[esp-r] Re: CFD NaNs

[Jon Hand]

A thought on this - there is an exceptionally thin accounting layer adjacent to the
boundary of a CFD domain with a thickness set via the parameter SMALL.  I had
a problem with several small values being multiplied in setting up the volumes of
cells.  

If the diffuser is at the boundary it might be 8e+11 results from a division by 
that small layer thickness.

Perhaps changing small in cfdat from E-20 to something like E-12 might
impact TEf.  

Regards, Jon Hand
________________________________________
From: esp-r-bounces at lists.strath.ac.uk [esp-r-bounces at lists.strath.ac.uk] On Behalf Of Andrew Cowie [cn06arc at leeds.ac.uk]
Sent: 10 May 2013 20:33
To: esp-r at lists.strath.ac.uk
Subject: [esp-r] Re: CFD NaNs

Hi all,

I've tracked the problem to MODED line 1783 (called from CALCED).  The first Inf is occuring because TEf(i,j,k) is about 8e+11, i take it this is not meant to go this high?  I can't make heads nor tail of the code to work out what TEf actually is, does anyone know?  And does anyone know why this might be happening?  It is occuring at a node adjacent to the inlet, so im guessing it has something to do with the diffuser modelling.  However i've played around with the gridding geometry and it doesnt seem to make much difference.

If anyone has any ideas at all please do contact me, i have a conference paper deadline in a week and no results now because of this problem.  I don't really have any other option but to have some kind of diffuser in there, and i need to keep the gridding quite coarse to keep computation time down.

Thanks in anticipation,
(and a little desperately)
Andy Cowie

_______________________________________________
esp-r mailing list
esp-r at lists.strath.ac.uk
http://lists.strath.ac.uk/mailman/listinfo/esp-r