CFdesign accuracy

[PAD74]

Hi,

Does any of you guys have experience with CFdesign when calculating both fluid flow and heat transfer?

I'm currently running a simulation with a small water cooler which needs to remove 130W. Now the problems arises after a converged laminar solution when I look at a integral energy balance. The temperature rise of the water should correspond to 130W but is actually in the vicinity of 280W.

How well should a solution be converged in order to get feasible results. The manual says a max. change of 5% of the residuals within the last 20% of iterations should be sufficient?

What are your experience with this?

Regards,
PAD

 

[IRstuff]

I have no experience with CFD, but I would think that a factor of 2 error is excessive

TTFN

 

[sailoday28]

What does converged laminar solution mean with regard to CFD integral energy balance?
Does manual have any sort of verification data by test/experiment that justifies "max. change of 5% of the residuals within the last 20% of iterations should be sufficient"?
Does formulation of CFD calculation include momentum distribution in flow?

 

[PAD74]

What I mean is that I get a converged laminar solution. I then look at temperatures and flow rates at the boundaries and I'm able to calculate a heat transfer of 280W based on the boundary values. However, only 130W is going into the fluid.

The manual says something like "5% during last 20% of iterations is sufficient in most cases"

/PAD

 

[sailoday28]

Ask the supplier/vendor of the manual/program my questions?

 

[drjrice]

I am quite familiar with CFD and have quite a bit of experience with convergence of software like CFDesign and I've seen this question come up numerous times. I would suggest that you monitor the convergence of the quantity that you are primarily interested in. For example, if it's the total heat transfer, then check that quantity to see how well it has converged or if it is still changing. This can be very informative.