Assessment of Hazards of Flammable and
Combustible Liquids in Composite IBC’s in
Operations Scenarios
Final Report
Prepared by:
Hughes Associates, Inc.
The Fire Protection Research Foundation
One Batterymarch Park
Quincy, MA, USA 02169-7471
Email: [email protected]
http://www.nfpa.org/foundation
© Copyright Fire Protection Research Foundation
October 2011
FOREWORD
NFPA 30 Flammable and Combustible Liquids Code provides specific guidance for both
containment and fire protection of Listed IBCs containing flammable and combustible liquids in
storage configurations. However, a common usage scenario involves the use of non-Listed,
composite IBCs containing flammable/combustible liquids in operations scenarios. The code
does not provide specific fire protection criteria for these applications. The only guidance
provided, to date, relates to the quantities of liquids that are permitted in such scenarios.
Historically, the protection schemes prescribed in NFPA 30 have been based on experimental
testing and engineering analysis. The goal of this Phase I project as described in this report is to
identify candidate protection strategies that can be evaluated in a future Phase in general
accordance with Appendix E of NFPA 30.
The content, opinions and conclusions contained in this report are solely those of the authors.
Assessment of Hazards of Flammable and Combustible Liquids
In Composite IBC’s in Operations Scenarios
Project Technical Panel
Christina Francis, Procter & Gamble
Phil Hooker, Health and Safety Laboratory
Brian Minnich, Schuetz Containers Systems
Louis Nash, U.S. Coast Guard
Jon Nisja, Minnesota State Fire Marshal Division
Keith Olson, Ansul/Tyco Fire Products
Tony Ordile, Haines Fire & Risk Consulting
Bob Benedetti, NFPA staff liaison
Property Insurance Research Group Sponsors
CNA Insurance
FM Global
Liberty Mutual
Tokio Marine Management, Inc.
Torus Insurance
Travelers Insurance
XL Group
Zurich NA
Project Contractor
Joseph Scheffey, Hughes Associates
HAI Project # 1JLS00021.000
ASSESSMENT OF HAZARDS OF FLAMMABLE AND COMBUSTIBLE LIQUIDS
IN COMPOSITE IBCs IN OPERATIONS SCENARIOS
Prepared for
The Fire Protection Research Foundation
Quincy, MA
Prepared by
Hughes Associates, Inc.
3610 Commerce Drive, Suite 817
Baltimore, MD 21227
Ph. (410) 737-8677 FAX (410) 737-8688
www.haifire.com
October 13, 2011
EXECUTIVE SUMMARY
NFPA 30 Flammable and Combustible Liquids Code [1] provides specific guidance for both
containment and fire protection of Listed IBCs containing flammable and combustible liquids in
storage configurations. However, a common usage scenario involves the use of non-Listed,
composite IBCs containing flammable/combustible liquids in operations scenarios. The code
does not provide specific fire protection criteria for these applications. The only guidance
provided, to date, relates to the quantities of liquids that are permitted in such scenarios.
Historically, the protection schemes prescribed in NFPA 30 have been based on experimental
testing and engineering analysis. In keeping with this approach, the goal of this work is to
identify protection strategies that can be evaluated in general accordance with Appendix E of
NFPA 30. If adequate performance is identified, it might be implemented into the code.
This assessment was limited to the hazards associated with the use of both flammable and
combustible liquids (i.e., Class IB–IIIB) in non-Listed/Approved IBCs (i.e., Type 31HA1\Y).
The IBCs were considered as being out in the open (i.e., no enclosure) under 3.1–9.1 m
(10–30 ft) ceilings. The liquids considered included heptane, isopropyl alcohol, and mineral seal
oil. Using these variables, a fire hazard analysis was performed to identify the range of fire
scenarios and associated hazards that could occur. In this analysis, maximum fire size, ceiling
temperatures, sprinkler activation times, and radiant heat fluxes were calculated for five IBC
configurations and ten fire scenarios.
Predicted fire sizes from 0.7–160 MW were calculated for the confined scenarios and from
2.7–6600 MW were calculated for unconfined scenarios. Maximum exposure durations to
prevent structural collapse of the overhead ceiling structure were calculated. Minimum
separation distances for both ignition of adjacent combustibles and human pain were also
calculated. Depending on the fuel and exposure duration, these distances ranged from unlimited
for relatively small, contained isopropyl alcohol fires to a minimum of 14 m (45 ft) for a
contained, unprotected heptane release. Different levels of protection were identified for the
various IBC configurations based on predicted thermal exposures to the structure and adjacent
combustibles. Containment was determined to be essential for all scenarios considered. The
level of protection required was determined to be dependent on several different parameters
including: the fuel being stored, the geometry of the space in which the IBC is being stored, the
expected fire sizes, and the available separation distances between the IBC and adjacent
combustibles.
A range of viable strategies to contain, detect, and suppress these events was explored. The
emphasis was on existing building scenarios, not new construction. A total of three containment,
six detection, and seven suppression systems were considered, as shown in Table E.1. The
advantages and disadvantages with respect to system cost, performance, and impact on
logistics/ITM were considered. A summary of these findings is provided in Table E.2.
ii
HUGHES ASSOCIATES, INC.
Table E.1 –– Summar y of Mitigation/Pr otection Str ategies Consider ed
Containment Systems
Detection Systems
Raised Sump
Overhead Thermal Link
Fixed Berm/Raised Sill Local Thermal Link
HazMat Storage Locker Local Linear Heat Detection
Optical Flame Detection
Video Fire Detection
Liquid Detection
Suppression Systems
Overhead Sprinklers
Overhead Foam-Water Sprinklers
Pre-Engineered Dry Chemical
Low-Level AFFF Foam
Compressed Air Foam (CAFS)
Local Application Water Mist
Aerosol Generators
The primary factor influencing the severity of the fire hazard was the required area of
containment. Two approaches to containment were explored: the use of large containment areas
to capture ejected liquids; or the use of enclosed, raised sumps to reduce the required
containment area while still preventing liquid release outside the containment area.
In configurations with enclosed, raised sump containment of one or two side-by-side IBCs or
IBCs containing alcohol based fuels (e.g., IPA), no additional protection other than fire resistive
containment and adequate separation may be needed. This assumes that the facility, in
accordance with insurer/fire code requirements, will at a minimum have an overhead water
sprinkler system designed for an ordinary hazard scenario. This also assumes that the
performance of the fire resistive containment system being used has been verified.
For other scenarios where the fire hazard is large enough to threaten the steel overhead or
where adequate separation distance is not provided, it is recommended that the local suppression
systems described in this assessment be considered in addition to the fire resistive containment.
A variety of systems were considered in this analysis with some systems (i.e., water mist, dry
chemical, CAFS) having proven performance (i.e., FM Approvals) for similar scenarios. Other
systems (i.e., aerosol, low-level AFFF, passive protection) have potential, but have not yet been
Approved or Listed for the application. In any case, the performance of the suppression system
should be verified via full-scale fire testing in a configuration comparable to that expected in an
operations scenario.
With respect to detection, the type of system required was primarily dependent on the
available separation distances. Scenarios where adequate separation distance is not available
require that local, automatic detection equipment be installed to activate the suppression system
quickly. In cases where adequate separation distance is provided but the threat to the overhead is
still present, the use of manual activation could be permitted. However, manual activation
equipment should be installed at appropriate distances (i.e., separation distance greater than that
required to prevent pain to humans).
iii
HUGHES ASSOCIATES, INC.
Table E.2 –– Compar ison of Advantages/Disadvantages of Fir e Suppr ession Systems
iv
Water
Usage
Efficiency1
Suppr ession
of 3D Fire
Inherent
Protection of
Neighbor ing
Combustibles
Protection
against
Re-ignition/
Burn-back
Obstructs
Movement in
Containment
Area
Low-Level
Piping
Req.
Insp.,
Testing,
and
Maint.
Requires
Local
Detection
System
UL Listed
or FM
Approved
System
Requir es
Dev. or
Veri.
Testing3
Overhead
Sprinkler
1.00
No
Yes
No
No
No
Minimal
No
Yes
Veri.
Overhead
AFFF
0.75 or less
No
Yes
Yes
No
No
Moderate
No
Yes
Veri.
0.17
No
No
Yes
Yes, Minimal
Yes
Moderate
Yes
No
Dev.
0.11
No
Yes/No
Yes
No
Yes,
Minimal
Moderate
Yes
Yes
Veri.
No
No
Yes, Minimal
Yes
Moderate
Yes
Yes
Veri.
No
Unlikely
No
No
Very Low
Yes
No
Dev.
Low-level
Application
of AFFF
Compressed
Air Foam
PreEngineered
Dry
Chemical
Aerosol
Generators
N/A
Potentially
yes,
Re-flash
possible
Potentially
yes,
Re-flash
possible
HUGHES ASSOCIATES, INC.
Local
Yes, for
Yes,
0.062
Yes, Minimal
Yes
Moderate
Yes
Yes
Veri.
Application
duration of
No
potentially
Water Mist
discharge
Passive Pool
Essentially
Fire
N/A
No
Yes
N/A
Yes
N/A
N/A
N/A
Veri.
None
Suppression
1 – Water usage efficiencies were normalized with respect to the system with the highest liquid volume discharge (i.e., overhead sprinkler system) over ten
minutes duration. Over this duration, the sprinkler system would have discharged approximately 8600 L (2300 gal.) over a 35 m2 (380 ft2) coverage area.
2 – Water mist water usage efficiency based on 5 minute discharge duration.
3 – Dev. – Development / Veri. – Verification
ACKNOWLEDGEMENTS
Thanks are extended to the project sponsors, including CNA Insurance , FM Global, Liberty
Mutual Property, Liberty Mutual Agency Markets, Tokio Marine, Torus Insurance, Travelers,
XL Gaps, and Zurich NA. Thanks are also extended to the project technical panel, and the staff
of the FPRF, particularly Kathleen Almand.
v
HUGHES ASSOCIATES, INC.
TABLE OF CONTENTS
Page
EXECUTIVE SUMMARY ............................................................................................................ ii
1.0
INTRODUCTION .............................................................................................................. 1
2.0
OBJECTIVES AND ASSUMPTIONS............................................................................... 1
3.0
HAZARD DESCRIPTION ................................................................................................. 3
3.1
3.2
3.3
3.4
4.0
FIRE HAZARD ANALYSIS ........................................................................................... 14
4.1
4.2
4.3
4.4
4.5
5.0
Identification of Fire Hazards ................................................................................. 4
Development of Fire Scenarios ............................................................................... 5
3.2.1 Operations Environment ............................................................................. 5
3.2.2 IBC Configurations ..................................................................................... 5
3.2.3 IBC Content ................................................................................................ 5
3.2.4 Liquid Release Scenarios ............................................................................ 7
Design Fire Scenarios ............................................................................................. 8
Critical Exposures ................................................................................................. 14
Design Fire Sizes .................................................................................................. 15
Structural Integrity Analysis ................................................................................. 16
Overhead Thermal Detection Analysis ................................................................. 20
Fire Spread Analysis ............................................................................................. 21
Fire Hazard Analysis Findings.............................................................................. 26
PROTECTION SYSTEM ANALYSIS ............................................................................ 27
5.1
5.2
5.3
Liquid Containment Systems ................................................................................ 31
5.1.1 Raised Sumps ............................................................................................ 32
5.1.2 Fixed Berm/Raised Sill Systems ............................................................... 33
5.1.3 Hazardous Materials Storage Lockers ...................................................... 34
5.1.4 Summary of Liquid Containment Approaches ......................................... 36
Fire Suppression Systems ..................................................................................... 36
5.2.1 Traditional Suppression System Approaches ........................................... 36
5.2.2 Novel Suppression System Approaches ................................................... 38
5.2.3 Summary of Fire Suppression System Approaches .................................. 50
Fire Detection Systems ......................................................................................... 52
5.3.1 Thermal Detection Systems ...................................................................... 53
5.3.2 Visual Detection (UVIR/VID) .................................................................. 55
5.3.3 Liquid Detection ....................................................................................... 57
5.3.4 Summary of Detection System Approaches ............................................. 58
6.0
CONCLUSIONS............................................................................................................... 59
7.0
RECOMMENDATIONS FOR ASSESSING OPTIONS ................................................. 60
8.0
REFERENCES ................................................................................................................. 62
APPENDIX A –– ENGINEERING TOOL ................................................................................ A-1
vi
HUGHES ASSOCIATES, INC.
TABLE OF FIGURES
Page
Figure 1 –– Example of UN31H1A/Y 1041 L (275 gal.) unlisted/non-hardened intermediate bulk
container .............................................................................................................................. 4
Figure 2 –– Illustration of calculated liquid discharge distances from IBCs in both single and
stacked configurations ........................................................................................................ 9
Figure 3 –– Illustration of fire size versus area of release for various fuels. Note that the DF#
identifiers on the figure represent the design fire scenarios described in Table 4 ............ 15
Figure 4 –– Thermal exposure to ceiling structure for various fire sizes ..................................... 18
Figure 5 –– Thermal exposure to ceiling structure for various fire sizes ..................................... 19
Figure 6 –– Photograph of plastic (left) and steel (right) raised sumps ........................................ 32
Figure 7 –– Photograph of Denios Hazberm® system installed on concrete pad ......................... 34
Figure 8 –– Photograph of Denios 2-hr fire-rated storage locker ................................................. 35
Figure 9 –– Example low-level AFFF foam system protecting Design Fire #7 ........................... 39
Figure 10 –– Example of fixed pipe CAFS systems protecting Design Fire #7 ........................... 40
Figure 11 –– Illustration of potential dry chemical suppression system installation.................... 43
Figure 12 –– Photograph of thermally-activated Stat-X aerosol generators [19] ......................... 45
Figure 13 –– Illustration of potential aerosol suppression system installation ............................. 45
Figure 14 –– Illustration of twin-fluid, low pressure water mist system based generally on FIRESCOPE2000® design parameters ..................................................................................... 47
Figure 15 –– Photographs of ‘raw’ cellular glass contained within a pan of heptane [24] (left) and
FOAMGLAS® product installed within a sump [25] (right) ........................................... 49
Figure 16 –– Measured radiative heat flux from LNG pool fires burning with and without
FOAMGLAS installed within the pool [23] ..................................................................... 49
Figure 17 –– Illustration of linear heat detector (left) and fiber optic heat detector (right) ......... 55
Figure 18 –– Example of commercially-available flame detectors .............................................. 56
Figure 19 –– Photograph of GEM LS-10® leak detection sensor ................................................. 57
Figure A.1 –– Liquid classification menu where the User identifies the liquid being contained
within the IBC from a list of known liquids or provides necessary combustion data for a
User Specified liquid....................................................................................................... A-2
Figure A.2 –– Scenario description and hazard calculation menu. The User inputs the physical
characteristics of the facility, IBC configuration, and available separation distances.
Certain fire hazard values are calculated based upon liquid classification and facility
characteristics .................................................................................................................. A-3
Figure A.3. –– Protection recommendation menu where system recommendations are made
based upon fire hazard and facility characteristics provided in previous two menus ..... A-4
vii
HUGHES ASSOCIATES, INC.
LIST OF TABLES
Page
Table E.1 –– Summary of Mitigation/Protection Strategies Considered ....................................... iii
Table E.2 –– Comparison of Advantages/Disadvantages of Fire Suppression Systems ............... iv
Table 1 –– Summary of Assumption and Limitations of Assessment ............................................ 2
Table 2 –– Summary of Possible IBC Configurations.................................................................... 6
Table 3 –– Summary of Fuel Fire Properties for Liquid Fuels ....................................................... 7
Table 4 –– Summary of Maximum Discharge Distances for Various IBC Configurations and
Liquid Contents ....................................................................................................................... 9
Table 5 –– Summary of Design Fire Scenarios ............................................................................ 10
Table 6 –– Summary of Equilibrium Spill Fire Areas for Different Liquid Release Scenarios ... 14
Table 7 –– Summary of Estimated Fire Sizes for Unconfined Design Fire Scenarios ................. 16
Table 8 –– Summary of Actuation Times for Various Design Fires and Sprinkler
Configurations....................................................................................................................... 22
Table 9 –– Summary of Fire Spread Exposure Conditions .......................................................... 23
Table 10 –– Summary of Minimum Separation Distances for All Exposure Durations .............. 24
Table 11 –– Summary of Protection Systems Considered for Design Fire Scenarios.................. 29
Table 12 –– Comparison of Advantages/Disadvantages of Liquid Containment Systems .......... 36
Table 13 –– Comparison of Advantages/Disadvantages of Fire Suppression Systems................ 51
Table 14 –– Comparison of Advantages/Disadvantages of Potential Detection Systems ............ 58
Table 15 –– Summary of Recommendations ................................................................................ 60
viii
HUGHES ASSOCIATES, INC.
1.0
INTRODUCTION
NFPA 30 Flammable and Combustible Liquids Code [1] provides specific guidance for both
containment and fire protection of Listed IBCs containing flammable and combustible liquids in
storage configurations. However, a common usage scenario involves the use of unlisted,
composite IBCs containing flammable/combustible liquids in operations scenarios. The code
does not provide specific fire protection criteria for these applications. The only guidance
provided to date relates to the quantities of liquids that are permitted. Historically, the protection
schemes prescribed in NFPA 30 have been based on experimental testing and engineering
analysis. In keeping with this approach, the goal of this work is to identify protection strategies
that can be evaluated in general accordance with Appendix E of NFPA 30. If adequate
performance is identified, the scheme might be implemented into the code for the protection of
an operations scenario.
In operations scenarios, limited quantities (i.e., up to 4 IBCs stacked one or two high) are
used within an open-air production environment. The NFPA 30 sections related to the transfer
and use of these liquids in operations scenarios (e.g., Chapter 18) provide criteria on the amount
of liquids that are permitted but no specific criteria as it relates to the protection and containment
of the IBC and its content. While some guidance is provided by the code on overhead sprinkler
protection for Class IIIB liquids, there are gaps in knowledge of other protection strategies and
liquids. Consequently, a fire hazard and protection system analysis was conducted to
characterize the hazards associated with this use scenario and identify one or more viable
protection strategies.
