Convergence in AcuSolve

Junta_20361

Hello expert,

I'm using AcuSolve to simulate an internal flow in HVAC system.

The inlet speed is very high (25m/s).

I'm having problem with the convergence of solution, especially with Eddy Viscosity ratio.

Below is my residual ratio plot:

Usually, my calculation converge around 40 time steps.

However, for this simulation, the result always look as above, couldn't achieve congerence with eddy viscosity despite of many mesh refinement.

My question is: if i only concern about the pressure & velocity result, is the velocity & pressure result reliable?

Thank you!

Adriano Koga_20259

you should probably need to give more informations about your model, such as turbulence model used, refinement zones, solution parameters...do you have your BL in the right places?

ydigit

The velocity and pressure residuals are low.

Turbulence can hard to converge especially at higher speeds.

Can you plot your surface Y plus on the walls? This should not exceed 70.

 

Also plot Solution Ratios of all V,P and EddyV. They will give an idea of how the solution is changing.

 

Junta_20361

Dear Koga & Ydigit,

Thank you for your response.

Unfortunately, due to confidential, i cannot show/share the model.

But i can share with you guys some information:

1. i used SA turbulence model >> i believe you have recognized that.

2. the 1st layer thickness is 0.1mm * 10 layers >> max surface Y+ is 18 >> good, right?

3. this is an internal flow, i used Inlet BC = normal velocity & viscosity ratio = 30.

4. maximum value for Eddy Viscosity Residual is 9E-7.

5. below is the Solution Ratio plot from my calculation:

Anyway, thanks for you help.
Hope can receive some ideas from you.

Have a nice weekend!

ydigit

The solution ratios are very high. Especially for velocity.

This means that the velocity in this case is varying by ~10% with every time step. So whatever solution you get at the end, cannot/should not be used, as it has bad convergence.

Corrections ?

1. Check the surface y plus on walls. It should be below 70 and preferrably lower.

2. Run with higher relaxation factor of 0.5

Send screenshots of mesh, if possible. Any large jumps in element size should be avoided, especially from the last BL element to the tetra core elements.

 

 

Junta_20361

ydigit_21275 said:

The solution ratios are very high. Especially for velocity.

This means that the velocity in this case is varying by ~10% with every time step. So whatever solution you get at the end, cannot/should not be used, as it has bad convergence.

Corrections ?

1. Check the surface y plus on walls. It should be below 70 and preferrably lower.

2. Run with higher relaxation factor of 0.5

Send screenshots of mesh, if possible. Any large jumps in element size should be avoided, especially from the last BL element to the tetra core elements.

 

 

hi Ydigit,

Thanks for your input.

The relaxation factor is already 0.6.

Max Y+ is 18.

I cannot show the whole model due to confidential rule.

However, i try to show you some portion of the mesh.

I think the mesh is very small already.

Do you have any suggestion?

Thanks!

Nazrin

ydigit_21275 said:

The solution ratios are very high. Especially for velocity.

This means that the velocity in this case is varying by ~10% with every time step. So whatever solution you get at the end, cannot/should not be used, as it has bad convergence.

Corrections ?

1. Check the surface y plus on walls. It should be below 70 and preferrably lower.

2. Run with higher relaxation factor of 0.5

Send screenshots of mesh, if possible. Any large jumps in element size should be avoided, especially from the last BL element to the tetra core elements.

 

 

Hi @ydigit,

"Can you briefly explain about solution ratios? What can the user get from the solution ratios? Is the pattern of solution ratios will be different for Steady sate simulation and transient simulations for same case (let say, flow over cylinder)?

I have been learning about this solution ratios multiple time on youtube, blog, etc, But still, i still dont know how to monitor and deduce from the solution ratios i get.