Question about MASS FLOW DEVC

834 views
Skip to first unread message

Fabian Braennstroem

unread,
Aug 21, 2015, 5:16:37 AM8/21/15
to fds...@googlegroups.com
Hello,

I have a short question about using the MASS FLOW device. My intention is to
measure the MASS FLOWs through a door e.g. of the Steckler case.
FDS allows to measure the total mass flow (MF), the positive (MFp) and negative (MFn) mass flow through the door.
My expection is that: MF = MFp - MFn
This should be independent of the ignition source, any reaction or temperature.
Unfortunately, I do not see this in the simulation Steckler014 when adding these lines:

&DEVC FYI = '2A', ID = 'MASS FLOW',       XB = 2.85, 2.85, -0.37, 0.37, 0., 1.83, QUANTITY = 'MASS FLOW',  /
&DEVC FYI = '2A', ID = 'MASS FLOW +',     XB = 2.85, 2.85, -0.37, 0.37, 0., 1.83, QUANTITY = 'MASS FLOW +',  /
&DEVC FYI = '2A', ID = 'MASS FLOW -',     XB = 2.85, 2.85, -0.37, 0.37, 0., 1.83, QUANTITY = 'MASS FLOW -',  /



The difference between the output of "MASS FLOW" and the calculation "MASS FLOW +" - "MASS FLOW -" is very small, but still not zero.
The difference varies in time as well.
For other cases with larger HRR of the burner this difference seems to get larger.

I tried to understand a bit the source. As far as I got it, the calculation is done in dump.f90 (around lines 4963 ):

               SELECT CASE(IND)
                  CASE(111:113)
                     FLOW = FLOW + VEL*HMFAC*AREA
                  CASE(114:116)
                     FLOW = FLOW + MAX(0._EB,VEL)*HMFAC*AREA
                  CASE(117:119)
                     FLOW = FLOW - MIN(0._EB,VEL)*HMFAC*AREA
               END SELECT

From this I still expect that there should be no difference...!?

Do you have an explanation for the difference in the simulation? Probably, I am doing something wrong...


Thanks in advance!
Best Regards
Fabian


Kevin

unread,
Aug 21, 2015, 6:21:05 AM8/21/15
to FDS and Smokeview Discussions
I'll look at this today. It might be related to the way these devices are time-averaged.

Randy McDermott

unread,
Aug 21, 2015, 7:51:04 AM8/21/15
to FDS and Smokeview Discussions
At a glance, it looks like we average neighboring cells for species concentration, then just grab the velocity from the "cell" where the DEVC is stored.  This will be different if you are on one side of a face or another (different IOR gives slightly different velocity and flow).  Further, note that this flow only tries to give the advective flux*area, not also the diffusive flux*area, so it does not give total mass flow if integrated for a given species.  The advective part will also be in error because the actual calculation uses a flux limiter which is not accounted for in GAS_PHASE_OUTPUT.  It would take a bit of work to get this all consistent, but it could be done.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/b712c19f-6739-4724-84f0-aafdc981435e%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Fabian Braennstroem

unread,
Aug 21, 2015, 10:25:58 AM8/21/15
to fds...@googlegroups.com
Hi Randy,
thank you for your quick response!

I am not sure, if this explains the difference between the "MASS FLOW" and the "MF" based on the pos. and neg. mass flows.

Even when the pos. mass flow takes other averages for the concentration than the negative mass flow, shouldn't the "MASS FLOW" be based on these slightly uncorrect mass flows as well?


Best Regards
Fabian



For more options, visit https://groups.google.com/d/optout.



--
Regards!
Fabian

Fabian

unread,
Aug 21, 2015, 11:57:07 AM8/21/15
to FDS and Smokeview Discussions
One side comment, the actual idea is to compare FDS for ventilation-limited scenarios with the empirical formula from Babrauskas M_in=0.5*A*sqrt(H).
Maybe you know of some work about this which was already published!?

Randy McDermott

unread,
Aug 21, 2015, 12:59:15 PM8/21/15
to fds...@googlegroups.com
Fabian,

I don't know other work published on this.

The problem with the +/- side flow values is that FDS is grabbing two different velocities. As I said, we could tighten this up. It is a first order error.

However, in my opinion, a more productive enhancement request would be to add a general linear interpolation for device quantities at their respective positions. But this would take some work.

Best,
Randy

For more options, visit https://groups.google.com/d/optout.


--
Sent from my iPhone

Kevin

unread,
Aug 21, 2015, 2:57:49 PM8/21/15
to FDS and Smokeview Discussions
Fabian, do you have a simple test case you can send?
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+unsubscribe@googlegroups.com.

To post to this group, send email to fds...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/46817149-e207-4454-808f-d5faf583030b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Fabian Braennstroem

unread,
Aug 23, 2015, 9:57:39 AM8/23/15
to fds...@googlegroups.com
Hello to you both,

yes, sure, the actual case is based on the Steckler case mentioned, where I changed slightly the mesh to 5 and 10cm and fitted the opening. I try to send it tomorrow.

With +/- side flow and having two different velocities, I think I understood this part, but I am still having trouble to see that this part should lead to a difference in the overall "MASS FLOW"...

Best Regards
Fabian

Kevin

unread,
Aug 24, 2015, 10:23:02 AM8/24/15
to FDS and Smokeview Discussions
I looked at the cases you sent. Since you have set

...
SIG_FIGS=4, SIG_FIGS_EXP=2
...

you can only expect about 5 digits of precision. I see that for 'MASS FLOW' and 'VOLUME FLOW', the +/- output sums to the full output to within about 10^-5. The 'HEAT FLOW' output is different, because we include the diffusive term as well as the convective in both the + and - output.

Fabian

unread,
Aug 25, 2015, 12:29:44 AM8/25/15
to FDS and Smokeview Discussions
Hello,

thank you! I missed this setting and the error gets much smaller.

Best Regards
Fabian
Reply all
Reply to author
Forward
0 new messages