You are on page 1of 44

A Comparison of FPETool Predictions to Experimental

Results: Comparison of Clean Agent and Sprinkler System


Performance on In-Cabinet Fires:
Mark L. Robin and Eric F. Forssell
Hughes Associates, Inc.
and
Steven T. Ginn
Great Lakes Chemical Corporation
Presented at:
2003 Halon Options Technical Working Conference
Sheraton Old Town
Albuquerque, New Mexico
May 13-15, 2003

Introduction

Telco and EDP Facility Fire Protection

Clean agents
Automatic sprinklers
Clean agent + Automatic sprinkler
Best Protection?

Objective

Compare performance of clean agent and automatic


sprinkler systems
Employ FPETool for design of test scenario

Automatic Sprinkler Systems:

Design objective is fire control

Contain fire to room of origin


Control ceiling temperature to avoid structural damage
Fire extinguishment NOT primary objective

Fire size at system activation relatively high

Sprinkler systems employ thermal response


Water not released until temperature at fusible link or
glass bulb in excess of 135 oF

Automatic Sprinkler Systems:


Design Objectives
- In general terms of property protection,
sprinkler systems are typically designed to
acheive fire control...
- Fire control can be described as limiting the fire
size by decreasing the rate of heat release and
pre-wetting adjacent combustibles, while
maintaining ceiling gas temperatures so as to
avoid structural damage
NFPA Fire Protection Handbook, 19th Edition, p. 10-193.

Fire Control
Heat
Release
Rate

sprinkler activation

Time

Automatic Sprinkler Systems: Standard


Preaction System

System piping charged with air under pressure


System equipped with a supplemental detection
system
Water held back by a preaction valve

Operation of the supplemental detection system allows the


preaction valve to open, admitting water into the pipe
network (sprinkler still closed)

Water discharged from pipe network when fire has


generated sufficient heat to activate one or more
sprinklers

Preaction System
Operation of detector trips valve
Water delivered when head T=T rating

Control
panel

Valve

Detector
Sprinkler head

Water supply

Clean Agent Suppression Systems:


Primary Advantages

Ability to extinguish shielded, obstructed or threedimensional fires in complex geometries

Ability, through the use of detection, to extinguish fires at a


very early stage

Clean agents are gases


Uniform distribution throughout an enclosure

Extinguishment well before direct or indirect fire/smoke damage


occurs

Cause no collateral damage due to agent discharge

Are clean
No residues

Clean Agent Systems:


Design Objective = Fire Extinguishment
Fire Control vs.
Fire Extinguishment
Heat
Release
Rate

Fire Control (Sprinklers)

Fire extinguishment
(clean agent)

Time

Clean Agent System vs. Sprinkler System


(Fire Extinguishment vs. Fire Control)
sprinkler activation

FIRE CONTROL

Heat
Release
Rate

clean agent
system activation
detection
FIRE EXTINGUISHMENT

Time

Comparison of Sprinkler and


Clean Agent Systems
Suppression agent

Sprinkler Systems

Clean Agent System

Water

Gas

Design Objective

Fire Control:
Confine fire
Control ceiling T
Protect structure

Activation

Sprinkler head T 135 oF

Fire size at activation

Can be 100's of kW
No. Water not three dimensional,
will not fill entire enclosure
No. Water damage, smoke
damage

Design Objective

Total Flooding
Cleanliness

Fire Extinguishment
Protect contents of structure
Automatic activation following
detection (air sampling, smoke
detectors)
Low as 0.1 kW with air
aspirating detection system
Yes. Agent distributed
uniformly throughout enclosure
Yes. No residues to clean up
following extinguishment

Comparison Testing of Preaction


Sprinkler and FM-200 Systems

FM-200 System

Preaction Sprinkler System

Designed and installed in accordance with NFPA 2001

Designed and installed in accordance with NFPA 13

Detection/Alarm Systems

Designed and installed in accordance with NFPA 72

Comparison Testing of Preaction


Sprinkler and FM-200 Systems

Facility

32.8 x 32.8 x 12 foot enclosure


1.5 ft deep subfloor; suspended ceiling 2 ft
below drywall ceiling
1/2 gypsum board over metal stud construction
Enclosure access via two doors
Eight polycarbonate windows
Motorized damper system in ceiling for post-test
exhaust

Comparison Testing of Preaction


Sprinkler and FM-200 Systems

FM-200 Tests

Sprinkler Test

Subfloor covered in plastic to prevent leakage of agent into


subfloor
Plastic formed into dike to contain water and facilitate
removal of water after testing

All Tests

Area 20 x 26 ft covered with floor tiles


Cabinet containing fire array and various articles of
furniture
arranged on floor tiles

20 ft
North
32 ft 10 in

26 ft
32 ft 10 in
Four Drawer
File Cabinet

IBM Cabinet

Cabinet Containing Fuel Array

Smaller Cabinet
Tables

Exhaust Vent (in Ceiling)


Figure 1 Chamber Schematic.

