CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   CHT problem (http://www.cfd-online.com/Forums/cfx/19760-cht-problem.html)

sch July 9, 2003 10:56

CHT problem
 
Hi all I am testing the case of a flow across a solid cylinder (Aluminium) to study the CHT in the solid. I imposed T=1000K at the inlet and at the exit (opening). I defined the cylinder as a solid subdomain with no heat source. As a initial guess I set T=1000K for the fluid and I intentionnaly imposed T=800K for the cylinder expecting at the calculation that it will evolve to 1000K. After 500 iterations the equations all converged to 1.e-8 except the temperature equation of the solid (stable at 1.e-4). At the post, the temperature at the solid remains around the initial guess (800K).

can anyone tell me what I made wrong? Thanks. Thanks a lot

Glenn Horrocks July 10, 2003 18:49

Re: CHT problem
 
Hi sch,

I have not got a lot of experience with CHT problems, but a major factor to consider is that in many configurations the timescale of conditions is slower in the solid than the flow timescale, often by orders of magnitude. To explain what I mean, image you suddenly changed the input flow conditions from 1000K to 500K. This temperature change would work through the fluid section quickly (in seconds), but you would imagine it to take much longer in the solid part before its temperature stabilised out (maybe minutes or hours).

What this means is that a timestep suitable for the fluid section of the flow is not suitable for the solid section. In convergence control, you have an option "Solid Timescale Factor". This sets the ratio of timesteps in the solid portion of the domain versus the fluid part. I have used a value of 10 for a geometry with the solid being long and thin, so variations are "fast", but I could see much higher values being suitable for your case where you are using a cylinder, and the lower surface area means variations react "slowly". Try a value of 20 to 100.

Also, set a monitor point in the solid so you can watch its temperature throughout the simulation.

Glenn

sch July 11, 2003 10:32

Re: CHT problem
 
Hi Glenn You are right. In my initial model I set the local timestep factor to 3 and the solid timestep Factor to 1. When I inscreased the solid timestep factor (up to 1.e4) nothing happened, then I set Auto Timestep for all equations. Only in this case the influence of the solid timestep factor (from 10 to 1.e4) plays. and the results are good.

Thanks a lot for your help

Sch


All times are GMT -4. The time now is 07:45.