Verification and Validation of Selected Fire Models for Applications

NUREG-1824
Draft for Comment
EPRI 1011999
Preliminary Report
Verification and Validation
of Selected Fire Models for
Nuclear Power Plant
Applications
Volume 5:
MAGIC
January 2006
U.S. Nuclear Regulatory Commission
Office of Nuclear Regulatory Research
Washington, DC 20555-0001
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94303
crPow r
ksw
R
Cluc
AVAILABILITY OF REFERENCE MATERIALS
IN NRC PUBLICATIONS
NRC Reference Material
Non-NRC Reference Material
As of November 1999, you may electronically access
NUREG-series publications and other NRC records at
NRC's Public Electronic Reading Room at
httD://www.nrc.aov/readinp-rm.html.
Publicly released records include, to name a few,
NUREG-series publications; Federal Register notices;
applicant, licensee, and vendor documents and
correspondence; NRC correspondence and internal
memoranda; bulletins and information notices;
inspection and investigative reports; licensee event
reports; and Commission papers and their attachments.
Documents available from public and special technical
libraries include all open literature items, such as
books, journal articles, and transactions, Federal
Register notices, Federal and State legislation, and
congressional reports. Such documents as theses,
dissertations, foreign reports and translations, and
non-NRC conference proceedings may be purchased
from their sponsoring organization.
NRC publications in the NUREG series, NRC
regulations, and Title 10, Energy, in the Code of
Federal Regulations may also be purchased from one
of these two sources.
1. The Superintendent of Documents
U.S. Government Printing Office
Mail Stop SSOP
Washington, DC 20402-0001
Internet: bookstore.gpo.gov
Telephone: 202-512-1800
Fax: 202-512-2250
2. The National Technical Information Service
Springfield, VA 22161-0002
www.ntis.gov
1-800-553-6847 or, locally, 703-605-6000
A single copy of each NRC draft report for comment is
available free, to the extent of supply, upon written
request as follows:
Address: Office of the Chief Information Officer,
Reproduction and Distribution
Services Section
U.S. Nuclear Regulatory Commission
Washington, DC 20555-0001
E-mail:
[email protected]
Facsimile: 301-415-2289
Some publications in the NUREG series that are
posted at NRC's Web site address
http://www.nrc.gov/reading-rm/doc-collections/nuregs
are updated periodically and may differ from the last
printed version. Although references to material found
on a Web site bear the date the material was accessed,
the material available on the date cited may
subsequently be removed from the site.
Copies of industry codes and standards used in a
substantive manner in the NRC regulatory process are
maintained atThe NRC Technical Library
Two White Flint North
11545 Rockville Pike
Rockville, MD 20852-2738
These standards are available in the library for
reference use by the public. Codes and standards are
usually copyrighted and may be purchased from the
originating organization or, if they are American
National Standards, fromAmerican National Standards Institute
11 West 42 nd Street
New York, NY 10036-8002
www.ansi.org
212-642-4900
Legally binding regulatory requirements are stated
only in laws; NRC regulations; licenses, including
technical specifications; or orders, not in
NUREG-series publications. The views expressed
in contractor-prepared publications in this series are
not necessarily those of the NRC.
The NUREG series comprises (1) technical and
administrative reports and books prepared by the
staff (NUREG-XXXX) or agency contractors
(NUREGICR-XXXX), (2) proceedings of
conferences (NUREG/CP-XXXX), (3) reports
resulting from international agreements
(NUREG/IA-XXXX), (4) brochures
(NUREG/BR-XXXX), and (5) compilations of legal
decisions and orders of the Commission and Atomic
and Safety Licensing Boards and of Directors'
decisions under Section 2.206 of NRC's regulations
(NUREG-0750).
Verification & Validation of Selected
Fire Models for Nuclear Power Plant
Applications
Volume 5: MAGIC
NUREG-1824
EPRI 1011999
January 2006
U.S. Nuclear Regulatory Commission
Office of Nuclear Regulatory Research (RES)
Division of Risk Analysis and Applications
Two White Flint North, 11545 Rockville Pike
Rockville, MD 20852-2738
Electric Power Research Institute (EPRI)
3412 Hillview Avenue
Palo Alto, CA 94303
U.S. NRC-RES Project Manager
M. H. Salley
EPRI Project Manager
R. P. Kassawara
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI NOR ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, OR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED,
(I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER
(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
ORGANIZATION(S) THAT PREPARED THIS DOCUMENT:
U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research
Science Applications International Corporation
National Institute of Standards and Technology
ORDERING INFORMATION
Requests for copies of this report should be directed to EPRI Orders and Conferences, 1355 Willow
Way, Suite 278, Concord, CA 94520, (800) 313-3774, press 2 or internally x5379, (925) 609-9169,
(925) 609-1310 (fax).
