1
2	Outline Background of R & D in Water Mist Research: issues and editorials spray characteristics extinguishing behavior modeling fire testing
3	R & D - Water Mist Fire Protection Systems 1987 to 2001 urgent need for new technology economics right fire science tools available Outcome advances in understanding new technology for fire suppression systems available new freedom to push fire suppression systems design to the limits (“designer sprays”)
4
5
6
7
8
9	Research - Comments Spray characteristics Extinguishment mysteries revealed Computer modeling Fire Testing
10	Spray Characteristics Drop size distribution Mass flow rate Velocity Directionality
11	Drop size distribution “Water Mist” or “Fine Water Spray” aerosol ~ 1 to 10 micron diameter fog - mist ~ 10 to 100 micron diameter fine spray ~ 10 to 1000 micron diameter sprinkler spray ~ 50 to 2000 micron diameter The term “water mist” was chosen by a committee “Water Mist”, defined by particle size, has 99 % of its volume in drops smaller than 1000 microns (1 mm) in diameter Prefer “cumulative percent volume” presentation over single point values (eg. Sauter MD) or numerical counts except computer modeling of sprays … ?
12	Drop size distribution
13	Relationships between drop size, nozzle discharge rate, design spacing, and power required for four commercial water mist nozzles.
14	Velocity & Directionality
15	Velocity & Directionality
16
17
18
19	Extinguishing Mechanisms Oxygen depletion Flame cooling Radiation attenuation Fuel wetting (class A combustibles) Flammable vapor dilution
20
21	Implications Above ~ 60 ^oC begin to see total-flooding effects explains why large fires easier to extinguish than small fires explains why cycled discharge accelerates extinguishment explains why WM successful in machinery room Class B fires explains why higher mist application rate may not equate to faster extinguishment Water mist release on early fire detection is not necessarily a good design option room is not hot enough to extinguish by total flooding contradicts commonly held view
22	The “unextinguishable fire” …
23	Modeling Why model ? Confirm basic physics of suppression and extinguishment Predict performance, optimize design parameters Reduce the need for full-scale fire testing Computer Models Zone models Numerical simulations (CFD) Physical Scaling Laws
24	Zone Models Back, G. G. (HAI); Wighus, R. (SINTEF), Vaari, J. (VTT) Single zone, well-mixed, ventilation, quasi-steady state percentage of spray evaporates & displaces oxygen; air is saturated; evaporation extracts heat energy; fire also consumes oxygen; accounts for vent size Back predicts time to extinguishment, compartment temperatures as functions of compartment size, ventilation opening, and fire size Back correlates with test data from Coast Guard machinery room tests
25	Field Models (Numerical Simulations) Prasad, Patnaik, and Kailasanath. “Advanced Simulation Tool for Improved Damage Assessment 2) Water-Mist Suppression of Large Scale Compartment Fires.” NRL/MR/6410--00-8507, Naval Research Laboratory, Washington, DC, 2000. Tieszen, and Lopez (Sandia). “Issues in Numerical Simulation of Fire Suppression.” Halon Options Technical Working Conference, Albuquerque, NM, 1999. Hadjisophocleous, Cao, and Kim (NRCC). “Modeling the Interaction Between Fine Watersprays and a Fire Plume.” Fourth International Conference on Advanced Computational Methods in Heat Transfer, Udine, Italy, 1996. McGrath, et. al. … (NIST) Large-eddy simulation model ….
