user fortran
 

hi,all

Do you try to compute the gradient of a scalar variable? Now, I use the utility routine GRADS to compute the temperature gradient at a conduction boundary. You know, the heat flux can be compute by the temperature gradient. OK! I acquired the heat flux by the kind of way.

However,a problem is always dazzling me.

Do you try to use "WALL HEAT FLUX" variable name to acquire the heat flux directly?

I try the two kinds ways to compute the wall heat flux, but the results don't coincide, why?

If you are interested in the problem, we can discuss this problem in here!

Re: user fortran
 

Hi all,

This sounds similar to a problem I had recently. If I remember correctly, I was attempting to calculate the mass flow rate out a pressure patch. When I integrated over the patch (ie summed the node face area * density * normal valocity over the patch), the figure I got didn't equal the "mass flow rate out patches" listed in the output file. The two answers were close, but significantly different (about 10% different).

I never really figured out why there was a difference. I assumed it was due to my use of what is effectively a second order scheme to perform the integration, whereas CFX uses a better scheme. Of course, this was only my assumption and they're usually wrong.

Any suggestions?

Glenn

Re: user fortran
 

dingding and Glenn,

Are you referring to CFX-5, CFX-TASCflow, or CFX-4? Please be specific.

Robin

Re: user fortran
 

hi,Robin

According to my experiences, CFX5 or CFXTASCflow can not use the user-subroutine. I use cfx4.3. I think that Glenn use the same software too.

Re: user fortran
 

Hi,

Yes, sorry about that - I was using CFX4.3

Glenn

Re: user fortran
 

You can use user source code in CFX-TASCflow just fine, but I don't think you need to use it to calculate a mass or heat flow. So, you should probably try your experiments again as they did not work correctly.

On the other hand, you are correct about CFX-5.4.1. It does not have a user source code interface.

Dan.