Electric Power Research Institute and EPRI are registered service marks of the Electric Power
Research Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power
Research Institute, Inc.
COMMENTS ON DRAFT NUREG-1824 REPORT
This report is being published jointly by the U.S. Nuclear Regulatory Commission (NRC) and
the Electric Power Research Institute (EPRI). Any interested party may submit comments on
this report for consideration by the NRC and EPRI staffs. Comments may be accompanied by
additional relevant information or supporting data. Please specify both the report number (Draft
NUREG-1 824) and the volume number in your comments, and send them by March 31, 2006, to
the following address:
Chief Rules Review and Directives Branch
U.S. Nuclear Regulatory Commission
Mail Stop T-6D59
Washington, DC 20555-0001
For any questions about the material in this report, please contact:
Mark Henry Salley
Mail Stop T-IOE50
U.S. Nuclear Regulatory Commission
Washington, DC 20555-0001
Phone: (301) 415-2840
Email: MXS3(dnrc.gov
If EPRI members also wish to provide comments to EPRI, they may send them to the following
address:
R.P. Kassawara
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
Phone: (650) 855-2775
Email: RKASSAWARepri.com
NUREG-1824, Volume 5, has been
reproduced from the best available copy.
iii
CITATIONS
This report was prepared by
U.S. Nuclear Regulatory Commission,
Office of Nuclear Regulatory Research (RES)
Two White Flint North, 11545 Rockville Pike
Rockville, IvID 20852-2738
Principal Investigators:
K. Hill
J. Dreisbach
Electric Power Research Institute (EPRI)
3412 Hillview Avenue
Palo Alto, CA 94303
Science Applications International Corp (SAIC)
4920 El Camino Real
Los Altos, CA 94022
Principal Investigators:
F. Joglar
B. Najafi
National Institute of Standards and Technology
Building Fire Research Laboratory (BFRL)
100 Bureau Drive, Stop 8600
Gaithersburg, MD 20899-8600
Principal Investigators:
K McGrattan
R. Peacock
A. Hamins
Volume 1, Main Report: B. Najafi, M.H. Salley, F. Joglar, J. Dreisbach
Volume 2, FDTS: K. Hill, J. Dreisbach
Volume 3, FIVE-REV. 1: F. Joglar
Volume 4, CFAST: R. Peacock, J. Dreisbach, P. Reneke (NIST)
Volume 5, MAGIC: F. Joglar, B. Guatier (EdF), L. Gay (EdF), J. Texeraud (EdF)
Volume 6, FDS: K. McGrattan, J. Dreisbach
Volume 7, Experimental Uncertainty: A. Hamins, K. McGrattan
This report describes research sponsored jointly by U.S. Nuclear Regulatory Commission, Office
of Nuclear Regulatory Research (RES) and Electric Power Research Institute (EPRI).
The report is a corporate document that should be cited in the literature in the following manner:
Verification and Validation of Selected FireModels for Nuclear PowerPlantApplications,
Volume 5: MAGIC, U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory
Research (RES), Rockville, MD: 2005 and Electric Power Research Institute (EPRI), Palo Alto,
CA. NUREG-1824 and EPRI 1011999.
v
ABSTRACT
There is a movement to introduce risk- and performance-based analyses into fire protection
engineering practice, both domestically and worldwide. This movement exists in the general fire
protection community, as well as the nuclear power plant (NPP) fire protection community.
In 2002, the National Fire Protection Association (NFPA) developed NFPA 805, PerformanceBased Standardfor Fire Protectionfor Light-Water Reactor Electric GeneratingPlants, 2001
Edition. In July 2004, the U.S. Nuclear Regulatory Commission (NRC) amended its fire
protection requirements in Title 10, Section 50.48, of the Code of FederalRegulations (10 CFR
50.48) to permit existing reactor licensees to voluntarily adopt fire protection requirements contained
in NFPA 805 as an alternative to the existing deterministic fire protection requirements. In
addition, the nuclear fire protection community wants to use risk-informed, performance-based
(RI/PB) approaches and insights to support fire protection decision-making in general.