2.0
OBJECTIVES AND ASSUMPTIONS
The primary objectives of this assessment were the identification of the fire hazards
associated with the storage/use of unlisted IBCs containing flammable/combustible liquids in
operations scenarios and the development of protection strategies to mitigate these hazards. In
developing protection strategies, the following protection objectives were considered and used to
evaluate the potential benefits of the system:
•
Confinement of fire to reasonable area;
•
Limiting thermal exposure to structural steel overhead;
•
Preventing ignition of neighboring combustibles; and
•
Preventing thermal hazard to personnel.
To address these protection objectives, many types of systems and system combinations were
considered. The system types considered were containment systems, fire detection systems, and
fire suppression systems. These systems provided a means of containing a liquid release,
detecting a fire resulting from a liquid release, and controlling/extinguishing an accidental fire.
Each system and/or combination of systems was evaluated with respect to the protection
objectives listed above and the most viable options were identified.
In this assessment, the fire hazards and protection systems developed were based on
operational assumptions and limitations as described in Table 1.
1
HUGHES ASSOCIATES, INC.
Table 1 –– Summar y of Assumption and Limitations of Assessment
Assumption / Limitation
Only considered IBCs with
a capacity of up to and
including1041L (275 gal.)
Only considered
unlisted/non-hardened
IBCs
Leak, ignition, and
involvement of adjacent
IBCs was assumed
Only considered ceiling
heights ranging from
3.1 m (10 ft) to 9.1 m
(30 ft)
Small, enclosed area not
considered
Gaseous systems
(i.e., clean agent, CO2, etc.)
were not considered
Containment systems
considered have sufficient
capacity to contain the
contents of all IBCs within
the containment area
It would be extremely
difficult to detect, actuate,
and totally suppress a fire
involving a leaking IBC
before adjacent IBCs are
breached
Bulk storage of IBCs is
adequately protected
Justification
IBC volume was identified as
most prevalent in industry (i.e.,
most representative of typical
hazard)
Representative of real-world
conditions
Identifying/characterizing the
myriad of potential leak/
ignition/growth scenarios was
beyond the scope of assessment
and irrelevant to design of
protection schemes
Ceiling heights identified as
most prevalent in industry (i.e.,
most representative of
operations scenarios)
Not representative of typical
operations scenarios
Exclusion of small, enclosed
areas in assessement as well as
personnel life safety issues with
CO2 discharges
Given the IBCs considered in
this analysis are unlisted/nonhardened, it is likely that a fire
involving a single IBC will
spread to all adjacent IBCs and
result in the release of the liquid
contents of all containers
The unlisted, non-hardened
IBCs considered in this analysis
are susceptible to thermal
failure due to the relatively thin,
plastic liner used to contain the
liquid
Scope of this analysis was
limited to limited quantities,
with containment, in an
operations scenario
2
Consequence
Design fire scenarios only
applicable to scenarios
involving IBC of this capacity
Design fire scenarios only
applicable to these types of
containers
Only worst-case scenarios
considered, therefore
protection schemes could be
overly conservative
Estimates of thermal impact to
overhead only applicable to
operations scenarios with
ceiling heights in this range
Exclusion of total flooding
protection systems
Available protection systems
not considered
Containment area and
containment walls will have to
be sized accordingly
Containment areas must be
designed with sufficient
capacity to contain all liquids
contained within all IBCs
Protection of bulk storage of
unlisted units has yet to be
identified – operations
facilities should not have bulk
storage of unlisted IBCs
HUGHES ASSOCIATES, INC.
3.0
HAZARD DESCRIPTION
Characterizing the hazards associated with flammable and combustible liquids with unlisted,
composite IBCs in operation scenarios required that representative worst case fire scenarios be
developed. These scenarios provide the basis for determining the hazards associated with these
fires as well as the required performance of the protection systems. In this assessment, the
development of design fire scenarios required that the following items be addressed:
•
Identify/describe the hazard (i.e., the composite IBC and its contents)
•
Quantify the burning characteristics of the selected IBC contents
•
Characterize the conditions in the vicinity of the IBCs
Each of these items is discussed in detail in the following sections.
In this assessment, the hazard was identified as composite intermediate bulk containers
(IBCs) and their contents (i.e., flammable/combustible liquids). The specific scenario considered
the use of these containers in operations scenarios more specifically, unlisted/unlabeled IBCs
having the UN designation UN31H1A/Y with a capacity of 1041 L (275 gallons). Although IBC
capacities can range from 450–3000 L (119–793 gallons), the 1041 L (275 gallon) size was
found to be the most prevalent in operations scenarios. The 1041 L (275 gallon) IBC has a
1.2 x 1.0 m (48 x 40 in.) foot print and a height of 1.2 m (46 in.).
The IBCs discussed in this analysis were considered to be non-hardened and not listed under
UL 2368 [29] which requires that IBCs not leak or lose structural integrity when subjected to a
full-scale sprinklered exposure fire scenario for 30 minutes. NFPA 30 [1] currently recognizes
only Listed IBCs storing Class II or III liquids in protected storage. Liquids stored in non-Listed
IBCs would be considered as unprotected, since no protection scheme has been developed.
Unlisted IBCs are generally comprised of an inner container constructed from extrusion blowmolded high density polyethylene (HDPE) and an outer container constructed from welded steel
cage or sheet metal cover. The inner container typically has a single 0.05 m (2 in.) diameter
discharge port located at the bottom edge of the wall of the IBC and a 0.15–0.23 m (6–9 in.)
diameter fill port centered on the top. An example of the IBC considered in this assessment is
provided in Figure 1.
3
HUGHES ASSOCIATES, INC.
Figure 1 –– Example of UN31H1A/Y 1041 L (275 gal.) unlisted/non-hardened
intermediate bulk container
In this assessment the release of IBC liquid was assumed to be contained. This
assumption was necessary given that the IBCs being considered contain a relatively large volume
of liquid. If multiple units used in operations released in an unconfined scenario, the spill could
potentially spread over thousands of square feet presenting a hazard too large for any retrofitted,
localized protection system to handle. This assumption is also reasonable in that both NFPA 30
and various environmental regulations may require containment strategies depending on the
volume of flammable/combustible liquids stored. Several uncontained IBC leak scenarios were
explored in the fire hazard analysis but only to illustrate the hazards associated with this type of
scenario. See Table 7 for the consequences associated with uncontained scenarios.
3.1
Identification of Fire Hazards
The fire hazards associated with the use of unlisted, non-hardened, composite IBCs are
described in this section. The hazards described below assume a leak and ignition of the fluid
contained in an IBC. It was assumed that adjacent, unlisted IBCs would be threatened, fail
relatively quickly, and contribute to the fire. This assumed ignition sequence inherently
addresses the threat and corresponding fire that could result from an external fire exposure
(i.e., fire involving adjacent processing operations or combustible storage). With this assumed
involvement of the initial IBC and all adjacent IBCs, the hazard associated with the ignition of
neighboring combustibles and the thermal degradation of the building structural system was
characterized. The threat to neighboring combustibles was characterized for different fire sizes,
severities, and durations. The proximity and combustibility of the neighboring materials were
also considered to understand the spread potential of the IBC fire and potential damage/impact to
adjacent operations. The vulnerability of the building structure was considered for various fire
severities in order to identify both the fire size and duration needed to compromise the structure
and cause potential collapse.
4
HUGHES ASSOCIATES, INC.
3.2
Development of Fire Scenarios
This section describes the general environment in which the IBCs are used in operations
scenarios. The information presented was used in the assessment of the fire hazards associated
with the ignition / burning of the composite IBCs and their contents.
3.2.1
Operations Environment
Operations scenarios are generally performed in enclosed, open-air production spaces with
ceiling heights ranging from 6.1–9.1 m (20–30 ft); however, ceiling heights as low as
3.0 m (10 ft) are possible and were considered in Appendix B of this assessment. For the
purposes of this assessment the footprint of the operations environment was assumed to be 46 m
(150 ft) square. The ceiling support structure was assumed to be exposed steel joists.
The combustible fuel load within an operations environment will vary widely depending on
the operation being performed. The composition and proximity of these fuels will dictate the
severity of the fire exposure that can be accommodated before ignition and fire spread beyond
the first item ignited occurs. The ignition characteristics of both cellulosic and plastic materials
in both packaged and loose forms were used to assess the potential for fire spread. These types
of fuels are representative of common combustibles that can be expected in an operation
environment.
3.2.2
IBC Configurations
A variety of different IBC configurations were considered with the number of IBCs ranging
from one to four. Although in the majority of operations scenarios only two IBCs are utilized at
any given time, an industry survey conducted as part of this assessment did show that in some
scenarios more than two IBCs were present. Consequently, two different configurations utilizing
a total of four IBCs were developed. In total, five different IBC configurations were developed:
a single IBC, a pair of IBCs side-by-side, a pair of IBCs stacked one on top of the other, four
IBCs in a 2 x 2 array side-by-side, and a 1 x 2 array stacked two high. A summary of these
configurations, with illustrations, is provided in Table 2.
3.2.3
IBC Content
NPFA 30 prohibits the storage of Class I liquids in composite IBCs. The maximum single
unit capacity is 3000 L (790 gal) of Class II and Class IIIA liquids. Based on this requirement,
the composite IBCs used in operations scenarios should only contain combustible liquids whose
flashpoints are greater than 38oC (100oF). However, these limitations may not be strictly
enforced, and there are operations scenarios in which Class IB and IC liquids may be stored and
used in composite IBCs. Additionally, many tests with IBCs have involved Class IB liquids as a
potential worst-case.
5
HUGHES ASSOCIATES, INC.
Table 2 –– Summar y of Possible IBC Configurations
Configuration
ID
No. of
IBCs
IBC
Arrangement
1
1
Single Unit
2
2
Side-by-Side
3
2
Stacked
4
4
2 x 2 Array
Side-by-Side
5
4
1 x 2 Array
Stacked 2 high
Illustration of
Arrangement
6
HUGHES ASSOCIATES, INC.
For the purposes of this assessment and to be representative of potential real-world scenarios,
both Class IB and IIIB liquids were considered as possible contents of the composite IBCs being
used in operations scenarios. Representative IB liquids assessed were heptane and isopropyl
alcohol (IPA). Heptane was selected because it is a pure fuel and has historically been used in
IBC design fires. IPA was selected due to the challenges that an alcohol fuel poses to various
types of suppression systems as well its prevalence in various types of operations scenarios.
Mineral seal oil was selected as a representative Class IIIB liquid to demonstrate the hazard
posed by a relatively low flashpoint, Class IIIB liquid. This liquid has also been used in
historical testing of IBC fire protection schemes. A summary of the relevant fire properties for
these fuels is provided in Table 3.
Table 3 –– Summar y of Fuel Fir e Pr oper ties for Liquid Fuels
Fuel
Heptane [2]
Mineral Seal Oil
Isopropyl Alcohol [2]
Heat of
Combustion
(MJ/kg)
44.6
43.81
30.4
Maximum Burning
Rate per Unit Area
(kg/s-m2)
0.101
0.0202
0.015
k β (m-1)
1.1
3.53
100
1 – Heat of combustion data from Tewarson [28]
2 – Burning rate data based on assumed similarity between mineral seal oil and kerosene
3 – kβ data for kerosene [2] used as surrogate for mineral seal oil
3.2.4
Liquid Release Scenarios
Liquid release scenarios were developed for each of the IBC configurations presented in
Table 2. The scenarios consisted of a 7.6 Lpm (2.0 gpm) continuous release, a 57 Lpm (15 gpm)
continuous release, and an instantaneous release of the IBC contents. The first release scenario,
7.6 Lpm (2.0 gpm), has been used to represent a leak in similar operational scenarios (e.g., a
forklift tine with the tine left in place). The second release scenario, 57 Lpm (15 gpm), has also
been used in past testing. These scenarios have been used in prior liquid testing starting in 1975
by Newmann [3] and continuing through current testing as described in Annex E of NFPA 30.
Finally, the instantaneous release was used as a bounding scenario representing either the
catastrophic failure of an IBC due to structural damage or due to the burning of a neighboring
IBC.
In the event of a continuous liquid release (i.e., 2.0–15 gpm) from an opening in the sidewall
of an IBC, it is likely that the liquid being discharged from the IBC will have some lateral
velocity due to the head pressure within the vessel. Under this assumption, the approach
typically used for environmental protection concerns (i.e., raised sump), in which the
containment area is only slightly larger than the IBC is not sufficient to capture the ejected liquid
and must be designed to have a larger ‘footprint’ to capture all discharged liquid.
This lateral discharge distance is dependent on the pressure in the ullage of the IBC, the head
pressure due to the height of liquid above the opening, the height of the opening above the floor,
and the fluid dynamics properties of the liquid. Some assumptions were made to simplify the
7
HUGHES ASSOCIATES, INC.
process. The pressure in the ullage of the IBC was assumed to be ambient (i.e., 100 kPa
[14.7 psi]). The impact of changes in liquid properties was assumed to be well represented by
changes in the discharge coefficients.
The worst-case leak scenario was identified as an opening in the IBC at the mid-height of the
container (i.e., 0.58 m [23 in.]). At points lower on the IBC, the discharged liquid reaches the
ground sooner due to the reduced height from which it is released. At points higher the head
pressure (i.e., pressure resulting from the weight of liquid) is reduced and the liquid is not
discharged with as much force. Based on these assumptions and the configurations, a series of
calculations were performed using Equations 1 and 2 to determine maximum discharge distances
under representative leak scenarios. These discharge distances were used to determine required
containment areas for the IBC configurations presented in Table 2.
vo , x = C (2 g∆z )
Eq. 1
x2
2
vo , x
Eq. 2
y = −0.5 g
where vo , x is the discharge velocity of the liquid exiting the opening in the IBC, C is the
discharge coefficient, g is gravitational force (i.e., 9.8 m/s2 [32 ft/s2]), Δz is the height of liquid in
the IBC that is above the opening (m [ft]), y is the height of the liquid stream at a given x
location (m[ft]), and x is the lateral distance from the sidewall of the IBC (m [ft]).
In order to account for the myriad of potential fluids and associated fluid dynamics properties
(i.e., density, surface tension, viscosity, etc.) that could be stored within an IBC, a range of
discharge coefficients were considered. The coefficients used ranged from 0.6–1.0. The height
of liquid above the opening and liquid density were held constant at 0.6 m (24 in.) and 800 kg/m3
(50 lbs/ft3) for all tests, respectively. The height of the opening above the floor used in these
calculations was either 0.6 m (2 ft) representative of a single IBC installed on the floor (i.e., IBC
configurations 1, 2, and 4 in Table 2) or 1.8 m (6 ft) representative of an IBC that is stacked on
top of another IBC (i.e., IBC configurations 3 and 5 in Table 2). A summary of both the lateral
and vertical liquid discharge distances under the range of scenarios described above is provided
in Table 4 and Figure 2 below.
Based on these calculations it was determined that in order to ensure capture of all liquid
discharged by a worst-case liquid release scenario, a spill barrier system would have to be a
minimum of 1.2 m (4 ft) from the side walls of the IBCs being contained for scenarios in which
IBCs are not stacked and a minimum of 2.1 m (7 ft) for scenarios in which IBCs are stacked two
high. These values were used later in the fire hazard analysis to develop appropriately sized spill
barrier systems and calculate peak fire sizes.
3.3
Design Fire Scenarios
Using the information presented above, ten different design fire scenarios were developed.
The variables considered in these design fires included the number of IBCs involved, their
8
HUGHES ASSOCIATES, INC.
configuration, their content, and the method by which the IBC and the associated liquid release
were confined, if at all. A summary of the ten design fire scenarios is provided in Table 5.
Table 4 –– Summar y of Maximum Dischar ge Distances for Var ious IBC Configur ations
and Liquid Contents
IBC
Configuration(s)
Elevation of Opening
Relative to Floor
m [ft]
1, 2, 4
0.6 [2.0]
3, 5
1.8 [6.0]
Discharge
Coefficient
0.6
0.7
0.8
0.9
1.0
0.6
0.7
0.8
0.9
1.0
Maximum Lateral
Discharge Distance
m [ft]
0.9 [3.0]
1.0 [3.3]
1.1 [3.5]
1.2 [3.8]
1.2 [4.0]
1.7 [5.5]
1.7 [5.8]
1.9 [6.3]
2.1 [6.8]
2.1 [7.0]
Figure 2 –– Illustration of calculated liquid discharge distances from IBCs
in both single and stacked configurations
9
HUGHES ASSOCIATES, INC.
Table 5 –– Summar y of Design Fir e Scenar ios
IBC
Arrangement
Containment
System
Width of
Containment
Area
(m [ft])
1
Single IBC
Raised Sump
1.5 [4.8]
1.2 [4]
1.8 [19]
2
2
Side-by-Side
Raised Sump
1.5 [4.8]
2.4 [7.9]
3.5 [38]
3
1
Single IBC
Spill
Containment/
Berm System
3.7 [12]
3.4 [11]
13 [136]
Design
Fire
ID
No.
of
IBCs
1
Length of
Containment
Area
(m [ft])
Containment/
Liquid
Release Area
(m2 [ft2])
Illustration of
Arrangement
10
HUGHES ASSOCIATES, INC.
Table 5 –– Summary of Design Fire Scenarios (Continued)
Design
Fire
ID
No.
of
IBCs
4
5
Width of
Containment
Area
(m [ft])
Length of
Containment
Area
(m [ft])
Containment/
Liquid
Release Area
(m2 [ft2])
Containment
System
2
Side-by-Side
Spill
Containment/
Berm System
4.9 [16]
3.4 [11]
17 [183]
2
Stacked
Spill
Containment/
Berm System
5.5 [18]
5.3 [17]
29 [312]
4
Side-by-Side
in Spill
Containment
System
Spill
Containment/
Berm System
4.9 [16]
4.4 [14.5]
22 [237]
11
IBC
Arrangement
HUGHES ASSOCIATES, INC.
6
Illustration of
Arrangement
Table 5 –– Summary of Design Fire Scenarios (Continued)
Design
Fire
ID
7
No.
of
IBCs
4
IBC
Arrangement
Containment
System
Stacked 2-high
in Spill
Containment
System
Spill
Containment/
Berm System
Width of
Containment
Area
(m [ft])
Length of
Containment
Area
(m [ft])
Containment/
Liquid
Release Area
(m2 [ft2])
6.7 [22]
6.2 [21]
42 [452]
12
8
9
10
1
Single IBC
None with 7.6
Lpm
(2 gpm) spill
None with
57 Lpm
(15 gpm) spill
Instantaneous
Release
N/A
(varies with
fuel type)
Illustration of
Arrangement
HUGHES ASSOCIATES, INC.
