Prepared by Enrico Da Riva Gennaro Bozza 12042012 G Bozza E Da Riva ISOLDE facility aerosol results No Simulation Status 1 Sampling Tube 45 Degrees Cut operation 3ms Done 2 Sampling ID: 802247
Download The PPT/PDF document "1 April 18, 2012 Aerosol Sampling" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
1
April 18, 2012
Aerosol Sampling
Prepared by: Enrico Da Riva, Gennaro Bozza12-04-2012
G. Bozza, E. Da Riva
ISOLDE facility aerosol results
Slide2No.
Simulation
Status
1
Sampling
Tube, 45 Degrees Cut, operation (3m/s)
Done
2Sampling Tube, 45 Degrees Cut, flush (8m/s)Done3Sampling Tube, coaxial, operation (3m/s)Done4Sampling Tube, coaxial, flush (8m/s)Done5Sampling Tube, shrouded, operation (3m/s)Not done6Sampling Tube, shrouded, flush (8m/s)Not done7Bend, 45 degrees and Rb=1Done8Bend, 45 degrees and Rb=3Done9Bend, 90 degrees and Rb=1Done10Bend, 90 degrees and Rb=3Done11Straight pipe, 10m long, D = 1.12m Almost done12Final Enlargement, 7cm longFirst conclusions available13Complete Isolde Ventilation ductDone
Remaining simulations
G. Bozza, E. Da Riva
2
April 18, 2012
Slide3G. Bozza, E. Da Riva
3
April 18, 2012
Table of ContentsTransport efficiency (90° bend)ISOLDE (stack + sampling pipe)
Slide4G. Bozza, E. Da Riva
4
April 18, 2012
Transport efficiency (90° bend)
Slide5Geometry
G. Bozza, E. Da Riva
5April 18, 2012
D = 10 cm, flow rate = 20 m3 h-1Different bending radius (R/D = 1, R
/D = 3)
All possible orientations
Slide6Settings sensitivity
G. Bozza, E. Da Riva
6April 18, 2012
Modifying Drag law, Saffman’s force, Step Length Factor, Integration schemeDefault settingsAnalytic
All other settings
Slide7StarCCM+
vs. Fluent results for the 90° bend
G. Bozza, E. Da Riva7April 18, 2012
0.1 μmx
y
R/D=3
R/D=1
Slide8StarCCM+
vs. Fluent results for the 90° bend
G. Bozza, E. Da Riva8April 18, 2012
1 μmxy
Slide9StarCCM+
vs. Fluent results for the 90° bend
G. Bozza, E. Da Riva9April 18, 2012
10 μmxy
long stretch
horizontal
No-gravity & long stretch vertical
Slide10Trapped particle distribution (example)
G. Bozza, E. Da Riva
10April 18, 2012
dp=10 μm,
Slide11Long stretch horizontal (R
/D=1)
G. Bozza, E. Da Riva11April 18, 2012
Transport Efficiency
Simulation StarCCM
Simulation Fluent
Correlation
Particle 0.1μm95%87%100%Particle 1μm40%86%99%Particle 10μm5%28%33%This table shows that Fluent results are much closer to the correlation than StarCCM+ results.The 15% of efficiency underestimation of Fluent, compared to the correlation, is explainable in the following way: Fluent overestimates the turbulent inertial deposition compared to the correlation.Simulation vs. correlation
Slide12G. Bozza, E. Da Riva
12
April 18, 2012
Turbophoresis
Fluent
StarCCM+
Unlike STARCCM+, in Fluent the effects of the
turbophoresis are not relevant.InletDownstream of the elbowAlmost the same (very uniform) distribution at every cross section
Slide13Conclusions 1/2
G. Bozza, E. Da Riva
13April 18, 2012
Turbulent inertial deposition:negligible in StarCCM and CorrelationRelevant in Fluent (15% with L ~ 100D)Gravity:StarCCM:
Fluent:
: Gravity deposition is the dominant phenomenon : Relevant influence of the elbow but not of the bending radius of the elbow.Negligible influence of the gravityStarCCM+ vs FluentTransport efficiency
Slide14G. Bozza, E. Da Riva
14
April 18, 2012