One key tool needed to support RIIPB fire protection is the availability of verified and validated
fire models that can reliably predict the consequences of fires. Section 2.4.1.2 of NFPA 805
requires that only fire models acceptable to the Authority Having Jurisdiction (AHJ) shall be
used in fire modeling calculations. Further, Sections 2.4.1.2.2 and 2.4.1.2.3 of NFPA 805 state
that fire models shall only be applied within the limitations of the given model, and shall be
verified and validated.
This report is the first effort to document the verification and validation (V&V) of five fire models
that are commonly used in NPP applications. The project was performed in accordance with the
guidelines that the American Society for Testing and Materials (ASTM) set forth in Standard
E1355-04, "Evaluating the Predictive Capabilityof Determninistic FireModels. " The results of
this V&V are reported in the form of ranges of accuracies for the fire model predictions.
vii
CONTENTS
1 INTRODUCTION ...........................
1-1
2 MODEL DEFINITION ...........................
2-1
2.1 Name and Version of the Model .........................
2-1
2.2 Type of Model .........................
2-1
2.3 Model Developers .........................
2-1
2.4 Relevant Publications .........................
2-1
2.5 Governing Equations and Assumptions ...................................................
2-1
2.6 Input Data Required To Run the Modell ...................................................
2-2
2.7 Property Data ...................................................
2-2
2.8 Model Results...................................................
2-3
3 THEORETICAL BASIS FOR MAGIC ...................................................
3-1
3.1 Introduction ...................................................
3-1
3.2 Theoretical Basis for MAGIC ...................................................
3-1
3.2.1 Combustion ...................................................
3.2.2 Hot Gas Layer Temperature and Height ...................................................
3.2.3 Walls, Ceiling and Floor ...................................................
3-3
3-3
3-3
3.2.4 Flame Height, Fire Plume & Ceiling Jets ...................................................
3-3
3.2.5 Natural & Mechanical Ventilation ...................................................
3-4
3.2.6 Radiation ...................................................
3-4
3.2.7 Targets ...................................................
3-4
3.2.8 Electrical Cables ...................................................
3-4
3.2.9 Sprinkler Suppression ...................................................
3-6
3.3 Concluding Remarks ...................................................
4 MATHEMATICAL AND NUMERICAL ROBUSTNESS ...................................................
3-6
4-1
4.1 Introduction ...................................................
4-1
4.2 Mathematical and Numerical Robustness Analyses for MAGIC ..................................... 4-1
ix
4.2.1 Comparison with Analytical Solutions.................................................
4-1
4.2.2 Code Checking and Code Quality .................................................
4-2
4.2.3 Numerical Tests .................................................
4-5
4.2.4 User Interface .................................................
4-5
4.3 MAGIC Improvements as a Result of the V&V Process ................................................. 4-5
4.4 Concluding remarks .................................................
5 MODEL SENSITIVITY .................................................
4-6
5-1
5.1 Definition of Base Case Scenario for Sensitivity Analysis .............................................. 5-1
5.2 Sensitivity Analysis .................................................
5-3
5.2.1 Hot Gas Layer Temperature and Height .................................................
5-4
5.2.2 Ceiling Jet Temperature .................................................
5-5
5.2.3 Plume Temperature .................................................
5-7
5.2.4 Flame Height .................................................
5-9
5.2.5 Oxygen Concentration .................................................
5.2.6 Smoke Concentration .................................................
5-9
5-10
5.2.7 Room Pressure .................................................
5-11
5.2.8 Target Temperature and Heat Flux .................................................
5-12
5.2.9 Wall Temperature .................................................
5-17
5.3 Concluding Remarks .................................................
6 MODEL VALIDATION .................................................
5-20
6-1
6.1 Hot Gas Layer Temperature and Height .................................................
6-4
6.2 Ceiling Jet Temperature .................................................
6-6
6.3 Plume Temperature .................................................
6-8
6.4 Flame Height .................................................
6-9
6.5 Oxygen Concentration .................................................
6-10
6.6 Smoke Concentration .................................................
6-11
6.7 Compartment Pressure .................................................
6-12
6.8 Radiation, Total Heat Flux and Target Temperature .................................................
6-14
6.9 Wall Heat Flux and Surface Temperature .................................................
6-18
6.10 Summary .................................................
6-20
7 REFERENCES .................................................
7-1
A TECHNICAL DETAILS FOR THE MAGIC VALIDATION STUDY ....................................... A-1
x
A.1.1 ICFMP BE #2 .......................................