Design fires 1–7 represent confined fire scenarios (i.e., scenarios where a containment
system is installed) while 8–10 represent unconfined release scenarios (i.e., no containment).
Candidate containment systems will be discussed in greater detail later in this document.
Design fires 1 and 2 assume that the release/ignition of liquid from either one or two IBCs
into a containment area with a footprint that is 50 percent larger than that of the actual IBCs (i.e.,
1.8 m2 [19 ft2] for a single IBC and 3.5 m2 [39 ft2] for two IBCs). This scenario represents the
containment scenario where the IBCs are being stored on a non-combustible, semi-enclosed,
raised sump constructed to have a footprint slightly larger than the IBC with sufficient depth to
capture the entire contents of the IBC.
Design fires 3 through 7 were developed to represent scenarios in which a fixed berm / raised
sill system is used to protect the IBCs. In these scenarios the containment system was sized
based on the area of the IBC configuration and the predicted liquid discharge distances described
in Table 4. This approach resulted in spill containment areas ranging in area from 13–42 m2
(140–450 ft2).
The unconfined scenarios, Design fires 8–10, were developed for three different types of
liquid release, two continuous and one instantaneous. Due to the liquid in these scenarios being
unconfined, a steady-state area is not achieved until the fuel is ignited. Once ignited, the fuel
area will come to equilibrium when the mass burning rate of the fuel equals the mass flow rate of
the leak. Using the liquid release rates specified in Section 3.2.4 for design fires 8 and 9 in
conjunction with the mass burning rates provided for the three different fuels specified in
Table 3, equilibrium liquid spill areas were calculated using Equation 3.
Aeq =
Vρ
m ′′
Eq. 3
where Aeq is the equilibrium area of the released liquid, V is the volumetric liquid release rate
 ′′ is the mass burning rate per unit area of the
(m3/s), ρ is the density of the liquid (kg/m3), and m
2
liquid fuel (kg/s-m ). A total of six equilibrium areas were calculated and these values are
presented in Table 6.
The equilibrium areas presented in Table 6 are based on the assumption that ignition of the
released liquid occurs immediately after release. Under this assumption the equilibrium area
represents the largest possible spill area for a given scenario. However, if immediate ignition is
not assumed, then the maximum spill area is a function of time before ignition and the rate of
liquid release. This resulting area will be the initial fire and will gradually regress back to the
equilibrium area as the mass burning rate consumes more fuel than is released.
13
HUGHES ASSOCIATES, INC.
Table 6 –– Summar y of Equilibr ium Spill Fir e Ar eas for Differ ent Liquid Release
Scenar ios
Fuel
Type
Liquid Release
Scenario
Equilibrium Area
(m2 [ft2])
2 gpm continuous spill
0.9 [9.7]
15 gpm continuous spill
6.4 [68.8]
2 gpm continuous spill
8 [86.0]
15 gpm continuous spill
59 [635]
2 gpm continuous spill
6.6 [71.0]
15 gpm continuous spill
50 [538]
Heptane
Mineral Seal
Oil
IPA
3.4
Critical Exposures
The focus of the fire hazard analysis was to assess the conditions that could be generated by a
fire involving composite IBCs and compare these conditions to critical exposures. These critical
exposures were selected to prevent the spread of the fire beyond the area of origin and limit
thermal exposure to overhead steel structures in order to prevent structural failure.
It is possible that a variety of combustible materials may be present in and around the area in
which the IBCs are placed and used. Consequently, the severity of the thermal exposure
required to melt/ignite these combustibles will vary broadly depending on both the composition
and proximity of the item to the exposure. A range of heat fluxes associated with the ignition of
various materials was used to evaluate the propensity of a given fire to ignite neighboring
combustibles (i.e., spread fire beyond the area of origin). The heat fluxes used in this assessment
ranged from 2.5–100 kW/m2. Further description of the specific values and rationale for these
selections are provided Section 4.4.
For the purposes of this assessment, structural failure of the ceiling structure of the building
was associated with upper layer temperatures of 538oC (1000oF) or direct flame impingement
that were maintained for longer than one minute [1]. An exposure of this severity would most
likely result in unprotected structural steel temperatures that are comparable to the gas/flame
temperatures. At these temperatures, steel has lost approximately 50 percent of its design
strength. Since the typical design safety factor for structures is two, any additional degradation
of the steel strength could result in the structural steel members being compromised [4].
4.0
FIRE HAZARD ANALYSIS
A fire hazard analysis was performed based on the parameters and critical exposures
described in Sections 3.2–3.4. In addition to the information provided above, in order to perform
some of the calculations it was necessary to assume a general volume (i.e., enclosure geometry)
in which the fires would occur. For this assessment, all design fire scenarios were assumed to
occur within a 46 m (150 ft) square enclosure lined with gypsum wallboard and having a total
14
HUGHES ASSOCIATES, INC.
vent area of 8.9 m2 (96 ft2). The total enclosure volume based on this assumption was 19,000 m3
(680,000 ft3). The vent area selected was deemed comparable to that of four standard doorways
or a single roll-up door being open.
4.1
Design Fire Sizes
The approximate fire sizes resulting from the various liquid release scenarios described
above were determined. Using the expected liquid release areas and fuel combustion properties,
expected fire sizes were calculated. Fire size was calculated using Equation 4.
q = ∆hc m ∞′′ (1 − e − kβD ) Arelease
Eq. 4
where q is the heat release rate of the fire (MW), ∆hc is the heat of combustion of the fuel
(MJ/kg), m ∞′′ is the maximum burning rate per unit area of the fuel (kg/m2-s), kβ is an empirical
constant specific to each fuel, D is the equivalent diameter of the release (m), and Arelease, is the
area of the release (m2). The specific combustion values used to generate the fire size trends
presented in Figure 3 were provided in Table 3.
Figure 3 –– Illustration of fire size versus area of release for various fuels. Note that the DF#
identifiers on the figure represent the design fire scenarios described in Table 4
15
HUGHES ASSOCIATES, INC.
As shown in Figure 3, the design fires, for the confined scenarios, ranged in size from
0.7–160 MW depending on the size of the release, containment scenario, and the combustibility
of the liquid. Design fires 8–10 are not included in Figure 3 because the liquid release areas
associated with these scenarios vary for each fuel. A summary of the estimated fire size for each
of the continuous spill, unconfined design fires as well as the instantaneous release scenario, are
provided in Table 7.
Table 7 –– Summar y of Estimated Fir e Sizes for Unconfined Design Fir e Scenar ios
Estimated Fire Size (MW)
Mineral
Isopropyl
Heptane
Seal Oil
Alcohol
Liquid
Release
Scenario
Liquid
Release
Area (m2)
8
7.6 Lpm
(2 gpm)
(varies with fuel)
2.8
4.4
2.7
9
57 Lpm
(15 gpm)
(varies with fuel)
28
33
20
10
Instantaneous
1500
6600
2500
590
Design
Fire ID
Design fires 8–10 are presented only to illustrate the hazards associated with unconfined
liquid releases. Although two different continuous release scenarios were considered, given that
the IBCs considered in this analysis are unlisted and non-hardened, it is reasonable to assume
that shortly after the initiation of these fires, the instantaneous release of both the initiating and
all adjacent containers would occur, resulting in conditions comparable to that of Scenario 10.
The time frame in which this transition would occur is expected to be relatively short and
consequently the only rational unconfined fire scenario that should be considered is an
instantaneous release, which is too large for any fire protection system to control and justifies the
need for containment in all operations scenarios.
4.2
Structural Integrity Analysis
Given the large volumes of liquid potentially stored in the IBC configurations considered, the
burning durations for these fires, if left unsuppressed, are expected to be at a minimum on the
order of up to ten minutes with much longer burning durations possible. The thermal exposure
to the ceiling structure (i.e., exposed steel joists) typically present in operations scenarios was
evaluated in order to determine under what design fire scenarios (i.e., fire size / exposure
duration) the structure would be compromised. A series of thermal exposures, whose range
encompassed the fire scenarios described in Table 5, were used in conjunction with Equation 5 to
approximate the temperature of the fire gases produced.


Q2
Tg = 6.85

 Av hv ( AT hk ) 
(
)
16
1/ 3
+ TA
Eq. 5
HUGHES ASSOCIATES, INC.
where Tg is the upper layer gas temperature rise above ambient (K), Q is the heat release rate
(kW), Av is the area of the ventilation opening (m2), hv is the height of the ventilation opening
(m), AT is the total area of the compartment enclosing surface boundaries excluding area of vent
openings (m2), hk is the convective heat transfer coefficient (kW/m2-K), and TA is the ambient
temperature (K). Other than the heat release rate ( Q ), all variables were held constant for the
calculations. A list of the variables and the values used are provided below:
•
Av = 8.9 m2
•
hv = 2.4 m
•
AT = 5844 m2 (assumes 46 m [150 ft] square building with 9.1 m [30 ft] ceiling]
•
hk = ranged from 0.01–0.05 kW/m2-K and varied with time as shown in Eq. 6
•
TA = ambient temperature (assumed to be 298K).
hk =
kρc
t
Eq. 6
where kρc is the thermal inertia of the surface boundary ((kW/m2-K)2-s) and t is time (s).
The results of these calculations, shown in Figure 4, show that for ceiling heights of
6.1 m (20 ft) and 9.1 m (30 ft), upper layer temperatures do not exceed the 538oC (1000oF)
temperature threshold for fires less than 25 MW in size when permitted to burn for up to an hour.
Figure 4 also shows that for fires ranging in size from 25–75 MW, the window of time between
the ignition of the fire and the potential compromise of the ceiling structure ranges from as long
as 60 minutes to as short as 1.5 minutes, respectively. However, these times assume a global
heating of the overhead support structure (i.e., the entire upper layer is heated to the exposure
threshold) and therefore may not be very conservative estimates. In general, the differences in
time to reach thermal exposure threshold did not vary significantly when comparing exposures to
a ceiling at 6.1 m (20 ft) and a ceiling at 9.1 m (30 ft). A ceiling height of 3.0 m (10 ft) was not
considered in this approach because for the majority of scenarios described in Table 5 the
average flame height was greater than that of the ceiling. Consequently, direct flame
impingement was identified as being the more appropriate measure of thermal insult to the
ceiling structure.
17
HUGHES ASSOCIATES, INC.
Figure 4 –– Thermal exposure to ceiling structure for various fire sizes
The fact that upper layer gas temperatures were found to not exceed 538oC (1000oF) for fires
less than 25 MW does not imply that fires of this severity, or smaller, do not pose a threat to the
structural steel overhead. In fact, the large flame heights associated with fires of this size (i.e.,
9.1–11 m [30–35 ft] for a 25 MW fire) are sufficient to locally heat structural steel members to
critical temperatures even if the average upper layer gas temperatures cannot be elevated to
threshold levels (i.e., 538oC [1000oF]). To account for this, a more conservative approach was
adopted and plume centerline temperatures were calculated using Equation 7 presented below.
Associating plume centerline temperatures with localized structural steel temperatures provides a
lower bound (i.e., conservative estimate) with respect to the fire sizes and exposure durations
that are expected to produce structural failure.
1/ 3

 

T
a
 Q 2 / 3 (z − z )−5 / 3  + T
T p = 9.1
c
o
 a
  gc p 2 ρ a 2 


Eq. 7
where Tp is the plume center line temperature (oC), Ta is the ambient air temperature (K), g is the
acceleration of gravity (m/s2), cp is the specific heat of air (kJ/kg-K), ρa is the density of
air (kg/m3), Qc is the convective heat release rate (kW), z is the distance from the top of the fuel
to the ceiling (m), and z0 is the hypothetical virtual origin of the fire (m).
Centerline temperatures were calculated for a range of fire sizes under ceiling heights
ranging from 3.0 m (10 ft) to 9.1 m (30 ft) to determine at what fire sizes plume temperatures at
the ceiling reached the thermal exposure threshold of 538oC (1000oF). A plot of the plume
18
HUGHES ASSOCIATES, INC.
centerline temperatures as a function of fire size is provided in Figure 5 for ceiling heights of
6.1 m (20 ft) and 9.1 m (30 ft). The temperatures presented in Figure 5 represent gas
temperatures at the ceiling elevation for a steady-state fire. Plume centerline temperatures are
not presented for the 3.0 m (10 ft) ceiling height because direct flame impingement (i.e.,
temperatures greater than 538oC [1000oF]) is assumed to occur almost immediately after
ignition.
600
o
Plume Centerline Temperature ( C)
700
500
400
300
200
H = 6.1m [20ft]
H = 9.1m [30ft]
Thermal Exposure Threshold
100
0
0
5
10
15
20
25
30
35
Fire Size (MW)
Figure 5 –– Thermal exposure to ceiling structure for various fire sizes
The impact of fire size and ceiling height were more evident when considering plume
centerline temperatures than when upper layer gas temperatures were considered. As shown in
Figure 5, a fire size of approximately 13 MW is capable of producing temperatures of 538oC
(1000oF) under a 6.1 m (20 ft) ceiling while a 9.1 m (30 ft) ceiling requires a fire that is
approximately 25 MW in size. This analysis shows the greater susceptibility of the 6.1 m (20 ft)
ceiling to structural failure due to proximity of the structural elements to the flame plume.
The time at which the temperatures presented in Figure 5 are reached at the ceiling is
primarily dependent on the time required for the liquid fuel fire to reach steady-state conditions.
This time depends on the volatility and associated flame spread rate of the fuel which dictates
how quickly the fuel area becomes involved and reaches steady-state burning conditions. For
highly volatile fuels (e.g., heptane and IPA) flame spread rates are such that full involvement and
the development of steady-state burning conditions are achieved in time frames on the order of
seconds to tens of seconds. For less volatile, combustible fuels (e.g., mineral seal oil) flame
spread rates are substantially slower resulting in times to full involvement and steady-state
burning on the order of minutes. These time frames, while being very conservative, provide a
lower bound with respect to the exposure duration required to fail the overhead structural steel.
19
HUGHES ASSOCIATES, INC.
4.3
Overhead Thermal Detection Analysis
The design fire scenarios were also used to characterize expected thermal detection system
actuation times. In this characterization both standard response and quick-response sprinklers
were used. The sprinklers used had temperature ratings of 135oC (275oF) and 74oC (165oF),
respectively. The response time index used for the standard response sprinkler was 130 (m-s)0.5
and for the quick response link a value of 34 (m-s)0.5 was used. This range of activation
temperatures and response time indices were chosen because they bound the performance of the
majority of thermal detection technologies and, in the case of the standard response link,
represent the technology that is currently prescribed in NFPA 30 [1].
The vertical distance between the design fire and the sprinklers was dependent on the design
fire scenario. For scenarios in which the IBCs were not stacked, a vertical separation distance of
7.6 m (25 ft) was used. For all scenarios presented in this section a ceiling height of 9.1 m (30 ft)
was used. In cases where IBCs are stacked, a distance of 6.4 m (21 ft) was used. Both of these
distances were calculated assuming sprinklers were located 8.8 m (29 ft) above the floor and that
the height of an IBC was 1.2 m (4 ft).
A sprinkler spacing of 3.0 m (10 ft) was used based on the guidance provided in NFPA 30
[1] for the overhead sprinkler protection of Listed IBCs. Based on this spacing a maximum
radial distance of 1.5 m (5 ft) was calculated. However, this distance assumes that the fire is a
point, when in reality the fire has an area that is known (i.e., an equivalent diameter) that
effectively reduces the distance between the thermal link and the edge of the fire. To account for
this, the distance between the sprinkler and the fire was calculated for each scenario by
subtracting one-half of the equivalent diameters of the liquid release from the maximum radial
distance (i.e., 1.5 m [5 ft]). These parameters, when applied to Equation 8, were used to estimate
thermal actuation times for both standard and quick-response sprinkler heads for each design fire
and each liquid fuel.
 RTI   T jet − Ta 
 ln

t activation = 
 u jet   T jet − Tactivation 


 
Eq. 8
where tactivation is the sprinkler response time (s), RTI is the sprinkler response time index (m-s)0.5
u jet is the ceiling jet velocity (m/s), Tjet is the ceiling jet temperature (oC), Ta is the ambient
temperature (oC), and Tactivation is the activation temperature of the sprinkler (oC). A summary of
the radial distances and corresponding predicted thermal detection times for both the standardand quick-response technologies in each of the design fires is provided in Table 8.
For the majority of the design fire scenarios considered, sprinkler activation times, both
standard response and quick-response, were relatively short. This is an artifact of the calculation
assuming a steady-state fire size as opposed to incorporating some period of fire development
(i.e., growth curve). For both the heptane and IPA liquids, this assumption is valid in that the
volatility of these two liquids would result in a rapidly developing, high temperature ceiling jet
that would activate sprinkler heads relatively quickly. However, assuming the immediate
development of a steady-state fire for the scenarios involving the mineral seal oil is less
20
HUGHES ASSOCIATES, INC.
appropriate and most likely results in less conservative activation times. Since mineral seal oil is
a combustible fuel and assumed to be at ambient temperature at the time of discharge, the spread
of flame over the liquid pool after being released will be significantly slower than that of the
heptane/IPA. Consequently, longer periods of time are required before a steady-state fire is
developed. This period of growth is not accounted for in the activation times presented in
Table 8 and could be on the order to tens of seconds in duration depending upon the size of the
liquid pool. Despite the potential over-prediction of sprinkler activation times for scenarios
involving mineral seal oil, there were several scenarios for both sprinkler types where the
sprinklers were not predicted to activate. This was true for five of the thirty scenarios in which
standard response sprinklers were used and in one case where the quick-response sprinkler was
used.
Given the relatively quick response times that were predicted for the 9.1 m (30 ft) ceiling
height, calculations for 3.0 m (10 ft) and 6.1 m (20 ft) ceilings were not performed because it is
reasonable to assume that activation times would only get shorter due to the sprinklers being
closer to the fire.
