not getting convergence
 

HI

iam working on simple natural convection problem. i have taken a copper plate of 1250*80*5mm and placed vertically in a steel enclosure of size 1630*510*150 and 2mm steel thickness.steel enclosure opened at bottom and closed at top.

for analysis i have considered only copper plate and air inside the steel encloser,i have given inflation mesh to air,which is having around 105000 elements.i have given copper-air interfaces. I have given opening boundary condition to bottom portion of air and all other outer faces declared as adiabatic walls.my model's vertical direction is y , i have given density as -9.81 in Y direction.tarbulence model used is k-e.I have tried both idealgas and air at 25 degC.i have tried different time steps for air and copper.

Plese help me

Thanks a lot

NVSD BABU

Re: not getting convergence
 

hi plz find the geometry and give the comments that i have to follow

http://img153.imageshack.us/img153/4...mainlr5.th.png

Re: not getting convergence
 

Hi,

Could you please tell what kind of convergence issue are you facing, is your problem diverging or you simply not achieving the residual lower limit?.

thanks,,

Re: not getting convergence
 

thank you pratik

i have given convergence as 1e-4, v-momentum not getting converged, it is oscillating and all other u,w and p getting 1e-4 convergence,but they are also oscillating.

Thank you NVSD BABU

Re: not getting convergence
 

Is it is high resolution scheme that you are running with or blend factor, kindly switch on the double precision tab as well, what is your global Y plus range and let me know more about the moniter points if at all you have

best of luck

Re: not getting convergence
 

Thanks pratik, i will try with double precision.

I have tried high resolution,upwind and blend factor1 schemes.

Maximum mesh expansion factor is 70.1

moniter point is temperature.

Thanks

Re: not getting convergence
 

hi prathik

global Y+ range is 0 to 3.412

Thanks

Re: not getting convergence
 

Sounds all basic settings seems good, are you solving it steady state or unsteady state . Try running with smaller time steps . First use K epsilon model and first order trasient case. Let me know what happens then

Re: not getting convergence
 

thanks prathik

I am solving the problem as steady state.If i do in trasient state,please tell me how to take total duration and time steps. practically, to get convergence our model took 4hrs. do i give 4hrs as total duration.

Thanks

Re: not getting convergence
 

Re: not getting convergence
 

You need to find out the suitable timescale of your problem in way from the velocity lenght scale and the integral lenght scale , but usually it is so straight forward for complex geometries , in that case you can iterate your time scale factor starting from the time scale factor that CFX uses during the steady state calculation , we can get this number from the output file.

So run trasient with Kepsilon n first order ... i hope your fluid properties are reasonably quoted , anyways try out this and lets see what you get