A-2
A.1.2 ICFMP BE # 3 .......................................
A-4
A.1.3 ICFMP BE #4 .......................................
A-11
A.1.4: ICFMP BE #5 .......................................
A-12
A.1.5 FM/SNL Test Series .......................................
A-1 3
A.1.6 The NBS Multi-Room Test Series .......................................
A-16
A.2 Ceiling Jet Temperature .......................................
A-23
A.2.1 ICFMP BE # 3 .......................................
A-23
A.2.2 The FM/SNL Test Series .......................................
A-27
A-29
A.3 Plume Temperature .......................................
A.3.1 ICFMP BE # 2 .......................................
A-29
A.3.2 The FM/SNL Test Series .......................................
A-31
A.4 Flame Height .......................................
A-32
A.4.1 ICFMP BE #2 .......................................
A-32
A.4.2 ICFMP BE #3 .......................................
A-34
A-37
A.5 Oxygen Concentration .......................................
A.5.1 ICFMP BE #3 .......................................
A-37
A.5.1 ICFMP BE #5 .......................................
A-41
A.6 Smoke Concentration .......................................
A-42
A.6.1 ICFMP BE #3 .......................................
A-42
A-46
A.7 Room Pressure .......................................
A.7.1 ICFMP BE #3 .......................................
A.8 Target Temperature and Heat Flux .......................................
A.8.1 ICFMP BE #3 .......................................
A.8.2 ICFMP BE #4 .......................................
A.8.3 ICFMP BE # 5 .......................................
A.9 Compartment Wall Temperature and Heat Flux ................
A-46
A-49
A-50
A-87
A-89
....................... A-92
A.9.1 ICFMP BE #3 .......................................
A-92
A.9.2 ICFMP BE #4 .......................................
A-1 13
A.9.3 ICFMP BE #5 .......................................
A-1 14
B MAGIC INPUT FILES .......................................
B-1
xi
FIGURES
Figure 3-1: Pictorial Representation of MAGIC's Features ....................................................... 3-2
Figure 3-2: Simplified Inner Structure of an Armored Electrical Cable ...................................... 3-5
Figure 3-3: Modeling multi-conductor cables in MAGIC ............................................................ 3-6
Figure 4-1: Simplified Functional Breakdown of the MAGIC Code .......................................... 4-3
4-5
Figure 4-2: V-Cycle Representation .................................................................
Figure 5-1: Selected Heat Release Rate Profiles ................................................................. 5-2
5-4
Figure 5-2: Problem Specification in MAGIC .................................................................
Figure 5-3: Hot Gas Layer Temperature Profiles .................................................................. 5-5
5-6
Figure 5-4: Ceiling Jet Temperature Profiles .................................................................
5-8
Figure 5-5: Plume Temperature Profiles .................................................................
5-9
Figure 5-6: Flame Height Profiles .................................................................
5-10
Figure 5-7: Oxygen Concentration Profiles ..................................................................
5-11
Figure 5-8: Smoke concentration profiles .................................................................
5-12
Figure 5-9: Room Pressure Profiles ........................
Figure 5-10: Equivalence between Cable and Targets ........................................................... 5-13
5-14
Figure 5-11: Target temperature profiles ..................................................................
Figure 5-12: Target heat flux profiles, 1MW fire .................................................................. 5-15
5-16
Figure 5-13: Target heat flux profiles, 5MW fire..................................................................
Figure 5-14: Target temperature vs. wall temperature, 1 MW fire ........................................... 5-18
Figure 5-15: Target temperature vs. wall temperature, 5MW fire............................................ 5-19
Figure 6-1: Scatter plot of relative differences for hot gas layer temperature and height in
ICFMP BE #2, 3, 4, 5, and the selected FM/SNL and NBS Tests. Experimental
uncertainties are 13% (HGL temp) and 9% (HGL height) ................................................. 6-6
Figure 6-2: Scatter plot of relative differences for ceiling jet temperatures in ICFMP BE
#3, and the selected FM/SNL Tests. Experimental uncertainty is 16% ............................ 6-7
Figure 6-3: Scatter plot of relative differences for plume temperatures in ICFMP BE #2,
and the selected FM/SNL Tests. Experimental uncertainty is 14% .................................. 6-8
Figure 6-4: Scatter plot of relative differences for oxygen concentration in ICFMP BE#3
and BE#5. Experimental uncertainty is 16% .................................................................. 6-11
Figure 6-5: Scatter plot of relative differences for smoke concentration in ICFMP BE#3.