4.4
Fire Spread Analysis
The spread of fire beyond the IBCs involved in the initial fire event was also considered.
The intent is to provide guidance on the maximum separation distance to combustible materials.
Using the design fire sizes calculated in Section 4.1 in combination with the heat flux values
discussed below, minimum separation distances were established for two different fuels and all
design fire scenarios. The fuels used in this analysis represent the lower and upper bound with
respect to predicted fire size in the given design fire scenarios.
Three different exposures were considered. The first evaluated the ignition of structural
solids (i.e., substrates such as paper-covered gypsum wallboard [GWB] or plywood). These
materials would typically be present as part of the structure of the building and would represent
materials that are the most challenging to ignite and representing the least conservative
perimeter. The second threat involved the ignition or melting of neighboring combustible
materials (i.e., cardboard, fabrics, plastics, etc.) representing the commodities or materials that
could be present in the vicinity of the IBCs in use. Finally, the third threat perimeter was
developed based on the potential that personnel could be in the area in which the fire occurs and
consequently could be exposed to the fire conditions.
21
HUGHES ASSOCIATES, INC.
Table 8 –– Summar y of Actuation Times for Var ious Design Fir es and Spr inkler Configur ations
22
Design
Fire
Scenario
1
2
3
4
5
6
7
Area
(m2)
1.8
3.5
12.6
17
29
22
35
Heptane
Fire
Size
(MW)
6.6
14
56
75
130
98
160
Mineral
Oil Fire
Size
(MW)
1.5
3.0
11
14
25
19
30
IPA
Fire
Size
(MW)
0.7
1.4
5.1
6.8
12
8.8
14
Max
Distance
between
Fire and
Sprinkler
(m)
1.4
1.1
0.2
0.0
0.0
0.0
0.0
SR Sprinkler Activation at
Max. Spacing (min.)*
Heptane
0.8
0.3
0.0
0.0
0.0
0.0
0.0
Mineral
Seal Oil
DNA
1.0
0.0
0.0
0.0
0.0
0.0
IPA
DNA
DNA
1.2
0.8
0.4
0.6
0.2
QR Activation at Max.
Spacing (min.)*
Heptane
0.1
0.0
0.0
0.0
0.0
0.0
0.0
*Activation times of zero indicate that actuation was predicted to occur in less than 0.1 minutes (i.e., less than 6 seconds).
DNA – Did not activate
SR – Standard Response
QR – Quick Response
Mineral
Seal Oil
0.2
0.1
0.0
0.0
0.0
0.0
0.0
IPA
DNA
0.5
0.1
0.0
0.0
0.0
0.0
HUGHES ASSOCIATES, INC.
The heat fluxes associated with each of the threats were determined for three different
exposure durations: 0.1, 1.0, and 10 minutes. These durations were selected to be representative
of a fire event that is detected early via a locally-installed detection system (i.e., 0.1 minutes), a
fire event that is detected via manual activation of the suppression system or an overhead thermal
detection system (i.e., 1.0 minute), and fire event that is either not detected or detected and not
suppressed (i.e., 10 minutes). For the cases in which the fire is detected, it is assumed that the
fire is quickly suppressed thereby reducing the fire spread hazard. Critical heat flux values (i.e.,
minimum flux required to ignite a material in a specified duration) for each material description
and exposure duration was determined using Equation 9.
π
tig = 4
kρc(Tig − Ta )
2
Eq. 9
′′ )2
(qe′′ − qcritical
where tig is the material ignition time (s), kρc is the thermal inertia of the material
([kW/m2-K]2-s), Tig is the material ignition temperature (oC), Ta is the ambient air temperature
′′
(oC), q e′′ is the exposure heat flux (kW/m2), and q critical
is the critical heat flux for ignition
2
(kW/m ). The incident heat fluxes and associated exposure durations for each material
description are provided in Table 9. The thermal inertia and ignition temperature data used for
the structural solids and neighboring solids were obtained from the literature [5,6].
Table 9 –– Summar y of Fir e Spr ead Exposur e Conditions
Exposure
Category
Material
Description
1
Ignition of Structural Solids
(i.e., GWB, Plywood, etc.)
2
3
Melting of Plastics /
Ignition of Neighboring
Solids
(i.e., cardboard, fabric, etc.)
Pain to Human Skin
Critical Heat
Flux
(kW/m2)
25
50
100
15
25
Exposure
Duration
(min.)
10
1
0.1
10
1
65
0.1
2.5
1
Reference
[5]
[6]
Based on the heat flux values in Table 9 and the design fire sizes discussed earlier, minimum
separation distances were calculated using Equation 10.
xsd =
D 
Qχ r
−  eq 
′′ 4π  2 
qinc
Eq. 10
where xsd is the distance between the edge of the fire and material being exposed (m), Q is the
heat release rate (kW), χ r is the radiative fraction (assumed to be 0.3 [7] for all calculations),
23
HUGHES ASSOCIATES, INC
′′ is the incident radiative heat flux on the material (kW/m2), and Deq is the equivalent diameter
qinc
of the liquid fuel fire (m). Separation distances calculated for both IPA and heptane under
design fire scenarios 1–7 are provided in Table 10. These fuels were selected because they
represent the bounding cases for radiation hazard from a pool fire in that heptane is a relatively
high soot yield fuel (i.e., highly radiative) and IPA has a negligible soot yield (i.e., minimally
radiative). All distances were calculated using the critical heat fluxes established for exposure
category 2 in Table 9. As expected, the distances associated with the prevention of pain to
human skin were the largest (i.e., most conservative) ranging from 6–39 m (20–130 ft). It should
be noted that the IPA design fires were not able to generate thermal exposures of 50 and
100 kW/m2.
Table 10 –– Summar y of Minimum Separ ation Distances for All Exposur e Dur ations
Design
Fire
ID
Containment
Area
(m2 [ft2])
Fuel
IPA
1
1.8 [19]
Heptane
IPA
2
3.5 [38]
Heptane
IPA
3
13 [136]
Heptane
Exposure
Duration
(min.)
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
24
Minimum Separation
Distance
(m [ft])
Unlimited1
Unlimited1
0.5 [1.5]
1.7 [5.5]
0.9 [3.0]
1.5 [5.0]
2.4 [8.0]
6.7 [22]
Unlimited1
Unlimited1
0.6 [2.0]
2.6 [8.5]
1.2 [4.0]
2.7 [9.0]
3.7 [12]
10 [34]
Unlimited1
0.3 [1.0]
1.1 [3.5]
5.7 [19]
2.6 [8.5]
5.2 [17]
7.3 [24]
22 [71]
HUGHES ASSOCIATES, INC
Table 10 –– Summary of Minimum Separation Distances for All Exposure Durations
(Continued)
Design
Fire
ID
Containment
Area
(m2 [ft2])
Fuel
IPA
4
17 [183]
Heptane
IPA
5
29 [312]
Heptane
IPA
6
22 [237]
Heptane
IPA
7
42 [452]
Heptane
1
Exposure
Duration
(min.)
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
0.1
1
10
Human Pain Threshold
Minimum Separation
Distance
(m [ft])
Unlimited1
0.5 [1.5]
1.2 [4.0]
6.5 [21]
3.0 [10]
6.0 [20]
8.5 [28]
25 [82]
Unlimited1
0.6 [2.0]
1.5 [5.0]
8.5 [28]
4.0 [13]
8.2 [27]
11 [36]
33 [110]
Unlimited1
0.5 [1.5]
1.4 [4.5]
7.4 [24]
3.3 [11]
7.0 [23]
9.8 [32]
28 [93]
Unlimited1
0.6 [2.0]
1.8 [6.0]
10 [34]
4.9 [16]
9.8 [32]
14 [45]
39 [130]
Although predicted heat fluxes from these fires may be less than the critical values established in this analysis,
adequate separation distance should be provided to prevent direct flame impingement.
25
HUGHES ASSOCIATES, INC
Minimum separation distances were determined from this analysis based on consideration of
seven different confined design fire scenarios, two different fuels, and three different exposure
durations. These calculations show that:
4.5
•
For scenarios in which localized, rapid detection (i.e., 0.1 minute exposure duration)
is expected, a minimum distance of 4.9 m (16 ft) is required to minimize the
likelihood of ignition/melting of any neighboring combustible materials for a sooting
fuel (e.g., heptane). For non-sooting fuels (e.g., IPA) the separation distance was
determined to be unlimited when considering incident heat flux only.
•
For scenarios in which remote detection (i.e., 1.0 minute exposure duration) is used,
minimum distances of 0.6 and 9.8 m (2.0 and 32 ft) are required to minimize the
likelihood of ignition/melting of any neighboring combustible materials for a nonsooting (i.e., IPA) and sooting (i.e., heptane) fuel, respectively.
•
For scenarios in which no detection is expected, minimum distances of 1.8 and 14 m
(6.0 and 45 ft) are required to minimize the likelihood of ignition/melting of any
neighboring combustible materials for a non-sooting (i.e., IPA) and sooting (i.e.,
heptane) fuel, respectively.
•
When considering the location of personnel operation stations and safety control
system (i.e., manual activation buttons, etc.) locations, separation distances ranged
from 1.7–10 m (6–34 ft) and 6.7–39 m (22–130 ft) for a non-sooting (i.e., IPA) and
sooting (i.e., heptane) fuel, respectively.
Fire Hazard Analysis Findings
A total of ten design fire scenarios, seven confined and three unconfined were evaluated.
Each scenario was evaluated using three different fuel types. Predicted fire sizes from the
confined scenarios ranged in size from 0.7–160 MW while for the unconfined scenarios a range
of 2.7–6600 MW was determined.
Based on these fire sizes, minimum fire sizes and maximum exposure durations to prevent
structural collapse were calculated. It was determined that for 6.1 m (20 ft) and 9.1 m (30 ft)
ceiling heights, minimum fire sizes between 12–25 MW are required to sufficiently heat
overhead steel joists to a temperature of 538oC (1000oF) depending upon the ceiling height and
safety factor adopted. For ceiling heights of 3.0 m (10 ft) it was determined that all of the design
fire scenarios developed in this assessment present an immediate threat to the overhead ceiling
structure due to direct flame impingement.
Overhead sprinkler activation times were also evaluated for the various design fire scenarios
considered. Both standard- and quick-response sprinkler technologies were considered. Based
on the estimated design fire sizes calculated, sprinklers activated relatively quickly with
responses as early as seconds after ignition to as long as 1.2 minutes for some of the smaller
design fires. In all cases the sprinkler spacing was assumed to be 3.0 m (10 ft).
Minimum distances at which ignition is unlikely were determined for the various design fire
scenarios. Depending on the fuel and exposure duration these distances ranged from unlimited
26
HUGHES ASSOCIATES, INC
to a minimum of 14 m (45 ft) to prevent ignition for an unprotected heptane release. The
findings of this minimum separation distance analysis also showed the benefits that automatic
detection systems can provide. The primary benefit is the reduction in minimum separation
distances with decreased detection response times. The location (e.g., overhead vs. local) and
type of detection system (e.g., thermal vs. optical) selected directly impacts the minimum
separation distances that are required within the operations scenario.
Based on these findings, certain IBC configurations that may not require additional
protection beyond the fundamental containment criteria were identified. These configurations
may be acceptable since:
•
There is no threat to overhead steel structures;
•
There is no threat to neighboring combustible materials provided an appropriate
stand-off distance is specified; and,
•
A standard wet-pipe automatic sprinkler system is usually present to protect the
facility against incidental fires.
IBC configurations with these characteristics typically consisted of relatively small quantities
of fuel stored in relatively small containment areas (i.e., Design Fire Scenarios 1–7 for IBCs
containing IPA or a similar fuel). Other factors leading to a fuel presenting a minimal hazard
could include the volatility and associated burning rate of the fuel. For these scenarios, no
additional protection other than containment and separation may be needed. This assumes that
the facility, in accordance with insurer/fire code requirements, will at a minimum have an
overhead water sprinkler system designed for an ordinary hazard scenario. In these scenarios,
the expected fire sizes are less than the conservative minimum fire size required to fail an
overhead steel structure with a ceiling height of 6.1 m (20 ft). The radiation threat from these
fires is minimal, when considered to include an offset distance of a minimum of 4.6 m (15 ft).
This distance may be reasonable for specific operations scenarios, e.g., where counter-balanced
forklifts are used and aisles ranging from 3.4–4.3 m (11–14 ft) may be required. For other
scenarios, it is recommended that the alternative suppression and detection systems described in
this assessment be considered in addition to containment and separation.
IBC configurations with the characteristics identified above could potentially be permitted
without the installation of additional protection systems other than a containment system and
ordinary hazard wet pipe sprinkler system. Alternatively, the findings of the fire hazard analysis
also show that certain fuel/IBC configurations (i.e., Design Fire 7 for IBCs containing heptane or
a similar fuel) should be prohibited without the use of Listed/Labeled IBCs and the installation
of NFPA 30 storage protection systems. Finally, the analysis shows the significant hazard
present for unconfined, unprotected IBC configurations (i.e., Design Fire 10).
5.0
PROTECTION SYSTEM ANALYSIS
Based on the findings of the fire hazard analysis, active protection may be required for
specific operations scenarios involving the use of IBCs containing flammable/combustible
liquids. The purpose of this section is to identify existing, new, or previously unrecognized
protection systems that might provide localized protection.
27
HUGHES ASSOCIATES, INC
The systems identified in this section were geared towards existing buildings where the
ability to retrofit fixed protection systems in accordance with NFPA 30 [1] storage protection
criteria (e.g., large volume water or AFFF) is not practical. To address this, alternative systems
or combinations of systems were considered. The protection systems considered were divided
into three groups: liquid containment, fire suppression, and fire detection. A summary of the
protection systems considered and the applicability of these systems to the various design fire
scenarios is provided in Table 11. An engineering tool was also developed to aid in this analysis.
This tool is presented in the form of an electronic attachment provided to the project sponsors.
The tool was created in Microsoft Excel 2007 under the file name IBC Hazard & Protection
Tool.xlsx. Screen shots of the tool User screens and a description of how the tool works is
provided in Appendix A.
As identified in Section 4.5, certain IBC operations scenarios may only require that fire
resistive containment and adequate separation distance be provided. These containment options
are discussed in Section 5.1 and in all cases, must be fire-resistive. In this analysis, fire-resistive
containment was expected to be capable of containing a burning pool of liquid for an extended
period of time (e.g., time required for fire department arrival). Both enclosed, and open-air
containment systems were evaluated. In addition to containment, one passive fire protection
method is proposed. This method requires containment but potentially no additional suppression
and is described in Section 5.2.2.6. For scenarios where containment was determined to be
insufficient, primarily scenarios involving non-alcohol based liquids, containment and
suppression systems are recommended. Suppression system options are discussed in Section 5.2.
In these scenarios, automatic detection is only required where separation distances are not
sufficient to prevent ignition within one minute of fire initiation.
28
HUGHES ASSOCIATES, INC
Table 11 –– Summar y of Pr otection Systems Considered for Design Fir e Scenar ios
Ceiling
Height
(m [ft])
Fuel
IPA
29
6 [20]
Mineral
Seal Oil
No.
IBCs
14
24
1
2
2
4
4
14
24
1
2
2
4
4
4
Configuration
Single
Side-by-Side
Single
Side-by-Side
Stacked
Side-by-Side
Stacked
Single
Side-by-Side
Single
Side-by-Side
Stacked
Side-by-Side
Fire Resistive
Containment
Only1,2
Passive Fire
Protection System
with FR
Containment Only
Hazardous
Materials
Storage
Locker
Local
Application
with FR
Containment
Adequate Separation
Distances Provided :
Type of Detection Needed
Yes
Potentially No Additional Protection Needed1,2
Required by Option:
No Detection Required
Yes
Potentially No Additional Protection Needed1,2
Required by Option:
No Detection Required
No
Potentially3
Yes
Yes
Stacked
HUGHES ASSOCIATES, INC.
1
Single
Yes
Potentially No Additional Protection Needed1,2
24
Side-by-Side
1
Single
Heptane
2
Side-by-Side
No
Potentially3
Yes
Yes
2
Stacked
4
Side-by-Side
4
Stacked
1 – Assumes that 'typical' (e.g., Ordinary Hazard Group 2) overhead sprinkler system is installed per insurer/fire code requirement
2 – Assumes adequate separation distances are provided for adjacent combustibles and personnel operation locations
3 – Requires verification testing to characterize passive fire protection provided by FOAMGLAS material
4 – Assumes an enclosed raised sump (example shown in Figure 6 [right])
No:
Local
Automatic
Detection
Necessary
Required by Option:
No Detection Required
Yes:
Manual
Activation
Adequate
Yes:
Manual
Activation
Adequate
No:
Local
Automatic
Detection
Necessary
Table 11 –– Summary of Protection Systems Considered for Design Fire Scenarios (Continued)
Ceiling
Height
(m [ft])
Fuel
IPA
30
9 [30]
Mineral
Seal Oil
No.
IBCs
14
24
1
2
2
4
4
14
24
1
2
2
4
4
Configuration
Single
Side-by-Side
Single
Side-by-Side
Stacked
Side-by-Side
Stacked
Single
Side-by-Side
Single
Side-by-Side
Stacked
Side-by-Side
Stacked
Fire Resistive
Containment
Only1,2
Passive Fire
Protection
System with FR
Containment3
Hazardous
Materials
Storage Locker
Local
Application
with FR
Containment3
Adequate Separation
Distances Provided :
Type of Detection
Needed
Yes
Potentially No Additional Protection Needed1,2
Required by Option:
No Detection Required
Yes
Potentially No Additional Protection Needed1,2
Required by Option:
No Detection Required
No
Potentially3
Yes
Yes
HUGHES ASSOCIATES, INC.
