you have the reynolds problem. the mesh is not compatible fo your simulation. the relaxation factor is just a coefficient which accelerates the convergency or divergency. the answer will not change by different relaxation factors.

best regards.

You should re-check your Boundary Conditions. May be some unrealistic BC's you have set in your problem.

The mesh resolution could be a factor of course, in addition to the time step value if teh solution is unsteady (try to reduce it) and of course the discretization scheme. The models might be another factor to consider. Some turbulence models require specific treatment in some cases to avoid divergence at the beginning of the solution, some other models would require the solution to proceed for some iterations with a different model.

mesh offcourse is the main reason , you can ckeck your bc condition try adaptive method instead of fixed in calculation tab

AMG solver error has many reasons.

make your geometry horizontally and consider  the gravitational acceleration in the negative side of X-coordinate.

Finer mesh and under-relaxation and initialization!

When is the divergence detected? If at the beginning or the first iterations, boundary conditions or model settings might be a reason. Also considering outlet boundary conditions if reversed flows emerge. High deviations between values at the boundary and initial values are critical.

thanks for your guides

the mesh ratio is 8 with boundry layer mesh.I think it's quite fine for the problem.I've also chaged the depth and first row of B.L mesh.but it doesn't help.

B.cs are typical.mass flow inlet,pressure outlet and a heat flux on the wall.Does the change in mass flow inlet values,can change the situation?

I've decreased URFs,but the trend of residuals don't change and continuity residual goes up to 1.00.I think decrease in URFs just pospone the divergency.

Mr Akbari

I've been modeling the tube always horizontally with gravity at X-coordinate.but faced divergency!

I don't know what factors remain to change!

Note that multiphase simulations run almost under the unsteady state conditions. so is it not expected to reach exact convergence.

Therefore, you must use puny time step size and control another parameter for access convergence(ex: mass flow at outlet).

** the temperature of second phase must be fixed to superheat vapor does not occur.

thanks for your notes Mr akbari.

but How can i fix the temperature of phase 2?my inlet condition is water and vapour at saturated temperature.but over heat flux exposure it rises to super heat.

You're welcome.

cell zone conditions>> phase2>>> fixed values on >>> fixed values tab>>> temperature change to constant and write saturation temperature.

the inlet temperature of two phase is at saturated temperature.when i fixed the vapor temperature,the energy residual of this phase doens't change either.up to when the energy residual of phase1 goes from 1.e-7 to 1.#IND!!and diverged!!!

 I've searched about the 1.#IND,but nothing useful i've found!what is the meaning of this number?

Hey Sahar, May I know what solvers are you using for this problem. Did you try with second order accurate ones.

Also, there might a problem with source term need to be included as UDF for mass conservation.

Let me know about these details, then I would be able to help you through this.

Cheers 

the solution scheme is Coupled. for each of equations,i use first order upwind. i've never try second order,because of dispersion.

I dont have any idea about using UDF. RPI model includes some experimental equations.

thanks for your help 

I hope you would first like to try with second order solver. For UDF, I do have a similar problem, but on condensation, in which the source terms to be used is given. Moreover, writing a UDF for this is not a big deal. If you want, I can share some of these material with a good tutorial problem on boiling.

I will try second order!

if it doesn't work,I would be appreciate it that you share your UDF tutorial about boiling.

thanks Mr Ashish

have u tried to change mesh size???

Dear sahar

What is the purpose of your simulation?

What do you want to see in your results?

What is your convergence criteria?

.

.

.

Be sure UDF isn't necessary.

I've changed mesh many times.

I want to discuss about heat transfer paramters,h,Nu.

for convergence criteria,I monitor mass flow rate and residuals of equations.but continuity goes up.

Yeah! I think it will work. In other case, just let me know.

Check:

• properties of water
• operating conditions
• mass transfer conditions
• step time size
• .
• .
• .

I got same problem as

Divergence detected in AMG solver: x-momentum

I changed setting as general - type - from pressure based to density based.

try for your case if it works

multi phase models is not compatible with density-based type.

Which boiling regime do you expect in your simulation/experimental setup? Have you ever tried to change the wall boundary condition to temperature instead of heat flux? What is happening with the momentum equation when you start the simulation with moderate wall temperatures (“relatively cold flow”)?

local scaled

Hello! I have the same problem when simulating a boiling flux through a vertical channel.

My BCs are: velocity inlet (with a defined profile), pressure outlet and Wall (temperature). Temperature of vapor is fixed to saturation temperature.

I have tried both SIMPLE and COUPLE methods, and tried reducing the under relaxation factors. Still, divergence is detected after a few iterations.

Could it be due to the choice of phase interaction methods? I don't know what else I can try.

Any advice will be apreciated.

use a suitable relaxation factor , depending on the type of the problem.

search for somethings useful in these types!!

let me know more about your problem.

you have the reynolds problem. the mesh is not compatible fo your simulation. the relaxation factor is just a coefficient which accelerates the convergency or divergency. the answer will not change by different relaxation factors.

best regards.

I had experienced this kind of error before and I solved it by choosing the correct model and I paid a great attention to the boundary conditions especially the boundaries that are main in the problem.

Sometimes if you are working on a transient you may need to reduce the time step size at the beginning of the solution and increase it later.

Hi

just change the pressure solver for standard .