Transport Losses %
Elbow
Straight stretch
L
=10mTOTALTurbulent depositionGravitational deposition0.1 μm~5%~10%~0%~15%1 μm~5%~10%~0%~15%10 μm horizontal~10%~35%~25%~70%10 μm vertical~10%~35%~0%~45%Conclusions 2/2Some rough numbersD = 10 cm, flow rate = 20 m3 h-1
Slide15G. Bozza, E. Da Riva
15
April 18, 2012
2. ISOLDE (stack + sampling pipe)
Slide16G. Bozza, E. Da Riva
16
April 18, 2012
ISOLDEMachine Mode: Qmain = 7500
m3/hFlush Mode
: Qmain
= 15000 m3/hStack (main pipe)Sampling pipe
Slide17G. Bozza, E. Da Riva
17
April 18, 2012
Tables4
3
To instrumentation
After the inlet
1Before the Sampling point2Entering the inletMachine Mode: Qmain = 7500 m3/h0.1μm1μm10μm1- Particles before the sampling point Nst125726212571081255855Particles supposed to enter the sampling inlet: 4191419041862- Particles entering the sampling inlet Ni4758473047663- Particles after the inlet, outside of the stack Nis3419366239714- Particles reaching the instrumentation Nout13361324375Aspiration efficiency: Ni/Ns114%
113%
114%
Transport efficiency:
Nout
/Ni
28%
28%
8%
Sampling
efficiency = Asp eff * Trans eff
32%
32%
9%
Machine Mode:
Q
main
= 7500 m
3
/h
0.1
μ
m
1
μ
m
10
μ
m
1- Particles
before the sampling point
Nst
1257262
1257108
1255855
4191
4190
4186
2- Particles entering the sampling inlet
Ni
4758
4730
4766
3- Particles
after the inlet, outside of the stack
Nis
3419
3662
3971
4- Particles
reaching the instrumentation
Nout
1336
1324
375
Aspiration efficiency:
Ni/Ns
114%
113%
114%
Transport efficiency:
Nout
/Ni
28%
28%
8%
Sampling
efficiency = Asp eff * Trans eff
32%
32%
9%
Slide18G. Bozza, E. Da Riva
18
April 18, 2012
Eddies (sampling inlet)Strong recirculation zone which might negatively affect the measurementWhat to do: avoid this geometry, and use a recommended one (coaxial or
shrouded)
Slide19G. Bozza, E. Da Riva
19
April 18, 2012
Eddies (enlargment)The diverging angle of the final enlargement is too large. Therefore, a separation of the boundary layer occurs and the result is a recirculation zone which might negatively affect the sampling efficiency.What to do: modify the geometry, and possibly eliminate the enlargement.
Slide20Conclusions
G. Bozza, E. Da Riva
20April 18, 2012
Simulations for the machine case (7500 m3 h-1 in the stack, 25 m3 h-1 in the sampling) were run.Generally, we can distinguish two groups of particles with different behaviour:0.1 ÷ 1 μm → ~ 30% Sampling Efficiency10 μm → ~ 10% Sampling EfficiencyRecirculation zones at the inlet and at the final enlargement which could negatively affect the measurement.Next steps: simulations for the flush mode, simulation for the only stack.
Slide21What can be already recommended: multiple shrouded probes for sampling
G. Bozza, E. Da Riva
21April 18, 2012
Because it is necessary to design sampling systems to operate under both normal and accident conditions (when proportionately higher concentrations of large particles could be present), it is customary to sample isokinetically. Under the ANSI standard, for large ducts it is recommended that rakes of isokinetic probes be used to span the duct cross section, ostensibly to collect representative samples. (A Predictive Model for aerosol transmission through a shrouded Probe, HONGRUI GONG, SUMIT CHANDRA, ANDREW R. MCFARLAND, AND N. K. ANAND).