momentum -> Increasing relaxation sweeps!

Error: Divergence detected in AMG solver: x-momentum
Error Object: #f

May i must increase from the solution controls---advanced---max fine relaxation ????
What shall i do??
Thank you in advance!

Hi, I had the same problem before,

1. check your mesh, make finer
2. check the initial conditions, usually this kind of errors appears because of wrong values of the initial conditions
3. good luck

Hi dude,

I have still the same problem...Maybe i should chande the under relaxation factors and scale better the mesh...pfff..

try to make the energy under relaxation factor 0.995
I am sure that eventually you will solve the problem.

good luck

the reason of the divergence is manifold: without any setting you get only speculative tips nothing more! Either you post your settings or you contact you Ansys support if you have a legal license.

-Boundary Conditions
-Velocity Profiles
-Numerical Setup

The fact that you give small URS til 0,001 and we still get failure that means your setup is wrong

Hi Zaktatir,
so do you want to send you my settings-project to take a look?
My email is jimromanas@hotmail.com
Thank you in advance!

same problem is there with me

Minor problem in geometry
 

Good day
I have a problem when I am simulate unsteady flow around a propeller, divergence detected in amg solver. x momentum, I reduce the URF but is the same problem

try to use coupled solver and also see the courant number for your case

Some tips.
1) Use Check Case button under Run Calculation menu in modern (12+) versions of FLUENT. Maybe you have forgotten some settings, FLUENT can point some of them out.
2) Check mesh (Mesh -> Check) - skewed elements, minimal element volume (should be positive).
Check visually in your geometry utility for double faces, wrong cell zone/ boundary zone assignment.
3) Use under-relaxation factors.
4) Try to start from 1st order solver.
5) Try to use coupled p-v scheme.
6) Try change the "Advanced" options (multigrid cycle -> F-cycle with
BCGSTAB, reduce termination criteria) - find such recommendations in tutorials for your type of problem.
7) As dinesh says, for unsteady tasks Courant number should be taken into account - generally, less than 1 for explicit coupled solver, try 0.1-0.5 if you have troubles (see Changing the Courant Number in FLUENT manual).

divergence detected in AMG solver
 

hi frns

i am doing combustion simulation of IC engine. i have used pressure based solver and under solution controls tab i hav selected PISO scheme similarly for spatial descritization i have selected "Green-Gaus node based" and PRESTO for pressure. i have even Previwed the dynamic mesh for whole 720 degress. but when i go for calculation it runs only for few degrees i.e i have started simulation from 348 deg it runs upto 871 and exactly at 871.25 it is showing error as shown below
"turbulent viscosity limited to viscosity ratio of 1.000000ee05 in 590 cells divergence detected # Divergence detected in AMG solver: x-momentum -> Increasing relaxation sweeps!

i have tried by reducing URFs but it didnt work. can anybody suggest or help me how can i overcome by this problem. it ll be great help if somebody help me out....

regards
sarf

Dear All,

I try "Solution Control > Explicit Relaxation Factor " reduce to 0.01, and then it works

Let's try yours

Best Regards,
Y S.

Coarse mesh is the culprit
 

Hi,
I had the same divergence problem. I just made the mesh finer, now the solution is going smoothly.

All the best

Sabin Mathew

try increasing the viscosity ratio to1e8 ' it may help, but the fine mesh is the right call!

Further seggestions for Divergence detected in AMG solver
 

The Under Relaxation Factors can help. Moreover, very often, this error is related to the initial solution that is too far from the final one.
If neither of these suggestions solves the problem....well, check accurately your boundary conditions.
Hope this helps.

question
 

Hello
I'm trying simulate a micro channel heat sink (laminar flow) with temperature dependence properties that I used one UDF for all properties. after interpreted UDF, I run and see this error:
"Divergence detected in AMG solver: x-momentum"
I initialized by fmg , before this initialization I saw this error ,too.
"Divergence detected in AMG solver: temperature"
I use simple method, checked mesh and I didn't find problem.
Maybe this all problem related to my UDF?
I would appreciate if anybody help me.
sorry for grammar, I'm weak in English.

you solved this problem? i am also facing same problem..any help would be great

Hi everyone !

