[aspect-devel] convection in a (virtual) box

[cedric thieulot]

hi all, 

for reasons too long to explain here I found myself this afternoon quickly installing and running the VirtualBox software on my Mac.
I quickly wanted to obtain the vrms and Nu values of the Blankenbach benchmark (‘convection-box.prm’) for various values of Ra.
Without even editing the prm file, I ran the model and plotted the results, only to find out they are off with regards to the manual (and the actual benchmark values from the paper.)
And not half a percent off, a good 10-20% off :)
I believe there probably is a simple explanation: since the thermal expansion coefficient is set to 1 in the prm file the true Boussinesq approximation should be used, yet it is absent from the prm file. 
I therefore suspect that the virtual box was updated in the Spring but not correctly ? (the .prm file on github does contain a Formulation).

Best,

Cedric.







-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20180104/fc4ae338/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: two.png
Type: image/png
Size: 908902 bytes
Desc: two.png
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20180104/fc4ae338/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: one.png
Type: image/png
Size: 679817 bytes
Desc: one.png
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20180104/fc4ae338/attachment-0003.png>