Problems using MPUA and AMPUA

[daserra]

I am stuck with a problem concerning MPUA and AMPUA.
I run the following problem and I do not understand the results. In
the xxx_devc.csv file the value for MPUA is zero and the value for
AMPUA rises up to 12.7 kg/m2. If I look at the BNDF data in smv, MPUA
seems to be zero and AMPUA is 0,85 kg/m2 maximum.

To my understanding MPUA is the actual mass of water on the floor
while AMPUA is the mass of water that has fallen on this location in
the course of time. Both should give more or less a circle.

I would also like to recommend that smv should be optimized such that
plots of data which are always greater or equal zero should start with
zero as minimum. Sometimes smv has negative minimumin in the scale,
which is impossible by physics.

&HEAD CHID='BNDF_test', TITLE='BNDF_test' /

&MESH IJK=25,25,40, XB=22.0,27.0,19.0,24.0,0.0,7.6 /

&TIME TWFIN=120. /

&PART ID='water drops',
WATER=.TRUE.,

QUANTITIES(1:3)='DROPLET_DIAMETER','DROPLET_TEMPERATURE','DROPLET_AGE',
DROPLETS_PER_SECOND=2000,
DIAMETER=750.,
SAMPLING_FACTOR=1 /

&PROP ID='nozzle',
PART_ID='water drops',
FLOW_RATE=2.0,
DROPLET_VELOCITY=5.,
SPRAY_ANGLE=0.,45.0 /

&VENT MB='XMIN', SURF_ID='OPEN' /
&VENT MB='XMAX', SURF_ID='OPEN' /
&VENT MB='YMIN', SURF_ID='OPEN' /
&VENT MB='YMAX', SURF_ID='OPEN' /

&DEVC XYZ=24.5, 21.5 , 7.5,
PROP_ID='nozzle',
ORIENTATION=0,0,-1,
QUANTITY='TIME',
SETPOINT=0.,
ID='noz_1' /

&BNDF QUANTITY='water drops_MPUA' /
&BNDF QUANTITY='water drops_AMPUA' /

&DEVC XYZ=24.5, 21.5 ,0.0, IOR=3, QUANTITY='water drops_MPUA' /
&DEVC XYZ=24.5, 21.5 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /

&TAIL /

[Kevin]

I understand why this is confusing. MPUA is the instantaneous mass of
water on the floor. Because the floor is assumed "POROUS", the
droplets simply pass through. This is the default because we don't
want to waste time tracking droplets on the floor if the user does not
think it important. So MPUA is zero on the floor. However, we still
measure the amount of water that has passed through the floor using
AMPUA. This is so we can do a "bucket test".

I will add text to the User's Guide to clarify this.

[daserra]

Thank you for your quick response. But unfortunately I still have
problems with MPUA and AMPUA. I have added the POROUS_FLOOR=.FALSE.
line and rerun the problem. When I look at the data at 100.1 s the
values from smv at the device position (bright little spot in the red
area) are
MPUA ~0.1?? kg/m2 (only 2 digits available) and AMPUA ~0.85 kg/m2
while the device file xxx_devc.csv values are 1.38E-2 kg/m2 and 1.206E
+1 kg/m2. At least for AMPUA there is a clear difference that I do not
understand.

By the way, is the integral AMPUA (all buckets) available somewhere in
the output?

In order to save the developpers' time, I would like to propose to
start a "fds/smv knowledge base" because fds/smv has become pretty
complex and it is hard to think at all relevant issues when preparing
the input.

[dr_jfloyd]

The default DT_BNDF is twice the default DT_DEVC. i.e. if you ask for
NFRAMES=1000 you will only get 500 frames of BNDF data. The
integration time for those quantities is slightly different, so the
instantaneous values may be a little different.

Try changing the boundary min and max bounds in smoke view to
something the bounds just slighty above and below the values you
reported and see if you still get the same results. (i.e. check to
see that smokeview is coloring the contours correctly).

The BNDF file data in smokeview can be obtained with the fds2ascii
tool.

> > > &TAIL /- Hide quoted text -
>
> - Show quoted text -

[Bryan Klein]

The users guide is intended to be the "knowledge base". As in the PDF
format it is searchable and now contains many internal links and the
next release will include a new index in the back.

As all of the issues and answers that are relevant to creating FDS
input files should eventually be documented there.

Other than that we have this group which can be searched and the Issue
Tracker which can also be searched.

Adding another online repository of information would be too many.
Instead of adding more, my goal is to create a single search function
on the fds-smv website that will return all results from the User
Guide, Tech Guide, Group and Issue tracker. A single search tool for
all four disconnected sources, returning results to one place.

But until that is running, if you can identify what information is
missing from the existing information sources and/or what would be a
better format to present it in. That will help us make the existing
sources better.

-Bryan

--
Bryan Klein
bryan...@nist.gov

National Institute of Standards and Technology
100 Bureau Drive, Mail Stop 8663
Gaithersburg MD 20899
Phone: 301 975 5171
Fax: 301 975 4052

[daserra]

Thanks fpr your quick answer. I checked the values in smv by choosing
adequate minimun and maximum as proposed. The good news is the values
for MPUA aggree perfectly. The bad news is that for AMPUA the smv -
value is 5.54 kg/m2 while the DEVC - value is 12.06 kg/m2. There seems
to be a factor of 2 missing. I do not think this is only caused by the
different time steps. I have looked into the source code but did not
reach any conclusions.

> bryan.kl...@nist.gov

[dr_jfloyd]

The DEVC uses just the value from a single wall cell. Smokeview needs
info at cell corners rather than cell centers, so the BNDF data is an
average over four wall cells to get the corner value. Try placing 4
DEVCs so that they lie one each a block of 4 wall cells. The BNDF
data would be the average of those.

> > Fax: 301 975 4052- Hide quoted text -

[daserra]

I am a little inquisitive and tried to check your explanation. I put
five &DEVC into the input files, for on the four corners ofa single
grid and one in the middle. The strange finding is that the value for
the middle point (0.1,0.1) and the point bottomm left (0.0,0.0) are
the same and the values for the middle point is not the avarage as
expected.

I tried to check the values used in smv for the presentation, but I
experienced a read error when using fds2ascii on BNDF. I have taken
the source code from the repository and I build fds2ascii new.

[dr_jfloyd]

The centermost floor cell (5 below) in the bndf file is obtained by
the average of the four corners (the .'s) of that cell. This means 9
locations contribute.
1 2 3
. .
4 5 6
. .
7 8 9

I added the following to your case:
&DUMP DT_BNDF=0.05,DT_DEVC=0.05/ Force both to dump at the same
timesteps

&DEVC XYZ=24.3, 21.3 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /
&DEVC XYZ=24.5, 21.3 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /
&DEVC XYZ=24.7, 21.3 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /
&DEVC XYZ=24.3, 21.5 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /

&DEVC XYZ=24.5, 21.5 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /

&DEVC XYZ=24.7, 21.5 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /
&DEVC XYZ=24.3, 21.7 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /
&DEVC XYZ=24.5, 21.7 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /
&DEVC XYZ=24.7, 21.7 ,0.0, IOR=3, QUANTITY='water drops_AMPUA' /

at 20 s Smokeview shows AMPUA is 1.17 and using the DEVC's to get the
4 corners and the averaging to get the center results in 1.167.

> > > - Show quoted text -- Hide quoted text -

[daserra]

Thank you for your quick and clear explanation. Now I understand how
the averaging is done. Thank you.