14
Single
Yes
Potentially No Additional Protection Needed1,2
24
Side-by-Side
1
Single
2
Side-by-Side
Heptane
2
Stacked
No
Potentially3
Yes
Yes
4
Side-by-Side
4
Stacked
1 – Assumes that 'typical' (e.g., Ordinary Hazard Group 2) overhead sprinkler system is installed per insurer/fire code requirements
2 – Assumes adequate separation distances are provided for adjacent combustibles and personnel operation locations
3 – Requires verification testing to characterize passive fire protection provided by FOAMGLAS material
4 – Assumes an enclosed raised sump (example shown in Figure 6 [right])
No:
Local
Automatic
Detection
Necessary
Required by Option:
No Detection Required
Yes:
Manual
Activation
Adequate
Yes:
Manual
Activation
Adequate
No:
Local
Automatic
Detection
Necessary
While identifying the systems considered in this assessment, one of the primary search
criteria used, other than protection performance, was the availability of the alternative system
and whether or not the system was Listed/Approved for this type of application. The intent was
to identify systems that were both Listed/Approved and available ‘off the shelf’ so that further
evaluation of selected systems could be geared more towards performance verification as
opposed to product development.
The extent to which systems were combined was dependent upon the system be considered.
Some suppression systems require that a separate detection system be installed to serve as a
means of activation (e.g., aerosol suppression). Other systems have integral detection
(e.g., overhead sprinklers). For certain design fire scenarios it is possible that containment alone
could be sufficient to achieve the protection objectives, provided that appropriate stand-off
distances and building fire protection (i.e., overhead sprinklers) are provided.
In general, conceptual designs were developed for the proposed local application systems.
These proposed systems were then evaluated against the worst-case design fire scenario
(i.e., Design Fire #7). A common design fire was used so that comparisons between expected
performances of the systems could be made. All local applications systems identified in this
assessment should have manual activation capabilities and the ability to locate these stations
remote from the discharging system (i.e., beyond the required personnel hazard separation
distances). The majority of the systems discussed, except in the case of the passive system, have
several commercial variants available.
Numerous fire suppression systems were identified in this assessment. However, certain
types of systems, like clean agents and carbon dioxide gaseous systems were not included.
These systems were excluded because of the lack of confinement in the scenarios considered.
This confinement is typically required to achieve extinguishment concentrations. It should also
be noted that for scenarios in which water miscible fuels are considered, the AFFF systems
would require the use of alcohol resistant (AR) AFFF.
5.1
Liquid Containment Systems
NFPA 30 [1] requires that in storage configurations where control of the spread of liquid is
required, a means to limit the spread of liquid to an area not greater than the design discharge
area of the ceiling sprinkler system shall be provided. Design discharge areas typically range
from 190–280 m2 (2000–3000 ft2). However, these containment areas are geared towards
storage occupancies in new construction. This is typically accomplished using trench or spot
drains that divide the floor of the storage area into rectangles having areas equal to or less than
the design area of the sprinkler system. Such an approach was not considered viable for most
existing structures. Consequently, the use of raised sumps, berms, or sills was considered as a
means of providing liquid containment for IBCs used in operations scenarios.
These systems are typically designed with a capacity to contain 10 percent of the volume of
containers or the volume of the largest container, whichever is greater [8]. They can be
constructed from both combustible and non-combustible materials. The fire hazard analysis
indicated that for fire protection purposes, the containment systems should have sufficient
capacity to contain all liquid contained with the IBCs (i.e., assuming release of all liquid from all
31
HUGHES ASSOCIATES, INC
IBCs) and the quantity of agent discharged during suppression activities. These systems were
also assumed to be constructed from non-combustible material with sufficient fire resistance to
prevent breach for an extended period of time. Currently, three general types of commercial spill
containment systems are available consisting of raised sumps, berm systems, and hazardous
materials storage lockers.
5.1.1
Raised Sumps
Raised sumps are probably the most common spill containment system used in operations
scenarios involving IBCs. A small-foot print relative to the size of the IBC and the flexibility
with respect to re-locating IBCs makes raised sumps a cost-effective means of providing liquid
containment. The intent of raised sumps is to contain an accidental spill from an IBC. These
sumps can be constructed to accommodate one or multiple IBCs. Raised sumps are typically
0.1–0.3 m (4–12 in.) larger than the IBC in all lateral directions. The height of the sump is
dependent on the expected containment capacity.
Raised sumps can be manufactured from a range of materials including both combustible and
non-combustible materials. Typical construction materials include, but are not limited to, high
density polyethylene (HDPE), molded fiberglass, steel, and stainless steel. For the purposes of
liquid containment, combustible materials are generally acceptable; however, from a fire hazard
standpoint, sumps would have to be non-combustible construction in order to maintain
containment during a fire scenario. Photographs of both plastic and metal raised sumps are
shown in Figure 6.
Figure 6 –– Photograph of plastic (left) and steel (right) raised sumps
The small footprint of the raised sump, while a benefit from a logistical standpoint, makes
the sump susceptible to liquid release overspray as described in Section 3.2.4. For standard
raised sumps, if a leak forms along a side wall of the IBC, the internal pressure created by the
liquid head could give the ejected liquid enough momentum to fall outside the containment area
of the sump. Such an event could result in a spill fire scenario outside the containment area if the
leaking liquid is ignited. To address this hazard, specially constructed raised sumps, as seen in
Figure 6 (right), can be constructed. These sumps use enclosure panels that extend upwards
32
HUGHES ASSOCIATES, INC
from the base of the sump to provide both impact protection to prevent damage to the IBC and
leak protection in the event that an IBC is punctured. The enclosed nature of the raised sump
shown in Figure 6 (right) requires that a hinged door be provided to allow IBCs to be
installed/removed. It also requires that sections of the enclosure be removed for liquid transfer
equipment.
In summary, the raised sump provides a relatively cheap, easy to install, mobile containment
system. However, a limitation of this type of system is that the fire resistance of the system is
not quantified and would require testing to verify that adequate containment could be maintained
during a fire event. The mobile nature of this system would also make enforcement of separation
distances challenging.
5.1.2
Fixed Berm/Raised Sill Systems
Fixed berm/raised sill systems were also considered as an option to mitigate the hazards
associated with a liquid release and resulting fire from an IBC. This containment strategy is
relatively commonplace and generally consists of a physical barrier serving as a perimeter
around the IBC(s) past which liquid cannot spread. Typically, these barriers are not as tall as the
raised sumps described in Section 5.1.1 and consequently the containment areas have a larger
footprint in order to accommodate the volume of liquid released in a worst-case scenario.
Depending on the operations scenario, these systems can be constructed to be permanent
using concrete forms or made to be temporary using large steel containment pans or specially
designed barrier systems. For existing facilities, the addition/construction of either option is not
expected to be cost prohibitive provided that fixed drainage to exterior storage systems are not
required. In other words, a plan for removing contained, accidentally spilled liquid would have
to be implemented. Currently, there are commercially-available concrete raised sill alternatives
consisting of modular systems that can be configured and re-configured in a variety of ways.
Access to these containment systems is generally accomplished using ramps.
In this assessment, the Denios Hazberm® spill barrier system was considered as an alternative
to the raised concrete sill approach. This system consists of 0.9 m (3 ft) polymer concrete
segments that are bolted to each other and to the floor on which they are being installed. The
segments are generally 0.45 m (18 in.) tall but can be designed to different heights. Once
mechanically attached, the seams of the berm are sealed using a specially selected chemical and
fire-resistant sealant. The material is identified as being both chemically- and fire-resistive;
however, the extent to which this system has been tested against pool fire scenarios is unknown
and additional testing would be required prior to recognizing this as a concrete alternative. An
example photograph of the Denios Hazberm® system is provided in Figure 7.
33
HUGHES ASSOCIATES, INC
Figure 7 –– Photograph of Denios Hazberm® system installed on concrete pad
The modular nature of the Hazberm® system makes this approach to spill containment
advantageous because it is relatively easily disassembled, re-configured, and re-assembled.
However, the ease with which a system can be installed and dis-assembled raises the concern as
to whether or not the berm is liquid tight. The integrity of the system would need to be checked
after installation and monitored over the lifetime of the system. Local protection systems are not
“modular,” so any fixed system design would have to be changed if and when the berms are
moved. Enforcement of stand-off requirements would also be an issue.
The fixed berm/raised sill systems provide a means of creating user specified containment
areas that can be constructed/re-constructed to adapt to a changing operations scenario. These
systems also provide a greater degree of overspray protection in that the perimeter of the system
can be determined based on the IBC configuration expected. Portable systems may be
susceptible to leaks given that they are modular and their resiliency to fire is yet to be
determined. The fire resistance of these types of systems would need to be verified via testing.
5.1.3
Hazardous Materials Storage Lockers
Hazardous materials storage lockers were also considered as a means of containing liquid
releases from IBCs being used in operations scenarios. Lockers of various sizes and shapes are
commercially available and can easily accommodate up to four IBCs. These lockers are
generally constructed from steel and can serve as both spill containment, via a raised sump
approach to liquid containment, as well as a berm system in that the walls of the locker mitigate
the hazards associated with overspray. These lockers also provide an additional level of fire
protection in that the lockers enclose the IBC. This enclosure limits the extent to which a fire
involving an IBC can grow as well as attenuating the radiant heat being emitted by the fire.
Another benefit of storage lockers is the relatively small footprint of the system relative to the
IBCs being stored. An example of a hazardous materials storage locker containing a single IBC
is provided in Figure 8.
34
HUGHES ASSOCIATES, INC
Figure 8 –– Photograph of Denios 2-hr fire-rated storage locker
These lockers can be constructed in a variety of ways with some being constructed simply to
provide shelter from the elements when in an outdoor environment and others being constructed
with UL-approved fire rated walls and integrated fire detection and suppression systems. The
lockers come with both hinged and roll-up doors that can be both manually or automatically
closed. The doors are UL-approved fire doors with a typical fire resistance rating of 1.5 hours.
The integrated fire suppression systems that may be installed in these storage lockers include:
clean agent, dry chemical, and sprinkler systems. The use of clean agent and dry chemical
systems allows for the storage locker to be relatively mobile allowing the location of the IBC to
change as the operations scenario changes. The alternative (i.e., installation of a sprinkler system
connected to the existing sprinkler system) would minimize the potential mobility of the locker.
A disadvantage of storage lockers is that, unless the design is modified, the door to the
storage locker must remain open while in use. Under this assumption, the benefits previously
described (i.e., fire containment, oxygen starvation, fire resistance, suppression) are potentially
compromised given that the system was designed as a closed system. The extent to which these
benefits are compromised is unknown and would have to be evaluated via an engineering
evaluation (i.e., is there fusible link release, can piping be accommodated) or testing. The
opening may substantially reduce the ability of the locker to protect neighboring combustibles
from the radiant heat of a fire and suppress the fire within.
35
HUGHES ASSOCIATES, INC
5.1.4
Summary of Liquid Containment Approaches
Three different liquid containment approaches were identified to address the hazard
associated with the release of liquid from an IBC being used in an operations scenario. The
approaches identified were: raised sumps, fixed berm / raised sill systems, and hazardous
materials storage lockers. A comparison of the advantages and disadvantages of these
approaches is provided in Table 12.
Table 12 –– Compar ison of Advantages/Disadvantages of Liquid Containment Systems
Relative
System
Cost1
0.3
Fixed
Location
No
Susceptible to
Overspray
Leak
Scenarios
Yes
Provides
Protection to
Neighboring
Combustibles
No
Additional
Detection /
Suppression
Potentially
Required
Yes
Inspection,
Testing, and
Maintenance
Minimal
Raised Sump
Berm System /
No
No
Yes
Minimal
0.4
Yes/No2
Raised Sill
Pre-fabricated
Storage Lockers
0.8
No
No
Yes
No
Moderate
w/o Suppression
Pre-fabricated
Storage Lockers
No
Yes
No
Moderate
1.0
No3
w/ Suppression
1
Costs were normalized with respect to the most costly approach (i.e., most costly approach will be equal to 1.0
with all other values being some fraction of this cost).
2
Raised concrete sills would be considered fixed location while modular berm systems (i.e., Denios Hazberm®)
would be considered not fixed.
3
Not fixed unless sprinkler system was used in place of clean agent/dry chemical suppression
5.2
Fire Suppression Systems
5.2.1
Traditional Suppression System Approaches
5.2.1.1 Overhead Water Sprinkler System
The use of traditional overhead water sprinkler systems is currently prescribed by NFPA 30
to protect Listed/Labeled composite IBC configurations. In non-stacked (i.e., one high)
configurations, NFPA 30 currently requires a water application rate of 18.5 Lpm/m2
(0.45 gpm/ft2) be delivered over a coverage area of 280 m2 (3000 ft2). The application rate for
stacked configurations (i.e., two-high) is 24.6 Lpm/m2 (0.60 gpm/ft2) over the same coverage
area. These application rates were developed through full-scale fire testing [9] geared towards
protecting non-metallic, Listed/Labeled IBCs in palletized storage configurations. The overhead
sprinklers protecting these types of IBCs are permitted to protect a maximum of 9.3 m2 (100 ft2)
per sprinkler for Class II and IIIA liquids and 11 m2 (120 ft2) per sprinkler for Class IIIB liquids.
The primary limitation in specifying this system is its effectiveness for a non-listed container
in an unbermed/uncontained situation. Tests with mineral oil indicate that containment would be
required with overhead water systems [10]. However, these water application rates result in a
36
HUGHES ASSOCIATES, INC
substantial volume of water being discharged. This large volume of water, while potentially
suppressing the burning fuel and limiting the extent to which neighboring combustibles become
involved, also poses the risk of overfilling a containment area if not properly accounted for. In
this case, any suppression liquid discharged in the vicinity of the containment area may result in
the containment system being overwhelmed and instead of a contained pool fire, a running fuel
fire spreading across the floor of the facility could result.
Preventing such a scenario would require that the design of the containment area to capture
the required volume of liquid released from the breached container(s) plus the volume of water
discharged into the containment area by the suppression system. A ten minute discharge
duration is reasonable for design purposes [1]. For the overhead water sprinkler systems
discharge densities described above (i.e., 18.5 and 24.6 Lpm/m2 [0.45 and 0.60 gpm/ft2]),
berm/sill heights would need to be increased by approximately 0.2 m (8 in.) and 0.25 (10 in.),
respectively. Lower water rates, i.e., those associated with ordinary hazard Group 1/2 design,
would require less containment volume.
Consequently, while the traditional overhead sprinkler system is the easiest and most
common-place, the efficiency of the system in suppressing and containing the hazards associated
with non-Listed IBC fires in operations scenarios is marginal. The potential overflow of the
containment area plus limited capacity of existing sprinkler systems limits the practicality of this
approach.
5.2.1.2 Overhead Foam-Water Sprinkler System
In addition to traditional sprinkler systems, NFPA 30 also recognizes the use of overhead
foam-water sprinkler systems to protect Listed/Labeled composite IBCs in storage
configurations. Based on testing, AFFF might be used without a berm [10], or a lower rate used
in conjunction with a berm.
Based on the relatively rapid rate at which control/extinguishment of the test fires was
achieved, it is possible that a reduced application rate could be viable. This reduced application
rate would require full-scale test verification. It is possible that application rates in the range of
4.4–13 Lpm/ m2 (0.10–0.30 gpm/ft2) could be adequate to control/suppress the fire hazard.
These application rates are used in NFPA 11 [11] to protect liquid fuel fire hazards and in
NFPA 16 [12] (i.e., 7.0 Lpm/ m2 [0.16 gpm/ft2]) as minimum application rates for foam-water
sprinkler/spray systems. Spacing requirements identical to those specified for the traditional
warehouse water sprinkler systems are required for the foam water systems as well. Foam-water
sprinkler systems are also required to have at least 15 minutes of foam concentrate, based on the
required design flow rate.
In general, the comments made in the previous section (Section 5.2.1.1) regarding the
hazards associated with relatively high application rates and the potential for overfilling
containment areas are applicable to overhead foam-water sprinkler systems as well. These
systems also require that the liquid fuel used in the operations scenario be defined at the time of
the system design. Foam-water sprinkler systems protecting alcohol-based fuel require that
alcohol resistant (AR-AFFF) be used in order to adequately protect against the liquid fuel fire
hazard.
37
HUGHES ASSOCIATES, INC
5.2.2
Novel Suppression System Approaches
5.2.2.1 Local Application Foam System
Based on the relative success of the foam-water sprinklers when installed in an overhead
configuration and the use of firefighting foam in numerous liquid fuel fire scenarios, a low-level
foam suppression system was considered. It is expected that a local application system would
improve the efficiency with which the foam is applied to the fuel surface and consequently
reduce the application rate required to achieve suppression. With nozzles installed near the fuel
surface (i.e., berm-mounted nozzles), the losses typically associated with overhead suppression
agent having to penetrate the fire plume in order to reach the fuel surface would be reduced.
This reduction would result in a larger fraction of the discharged foam being delivered to the fuel
surface.
The concept of low-level application of AFFF is described by Scheffey [13]. The concept
was originally developed for the application of AFFF on aircraft carrier flight decks through
flush deck water wash down nozzles. These nozzles come up through the flight deck and
discharge in a relatively low pattern. These non-aspirated nozzles have been demonstrated to
achieve extinguishment of JP-4 and JP-5 pool fires at application rates as low as
1.61–2.41 Lpm/m2 (0.04–0.06 gpm/ft2). AFFF is effective at these low rates since the foam does
not have to penetrate through the fire plume. This concept has been extended to protect: military
aircraft hangars with low level monitor nozzles; aircraft hangars with flush deck nozzles installed
in floor trench drains; and shipboard special hazard areas using edge-mounted spray nozzles.
Currently, there are no Listed/Approved systems for this type of installation. However, for
the purposes of this assessment an example system design was developed. A low-level AFFF
system could be installed around the perimeter of the IBC operations area. In scenarios where
the hazard being protected is an alcohol or alcohol-based liquid, alcohol resistant AFFF
(AR-AFFF) could be used.
For an operations scenario, edge mounted nozzles could be installed around the perimeter of
the protected area. Figure 9 shows an example protection scheme for Design Fire #7. A design
application rate of 2.41 Lpm/m2 (0.06 gpm/ft2) of AFFF should be sufficient based on the data
provided by Scheffey [13]. A design rate of 4.03 Lpm/m2 (0.10 gpm/ft2) is used to include a
factor of safety. At this flow rate, in order to provide a minimum of ten minutes of discharge, a
supply of at least 1500 L (400 gal.) of pre-mixed AFFF solution would be required. Locating the
premixed supply near the operations area, as shown in Figure 9, could be done to minimize the
amount of piping needed to install the system. The tank and manual actuation device would
have to be located outside of the threat zone (see Table 10 or the engineering tool developed in
this work). The supply tank could be pressurized with nitrogen and discharge through open,
edge-mounted nozzles protecting the perimeter of the area. The system would be activated either
manually or by a local detection system.
