firemodels
Remove excessive warning messages "DEVC... has no values"
Thanks as ever for the updated FDS6.10.0.
In my current cases, I like to add monitors of minimum, maximum and average temperature in various locations. These locations include doors and other HOLEs that switch on and off during the simulation. Some of these doors etc are closed for the majority of the simulation. Here is an example:
&DEVC XB = 7.079, 7.081, -1.500, -0.600, 0.000, 2.100, QUANTITY='TEMPERATURE', ID='tave_door_lobby_gap_', STATISTICS='MEAN', / door_lobby_gap_
There is a new warning that appears when a DEVC has some SPATIAL_STATISTIC to calculate, but there are no fluid cells within its scope:
WARNING: DEVC tave_door_lobby_gap_ has no values.
This warning appears for every timestep, for every such DEVC -- hence I now get hundreds of thousands of these warnings, and they overwhelm all other output. (This is new in FDS6.10.0.) Please could there be a way to limit the number or frequency of these warnings? For example, a parameter in &DUMP? (I would prefer this to be separate from &DUMP SUPPRESS_DIAGNOSTICS, which I override to T.) Thanks.
One alternative to the warning message would be for the values of the DEVC to make it clear when there is no valid value, using an error value such as NaN. Currently, DEVCs are reported with output 0.0 in timesteps with no valid values. However, the spurious 0.0 values are sometimes convenient in plotting, because some plotting programs assume that a column does not contain numeric data if its first and last values are not easily-interpreted values.
As a workaround, I guess I could extend the DEVC extents by a cell in each direction, so that there is always some fluid involved. Or maybe I'll just pipe the output through a "grep -v" command.
I removed the warning message. It is possible to print out NaN, but I have to check to see if the way to do it is standard and accepted by all Fortran compilers. There are various ways of representing "bad" numbers, and I want to make sure that I don't do something that is not standard.
The warning messages are gone, as is the variable that tracks the number of values used by a spatially-integrating device. I considered adding NaN to the output files, but it seemed to me that the current value of 0 is preferable.