6-12
Experimental uncertainty is 33% ...................................................................
Figure 6-6: Scatter plot of relative differences for room pressure in ICFMP BE#3.
Experimental uncertainties are 40% (no forced vent) and 80% (forced vent) .................. 6-13
xiii
Figure 6-7: Scatter plot of relative differences for target temperature in ICFMP BE#3.
6-15
Experimental uncertainty is 14% ....................................................................
Figure 6-8: Scatter plot of relative differences for target temperature in ICFMP BE#3.
6-16
Experimental uncertainty is 14% ....................................................................
Figure 6-9: Scatter plot of relative differences for radiant heat flux in ICFMP BE#3.
6-16
Experimental uncertainty is 20% ....................................................................
Figure 6-10: Scatter plot of relative differences for total heat flux in ICFMP BE#3.
6-17
Experimental uncertainty is 20% ....................................................................
Figure 6-11: Scatter plot of relative differences for target temperature in ICFMP BE#3, 4,
6-19
and 5. Experimental uncertainty is 14% ...................................................................
Figure 6-12: Scatter plot of relative differences for heat flux in ICFMP BE# 3, 4 and 5.
6-19
Experimental uncertainty is 20% ...................................................................
Figure A-1: Cut-away view of the MAGIC simulation of ICFMP BE #2, Case 2 ....................... A-2
Figure A-2: Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #2 ............................. A-3
Figure A-3: Snapshot of the MAGIC simulation of ICFMP BE #3, Test 3 ................................. A-4
Figure A-4. Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #3, closed door
tests................................................................................................................................ A-6
Figure A-5. Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #3, closed door
tests................................................................................................................................ A-7
Figure A-6. Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #3, open door
tests................................................................................................................................ A-8
Figure A-7: Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #3, open door
tests................................................................................................................................ A-9
Figure A-8: Snapshot of the MAGIC simulation of ICFMP BE #4, Test 1............................... A-i 1
Figure A-9: Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #4, Test 1. ............. A-i 1
Figure A-10: Snapshot of the MAGIC simulation of ICFMP BE #5, Test 4 ............................. A-12
Figure A-1 1: Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #5, Test 4 ............. A-1 3
Figure A-12: Snapshot from the MAGIC simulation of FM/SNL Test 5 ................................. A-14
Figure A-13: Hot Gas Layer (HGL) Temperature and Height, FMWSNL Series ....................... A-15
Figure A-1 4: Snapshot from the MAGIC simulation of NBS Multi-Room Test 10OZ ............... A-1 6
Figure A-15: Hot Gas Layer (HGL) Temperature and Height, NBS Multiroom, Test 10OA ..... A-17
Figure A-1 6: Hot Gas Layer (HGL) Temperature and Height, NBS Multiroom, Test 1000. ... A-1 9
Figure A-1 7. Hot Gas Layer (HGL) Temperature and Height, NBS Multiroom, Test 1OZ .OOZ....
A-22
Figure A-1 8: Near-ceiling gas (ceiling jet) temperatures, ICFMP BE #3, closed door
tests.............................................................................................................................. A-24
Figure A-1 9: Near-ceiling gas (ceiling jet) temperatures, ICFMP BE #3, open door tests ...... A-25
Figure A-20: Near-ceiling gas (ceiling jet) temperatures, FM/SNL Series, Sectors 1 and
A-27
3....................................................................
Figure A-21: Fire plumes in ICFMP BE #2. Courtesy Simo Hostikka, VTT Building and
A-29
Transport, Espoo, Finland....................................................................