The low-level AFFF foam system shown in Figure 9 consists of four Bete™ NF100 nozzles
[14] positioned around the perimeter of the containment area. These nozzles, when operating at
2.8 bar (40 psi) will each deliver 37.8 Lpm (10 gpm) providing an average application rate of
approximately 4.03 Lpm/m2 (0.10 gpm/ft2) over the containment area. The discharge spray from
38
HUGHES ASSOCIATES, INC
these nozzles is a 120 degree flat pattern, as shown in Figure 9. The pattern was selected to
provide the largest area of coverage while minimizing the number of nozzles needed.
While this system should be effective on the two-dimensional pool fire, it will not extinguish
any three-dimensional, running fuel fire associated with an IBC leak/rupture.
Figure 9 –– Example low-level AFFF foam system protecting Design Fire #7
Although low-level AFFF foam systems are often used to protect against liquid fuel fire
scenarios, currently there is not a Listed/Approved system designed specifically for the scenarios
described above. However, this system could be readily fabricated and tested in order to perform
verification tests and ultimately developed by a vendor for commercial availability.
An alternative foam system was also considered in this assessment. The alternative system
would consist of an array of overhead nozzles mounted directly above the contained IBCs,
similar in construction to that shown in Figure 10. Foam solution would be discharged through
an array of air aspirated foam-water nozzles (e.g., Ansul Model B-1 nozzles). Since these
nozzles would be mounted in close proximity (i.e., lower elevation than that discussed in
Section 5.2.1.2) to the IBC hazard, it is possible that lower application rates could be used. The
system could be supplied using a pre-mixed system as described for the low-level AFFF system
or could be supplied by the building water supply accompanied by an appropriately sized foam
concentrate tank and proportioning system.
39
HUGHES ASSOCIATES, INC
Figure 10 –– Example of fixed pipe CAFS systems protecting Design Fire #7
5.2.2.2 Compressed Air Foam System (CAFS)
In addition to the non-aerated foam systems described above, a compressed-air foam system
(CAFS) was also considered as a possible means of protecting IBCs in operations scenarios.
Compressed air foam is created by injecting air/nitrogen into an AFFF foam solution. This
forced aeration results in a small-bubbled, stable, uniform foam that can be discharged in a
relatively high momentum jet. The stability of the agent combined with the high energy
discharge streams reportedly allows for better penetration of the fire plume and more effective
blanketing the burning fuel surface. When compared to aspirated foams, compressed air foam
has a more controllable bubble structure which supposedly improves fire extinguishing
performance. The injection of compressed air into the foam solution creates smaller, more rigid
bubbles giving the foam longer drain times. The advantages of CAFS compared to non-aspirated
AFFF are:
•
Reduced water demand;
•
Reduced foam concentrate demand;
•
Potential for enhanced fire plume penetration providing faster knock-down;
•
Potential for improved burn-back resistance; and,
•
Listed systems available for local application scenarios.
However, these benefits come with associated costs resulting from the complexity of the
system and the foam structure within the pipe network. The CAFS piping systems must
40
HUGHES ASSOCIATES, INC
accommodate a two-phase flow scenario (i.e., the injected air/nitrogen and the foam/water
solution). The piping systems must be calculated to be balanced so that two-phase flow is
accommodated. For this reason most systems are pre-engineered. Two-phase flow friction is
greater than single-phase friction for the same conduit dimensions and mass flow rate. The
increased fiction losses result from increased flow speeds and require that the piping be of a
larger diameter than if a single phase flow system were to be used. In addition to this complexity
within the piping network, the system supplying the piping network is also complex and requires
additional floor area to accommodate the supply equipment as well as additional manpower to
perform inspection, testing, and maintenance activities.
Currently, there are several fixed pipe compressed air foam systems that are approved under
FM Approval 5130 for Foam Extinguishing Systems which requires the system to extinguish a
4.7 m2 (50 ft2) pool fire with nozzles at the furthest spacing. In this assessment, the fixed pipe
CAF system provided by FireFlex Inc. was used to provide an example system that could be
used to protect unlisted IBCs in operations scenarios [15].
The FireFlex fixed pipe CAF system is a deluge type system that can be activated both
manually as well as automatically based on the alarm from a local detection system. The supply
system (i.e., concentrate tank, proportioner, compressed gas supply, and all controls) are
manufacturer assembled and tested in cabinets or skids. The system requires water supply
pressures between 3.4–12 bar (50–180 psi) and air pressure supplied by cylinders at 170 bar
(2400 psi). The concentrate supply tank provided is sufficient for discharge duration of ten
minutes.
Areas of coverage for this system are 9.3 and 14 m2 (100 and 150 ft2) per nozzle for
alcohol/ketone and hydrocarbon fuels, respectively. The minimum design application rate is
also fuel dependent and is 1.6 Lpm/m2 (0.04 gpm/ft2) for hydrocarbons and 2.3 Lpm/m2
(0.06 gpm/ft2) for alcohol/ketones. System nozzles must be at least 2.4 m (8 ft) above the
protected surface and have a maximum height of 14 m (45 ft) for hydrocarbons and 11 m (35 ft)
for alcohol/ketones.
Based on these design specifications, protecting a 42 m2 (450 ft2) containment area
(i.e., Design Fire #7) with a hydrocarbon fuel would require only three nozzles and a minimum
of 570 L (150 gal.) of foam solution to achieve extinguishment. However, to achieve a more
uniform application of foam throughout the containment area as well as provide a factor of
safety, it is recommended that an additional nozzle be added, as shown in Figure 10. The
addition of the nozzle would increase the minimum foam solution to 760 L (200 gal.).
CAFS, which uses AFFF concentrate, has the same limitations on three-dimensional fires as
the AFFF low level system. The overhead approach may provide some additional protection.
The storage/control system would have to be located outside the threat zone.
5.2.2.3 Pre-Engineered Dry Chemical System (Local Application)
Dry chemical fire suppression systems are commonly used in petroleum and petrochemical
loading racks, refinery processing equipment, product transfer and storage areas and large
vehicle engine enclosures. Suppression is achieved by coating the surface of the burning
41
HUGHES ASSOCIATES, INC
material and, primarily, reducing the concentration of oxygen in the vicinity of the fire. There
are two types of dry chemical suppression systems, a total flooding system and a local
application system. However, the open-air environment in which the unlisted composite IBCs
are used prevents a total flooding system from being a viable option. Only local application
systems were considered.
Requirements for the installation, operation, suppression performance, and maintenance of
Class B local application protection systems are outlined in the UL 1254 [16] and NFPA 17 [17].
Class B local application systems Listed under UL 1254 are required to extinguish a heptane
pool fire after a 30-second pre-burn. These systems are tested at both the minimum and
maximum discharge heights under maximum and minimum extinguishment agent flow
conditions, respectively. In this Listing, the size of the pan fire corresponds to the protection
area limitations for the system being evaluated. During suppression testing the discharging
nozzle cannot splash the burning fuel out of the test pan. In both tests, the protection system
must completely extinguish the test fire. For local application systems the suppression nozzles
discharging the dry chemical can be mounted tank side or above the fuel being protected.
Local application dry chemical suppression systems are typically activated pneumatically
using a nitrogen cartridge that is triggered by an electrical impulse from either a manual
discharge switch or local detection system. For local applications, a local thermal detection
system (e.g., fusible links) are typically installed in the expected fire area and used to trigger the
nitrogen cartridge.
For the purposes of this assessment and evaluating the potential advantages and
disadvantages of different types of fire protection systems, a commercially-available Class B
local application protection system was selected and a system design was developed. The
system selected was done so at random, and does not necessarily indicate the type of system that
should be specified for future testing if dry chemical systems are identified as a viable fire
protection option.
The system selected was a Kidde Model IND-50 dry chemical fire suppression system. The
UL 1254 [16] listing (EX2153) for this system indicates that it is a stored pressure type
extinguishing system unit designed to discharge Kidde-Fenwal sodium bicarbonate from fixed,
open discharge nozzles. The system is Listed for the extinguishment of Class B fires by tankside or local overhead applications. The recommended design minimum for fully exposed
hydrocarbon spill/diked fires is 0.15 kg/s-m2 (0.03 lbs/s-ft2). This design criterion was based on
correlation of test data on dry chemical application rates required to extinguish liquid
hydrocarbon spill and diked fires.
The IND-50 system has a 23 kg (50 lbs.) system capacity and can supply up to four nozzles
protecting an area of 27 m2 (290 ft2). Based on these design parameters, in order to protect a
42 m2 (450 ft2) containment area containing a hydrocarbon fuel (e.g., heptane), two IND-50
systems would be required. The nozzles for these systems could be evenly distributed around the
perimeter of the containment area, as shown in Figure 11, in order to provide uniform coverage
throughout.
42
HUGHES ASSOCIATES, INC
Figure 11 –– Illustration of potential dry chemical suppression system installation
In order to reduce the footprint for this type of pre-engineered system, the dry chemical
cylinders could be located near to the containment area but outside the threat zone. This location
would reduce the size of the piping network and prevent the piping from impeding movement
around the containment area. Berm-mounted nozzles would allow for easy access inside the
containment area for IBC maintenance or replacement. The inspection, testing, and maintenance
procedures associated with the use of a dry chemical suppression system include the semi-annual
inspection of all system components and verification that suppression agent quantities are
sufficient.
The advantages of the dry chemical suppression system, if used to protect IBCs in operations
scenarios, include the fact that Listed systems are available and generally easy to install and
maintain. These systems can also be installed with relative ease and have a relatively small
footprint. One limitation of dry chemical systems with respect to this application is the
susceptibility of the system to re-flash (i.e., re-ignition of the liquid fuel after initial
extinguishment). Another disadvantage is that in the event of a discharge, the dry chemical
agent, once airborne, can migrate throughout the facility. This could, depending on the process,
result in collateral damage / loss of product in areas not intimate with the fire.
43
HUGHES ASSOCIATES, INC
5.2.2.4 Aerosol Suppression System (Local Application)
Aerosol-based suppression systems were also considered as an alternative suppression
system to protect IBCs in operations scenarios. These systems typically consist of a propellant
that, once ignited, burns and produces an ultra-fine potassium-based aerosol at very high
generation rates. Extinguishment of the fire hazard is achieved exclusively in the vapor phase
through a combination of oxygen displacement and chemical interference between the ultra-fine
aerosol particulate and the free radicals of the flame. The chemical interference may be more
efficient with aerosol suppression systems due to the smaller particles that are generated during
discharge. The fact that suppression is achieved exclusively in the vapor phase means that the
suppression performance of these systems is independent of fuel type and can be used for
hydrocarbons, polar solvents, alcohols, etc.
As an example case in this assessment, the Stat-X fire suppression system manufactured by
Fireaway was selected. These systems are typically discharged via thermal detection or manual
activation. For scenarios in which multiple aerosol generators are being used to protect a
common hazard, all generators can be activated simultaneously using an integrated, miniature
fire alarm control panel (FACP). In this case, thermal detection is accomplished using a linear
heat detection system which communicates with the FACP. Once an alarm is received, the
FACP sends an electrical impulse to the generators to activate suppression. Once activated the
suppression agent is dispersed within 8–37 seconds depending on the size of generator used.
Typical applications for these types of systems include the protection of electronics cabinets,
flammable liquid storage, hazmat storage, and marine engine rooms. It should be noted that in
all of these applications, the aerosol is being dispersed into an enclosure. This confinement
allows a relatively high concentration of aerosol to be maintained in the vicinity of the fire
thereby aiding with suppression. Consequently, for the scenarios being considered in this
assessment (i.e., open-air fire scenarios), suppression is likely to be achieved with higher aerosol
concentrations (i.e., more generators will be required).
Currently, there are several manufacturers of aerosol-based suppression systems. While
there are no Listed systems for local application, there are Listed systems for total flooding
applications. The Stat-X product is Listed under UL 2775 [18] for total flooding applications. A
photograph of the Stat-X generator is provided in Figure 12.
The Stat-X generators range in size from 30–2500 g (0.07–5.5 lbs) with coverage areas
ranging from 0.5–23.8 m2 (5.4–256 ft2), respectively. As noted earlier, these coverage areas are
based on enclosed environments and will most likely have to be reduced given the open-air
scenario being considered. Thermally activated units have activation temperatures of 70, 95, and
123oC (158, 203, and 254oF) but are only available with aerosol masses up to 500 g (1.1 lbs).
Electrically-activated units are available in sizes up to 2500 g (5.5 lbs.). These units require a
minimum of 1A electrical pulse for at least 50 milliseconds in order to initiate the discharge of
the aerosol.
44
HUGHES ASSOCIATES, INC
Figure 12 –– Photograph of thermally-activated Stat-X aerosol generators [19]
Based on these design parameters and a safety factor of two to account for the lack of
confinement around the containment area being protected, an aerosol suppression system capable
of protecting a 42 m2 (450 ft2) area containing a hydrocarbon fuel (e.g., heptane) would require a
minimum of four 2500 g (5.5 lbs.) aerosol generators and two actuators. The actuators would be
interconnected and would each activate two generators. Due to the size of the generators
required and the fact that there are multiple generators that would have to be activated
simultaneously, electrically-activated generators were selected. Generators would be installed in
the corners of the containment area as shown in Figure 13. These systems would not require any
additional piping, making installation and replacement costs minimal.
Figure 13 –– Illustration of potential aerosol suppression system installation
45
HUGHES ASSOCIATES, INC
Aerosol-based suppression systems are also virtually maintenance free and have a shelf life
of over 10 years. When coupled with their very low installation cost, this makes them an
extremely cost effective fire protection solution. The generators can also be berm-mounted
which provides easy access to the containment area when IBC maintenance/replacement
activities are required. The primary disadvantage of these systems at this time is the unknown
fire suppression performance when discharged in an open environment. It is possible that when
discharged in the open, into a highly turbulent fire plume, the aerosol will be entrained into the
plume and immediately carried away from the burning fuel surface which would minimize the
effectiveness of the aerosol in suppressing the fire. These scenarios are currently being
evaluated by the Gas Technology Institute to protect gas workers in gas line trenches but this
testing has not been completed nor results released to the public. Re-flash/burn-back protection
is also likely to be limited. Another potential disadvantage of the aerosol suppression systems,
when used in an indoor, open-air environment (i.e., operations scenario) is that in the event of a
discharge the fine particulate can migrate throughout the facility which could, depending upon
the process, result in collateral damage / loss of product in areas not intimate with the fire.
5.2.2.5 Local Application Water Mist System
Although water mist suppression systems are generally more efficient when installed in
enclosed areas, these types of systems, when properly designed and installed, can serve as an
effective local fire suppression technique. In the absence of confinement to contain the water
vapor being generated by the interaction of the fire and water mist, the system nozzles must be
oriented such that the water being discharged is directed at the seat of the fire. In this type of
application (i.e., local application), the primary mechanisms for extinguishment are gas phase
cooling and/or the wetting and cooling of the fuel surface.
NFPA 750 [20] classifies water mist systems into two media system types: single fluid and
twin fluid. Single fluid systems are defined as using a single piping system to supply each
nozzle. Twin fluid systems are defined as systems in which water and atomizing media,
typically air or nitrogen, are separately supplied to and mixed at the water mist nozzle. Typically
twin fluid systems are utilized to take advantage of the low pressure water supply requirements
of these systems. Twin-fluid, low-pressure water mist systems can be designed such that facility
water and air supplies can be utilized, eliminating the need for water tanks and compressed gas
bottles and enabling suppression efforts to continue for extended periods of time. Although
single fluid systems can be used in both low and high pressure applications, these systems are
typically designed for high pressure allowing for reduced pipe/tubing diameters and more
efficient atomization of the suppression water at the nozzle.
For this assessment, a low pressure system was considered. The system selected was the
FIRE-SCOPE 2000® [21] manufactured by Securiplex LLC. This system uses twin-fluid
technology and large water pathway nozzles to generate water mist. The system is typically
supplied as a fully-integrated system including control, detection, and extinguishment but can be
integrated with existing detection systems. The FIRE-SCOPE 2000® system was selected
because it has an FM-approval for local application use in accordance with Appendix K of
FM 5560 [22]. The local application fire tests for FM-approved water mist systems consist of
both liquid fuel pan and spray fires in both obstructed and re-flash conditions. The systems are
evaluated at minimum flow rates and maximum spacing.
46
HUGHES ASSOCIATES, INC
The FIRE-SCOPE 2000® system operates at water pressures of 5.2–5.9 bar (75–85 psig) and
require an air supply at 6.6 bar (95 psig). Although system designs for the specific scenarios
being considered in this assessment have not been developed, for similar scenarios (i.e.,
conveyor belt systems using oil quenching tanks), a nozzle separation distance of 1.2 m (4 ft)
was adopted. Based on this information an example system design for the 42 m2 (450 ft2)
containment area scenario (i.e., Design Fire #7) was developed. An illustration of this system is
provided in Figure 14. Using a nozzle spacing of 1.2 m (4 ft), a total of 20 nozzles would be
required to protect the containment area calculated for Design Fire #7. This system would
require a minimum of water supply of 98 Lpm (26 gpm) and a minimum air supply of 0.1 m3/s
(204 scfm).
Figure 14 –– Illustration of twin-fluid, low pressure water mist system based generally on
FIRE-SCOPE2000® design parameters
The advantages of the water low pressure water mist system described above include the fact
that the system identified is FM-approved for local applications, the system is available in a selfcontained package, a minimal amount of water that is discharged during suppression, and the
ability of the system to potentially utilize the facility water and air supply which allows for
extended discharge durations, if needed. However, the twin-fluid system is relatively complex in
that dual sets of piping and supplies must be installed and maintained. Even though the system is
FM-approved, full-scale fire testing should be conducted to verify the suppression capabilities of
the system for the hazards identified in this assessment.
47
HUGHES ASSOCIATES, INC
Other water mist systems, with overhead nozzles or high pressure nozzles, might also be
considered for this application.
