ck height=1 colspan=2>
DuPont Product/Presentation Title March 23, 2007
Suppression of Class C Fires
DEVELOPMENT OF A STANDARD PROCEDURE FOR THE
EVALUATION OF THE PERFORMANCE OF CLEAN AGENTS
Mark L. Robin, Ph.D., DuPont Fluoroproducts
Bon Shaw and Brad Stilwell, Fike Corporation
Presented at: Suppression & Detection Research & Applications
(SUPDET 2007), March 6, 2007
Wyndham Orlando Resort, Orlando, Florida 3/23/2007
2
Introduction: Class C Fire What is it?
NFPA 10 3.3.4.3 Class C Fires.
Fires that involve energized electrical equipment.
NFPA 2001 3.3.5 Class C Fires.
Fires that involve energized electrical equipment
where the electrical nonconductivity of the extinguishing media is of importance. 3/23/2007
3
Class C Fire Why do we care?
In some applications, power disconnection is highly
undesirable because:
Time = Money
Downtime impact for a typical computing infrastructure
is estimated to be $42,000 per hour 3/23/2007
4
Downtime Impact per Minute
Business
Application
Downtime Cost
per Minute
Supply Chain Management
$11,000
e Commerce
$10,000
Customer Service Center
$3,700
ATM
$3,500
Financial Management
$1,500
Messaging
$1,000
Infrastructure
$700
Source: Alienan ROI Report, 2004 3/23/2007
5
Class C Fires and Clean Agents NFPA 2001, 2004 edition
5.4.2.5.
Minimum design concentration for Class C hazards
shall be at least that for Class A surface fire. Class A minimum design concentration equals 1.2 times the minimum
extinguishing concentration determined by test as part of a listing program,
where the listing program at a minimum must conform to UL 2127 or UL 2166. 3/23/2007
6
Class C Fires and Clean Agents Report on Comments A2007 ROC Comment 2001-61a (log #CC7) Minimum design concentration for Class C hazard shall be 1.6 times
the Class A minimum extinguishing concentration Committee Action: Accept Substantiation: Laboratory testing indicates that the agent concentration
required to extinguish a fire in energized electrical equipment typically
increases with increased electrical power input. 3/23/2007
7
ROC 2001-61a: Implications For Class C hazards, the minimum design concentrations
for all clean agents increased by 33% As a result, system costs increase by at least 33% Several clean agents basically eliminated from the market 3/23/2007
8
Reality Check:
Is Such Drastic Action Warranted?
Q:
Number of documented failures of any clean agent system
to extinguish fires involving energized electrical equipment
since the initial installation of clean agent systems
approximately 15 years ago?
A:
Zero 3/23/2007
9
Reality Check:
Does laboratory testing warrant such a drastic change? Thirteen reports reviewed by NFPA 2001 Committee Energized Ni foil in an ethylene flame Energized Nichrome wire in a heptane flame Tests involving no electrically energized equipment Numerous reports on PMMA wrapped in Nichrome Conductive heating, ohmic heating, PC board failures (arc)
Source: 3M, Modular Protection, Hughes Assoc., NIST 3/23/2007
10
PMMA, Nichrome and Class C Hazards
Typical Class C Hazards Telecommunications & EDP facilities Cables, electronic equipment
Power Conduction Almost always copper Small percentage aluminum Not Nichrome
Power Cable Insulation Majority PVC, FR-PVC, PE, HDPE Some Hypalon, XLPO Not PMMA 3/23/2007
11
PMMA, Nichrome and Class C Hazards
Nichrome wire Alloy of Ni and Cr High electrical resistancewidely employed for resistive heating Never employed for power or data conduction Extremely corrosion resistance High mechanical strength at temperatures up to 2500
o
F 3/23/2007
12
Nichrome vs Copper
Nichrome
Copper
Ni/Cr alloy
Cu
High mechanical strength
Low mechanical strength
Highly resistant to corrosion
Low resistance to corrosion
High electrical resistance
Low electrical resistance
Use: resistive heating
Use: power and data conduction
Max use T = 2000
o
F
Max use T = 1000
o
F 3/23/2007
13
Nichrome vs Copper
70000
60000
50000
40000
30000
20000
10000
0 0
500
1000
1500
2000
2500
Nichrome 60/15 wire
Hot Drawn Copper wire
Strength versus Temperature
Tensile strength (psi)
Temperature (
o
F
) 3/23/2007
14
Overloaded Copper Wire
24 AWG Bare Copper Wire
Current (A)
Temperature (F)
Duration
21
700
> 10 min
> 10 min
> 10 min
8 min
3:23
5:13
6:02
23
800825
25
925950
26
1000
27
1050 3/23/2007