26	Numerical Modeling Micro-scale small diffusion flames by mono-disperse sprays (Prasad) opposed-jet turbulent pre-mixed flames (Mesli; Abbud-Madrid) this work advances numerical modeling and modeling of extinction mechanisms Issues over-simplification of spray characteristics difficult to close the gap between micro-scale model environments and complexity of full-scale fire
27	Numerical Modeling Macro-scale (full-scale compartments) NIST Large eddy model - may be adapted to water mist commercial CFD platforms already useful for after-the-fact analysis Issues difficulty in realistic modeling of sprays injection density = 0.20 gm/cm³ (real = 2 (10^-6) gm/cm³) injection velocity 1 m/s (real = 20 m/s) 4-nozzle mist flow rate: 200 kg/s (= 12,000 L/min !) (realistic 4 nozzle flow rate = 80 L/min) need experimental data for validation time intensive and costly
28	Physical Scaling Laws VTEC project - University of Maryland apply theoretical scaling laws (Heskestad’s) scale fire size, heat release rate, flame height, plume velocity, spray characteristics based on a standardized test scenario
29	Modeling Current modeling efforts contribute to analytical advances in numerical simulation single zone model produces useful macro-scale results CFD work is difficult to validate may not reflect observed differences in performance between manufacturers Potential still the best option to reduce reliance on full-scale testing suggest CFD model of a standardized test protocol
30	Fire Testing “Design” of water mist systems NFPA 750 - Design Objectives and Fire Test Protocols design must be based on full-scale fire testing based on a protocol for each specific application
31	Fire Testing Fire Test Protocols analyze hazard & generalize protection objectives build mock-up: geometry, scale, ventilation select representative fuel package, scenarios set performance objectives (pass/fail criteria) tests conducted by recognized laboratory (credibility) confirm design criteria, write report sell systems for applications similar to protocol Consensus type (IMO MSC/Circ.668; FMRC) Ad-hoc type (custom)
32	Consensus Test Protocols Class B Liquid Fuels Gas turbine enclosures Special hazard spaces Local application in machinery spaces Class A Combustibles Crews quarters, cabins and corridors Light hazard public spaces Public spaces, shops, storage areas Ordinary Hazard Groups I and II Limited Listings, limited ceiling heights
33	Consensus Test Protocols Semi-conductor Wet Benches Enclosed compartments, liquid fuels Aircraft - passenger or cargo compartments Telecom switch-gear Computer / control rooms Libraries - fixed shelving Archives - mobile shelving Heritage buildings
34	Can fire testing be eliminated? Five commercial systems - all passed machinery space protocols Five independent solutions to fire test protocol flux densities varied 1.2 to 3.0 L/min/m² volume concentrations 0.4 to 1.0 L/min/m³ Design parameters not consistent between manufacturers criteria unique to each system performance depends on specific nozzles imposing one set of design criteria on all systems would not necessarily be conservative
35	IMO Machinery Space Protocol #. Fire Scenario (must extinguish all fires in 15 min or less) Manuf 1 Manuf 2 (min) (min) 1 LP horizontal diesel spray… 2.2 12.6 2 LP diesel spray on top of simulated engine centered 2.5 9.3 3 LP concealed horizontal diesel spray fire… 3.0 14.9 4 Combination worst spray fire from tests 1-3 and 2.6 10.0 5 HP horizontal diesel spray fire on top of the simulated engine 3.7 10.7 6 LP low flow concealed horizontal diesel spray fire 5.0 11.1 7 0.5 m² heptane pool central under mock-up 12.4 13.8 8 0.5 m² 10W30 lube oil pool central under mock-up 8.6 13.5 9 0.1 m² heptane pool on top of bilge plate 16.5 (fail) 15.0 (pass) 10 Flowing heptane fire 0.25 kg/s from top of mock-up 3.5 12.8 11 Class A fires wood crib in 2 m² heptane pool fire 13.9 14.3 12 A steel plate offset 20º to a heptane spray heated to 350ºC 5.0 12.6 13 4 m2 diesel pool tray under mock-up 8.1 12.3
36	Fire Testing Problems with test-based design too few consensus protocols (serve largest markets only)
37	Fire Testing Problems with test-based design too few consensus protocols (serve largest markets only) too many consensus protocols (duplication)
38	Fire Testing Problems with test-based design too few consensus protocols (serve largest markets only) too many consensus protocols (duplication) design criteria cannot be generalized, therefore we will never escape need for testing
39	Fire Testing Problems with test-based design too few consensus protocols (serve largest markets only) too many consensus protocols (duplication) design criteria cannot be generalized, therefore we will never escape need for testing test protocol is a representation of an idealized scenario (but criteria will be applied to anything vaguely similar)
40	Fire Testing Problems with test-based design too few consensus protocols (serve largest markets only) too many consensus protocols (duplication) design criteria cannot be generalized, therefore we will never escape need for testing test protocol is a representation of an idealized scenario (but criteria will be applied to anything vaguely similar) passing the test becomes the target safety margins not consistent (or known)
41	R & D - Water Mist Fire Protection Overlapping Domains Different objectives Complexity of the problem
42	After a decade of R & D in Water Mist ...
43