Figure A-22: Near-ceiling gas temperatures, FM/SNL Series, Sectors 1 and 3 ..................... A-30
Figure A-23: Near-plume temperatures, FM/SNL Series, Sectors 13 .................................... A-31
Xiv
Figure A-24: Flame heights for ICFMP BE # 2 .................................................................. A-32
Figure A-25: Photographs of heptane pan fires, ICFMP BE #2, Case 2. Courtesy, Simo
Hostikka, VTT Building and Transport, Espoo, Finland .................................................. A-33
Figure A-26: Photograph and simulation of ICFMP BE #3, Test 3, as seen through the 2
m by 2 m doorway. Photo courtesy of Francisco Joglar, SAIC ...................................... A-34
Figure A-27: Near-ceiling gas temperatures, ICFMP BE #3, closed door tests ...................... A-35
......................... A-36
Figure A-28: Flame heights, ICFMP BE #3, open door tests ........................
Figure A-29: 02 concentration, ICFMP BE #3, closed door tests........................................... A-38
Figure A-30: 02, ICFMP BE #3, open door tests. .................................................................. A-39
A-41
Figure A-31: 02, ICFMP BE #5, Test 1..................................................................
Figure A-32: Smoke concentration in ICFMP BE #3, closed door tests ................................. A-43
Figure A-33: Smoke concentration in ICFMP B3E #3, open door tests ................................... A-44
Figure A-34: Compartment pressure in ICFMP BE #3, closed door tests .............................. A-47
Figure A-35: Compartment pressure in ICFMP BE #3, open door tests ................................ A-48
Figure A-36: Thermal environment near Cable B, ICFMP BE #3, Tests 1 and 7 ................... A-51
Figure A-37: Thermal environment near Cable B, ICFMP BE #3, Tests 2 and 8 ................... A-52
Figure A-38: Thermal environment near Cable B, ICFMP BE #3, Tests 4 and 10 ................. A-53
Figure A-39: Thermal environment near Cable B, ICFMP BE #3, Tests 13 and 16 ...............A-54
Figure A-40: Thermal environment near Cable! B, ICFMP BE #3, Test 17 ............................. A-55
Figure A-41: Thermal environment near Cable B, ICFMP BE #3, Tests 3 and 9 ................... A-56
Figure A-42: Thermal environment near Cable' B, ICFMP BE #3, Tests 5 and 14 ................. A-57
Figure A-43: Thermal environment near Cable B, ICFMP BE #3, Tests 15 and 18 ............... A-58
Figure A-44: Thermal environment near Cable D, ICFMP BE #3, Tests 1 and 7 . ................. A-60
Figure A-45: Thermal environment near Cable D, ICFMP BE #3, Tests 2 and 8 . ................. A-61
Figure A-46: Thermal environment near Cable D, ICFMP BE #3, Tests 4 and 10 . ............... A-62
Figure A-47: Thermal environment near Cable D, ICFMP BE #3, Tests 13 and 16 ............... A-63
Figure A-48: Thermal environment near Cable D, ICFMP BE #3, Test 17............................. A-64
Figure A-49: Thermal environment near Cable D, ICFMP BE #3, Tests 3 and 9 ................... A-65
Figure A-50: Thermal environment near Cable D, ICFMP BE #3, Tests 5 and 14 ................. A-66
Figure A-51: Thermal environment near Cable D, ICFMP BE #3, Tests 15 and 18 ............... A-67
Figure A-52: Thermal environment near Cable F, ICFMP BE #3, Tests 1 and 7.................... A-69
Figure A-53: Thermal environment near Cable F, ICFMP BE #3, Tests 2 and 8 .................... A-70
Figure A-54: Thermal environment near Cable F, ICFMP BE #3, Tests 4 and 10 .................. A-71
Figure A-55: Thermal environment near Cable F, ICFMP BE #3, Tests 13 and 16 ................ A-72
Figure A-56: Thermal environment near Cable F, ICFMP BE #3, Test 17 ............................. A-73
Figure A-57: Thermal environment near Cable F, ICFMP BE #3, Tests 3 and 9.................... A-74
Figure A-58: Thermal environment near Cable F, ICFMP BE #3, Tests 5 and 14.................. A-75
Figure A-59: Thermal environment near Cable F, ICFMP BE #3, Tests 15 and 18 ................ A-76
Figure A-60: Thermal environment near Cable G, ICFMP BE #3, Tests 1 and 7................... A-78
xv
Figure A-61: Thermal environment near Cable G, ICFMP BE #3, Tests 2 and 8................... A-79
Figure A-62: Thermal environment near Cable G, ICFMP BE #3, Tests 4 and 10 . ............... A-80
Figure A-63: Thermal environment near Cable G, ICFMP BE #3, Tests 13 and 16 ............... A-81
Figure A-64: Thermal environment near Cable G, ICFMP BE #3, Test 17............................. A-82
Figure A-65: Thermal environment near Cable G, ICFMP BE #3, Tests 3 and 9 ................... A-83
Figure A-66: Thermal environment near Cable G, ICFMP BE #3, Tests 5 and 14 ................. A-84
Figure A-67: Thermal environment near Cable G, ICFMP BE #3, Tests 15 and 18 ............... A-85
Figure A-68: Heat Flux and Surface Temperatures of Target Slabs, ICFMP BE #4, Test
A-88
1..................................................................