5.2.2.6 Passive Pool Fire Suppression
In addition to the active fire suppression systems discussed above, a passive fire suppression
method was also explored. This approach was defined as being passive because it provides a
means of mitigating the development of the liquid fuel fire without requiring any action to occur.
This method consists of using an inorganic, closed-cell, cellular glass insulation material to
effectively reduce the total area of fuel available for combustion. This approach provides a
relatively simple and reliable form of passive fire protection. When used in conjunction with a
spill containment system, it provides a means of reducing the burning area and burning rate of a
pool fire which in turn reduces the thermal insult produced by the fire. The glass insulation is
installed within the containment area. In the event of a liquid release, the material will float on
top of the liquid surface to form a solid foam layer.
This solid foam layer provides a number of potential fire hazard reducing benefits including
the following:
•
Reducing the extent to which a fuel, if volatile at ambient conditions, can evaporate;
•
Reducing fire intensity by reducing available surface area for combustion; and
•
Providing an instantaneous, almost pre-emptive, suppression of the fire.
The cellular glass insulation also has several logistical and maintenance benefits when
compared to conventional means of fire protection, including:
•
Lightweight material allows for easy installation and removal;
•
Material can be re-used because closed cell form prevents absorption of liquid/vapor;
•
Relatively temperature resistant (i.e., thermal degradation does not occur until
approximately 730oC [1346oF]); and
•
Negligible maintenance, other than ensuring a relatively uniform layer is maintained.
There is currently only one commercially available product in the United States that is
marketed for the purpose of pool fire suppression. The product, called FOAMGLAS® PFS, is
shown in Figure 15. It is manufactured by the Pittsburgh Corning Corporation. An overseas
company, Hasopor, was also identified as being a manufacturer of lightweight foam glass
aggregate but a commercial product is currently not available from this company. Consequently,
the design information and testing data presented for this passive fire suppression approach was
solely based on data for the FOAMGLAS® PFS.
FOAMGLAS® PFS was developed and tested for use in the liquefied natural gas (LNG)
industry to mitigate the thermal hazards to neighboring objects associated with fires in
impounding areas. Original testing by Lev [23] investigated LNG fires in both a 2.5 m2 (27 ft2)
and a 36 m2 (400 ft2) bund scenario. Approximately 100–250 mm (3.9–9.8 in.) of
FOAMGLAS® was used to line the bottom of the bund. The bund was then filled with LNG
and ignited. The mass burning rate and radiative heat fluxes from these fires were compared to
48
HUGHES ASSOCIATES, INC
that measured for a free-burning LNG fire of the same size. Data from these tests show that both
the mass burning rate and radiative heat flux were reduced by more than ninety percent with the
cellular glass layer in place. Heat flux data measured in these tests for two different cellular
glass layer thicknesses and a representative free-burning scenario is presented in Figure 16.
Figure 15 –– Photographs of ‘raw’ cellular glass contained within a pan of heptane [24] (left) and
FOAMGLAS® product installed within a sump [25] (right)
12
11
Free-Burn_1
100mm FOAMGLAS
Free-Burn_2
250mm FOAMGLAS
10
2
Heat Flux (kW/m )
9
8
7
6
5
4
3
2
1
0
0
100
200
300
400
500
600
Time (s)
Figure 16 –– Measured radiative heat flux from LNG pool fires burning with and without
FOAMGLAS installed within the pool [23]
In addition to these tests, more recent testing conducted at the University of Texas A&M
Brayton Fire Training Field showed similar performance of the material for larger size fire
scenarios [26]. In these tests, the fire suppression performance of the material was studied for
LNG pool fires ranging in size from 2.5–65 m2 (27–700 ft2). This testing showed that the
49
HUGHES ASSOCIATES, INC
material could provide consistent coverage which stabilized the fire with no fluctuations during
suppression efforts.
Despite the numerous benefits and supporting data listed above, the use of cellular glass
material within containment areas does not provide a means of complete suppression. It only
provides a means of mitigating the hazards associated with the fire. The effectiveness of
suppressing a 3-D fire is unknown but likely limited. However, with the substantial reduction in
burning rate and thermal radiation, it could be possible that existing fire suppression systems
(i.e., overhead sprinklers in accordance with NFPA 13 [27] or NFPA 30 [1]) could be sufficient.
Such a reduction would require testing to verify this possibility. The performance of the cellular
glass over the lifetime of the product would need to be determined.
The installation of cellular glass may also cause logistical issues for the operations scenarios
since IBCs being protected will need to be exchanged periodically. This exchange process may
require that personnel and/or machinery enter the containment area and have access to the
immediate area around the IBC in order to remove/replace the existing container with a fresh
container. If passive protection via cellular glass is used, a support structure (e.g., open-mesh
metal grating, etc.) may need to be installed above the cellular glass. This should support the
load from the personnel/equipment and assure the cellular glass is not damaged. An additional
limitation of the use of FOAMGLAS is the potential high cost associated with providing a
100–250 mm (4–10 in.) deep layer within the containment areas, some of which were determined
to be as large as 42 m2 (450 ft2).
5.2.3
Summary of Fire Suppression System Approaches
A total of seven active and one passive fire suppression systems were considered as
candidates for the protection of unlisted IBCs being used in operations scenarios. Of the seven
active systems, two were traditional water-based overhead suppression systems, three were novel
approaches using existing water-based suppression system technologies, and two were novel,
non-water-based alternatives. The passive fire suppression system considered was a closed-cell
cellular glass material used to minimize fuel combustion and fire radiation.
Two of the seven active fire suppression systems considered were overhead systems. Due to
their location, these systems offer the benefit of minimal impact on logistics within the facility
being protected. In general, these overhead systems are at a disadvantage relative to suppression
efficiency (i.e., volume of agent needed to suppress a fire). These are also impractical in retrofit
scenarios because they require high volume flows which are probably not readily available. The
remaining five active suppression systems were all local application approaches, three waterbased and two, non-water based. Given the proximity of these systems to the fire, they typically
require lower agent application rates but can have an impact on logistics in and around the area
being protected. A summary of these systems and the advantages and disadvantages of each are
provided in Table 13.
50
HUGHES ASSOCIATES, INC
Table 13 –– Compar ison of Advantages/Disadvantages of Fir e Suppr ession Systems
Protection
against
Re-ignition /
Burn-back
Obstructs
Movement
in
Containme
nt Area
LowLevel
Piping
Req.
Insp.,
Testing,
and
Maint.
Requires
Local
Detection
System
UL Listed
or FM
Approved
System
Requir es
Dev. or
Veri.
Testing
51
Water
Usage
Efficiency1
Suppr ession
of 3D Fire
Inherent
Protection of
Neighbor ing
Combustibles
Overhead
Sprinkler
1.00
No
Yes
No
No
No
Minimal
No
Yes
Veri.
Overhead
AFFF
0.75 or less
No
Yes
Yes
No
No
Moderate
No
Yes
Veri.
Low-level
Application
of AFFF
0.17
No
No
Yes
Yes,
Minimal
Yes
Moderate
Yes
No
Dev.
Compressed
Air Foam
0.11
No
Yes/No
Yes
No
Yes,
Minimal
Moderate
Yes
Yes
Veri.
No
No
Yes,
Minimal
Yes
Moderate
Yes
Yes
Veri.
N/A
Potentially
yes, Reflash
possible
Potentially
yes, Reflash
possible
No
Unlikely
No
No
Very Low
Yes
No
Dev.
PreEngineered
Dry
Chemical
Aerosol
Generators
HUGHES ASSOCIATES, INC.
Local
Yes, for
Yes,
Yes,
Application
0.062
No
duration of
Yes
Moderate
Yes
Yes
Veri.
potentially
Minimal
Water Mist
discharge
Passive Pool
Essentially
Fire
N/A
No
Yes
N/A
Yes
N/A
N/A
N/A
Veri.
None
Suppression
1 – Water usage efficiencies were normalized with respect to the system with the highest liquid volume discharge (i.e., overhead sprinkler system) over ten
minutes duration. Over this duration, the sprinkler system would have discharged approximately 8700 L (2300 gal.) over a 35 m2 (380 ft2) coverage area.
2 – Water mist water usage efficiency based on 5 minute discharge duration.
Dev. – Development
Veri. - Verification
As shown in Table 13, these systems were compared using three main criteria: the
suppression capabilities of the system, the impact of the system on ‘typical’ operation scenario
activities, and availability/viability of the system. With respect to suppression capabilities the
areas considered included the following:
–
Water Usage Efficiency – used to compare the performance of water-based systems
based on the volume of water required to suppress a given fire.
–
Suppression of 3D Fire – assess whether the suppression system has the capability
to extinguish running fuel fire and/or burning IBCs.
–
Protection of Neighboring Combustibles – assess whether or not the system provides
‘wetting’ or radiation attenuation to minimize thermal exposure to adjacent
combustibles.
The impact of the fire suppression system on the ‘typical’ operations scenario activities was
evaluated based on:
–
Obstructs Movement in Containment Area – installation of suppression system would
limit/prevent access to the IBC or movement within the containment area.
–
Low Level Piping Required – installation of the suppression system would require the
installation of piping that would run across the facility floor thus potentially obstructing
logistics within the facility.
The availability/viability of the fire suppression system was evaluated based on whether or
not the system was UL Listed/FM Approved and whether the system was designed for this type
of application. In this case, verification tests might be performed to demonstrate system
capability on the IBC operations scenarios. Other non-Listed/Approved systems would require
developmental tests to establish appropriate design parameters.
5.3
Fire Detection Systems
Some of the suppression options identified in Section 5.2 require that a fire detection system
be installed in order to initiate the discharge of the suppression agent. The fire growth rates
associated with liquid release fire scenarios involving these liquids can be relatively rapid. If
suppression systems are not activated soon after ignition, the threat to the overhead structure and
neighboring combustibles can become severe and the design area for the system could be
overcome, resulting in a system failure. This consequence becomes less severe if an
appropriately sized containment area is installed and the suppression system is designed
accordingly.
In selecting a fire detection system to protect IBCs in operations scenarios, the following
performance objectives were considered: provide early warning detection to minimize the
thermal threat to the overhead and neighboring combustibles where appropriate and minimize the
frequency of false alarms without decreasing the system sensitivity. For the purposes of this
assessment, three different means of detection were considered. These included thermal, optical,
liquid, and manual detection.
52
HUGHES ASSOCIATES, INC
The type of detection identified for a given operations scenario was dependent on whether or
not adequate separation distance is provided between the IBC containment area and neighboring
combustibles/safety control systems. The type of detection system selected was also dependent
on whether or not the IBC scenario threatens the steel overhead. In general, local, automatic
detection would be required if inadequate separation distance is provided or the fire scenario
poses an immediate threat to the steel overhead. Manual detection would be required if adequate
separation distance for a one minute exposure is provided and the safety control system is
located with an appropriate separation distance. Finally, no detection system would be required
if the fire scenario presents no threat to the steel overhead and neighboring combustibles for a ten
minute exposure.
Each method of detection considered offers both benefits and disadvantages with respect to
their installation, detection performance, and inspection, testing, and maintenance (ITM) over the
lifetime of the system.
5.3.1
Thermal Detection Systems
In this assessment, thermal detection will be discussed from both area detection as well as
local detection standpoint. In these discussions, area detection was considered as detection
equipment that is remote from the fire (i.e., overhead installations) and responds primarily to
heating from the combustion gases produced by the fire. Local detection was considered as
devices installed in the expected area of the fire thereby being exposed to direct flame
impingement and radiant heating.
5.3.1.1 Overhead Thermal Detection
Traditional means of thermal fire detection, when considering large, open-air, industrial/
operational scenarios, consists of overhead thermal detection via the frangible bulb or thermal
link that is integral to the overhead sprinkler heads. Given their location in the overhead, these
types of systems are generally preferred / adopted because they are less prone to nuisance
activation and the associated inspection, testing, and maintenance (ITM) is minimal when
compared to other fire detection technologies. However, when compared to other means of fire
detection, overhead thermal detection can be substantially slower at detecting a fire.
Currently, NFPA 30 [1] prescribes the use of either standard or high temperature overhead
sprinkler heads to protect palletized storage of Class II and III liquids in Listed/Labeled rigid,
non-metallic IBCs. Quick-response sprinkler heads are currently permitted in rack storage
configurations. Typically, high temperature/standard response sprinkler heads will activate in a
temperature range of 121oC–149oC (250oF–300oF) [27] with response time indices (RTI) in the
range of 80–350 s0.5m0.5 (180–650 s0.5ft0.5). However, given the potential for the relatively rapid
growth of the liquid release scenarios considered and the large fire sizes expected, earlier
overhead thermal detection is desirable. To address this, overhead thermal detection using
ordinary temperature classification sprinklers (i.e., 57oC–77oC [135oF–170oF]) [27] could be
explored. These more sensitive thermal elements (i.e., frangible bulb / fusible link) will provide
a faster response to the fire. However, the response of the elements cannot be so sensitive that a
large number of sprinkler heads activate and put to great of a demand on the supply system.
53
HUGHES ASSOCIATES, INC
5.3.1.2 Berm-mounted Local Thermal Detection (Frangible Bulb-Fusible Link)
The entrainment of fresh air into the fire plume effectively reduces the bulk gas temperature
being transported to the upper layer within an enclosure. This dilution creates an inherent delay
in the response of overhead thermal detectors to a fire scenario. To reduce this delay, bermmounted detection systems, such as locally installed frangible bulbs/thermal links or linear heat
detection, were explored.
In these scenarios, the activation of the detection system serves as the release mechanism for
the suppression agent and provides both rapid detection as well as suppression agent discharge in
a single action. This integration minimizes the number of components that are required to work
within a system to achieve detection and suppression thereby minimizing the likelihood that the
system does not activate when needed. This reduction in ‘moving pieces’ within the system also
reduces the need for inspection, testing, maintenance activities also. With the exception of the
low-level AFFF suppression system, all of the local application systems described in Section 5.2
have integral detection systems that operate in a manner similar to that just described.
Despite the current absence of an integrated detection system for the low-level AFFF
application, it would be relatively easy to develop such a system using the methods currently
utilized in wet/dry chemical suppression systems that protect commercial kitchen hoods, large
vehicle engine compartments, petroleum loading racks, etc. In these systems, agent discharge is
typically initiated using a fusible link that is tied to a pneumatic piston which punctures a
pressurized agent cylinder thus allowing flow of agent to the discharge nozzles.
The disadvantages of localized thermal detection using the frangible bulb/fusible link
products include the relatively small area of detection (i.e., spot detection) due to the proximity
of the element to the containment system. This limitation requires that these devices be installed
around the entire containment perimeter in order to protect the entire area. However, even with
detection points around the perimeter it is possible that for large containment areas and relatively
small liquid releases, the thermal elements may be far enough removed that rapid detection
cannot be achieved. Another disadvantage of a localized thermal detection system is the
susceptibility of the units to potential damage / false activation due to the necessary floor level
installation location and the assumption that the area being protected is expected to have a
relatively high level of traffic from both personnel and machinery (i.e., operations scenario).
5.3.1.3 Local Thermal Detection (Linear Heat Detector)
The installation of linear heat detection in/around the containment area could also be used as
a means of locally detecting fire. The response of these systems to heat is comparable to that of
spot type detection devices (i.e., similar activation temperatures, similar response time indices,
etc.). However, the benefit of these systems is that continuous lengths of the wire can be
installed throughout the containment area, providing a continuous run of thermal detection and
total area coverage. This method of detection allows large areas to be protected in a relatively
non-intrusive fashion because continuous lengths of the linear heat detection wire, shown in
Figure 17, can be installed along the edges or on the internal walls of the containment area. This
should provide rapid detection of a fire within the containment area while at the same time
having minimal impact on the access in/around the IBC/containment area.
54
HUGHES ASSOCIATES, INC
Figure 17 –– Illustration of linear heat detector (left) and fiber optic heat detector (right)
These systems detect heat using one of two detection principles: the first is via the thermal
degradation of wire insulation which allows the two live wires to short and trigger an alarm; the
second utilizes fiber optic strands and the principles of Raman scattering to detect temperature
change. Typical linear heat detection technologies have RTI values comparable to quickresponse sprinkler heads while fiber optic technologies can have sampling rates as fast as 0.1 Hz.
Both methods of detection are capable of detecting/responding a neighboring fire and sending an
alarm signal.
Just as with the local thermal detection using thermal links, due to the location on/within the
containment area the linear heat detectors are susceptible to potential damage. The low level
installation subjects the detection devices to a relatively high level of traffic from both personnel
and machinery (i.e., operations scenario). However, neither of these technologies lose all
functionality when damaged (i.e., severed). For the standard linear heat detection systems, only
the section of detector downstream from the break is lost. For the fiber optic detection system,
only the point at which the line is broken is lost. Since the system relies on light scattering, if the
cable is broken, then temperature measurements can still be made from the break to the laser and
the location of the break will be known. If both ends of the cable are connected to a controller,
then temperature measurements will be made everywhere except for the length of damaged
cable. Consequently, as long as the control panel is undamaged, the fiber optic detection system
will remain viable. For the standard linear heat detection systems repair/replacement of
damaged/activated section of wire is a simple matter of splicing in a new segment to replace the
broken length of wire making turn-around after a fire event relatively quick and easy compared
to the replacement of spot type detection devices.
5.3.2
Visual Detection (UVIR/VID)
Visual detection systems sense a fire by utilizing the light, heat, and/ or smoke generated by a
fire. Visual systems could be installed within the operations area and potentially provide
protection over a large area from relatively few locations. However, these systems can be prone
to false alarms and the effectiveness of these systems may be reduced by the extent to which the
operations area is filled with obstructions due to the fact that these systems generally require
direct line-of-sight.
55
HUGHES ASSOCIATES, INC
Optical flame detectors (OFD), as shown in Figure 18, provide a rapid means of detecting
fire events without requiring the presence of substantial amounts of heat being released /
dispersed throughout the building being protected. Flame detection systems utilize a visual
sensor that is sensitive to IR and or UV light in the frequency ranges emitted by flames. These
sensors are typically threshold type sensors that merely look for a signal strength that exceeds
some threshold. Typically, these types of sensors have a field of view of ± 45 degrees in both
the horizontal and vertical axes and are capable of detecting relatively small fires at distances of
up to 30 m (100 ft).
