Problems with 3D
 

Hi! Up till now I have been doing simulations in CFX with only 1 cell in the third coordinate direction. My results were pretty good so i wanted to check what would happen if I extend my model so as to make it three dimensional. I increased the thickness and added more cells in this direction. Using the same setup as before (timestep, accuracy ...) I have tried running this simulation but it always crashes on the first timestep in the first coefficient loop (floating point exception: overflow). Any ideas on what I should do to enable this 3D run?

Hi,
Just the first step check - are you sure about the integrity of your 3D mesh ?
I mean do you see any error or warning messages for the new mesh ?

No, the mesh is is fine. I get no warnings or anything like that. It just crashes straight away. Should I change the setup for the 3D mesh or should it work with the same settings as before?

Please post the output file around where it crashes.

Hi! The error looks like this:

+--------------------------------------------------------------------+
| Convergence History |
+--------------------------------------------------------------------+
+--------------------------------------------------------------------+
| Writing transient file 0.trn |
| Name : Transient Results 1 |
| Type : Selected Variables |
| Option : Time Interval |
+--------------------------------------------------------------------+

================================================== ====================
| Timestepping Information |
----------------------------------------------------------------------
| Timestep | RMS Courant Number | Max Courant Number |
+----------------------+----------------------+----------------------+
| 5.0000E-03 | 0.39 | 7.60 |
----------------------------------------------------------------------
================================================== ====================
TIME STEP = 1 SIMULATION TIME = 5.0000E-03 CPU SECONDS = 2.666E+01
----------------------------------------------------------------------
| SOLVING : Wall Scale |
----------------------------------------------------------------------
| Equation | Rate | RMS Res | Max Res | Linear Solution |
+----------------------+------+---------+---------+------------------+
| Wallscale-Bulk | 0.00 | 2.2E-05 | 1.1E-04 | 43.0 6.1E-01 ok|
+----------------------+------+---------+---------+------------------+
| Wallscale-Bulk | 3.25 | 7.1E-05 | 2.2E-03 | 33.6 9.0E-02 OK|
+----------------------+------+---------+---------+------------------+
| Wallscale-Bulk | 0.62 | 4.4E-05 | 1.6E-03 | 29.0 9.7E-02 OK|
+----------------------+------+---------+---------+------------------+
| Wallscale-Bulk | 0.87 | 3.8E-05 | 1.5E-03 | 29.0 8.8E-02 OK|
+----------------------+------+---------+---------+------------------+
| Wallscale-Bulk | 0.84 | 3.2E-05 | 1.2E-03 | 29.0 8.3E-02 OK|
+----------------------+------+---------+---------+------------------+
| Wallscale-Bulk | 0.90 | 2.9E-05 | 1.1E-03 | 24.3 9.8E-02 OK|
+----------------------+------+---------+---------+------------------+
----------------------------------------------------------------------
COEFFICIENT LOOP ITERATION = 1 CPU SECONDS = 7.928E+01
----------------------------------------------------------------------
| Equation | Rate | RMS Res | Max Res | Linear Solution |
+----------------------+------+---------+---------+------------------+

+--------------------------------------------------------------------+
| ERROR #001100279 has occurred in subroutine ErrAction. |
| Message: |
| Floating point exception: Overflow |
| |
| |
| |
| |
| |
+--------------------------------------------------------------------+
+--------------------------------------------------------------------+
| Writing crash recovery file |
+--------------------------------------------------------------------+

Details of error:-
----------------
Error detected by routine POPDIR
CRESLT = ILEG

Current Directory : /FLOW/NAMEMAP

Start with a smaller initial time step or a more representative initial condition. This assumes you have set the simulation up correctly, of course.