15
Overloaded Copper Wire
24 AWG Jacketed Copper Wire
Current (A)
Temperature (F)
Duration
20.5
700
> 10 min
0:24
0:28
0:10
21.5
725
23.5
850
27
1050 3/23/2007
16
Overloaded Bare Copper Wire
AWG
Current (A)
Temperature (F)
Duration
1000
8:00
3:18
7:07
1085
800
24
26
22
37
18
60 3/23/2007
17
Copper Wire Bare copper wire can withstand a 10-minute overcurrent
only when the wire temperature is limited to 1000
o
F Insulated copper wire can withstand a 10-minute overcurrent
only when the wire temperature is limited to 700
o
F Larger gauge wires require more current but behave similarly
to smaller gauge wires at similar temperatures Stranded cables and single conductor cables behave similarly 3/23/2007
18
Copper Wire Copper wire heated to 7501000
o
F is sustainable for
10 minutes ONLY if these temperatures are not exceeded
anywhere along the length of wire Once copper wire is heated to above 700
o
F, corrosion
is accelerated, and this corrosion is the primary reason
for failure at these temperatures 3/23/2007
19
Nichrome, PMMA and Class C Fires
PMMA/Nichrome studies Conducted at wire temperatures 1800
o
F At 1800
o
F, bare copper wire sustainable for < 10 seconds At 1800
o
F, insulated copper wire sustainable for even less time
These tests would be impossible to conduct
with the conductor employed in 99.9% of
all power transmission cables 3/23/2007
20
PMMA
Polymethylmethacrylate Lucite, Plexiglas Clear, shatterproof, excellent optical properties Hockey rink barriers Lenses, optical instruments, stop lights, auto lights Expensive Not employed for power conduction Combusts quantitatively to monomer 3/23/2007
21
PMMA/Nichrome Wire Tests
Not representative of real world hazard
Materials NiCr not employed for power conduction PMMA not employed for insulation
Conditions Cu wire able to withstand 1800
o
F for < 10 s before failing 3/23/2007
22
Conductive Heating Tests
Hughes Associates Report Simulate overloaded connection 350 MCM Cu cable; PVC, Hypalon insulation Cable heated to 590
o
F (310
o
C); 1000 W ring heater Pilot flame ignition, 60 s preburn, system actuation 5 min soak with current applied
McKenna, et al., 1998 HOTWC 3/23/2007
23
Printed Wire Board Failure (Arc) Tests
Hughes Associates Report Simulate PC board failure Test boards of FR-2 and FR-4 Parallel copper tracks Arc develops between tracks, travels along board After arc and flame travel 130 mm, actuate system
McKenna, et al., 1998 HOTWC 3/23/2007
24
Ohmic Heating Tests
Hughes Associates Report Simulate overheated wire Wire mounted between Cu buses connected to a 600 A arc welder Butane flame pilot Bundled cable: Cables bundled around a central energized wire Preburn until flame judged to be well sustained
McKenna, et al., 1998 HOTWC 3/23/2007
25
Hughes Associates Report Testing on representative materials Conductive heating tests Printed wire board failure (arc) tests Ohmic heating tests Concluded that Fires initiated by and involving
energized electrical circuits can be controlled
by HFC-227ea at concentrations below 7%
McKenna, et al., 1998 HOTWC 3/23/2007
26
Modified Conductive Heating Tests Introduces continuous arc into conductive heating tests Poor Reproducibility Small flame at top of cable sometimes ignites vapors
rising up sides of cable Flame sometimes disappears when enclosure sealed Difficult to determine point of extinguishment due in part
to obscuration during discharge 3/23/2007
27
Modified Conductive Heating Tests
HFC-227ea Poor Reproducibility: KS-5482L28F cable 7.0% extinguished flame/reignition 8.0% no extinguishment/reignition 8.0% extinguished flame/no reignition 11.0% extinguished flame/reignition 3/23/2007
28
Development of a Standard Class C Test Objectives Replicate real world scenario with
materials choice and conditions Conservative Reproducible Simple to perform 3/23/2007
29
Class C Test Protocol Replicate real word materials Conductor: Copper Insulation: PVC, PE, Hypalon, XLPO Replicate real world conditions But conservative (challenging) test 3/23/2007
30
Class C Test Protocol Conditions Copper only able to withstand temperatures
up to 1000
o
F for extended periods Employ NiCr wire as ignition source and as
heat source during preburn and soak periods Ignition: NiCr @ 1800
o
F Preburn and soak: NiCr @ 1200
o
F