Because (in the past?) new users were told in training that aspect ratios <20 are good, 20-200 are bad and everything >200 is probably killing the convergence.
The worst aspect ratio I ever had was around 10000 and I solved the case without any problems. I would definitely not worry about an aspect ratio of 60 (peanuts ). Even >200 might be ok.
What can be done?
- create geometry with a reasonable tolerance, avoid very small gaps where possible.
- set an appropriate minimum grid size - some keypointed grid lines will be removed. Can be dangerous!
- add more grid. Use localized grids.
If in doubt, raise an SR!
You are probably right. In general purpose CFD where using a distorted hex mesh you would need to have a much lower aspect ratio than 60. Using Cartesian based meshes allows much higher aspect ratios to be used in both solid and fluid regions. I have solved a case with an aspect ratio of 1000:1 in FloTHERM with no numerical artifacts in the pressure and flow field.
Unless there is a specific issue in this case (which is why I asked the reason for the concern) I would not worry at all.
Thanks Dirk and John for reply and inputs.
I fully agree with Dirk, the instructions during the training period (in the past and even now) will get embedded in the users memory as a rule of thumb and will be difficult to erase. And such discussions helps to come out of such mindset.
Like John I have also come across case solved with aspect ratio >200 and not convinced with results because of above mindset.
It will be good if more experts share their views.