Figure 18 –– Example of commercially-available flame detectors
These technologies are generally very reliable and are used extensively by the oil and gas
industry at production and refining facilities. Replacement of a failed sensor would be an easy
task given that they could be mounted in accessible locations where the IBC being protected is in
view. However, since flame detectors respond to certain spectra from a flame, they are not
capable of distinguishing between intentional flames and unintentional flames. Consequently,
these types of detectors could alarm to any repair or maintenance work involving open flames
(i.e., cutting torches, etc.). In addition to this propensity to false alarm, the complexity and
sensitivity of these detectors also requires that additional ITM procedures be developed in order
to ensure appropriate and rapid response to fire. These procedures include, but are not limited to,
the periodic verification of line of sight, verification of clear field of view, etc.
Another option with respect to visual fire detection is a video fire detection system. These
systems utilize standard video cameras that send real-time image data to a computer, which
processes the image and ‘looks’ for signatures to indicate a fire. These signatures can include
looking for flickering which indicates the possibility of a flame or looking for movement that
indicates a smoke plume. This ability to look for several different fire indicators allows the
video detection system to detect fire event with an obstructed view, which the OFD cannot. The
video cameras used in these systems can be the standard CCTV cameras (i.e., those typically
used for security surveillance) so if the facility requiring protection already has onsite
surveillance the detection system can be integrated into the existing system. Reliability and
maintainability of these systems is similar to that of flame detection systems.
In general, these visual fire detectors would be mounted at elevated locations along the wall
of the facility. This fixed installation allows the detector to remain relatively immune to the
56
HUGHES ASSOCIATES, INC
damage hazards associated with the general processes occurring on the facility floor but also
requires that the IBC remain in a fixed location. Given the potential for false alarms with visual
detection systems and the fact that these systems will be used as the means of activating a
suppression system; a minimum of two detectors that are cross-zoned should be used for system
activation. Using a cross-zoned approach (i.e., alarms from both detectors required to activate
suppression) provides redundancy in the detection system and reduces the likelihood of false
alarms resulting in suppression system activation.
5.3.3
Liquid Detection
Under worst case conditions, the fires associated with the release of the liquids typically
stored in IBCs would develop very rapidly. Consequently, conventional fire detection systems
(i.e., thermal and optical), no matter how ‘early’ the event is detected, require that a fire be
present and in these scenarios this fire would be expected to be large. An alternative to fire
detection is a liquid detection system. This type of system could be installed in the vicinity of
the IBC and/or within the containment area and would provide a means of alerting building
occupants that a potential hazard is developing. Ideally this alert would be before the ignition of
the liquid thereby potentially mitigating the fire hazard before the actual event occurs.
Sensors capable of performing this task are currently commercially available. These sensors
are designed to detect the presence of liquid in containment sumps, piping sumps, annular
spaces, etc. Because of the potential use in locations containing flammable/combustible liquids,
these devices (i.e., GEM LS-10®) are designed to be non-voltage producing and do not contain
energy storing components. Since the IBC scenario is a hazardous location, an appropriate
intrinsically safe interface device is required. An example of this type of sensor is the LS-10
leak detection sensor manufactured by GEM Sensors, shown in Figure 19.
Figure 19 –– Photograph of GEM LS-10® leak detection sensor
This sensor uses dry contact switching to detect the presence of liquids and ensure
dependability throughout the products lifetime. The sensor detects the presence of liquid using a
float within the sensor that when elevated, by the flowing liquid, closes the contacts and
transmits an alarm signal. This sensor activates when the float is raised by 10 mm (0.4 in.). It is
assumed that the sensors would be installed near the IBCs such that a 10 mm (0.4 in.) pool of
liquid would not have to be developed throughout the containment area before detection occurs.
The installation of several of these sensors around the base of the IBC within the containment
57
HUGHES ASSOCIATES, INC
area (i.e., one device on each side of the IBC) would provide early-warning liquid release
detection.
The only caveat identified for this type of ‘early-warning’ detection system is that in addition
to the installation of this system, it would also require the installation of a standard fire detection
system capable of the detecting / activating fire suppression systems in the event that a liquid
fuel fire does occur. The inclusion of this liquid detection system, while being an added cost,
provides the facility with a means of identifying a potential hazard very early in its development.
This very early warning enables facility personnel to take precautionary measures (i.e., patching
of leak, removal of released liquid, etc.) and mitigate the hazard before it has really any effect on
the operations process. If liquid detection was not present, the hazard may not be identified until
the liquid was ignited, resulting in operational down-time to suppress the fire, perform clean-up
activities, re-install detection/suppression systems, etc. The cost of the ‘added’ detection system
may be relatively minimal when compared the costs associated with the absence of the system.
5.3.4
Summary of Detection System Approaches
Five different fire detection approaches were identified. Five of these systems were fire
detection systems and the sixth system was a liquid detection system to be used in conjunction
with fire detection equipment. A summary of the advantages and disadvantages of the various
detection systems is provide in Table 14.
Table 14 –– Compar ison of Advantages/Disadvantages of Potential Detection Systems
Overhead
Thermal
Local
Thermal
(Fusible
Link /
Frangible
Bulb)
Local
Thermal
(Linear
Heat)
Visual Fire
Detection
(UVIR /
VID)
Liquid
Detection
Relative
System
Cost
Relative
Detection
Performance
Relative
Propensity
for False
Alarm
Require
Installation
of Alarm
Panel
Relative
Speed of
Detection
Inspection,
Testing, and
Maintenance
Low
Low
Low
No
1.0 min.
Low
Low
Moderate
Low
No
0.1 min.
Low
Moderate
Moderate
Low
Yes
0.1 min.
Low
High
High
High
Yes
0.1 min.
High
Low
High
Low
Yes
0.1 min.
Low
58
HUGHES ASSOCIATES, INC
The benefit of a local fire detection system is evident in Table 14 in that it provides a
shorter relative speed of detection which, when considered in conjunction with the findings of
Section 4.4, allow for smaller separation distances between the IBC containment area and
adjacent combustibles.
6.0
CONCLUSIONS
A fire hazard analysis was performed to identify/characterize the fire scenarios that could
result from IBCs in operations scenarios. Different levels of protection were specified depending
on the predicted hazard to the building, neighboring combustibles, and facility personnel. The
primary factor influencing the fire hazard was the required area of containment which drives the
potential fire size and the associated threat to both the overhead steel structure and adjacent
combustibles. Based on this finding two approaches to containment were explored. The first
approach utilized large containment areas to ensure capture of ejected liquids while the second
approach used enclosed, raised sumps to reduce the required containment area while still
preventing liquid release outside the containment area.
In configurations with enclosed, raised sump containment of one or two side-by-side IBCs or
IBCs containing alcohol based fuels (e.g., IPA), no additional protection other than fire resistive
containment and adequate separation may be needed. This assumes that the facility, in
accordance with insurer/fire code requirements, will at a minimum have an overhead water
sprinkler system designed for an ordinary hazard scenario. This also assumes that the
performance of the fire resistive containment system being used has been verified.
For other scenarios where the fire hazard is large enough to threaten the steel overhead or
where adequate separation distance is not provided, it is recommended that the local suppression
systems described in this assessment be considered in addition to the fire resistive containment.
A variety of systems were considered in this analysis with some systems (i.e., water mist, dry
chemical, CAFS) having proven performance (i.e., FM Approvals) for similar scenarios and
other systems (i.e., aerosol, low-level AFFF, passive protection) having potential, but not yet
Approved. Verification testing must be performed for any of the systems described above
because none have been tested against this specific type of fire scenario.
As described in Section 5.3, the type of detection required for these local suppression
scenarios is dependent on whether or not the separation distances available are sufficient for
either a 0.1 minute of 1.0 minute exposure to neighboring combustibles or whether or not the
safety control system has sufficient separation distance to permit manual activation.
This assessment presumes that flammable and combustible liquids stored in non-listed
composite IBCs are only used in open-air production spaces. Scenarios with a limited number of
units, having appropriate containment and protection strategies are proposed. While the
transport of flammable and combustible liquids in these IBCs is generally permitted in the US,
storage of liquid-filled, non-listed composite IBCs is essentially prohibited by building/fire
codes. The challenge of protecting indoor storage of non-Listed IBCs remains.
59
HUGHES ASSOCIATES, INC
7.0
RECOMMENDATIONS FOR ASSESSING OPTIONS
The recommendations provided in Table 15 are based on the findings of the fire hazard and
fire protection system analyses presented above as well as input from the technical panel. The
recommendations are listed in order of priority. The dry chemical and aerosol suppression
systems were not included in the recommendations due to concerns with the suppression
capabilities of these systems in the scenarios as well as the post-discharge contamination of
uninvolved areas within the facility. The low-level AFFF system was not included because it
was identified as a developmental system that would require additional testing beyond the scope
of this work. A fixed overhead AFFF system, using either a pre-mixed tank or proportioner
could be tested but was not identified as a priority.
Table 15 –– Summar y of Recommendations
Identifier
Recommendation
Approach
–
1
Conduct puncture
test to verify IBC
discharge distances
Liquid: Water
–
2
Conduct verification
testing of fire
resistive
containment
Liquid: Mineral seal oil
Containment Area: ≥1m2
Containment Type: Alternative
(e.g., Denios Hazberm)
Burning Duration: ≥ 10 min.
–
–
3
Conduct verification
testing of passive
fire protection
system in fire
resistive
containment
Liquid: Mineral seal oil
Containment Area: ≥1m 2
Containment Type: Steel
Burning Duration: ≥ 10 min.
60
–
Rationale
Distances used in analysis are
based on theory and
potentially too conservative.
Testing will serve as a
technical basis to potentially
reduce discharge distances
and associated minimum
containment areas
Demonstrates ability of
containment system to
contain a burning liquid pool
for an extended period of
time
If calorimeter is available,
testing will provide full-scale
burning rate data for mineral
seal oil. If unavailable,
heptane will be used in place
of mineral seal oil.
Approach is identical to that
used in second
recommendation thus direct
comparison between ‘freeburning’ and ‘protected’ pool
fire can be made
Provides technical support for
the idea that separation
distances can be reduced if
passive protection is provided
within containment area
HUGHES ASSOCIATES, INC
Table 15 –– Summary of Recommendations (Continued)
Identifier
Recommendation
Approach
–
4
Conduct verification
testing of enclosed
raised sump
containment system
Liquid: Heptane
Single IBC
Construction Type: Steel
–
–
5
Conduct verification
test of compressed
air foam system
against
Design Fire #3
Liquid: Heptane
Single IBC
Containment Area: 13m2
(140ft2)
–
–
6
Conduct verification
test of water mist
suppression system
against
Design Fire #3
Liquid: Heptane
Single IBC
Containment Area: 13m2
(140ft2)
–
–
7
Conduct verification
test of ‘best
performing system’
on array of two
IBCs, in stacked
configuration,
containing mineral
seal oil
Liquid: Mineral Seal Oil
Number of IBCs: Two
Configuration: Stacked
Containment Area: TBD
61
–
Rationale
Provides verification that this
type of system can withstand
thermal exposure without
being compromised (i.e.,
warping, etc.)
Heptane used as ‘worst-case’
fire exposure scenario
Provide verification that FMapproved foam suppression
system can extinguish the
fires associated with IBCs in
operations scenarios
Heptane used as ‘worst-case’
fire exposure scenario
Provide verification that FMapproved water mist
suppression system can
extinguish the fires associated
with IBCs in operations
scenarios
Heptane used as ‘worst-case’
fire exposure scenario
Proof of concept that the
system selected has the
capability to adequately
suppress a fire involving
IBCs in a complex, 3dimensional fire scenario
Mineral seal oil used because
it representative of liquids
typically used in operations
scenarios
HUGHES ASSOCIATES, INC
8.0
REFERENCES
1.
NFPA 30, Flammable and Combustible Liquids Code, 2012 Edition, National Fire
Protection Association Report, Quincy, MA.
2.
Babrauskas, V., “Heat Release Rates,” SFPE Handbook of Fire Protection Engineering,
4th Edition, DiNenno, P.J. (ed.), National Fire Protection Association, Quincy, MA,
2008, pp. 3.1–3.59.
3.
Newman, R.M., Fitzgerald, P.M., and Young, J.R., “Fire Protection of Drum Storage
Using ‘Light Water’ Brand AFFF in Closed Head Sprinkler Systems,” Factory Mutual
Research Corporation Report FMRC Ser. No. 22464, RC75-T-16, Norwood, MA,
March 1975.
4.
Kodur, V.R. and Harmathy, T.Z. “Properties of Building Materials,” SFPE Handbook of
Fire Protection Engineering, 4th Edition, DiNenno, P.J. (ed.), National Fire Protection
Association, Quincy, MA, 2008, pp. 1.167–1.195.
5.
Society of Fire Protection Engineers, Society of Fire Protection Engineering Guide:
Piloted Ignition of Solid Materials under Radiant Exposure, 2004.
6.
International Standards Organization 13571, Life Threat From Fires — Guidance on the
Estimation of Time Available for Escape Using Fire Data, International Standards
Organization (ISO), 2001.
7.
Beyler, C., “Fire Hazard Calculations for Large, Open Hydrocarbon Pool Fires,” SFPE
Handbook of Fire Protection Engineering, 4th Edition, DiNenno, P.J. (ed.), National Fire
Protection Association, Quincy, MA, 2008, pp. 3.271–3.319.
8.
Environmental Protection Agency (EPA), 40CFR264.175, Protection Of Environment,
Chapter I -- Environmental Protection Agency (Continued), Part 264 Standards for
Owners and Operators Of Hazardous Waste Treatment, Storage, and Disposal Facilities,
Code of Federal Regulations (CFR), Title 40, 25, TITLE 40--Subpart I Use and
Management of Containers Sec. 264.175 Containment. Revised as of July 1, 2010.
9.
Scheffey, J.L., Pabich, M., and, Sheppard, D., “Fire Testing of Intermediate Bulk
Containers – Phase IIA Evaluation of Large Arrays,” National Fire Protection Research
Foundation Report, Quincy, MA, April 1998.
10.
Scheffey, J.L. and Pabich, M., “Protection of Combustible Liquids Stored in Composite
Intermediate Bulk Containers – Research Project Phase II Final Report,” National Fire
Protection Research Foundation Report, Quincy, MA, August 2007.
11.
NFPA 11, Standard for Low-, Medium-, and High-Expansion Foam, 2010 Edition,
National Fire Protection Association Report, Quincy, MA.
12.
NFPA 16, Standard for the Installation of Foam-Water Sprinkler and Foam-Water Spray
Systems, 2011 Edition, National Fire Protection Association Report, Quincy, MA.
62
HUGHES ASSOCIATES, INC
13.
Scheffey, J., “Foam Agents and AFFF System Design Considerations,” SFPE Handbook
of Fire Protection Engineering, 4th Edition, DiNenno, P.J. (ed.), National Fire Protection
Association, Quincy, MA, 2008, pp. 4.89–4.127.
14.
Bete Spray Nozzles Company, Standard Fan NF Nozzle Datasheet,
http://www.bete.com/pdfs/BETE_NF.pdf, 2011.
15.
FireFlex Systems Incorporated, “Integrated Compressed Air Foam Systems for Fixed
Piping Networks – Design Manual,” FM-090M-0-1D, May 2007.
16.
UL 1254, Pre-Engineered Dry Chemical Extinguishing System Units, Underwriters
Laboratories, 2010.
17.
NFPA 17, Standard for Dry Chemical Extinguishing Systems, 2009 Edition, National Fire
Protection Association Report, Quincy, MA.
18.
UL 2775, Fixed Condensed Aerosol Extinguishing System Units, Underwriters
Laboratories, 2008.
19.
Fireaway, Stat-X Fire Protection Systems Data Sheet – Electrically Operated Units,
http://www.statx.com/pdf/Electrical_Data_Sheet_2_11.pdf , 2011.
20.
NFPA 750, Standard on Water Mist Fire Protection Systems, 2010 Edition, National Fire
Protection Association Report, Quincy, MA.
21.
Securiplex LLC, FIRE-SCOPE 2000® Low Pressure Water Mist System,
http://securiplexllc.com/low_pressure_water_mist_systems.html, 2011.
22.
FM 5560, American National Standard for Water Mist Systems, FM Approvals LLC,
2009.
23.
Lev, Y., “Novel method for controlling LNG pool fires,” Fire Technology, 4 (1981),
pp. 275–284.
24.
Persson, H., 2010, “Foam Glass Provides Effective Fire Protection in Bunded Areas,”
SP Fire Technology, Brandspoten.
25.
LNGIndustries.com, “Feeling the Burn,” 2009.
26.
Suardin, J.A. et al., “FOAMGLAS Pool Fire Suppression (PFS) Application on LNG,”
Mary Kay O’Conner Process Safety Center, Texas A&M University System, 2008.
27.
NFPA 13, Standard for the Installation of Sprinkler Systems, 2010 Edition, National Fire
Protection Association Report, Quincy, MA.
28.
Tewarson, A., “Generation of Heat and Gaseous, Liquid, and Solid Products in Fires,”
SFPE Handbook of Fire Protection Engineering, 4th Edition, DiNenno, P.J. (ed.),
National Fire Protection Association, Quincy, MA, 2008, pp. 3.109–3.194.
29.
UL 2368, Fire Exposure Testing of Intermediate Bulk Containers for Flammable and
Combustible Liquids, Underwriters Laboratories, 2001.
63
HUGHES ASSOCIATES, INC
APPENDIX A –– ENGINEERING TOOL
A-1
HUGHES ASSOCIATES, INC
A-2
Figure A.1 –– Liquid classification menu where the User identifies the liquid being contained within the IBC from a list of known
liquids or provides necessary combustion data for a User Specified liquid
HUGHES ASSOCIATES, INC.
A-3
Figure A.2 –– Scenario description and hazard calculation menu. The User inputs the physical characteristics of the facility,
IBC configuration, and available separation distances. Certain fire hazard values are calculated based upon liquid
classification and facility characteristics
HUGHES ASSOCIATES, INC.
A-4
Figure A.3. –– Protection recommendation menu where system recommendations are made based upon fire hazard and facility
characteristics provided in previous two menus
HUGHES ASSOCIATES, INC.