Window

Comparison Testing of Preaction


Sprinkler and FM-200 Systems

Test Fire

Eight sheets of 8 x 16 x 0.375 inch ABS


Vertically arranged in two rows of four sheets
Mounted on all-thread rods in unistrut stand
Fire array placed in a equipment cabinet
Ignition: 3 mL heptane in 2 inch square pan
Similar to plastic sheet fire test of UL 2166, UL
2127

Fuel Array

Eight Sheets of
8 x 16 x 3/8 in
ABS

76
16

16

1
7
21

38
2 in pan with
3 ml n-Heptane

Heat Release Rate for In-Cabinet Fire


500
450

Heat Release, kW

400
350
300
250
200
150
100
50
0
0

100

200

300

Tim e, seconds

400

500

Comparison of Preaction Sprinkler


and FM-200 Systems
Smoke Detection Systems (FM-200 System)

Fenwal AnaLASER II

Air-sampling system, designed and installed by


distributor
1 inch PVC tubing main, 0.75 inch PVC branch lines
Flow rate through system 0.0624 ft3/s
0.061% obscuration per foot alarm threshold
Nine sampling holes, 11 ft spacings

Exceeds NFPA 72 requirements


Based on FM Loss Prevention Data Sheet 5-32 on
Electronic Data Processing Systems

Comparison of Preaction Sprinkler


and FM-200 Systems

Smoke Detection Systems (FM-200 and


Sprinkler)

Simplex 4098 series True Alarm detectors


Ionization and photoelectric

Ionization: 1.3% obscuration/ft


Photoelectric: 2.5% obscuration/ft

Located at three air sampling points nearest fire


location

FM-200 Suppression System


Designed in accordance with NFPA 2001

7% by volume FM-200

Discharge time: 9.5 seconds

Hygood Ltd cylinder

Hygood Ltd 8-port aluminum nozzle


orifice area 1.57 in 2

System design via Hygood Ltds design software

30 s delay employed from detection to system activation


Maximum delay time allowed under recommendations of FM Global
Property Loss Prevention Sheet 5-14 on Telecommunication Facilities

Preaction Sprinkler System


Designed in accordance with NFPA 13

Design and installation based on Ordinary Hazard Class I


Nine sprinkler heads in main space
Nine sprinkler heads above suspended ceiling
11 ft spacing for area coverage of 121 ft2

Recessed pendant standard response glass bulb sprinklers

Maximum spacing allowed under NFPA 13 is 15 ft


Temperature rating 155 oF

Application density of 0.15 gpm/ft2 required

Preaction Sprinkler System


Water supply

NFPA 13 requires water supply to be adequate to supply all


sprinklers within the design area for a minimum of 60 minutes
18.2 gpm/head x 18 heads x 60 min = 19,600 gallons water

Test facility unable to handle this quantity of water

Water supply designed to supply the two sprinklers nearest the fire
location for a period of 30 minutes at the required design flow rate
of 18.2 gpm (application density 0.15 gpm/ft2)

Enclosure Instrumentation

Four thermocouple trees (Type K)


Optical density
O2, CO, CO2

FTIR: FM-200 and HF concentrations


Enclosure pressure
Nozzle pressure
IR camera

Test Procedure

Inp
Inp ut
ut for
forFF PE
PETo
To ol
ol SSim
im ula
ula tion
tion ss

Input Parameter
Minimum Oxygen Level
21 oC
600 oC
Heat Transfer Factor
Radiant fraction
Max energy loss internal
Sprinkler, NE location
Distance from center of fire (ft)
RTI (ft s)1/2
Activation temperature, oF
Sprinkler, N location
Distance from center of fire (ft)
RTI (ft s)1/2
Activation temperature, oF

10.0
2.0
0.35
0.90
10.6
144
155
11.3
144
155

In
In put
put for
for FPE
FPETo
To ol
ol Sim
Sim ula
ula tion
tion ss

Input Parameter
Room dimensions
Ceiling height (ft)
Length (ft)
Width (ft)
Ceiling Material
Thickness (in)
Thermal conductivity (kW/mK)
Density (kg/m3)
Specific heat (KJ/kg K)
Ceiling Material
Thickness (in)
Thermal conductivity (kW/mK)
Density (kg/m3)
Specific heat (KJ/kg K)

