Using FDS Modelling to Establish Performance Criteria for Water Mist Systems on Very Large Fires in Tunnels Jack R. Mawhinney, P. Eng., FSFPE Javier J. Trelles, Ph.D.
Authors & acknowledgement J. R. Mawhinney - full-scale fire testing 1990 to present Javier Trelles, Ph.D. - CFD modelling Modelling fire suppression in tunnels (NIST FDS) Marioff Corporation Water Mist spray characteristics Access to fire test data
Water Mist Systems in Tunnels Authorities require full-scale fire testing HGV fires 5 to 10 times larger than most testing experience Large fires = dynamic, high variability Need performance criteria appropriate for scale and complexity of the fire Fire tests (unsuppressed fires): 203 MW 158 MW 125 MW 70 MW
Performance Criteria for control Single-point measurements based on experience with smaller fires, e.g. Within 2 min, Temp at end of zone < 50 o C Within 5 min, Temp 5-m from truck < 250 o C Temperature at ceiling above fire: < 500 o C The apparent precision is out of scale with the chaotic variability of large fires We can never measure enough to communicate all dimensions of the performance
Computational Fluid Dynamics Modelling of Tunnel Fires CFD is a powerful tool to examine more than is measured NIST Fire Dynamics Simulator (FDS) version 4 Use fire test data to evaluate model Use the model to extend the understanding of the fire/mist interaction
Marioff Water Mist System Tests San Pedro de Anes TST Test Facility, Spain Ceiling Jet Fans Longitudinal Ventilation Fire Test Zone 0+350 to 0+430 Oxygen Analyzer Velocity Probes 0+610 Section modeled 140-m 0+320 0+460 600-m
Forced flow boundary Computational Domain Tunnel curvature Tunnel centerline + stations at 10-m intervals Four thermocouple trees Ceiling TC s on centerline at 5-m spacing Nozzles on three parallel lines A composite fuel load Figure 2.
FDS4 Input Parameters 25 23 21.5 Category Parameter Value CFD Domain Facility San Pedro de Anes Research Tunnel Simulation dimensions 140 m 23 m 5.17 m Numerical Grid dimensions 560 100 24 cells Cell size Total # of cells 1,344,000 Wall boundary conditions Floor boundary conditions Ceiling boundary conditions 25.0 cm 23.0 cm 21.5 cm Concrete Concrete Concrete/Promat Promatect-H Gravity vector (-1%) (0.0981, 0.0, -9.8095) m/s 2 Spray Nozzle Type Marioff 4S1MD6MD(1000,10RE) water mist Configuration Activation criteria 4 m 3.3 m grid Times as determined from test data
Water Mist Spray Characterization Drop size distribution F( d) = 1 ln[ d / d 1 + erf 2 2σ γ d ln 2 dm 1 e m ] if if d d d > d m m (Log - Normal) (Rosin - Rammler). (1) 0 25 50 75 100 125 150 175 200 225 250 275 300 1.0 1.0 0.9 0.9 0.8 0.8 Cumulative Fraction 0.7 0.7 0.6 0.6 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 Volume 0.1 0.0 Number 0.0 0 25 50 75 100 125 150 175 200 225 250 275 300 d (μm) Discharge Volume Radius = 200 mm
Water mist spray characterization Using measured values of Dv10, Dv50 and Dv90 Dv50 = d m = volumetric median drop diameter Dv90 = 171 μm, Dv50 = 89 μm, Dv10 = 35 μm Obtain σ = 0.728; γ = 1.84 > + =. e 1 2 ] / ln[ erf 1 2 1 ) ( ln 2 (Rosin - Rammler) if (Log - Normal) if m d d m m d d d d d d d F m γ σ
Spray inputs: velocity, mass flow/orifice Test Identifier Run Pressure (bar) T H2 O ( C) u (m/s) 4S 1MD 6MD 1000 m& (a) c m& (a) r (kg/s/m 2 ) (kg/s/m 2 ) 2N 1MD 6MD 10RE u (m/s) m& (a) c m& (a) r (kg/s/m 2 ) (kg/s/m 2 ) 1 100 50 123.5 51.2 199.7 116.8 48.4 188.8 2 80 50 110.5 45.8 178.6 104.4 43.3 168.9 3 78 80 103.1 42.7 166.8 10 90 80 110.8 45.9 179.1 13 80 50 110.5 45.8 178.6 104.4 43.3 168.9 a.) The subscript c stands for center orifice; subscript r for ring orifices.
