Particle tracking in an air flow

170 views
Skip to first unread message

R. Meyrand

unread,
Jan 24, 2012, 4:07:36 AM1/24/12
to FDS and Smokeview Discussions

I'm currently working on the propagation of particles in an air duct
network, and thus trying to figure out how the different particle
tools work in FDS (version 5.5.3).

I designed a test case consisting in one duct 9 m long and 0,2 m wide
(square section). One one end, I inject an air flow (1m/s) charged
with particles of 10µm, density 2000 kg/m3. The particles are light
enough to cross the entire length of the duct. The entry mass flow,
set up in the Intake SURF is calculated to have 1000 particles per
second injected in the flow.

My objective is to be able to mesure the mass flow of particles or to
count the flow of particles at different length of the duct. I already
figured out the working of the AMPUA command, to see how much
particles deposit on the "floor" of my duct.

1) I'm currently trying to use the PDPA sensor, because I read in User
guide that there is an output quantity "NUMBER CONCENTRATION", but
lacking any knowledge of the formula 14.15 presented on page 167 I
can't use it and being certain of what I actually mesure and where I
mesure it.

Is there any more in depth description of the PDPA command, or any
document to explain this 14.15 formula ?

2) I also wanted to know if there is any way to extract the data of a
SLCF line (that I use to evaluate the particle mass flow at different
sections of my duct) without time averaging it with the fds2ascii
command. For example to use the data in matlab or even Excel.

Thanks in advance for your response.

dr_jfloyd

unread,
Jan 24, 2012, 8:54:51 AM1/24/12
to fds...@googlegroups.com
1) All the terms are defined in the text below the equation.  Which term is unclear? 

2) See 14.3.10 for using DEVC to obtain integrated outputs in a csv file

R. Meyrand

unread,
Jan 26, 2012, 4:19:40 AM1/26/12
to FDS and Smokeview Discussions
1) It will simpler if I show you the lines

&PROP ID='pdpa_adm'
PART_ID='CHARGE'
QUANTITY='NUMBER CONCENTRATION'
PDPA_M=1
PDPA_N=1
PDPA_RADIUS=0.1
PDPA_START=2.0 /

&DEVC XYZ=0.0,0.0,1.8, QUANTITY='PDPA', PROP_ID='pdpa_adm'
ID='PDPA_Admin'/

With this, if I am correct, The sensor 'PDPA_admin' returns the number
of particles 'CHARGE' in a sphere of 0.2m of radius (centered on
[0,0,1.8]), and the measurment starts at the 2nd second of the
simulation ?

My problem in that case is that these lines give me a count of
millions of particles. Yet I only introduce arround 1000 particles per
second, and the velocity of the flow is sufficient to bring the
particles to the exit of the duct, so I should have results of the
10^3 order. I don't understand what my error is.

2) Thank you for this answer, I overlooked this part :).

dr_jfloyd

unread,
Jan 26, 2012, 8:19:29 AM1/26/12
to fds...@googlegroups.com
FDS considers each particle to be multiple particles.  Since a typical sprinkler outputs far too many droplets to try and simulate each one, FDS uses the flow rate to weight the individual particles.  For example, if you were to specify an injection of 100 particles per second with each particle having a mass of 1 g and wanted 1 kg/s of flow, then FDS would assign a weight of 10 for each particle (100 particles/s x 1 g/particle x 10 = 1 kg/s).  The NUMBER CONCENTRATION reflects this weighted value. 

R. Meyrand

unread,
Jan 26, 2012, 10:19:50 AM1/26/12
to FDS and Smokeview Discussions
Thanks for this information.

For information, I inject my particles via a SURF line. The intake
surface of my duct has a specified PARTICLE_MASS_FLUX of 1 kg/s for
example. Therefore, if I adjust the particle density and their
diameter to have 1g particles, FDS will only compute 100 particles of
10g per second, but the PDPA sensor will measure a flow of 100
particles of 10g => 1 kg/s ?

Can FDS track up to 100 particles per second (representing more
particles) or is there a rule for this multiple particles/real
particles ratio ?

And moreover, doesn't this regrouping of particles alter the particles
deposition on floor surfaces (as heavier particles will be more likely
to fall on the floor due to gravity) and bias the AMPUA measurement
for example ?


-

dr_jfloyd

unread,
Jan 26, 2012, 10:36:32 AM1/26/12
to fds...@googlegroups.com
The numbers I gave were a completely fictitious example to illustrate the concept of particle weighting and not intended to be inputs for your analysis.

A typical sprinkler injects far too many droplets to track.  You as the user need to decide the level of resources you wish to spend on particle tracking (e.g. the number of particles).  In your inputs you specify some number of particles per second with some size distribution along with a total mass flow. FDS will adjust the particle weighting so that when you sum over all particles being injected the particle weight times the particle mass you will get the correct mass flow rate.  The PDPA sensor includes the weighting factor.

Yes, of course tracking tens of thousands of weight particles will be different than tracking millions of unweighted particles.  We cannot decide for you if this assumption is appropriate for your application.

If you are trying to simulate aerosol deposition with FDS, you should review the technical manual and decide if FDS has the necessary particle physics for your application. 

R. Meyrand

unread,
Jan 26, 2012, 11:16:45 AM1/26/12
to FDS and Smokeview Discussions
> The numbers I gave were a completely fictitious example to illustrate
> the concept of particle weighting and not intended to be inputs for
> your analysis

Me too, it was just to see if I understood your point :).

Thank you for all your answers, I understand better how FDS works with
this.
Reply all
Reply to author
Forward
0 new messages