Figure A-69: Thermal environment near Vertical Cable Tray, ICFMP BE #5, Test 4 ............. A-90
Figure A-70: Long wall heat flux and surface temperature, ICFMP BE #3, closed door
tests.............................................................................................................................. A-93
Figure A-71: Long wall heat flux and surface temperature, ICFMP BE #3, closed door
tests.............................................................................................................................. A-94
Figure A-72: Long wall heat flux and surface temperature, ICFMP BE #3, closed door
tests.............................................................................................................................. A-95
Figure A-73: Long wall heat flux and surface temperature, ICFMP BE #3, open door
tests.............................................................................................................................. A-96
Figure A-74: Short wall heat flux and surface temperature, ICFMP BE #3, closed door
tests.............................................................................................................................. A-98
Figure A-75: Short wall heat flux and surface temperature, ICFMP BE #3, closed door
tests.............................................................................................................................. A-99
Figure A-76: Short wall heat flux and surface temperature, ICFMP BE #3, closed door
tests............................................................................................................................ A-100
Figure A-77: Short wall heat flux and surface temperature, ICFMP BE #3, open door
tests............................................................................................................................ A-101
Figure A-78: Ceiling heat flux and surface temperature, ICFMP BE #3, closed door tests. . A-1 03
Figure A-79: Ceiling heat flux and surface temperature, ICFMP BE #3, closed door tests. . A-1 04
Figure A-80: Ceiling heat flux and surface temperature, ICFMP BE #3, open door tests..... A-1 05
Figure A-81: Ceiling heat flux and surface temperature, ICFMP BE #3, open door tests ..... A-1 06
Figure A-82: Floor heat flux and surface temperature, ICFMP BE #3, closed door tests..... A-1 08
Figure A-83: Floor heat flux and surface temperature, ICFMP BE #3, closed door tests ..... A-109
Figure A-84: Floor heat flux and surface temperature, ICFMP BE #3, open door tests ........ A-110
Figure A-85: Floor heat flux and surface temperature, ICFMP BE #3, open door tests ........ A-111
Figure A-86: Back wall surface temperatures, ICFMP BE #4 .
............................................. A-1 13
Figure A-87: Back and side wall surface temperatures, ICFMP BE #5, Test 4 .................... A-1 14
xvi
TABLES
Table 5-1: Material Properties for Cables ..................................................................
5-2
Table 5-2: Summary of MAGIC Simulations Selected for Sensitivity Analysis .......................... 5-3
Table A-1: Relative differences of hot gas layer temperature and height in ICFMP BE# 2 ...... A-4
Table A-2: Relative differences of hot gas layer temperature and height in ICFMP BE# 3 .... A-1 0
Table A-3: Relative differences of hot gas layer temperature and height in ICFMP BE# 4 .... A-12
Table A-4: Relative differences of hot gas layer temperature and height in ICFMP BE# 5 .... A-1 3
Table A-5: Relative differences of hot gas layer temperature and height in FM/SNL .............A-15
Table A-6: Relative differences of hot gas layer temperature and height in NBS Tests ......... A-1 8
Table A-7: Relative differences of hot gas layer temperature and height in NBS Tests ......... A-20
Table A-8: Relative differences of hot gas layer temperature and height in NBS Tests ......... A-22
Table A-9: Relative differences for ceiling jet temperature in ICFMP BE #3 .......................... A-26
Table A-1 0: Relative differences for ceiling jet temperatures in FM/SNL Tests ..........
........... A-28
Table A-1 1: Relative differences for plume temperature in ICFMP BE #2 ............................. A-30
Table A-12: Relative differences of plume temperature in FM/SNL tests............................... A-31
Table A-1 3: Relative differences of oxygen concentration in ICFMP BE #3 tests .........