Lagrangian droplet tracking Fire peak: 56 MW in 14 minutes Ventilation velocity: 3.5 m/s
Spray activation time determined by test data Time start t = 0.0; time ignition t ig = 60 s; time water mist t wm = 450 s
Modelling the fire For FDS, all you need is the HRR curve! Need to capture other important phenomena Flame spread rate through wood pallets Air flow through the porous array Flame volume, angle, length Collapse of consumed sections changing ventilation
Heat Release Rate suppressed fire 30,000 Test 1 02-02-2006 25,000 20,000 HRR, kw 15,000 10,000 5,000 Water mist on 0 0 5 10 15 20 25 30 35 Time, min utes
Relating appearance to HRR Test 1 HRR smoothed & shifted 30,000 25,000 Water mist on full 20,000 HRR, kw 15,000 10,000 Ignition at 1:00 min 5,000 0 0 5 10 15 20 25 30 35 40 Time, minutes I + 4:20 min I + 1:08 min I + 2:00 min I + 3:00 min
Heat Release Rate not suppressed From T 10: T Representative 10-02-2006 HRR, kw 60,000 55,000 50,000 45,000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 0 0 5 10 15 20 25 30 35 Time, minutes
Modelling the fire 25 23 21.5 Wood pallet stacks Porous fuel package Air movement through support platform
Input Heat Release Rates 30,000 25,000 Test 1 02-02-2006 Heat release rate per unit area Open cells burn as horizontal bed Cells with X not yet contributing HRR, kw 20,000 15,000 10,000 5,000 0 0 5 10 15 20 25 30 35 Time, minutes
Half-height, top cell method
FDS4 computed HRR based on model Oxygen depletion algorithm turned off
FDS Simulation of HRR - unsuppressed
Flames.wmv: Suppressed & Unsuppressed Suppressed Unsuppressed Even suppressed, there is a region where flames touch ceiling
Temperature Control Test Data Temperature profile along tunnel 1000 900 800 02-02-2006 Test 1 TC Profile @ Ceiling 2/2/ 2006 10:55 2/2/ 2006 11:02 2/2/ 2006 11:07 11:21:03 AM Temperature, Deg C 700 600 500 400 C22 C21 C20 C19 C18 C1 7 C1 6 C1 5 C1 4 C 13 C 12 C 11 C 10 C 09 C08 C07 C06 C05 C04 C0 3 C0 2 C01 300 200 100 0 345 360 375 390 405 420 435 450 Tunnel Positi on (m)
Temperature Profiles at 420 Seconds (pre-mist) 1000 900 Ceiling Temperature, deg C 800 700 600 500 400 300 200 100 0 350 360 370 380 390 400 410 420 430 440 450 Position along Tunnel, m Model Test 1
Temperature Profiles at 720 Seconds 1000 900 Ceiling Temperature, deg C 800 700 600 500 400 300 200 100 0 350 360 370 380 390 400 410 420 430 440 450 Position along Tunnel, m Model Test 1
Temperature Profiles at 1020 Seconds 1000 900 Ceiling Temperature, deg C 800 700 600 500 400 300 200 100 0 350 360 370 380 390 400 410 420 430 440 450 Position along Tunnel, m Model Test 1
Modelling an unsuppressed fire Simulations agreed with test data Utilized model to simulate an un-suppressed fire (not tested) Examine the contrast between with and without an active suppression system
Animations Flames.wmv ISO.wmv Flux.avi
Temperature iso-surfaces near ceiling Back layering both cases Before mist Envelopes: >100 o C, > 350 o C, and > 500 o C at 7 min 22 s Unsuppressed
Temperature envelopes near ceiling Back layering gone 1 min after mist Envelopes: >100 o C, > 350 o C, and > 500 o C at 8 min 22 s Unsuppressed
Heat flux on floor of tunnel before mist Suppressed Heat flux > 5 kw/m 2 : Time = 7 min:22s Unsuppressed
Heat flux on floor of tunnel: 1 min after mist Suppressed Heat flux > 5 kw/m 2 : Time = 8 min:22s Unsuppressed
Heat flux on floor of tunnel, 16 minutes + Suppressed Heat flux > 5 kw/m 2 : Time = 16 min:05s Unsuppressed Without suppression, propagation in both directions assured
Global performance benefit of water mist Stop back layering Reduced area of direct flame impingement on ceiling Life-safety and Property Protection benefits evident
Conclusions Very large fires in tunnels are complex, chaotic, highly variable events Single-point measurements do not reveal the global benefits of the suppression system CFD modelling, validated by test data, reveals global performance benefits beyond what can be measured
Acknowledgements Marioff Corporation Oy