This may sound obvious but if you work an an unsteady flow, did you first initiate the solving of the equation with a steady flow ?

If your model is axisymmetric make sure that your geometry is adjusted in the first quarter of fluent coordinate system.

Dear sir

I am using 2D planer model with checmkin file to solve the model for ignitoion time.
I am unable to removef thr divergence error even after reducing the relaxation values.
I made the initialisation zero for turbulence

But unable to remove the error

Just ran into this divergence problem when trying to run a simple 2D natural convection problem (validation of the experiment done by P.L. Betts and I.H. Bokhari (2000), “Experiments on Turbulent Natural Convection in an Enclosed
Tall Cavity,” Int. J. Heat & Fluid Flow, Vol. 21, pp. 675‐683.).

Even though I was sure I had set everything correctly, the density of air under Boussinesqe density formulation was 0 (under materials - fluid - air), even though I've set an operating density of 1.225 kg/m3 in the operating conditions for the cell zone (an though this superseeds the setting for the material). The simulation wouldn't even start after. Initialization looked weird too (like artifacts on the screen when I tried to display the starting temperature of 320 K in the air domain).. So, if anyone runs into this problem - the fix was to set the density in the material creation after choosing Boussinesq density formulation (when you first pick it boussinesqe the density it set to zero). Now everything is fine.. URFs don't matter, SIMPLE/Coupled doesn't matter..

Hi everybody

I am trying to investigate the effect of Brownian motion and thermophoresis in microchannel by Ansys Fluent (single phase).

I prepared UDS for transport equation include source term and diffusion term.
Finally I set boundary condition for flux of particle.
But it doesn't converge even for 1 iteration by this error:
divergence detected in Amg solver-x momentum:confused:
(I tried many ways:
different URF+different solution methods+AMG solver setting checked but no change )

Thankyou in advance for your help ,I'm stuck in my project:(.
what should I do?

by the way my code was correct without UDS in this case(only udfs for temperature dependent properties) and UDS was correct for 2 another 2d-article,too. :)

Regards

Adaptive time stepping method
 

I didn't receive any divergence error after I changed the time stepping method to adaptive while running calculations. :)

Kindly check your mesh,if mesh is ok then this problem will be solved.

بسم الله الرحمن الرحيم

I answer with my simple knowledge

1) Reduce the parametres of the dynamic mesh (Spring / layering / remeshing)

2) Reduce the time step of the transient model

3) check that the density is not (incompressible) because the compression ratio is high

4) try to use coupled scheme and reduce coarant number

5) Reduce the under relaxation factor

6) reduce the stroke length or crank radius


Can you specifies clearly What is the condition of ignition you enters?

you can mailing me on (deyaa.nabil@gmail.com)


Asking Allah for helping you

Good luck

i had the same problem, i reduced the scale of my drawing and the problem was solve

Divergence detected in AMG for x-momentum: protective actions enabled
 

I am trying to analyze heat flow in an annulus of hollow cylinder keeping water as a medium in the annulus.

1) tried to model in 2 D axisymmetry keeping id 30mm and od 152 mm height of cylinder as 124 mm.

the boundary condition applied are: inner dia with temp 304 K and outer dia 300 K, top and bottom surface isolated means adiabatic

2) tried keeping in 2 D planar by treating it as a square of size 124 by 124 mm.

the boundary condition applied to the four sides of this square with temp 304 K from the bottom surface and 300 K to top surface, left and right walls isolated means adiabatic.

when i use air as a medium i get some kind of solution without error
but when i replace it with water, i get the error "Divergence detected in AMG for x-momentum: protective actions enabled"

can someone help me to figure out the mistake

Hello Dears

I have this problem in my model. I try to simulate an 2d axisymmetric tube that a ball move along the axis boundary. in order to move this ball I enabled Dynamic mesh setting. so the mesh must be decreased/increased in number in the boundary as the picture I have attached here.

I face with error (AMG solver) as the second picture when one of the nodes on the boundary is remove.

can anyone help me how I can solve this problem?


ThanksAttachment 84212