8.0
32.8
32.8
Glass fiber
0.5
0.00037
60
0.8
Gypsum board
0.5
0.00017
960
1.1

Heat Release Rate Required for


Sprinkler Activation

Input

Experimental HRR
Location of fire
Enclosure dimensions
Location of sprinkler heads
Properties of sprinkler heads

Predict activation of NE sprinkler head at 249 seconds

Ceiling jet temperature 249 oF


Fire size approximately 210 kW

Smoke
Detector
Activation
Time
Input

Experimental HRR
Location of fire
Enclosure dimensions
Location of smoke detector
FPETool: default activation temperature of 23 oF above initial T

Predict activation of NE smoke detector at 95 seconds

Ceiling jet temperature 93 oF


Fire size approximately 15 kW

Results: FM-200 System

AnaLASER II alarmed at 78 seconds from ignition


FM-200 system activated at 108 seconds from ignition
Fire extinguished at 125 seconds from ignition

Fire damage

7 seconds from end of system discharge


Maximum ceiling T of 85 oF observed
slight scorching of inside of test cabinet

Non-fire damage

several ceiling tiles displaced


ceiling runner slightly bent

500
450

Heat Release, kW

400
350
300

system activation

250
200

detection

150

extinguishment

100
50
0
0

100

200

300

400

Tim e, seconds

Extinguishment of In-Cabinet Fire by FM-200

500

Results: Preaction Sprinkler System

Photoelectric detector in NE corner in full alarm


at 94 seconds from ignition
Ionization detector in NE corner in full alarm at
112 seconds from ignition (FPETool: 95 s)
Complete obscuration due to smoke at
approximately 240 seconds from ignition

Results: Preaction Sprinkler System

Sprinkler head in NE corner actuated at 273


seconds from ignition (FPETool: 249 s)
Sprinkler head in N corner actuated at 347
seconds from ignition
Fire not extinguished by sprinkler system

IR camera shows fire burning through entirety of test


Fire contained to source cabinet

Max ceiling temperature of 560 oF observed at


thermocouple tree nearest fire

500
450

Heat Release, kW

400
350

system activation

300
250
200

detection

150
100
50
0
0

100

200

300

400

500

Tim e, seconds

Control of In-Cabinet Fire with Preaction System

Table 3. Sprinkler Activation Times

Sprinkler
NE
N

Radial
distance
to fire,
feet

Measured
Activation
Time,
seconds

10.6
11.3

278
352

Predicted
Measured
Activation Ceiling Jet T
Time,
at bulb at
seconds
activation
o
F
249
288
253
276

Predicted
Ceiling Jet T
at bulb at
activation,
o
F
249
250

Results: Preaction Sprinkler System

Fire Damage

Test cabinet suffered extensive scorching

Non-Fire Damage

Black ring around entire enclosure


Ceiling tiles discolored
Soot particles scrubbed from smoke layer cover floor,
horizontal surfaces
Walls discolored from smoke damage
Water damage to paper goods

Conclusions: System Performance

Preaction Sprinkler System

Design objective attained: Fire was controlled


System performed exactly as expected
Fire contained to room of origin
Ceiling temperatures managed such that structural
damage and/or collapse did not occur
Structure saved

FM-200 System

Design objective attained: Fire extinguished


System performed exactly as expected
Contents of structure saved

Conclusions

Clean agent and preaction sprinkler systems vastly different

Fundamental design objective different: Control vs Extinguishment


Preaction systems best suited to protection of structure
Clean agent systems best suited to protection of contents of structure
Sprinkler systems alone inappropriate for protection of high value assets
Clean agents not ideally suited for structural protection
FPETool useful for predicting performance of sprinkler system and
estimation of ceiling jet temperatures, smoke detector activation and for
test fire selection

Conclusions

Applications involving expensive, sensitive assets

Use of clean agents justified


Clean agents offer unparalled peformance for very early extinguishment
of fires

Maximum fire protection provided by use of a clean agent


system in combination with a sprinkler system

Substantial additional risk reduction at high benefit/cost ratios can be


realized by protecting such assets with both a clean agent system and a
sprinkler system

Acknowledgements

Great Lakes Chemical Corporation


- Steve Ginn
Hughes Associates, Inc.
- Eric Forssell
- Jason Ouellette
- Matt Harrison
- Ralph Ouellette
- Art Hammett
- Jason Kennedy

You might also like