......... A-40
Table A-14: Relative differences of oxygen concentration in ICFMP BE #3 tests .........
......... A-41
Table A-15: Relative differences of smoke concentration in ICFMP BE #3 tests .........
.......... A-45
Table A-1 6: Relative differences for compartment pressure in ICFMP BE #3 tests .........
...... A-49
Table A-1 7: Relative differences for surface temperature to cable B ...............
...................... A-59
Table A-1 8: Relative differences for radiative and total heat flux to cable B ...........
............... A-59
Table A-19: Relative differences for surface temperature to cable D..................................... A-68
Table A-20: Relative differences for radiative and total heat flux to cable D ...........
............... A-68
Table A-21: Relative differences for surface temperature to cable F ...............
...................... A-77
Table A-22: Relative differences for radiative and total heat flux to cable F........................... A-77
Table A-23: Relative differences for surface temperature to cable G..................................... A-86
Table A-24: Relative differences for radiative and total heat flux to cable G ...........
............... A-86
Table A-25: Relative differences for surface temperature, and total heat flux to targets ........ A-88
Table A-26: Relative differences for surface temperature ...................................................... A-90
Table A-27: Relative differences for total heat flux ................................................................ A-91
Table A-28: Relative differences for temperature and total heat flux corresponding to the
long wall ..................................................................
A-97
xvii
Table A-29: Relative differences for temperature and total heat flux corresponding to the
short wall ................................................................... A-1 02
Table A-30: Relative differences for temperature and total heat flux corresponding to the
A-107
ceiling ..................................................................
Table A-31: Relative differences for temperature and total heat flux corresponding to the
A-112
floor ..................................................................
Table A-32: Relative differences for wall temperature ......................................................... A-1 13
Table A-33: Relative differences for wall temperature ......................................................... A-1 14
xviii
REPORT SUMMARY
This report documents the verification and validation (V&V) of five selected fire models
commonly used in support of risk-informed and performance-based (RI/PB) fire protection at
nuclear power plants (NPPs).
Background
Over the past decade, there has been a considerable movement in the nuclear power industry to
transition from prescriptive rules and practices towards the use of risk information to supplement
decision-making. In the area of fire protection, this movement is evidenced by numerous
initiatives by the U.S. Nuclear Regulatory Commission (NRC) and the nuclear community
worldwide. In 2001, the National Fire Protection Association (NFPA) completed the
development of NFPA Standard 805, "Performance-Based Standard for Fire Protection for Light
Water Reactor Electric Generating Plants 2001 Edition." Effective July, 16, 2004, the NRC
amended its fire protection requirements in 10 CFR 50.48(c) to permit existing reactor licensees
to voluntarily adopt fire protection requirements contained in NFPA 805 as an alternative to the
existing deterministic fire protection requirements. RI/PB fire protection relies on fire modeling
for determining the consequence of fires. NFPA 805 requires that the "fire models shall be
verified and validated," and "only fire models that are acceptable to the Authority Having
Jurisdiction (AHJ) shall be used in fire modeling calculations."
Objectives
The objective of this project is to examine the predictive capabilities of selected fire models.
These models may be used to demonstrate compliance with the requirements of 10 CFR 50.48(c)
and the referenced NFPA 805, or support other performance-based evaluations in NPP fire
protection applications. In addition to NFPA 805 requiring that only verified and validated fire
models acceptable to the AHJ be used, the standard also requires that fire models only be applied
within their limitations. The V&V of specific models is important in establishing acceptable
uses and limitations of fire models. Specific objectives of this project are:
*
Perform V&V study of selected fire models using a consistent methodology (ASTM E1355)
and issue a report to be prepared by U.S. Nuclear Regulatory Commission Office of Nuclear
Regulatory Research (RES) and Electric Power Research Institute (EPRI).
*
Investigate the specific fire modeling issues of interest to the NPP fire protection
applications.
*
Quantify fire model predictive capabilities to the extent that can be supported by comparison
with selected and available experimental data.
xix
The following fire models were selected for this evaluation: (i) NRC's NUREG-1805 Fire
Dynamics Tools (FDTS), (ii) EPRI's Fire-Induced Vulnerability Evaluation Revision 1 (FIVERev. 1), (iii) National Institute of Standards and Technology's (NIST) Consolidated Model of
Fire Growth and Smoke Transport (CFAST), (iv) Electricite de France's (EdF) MAGIC, and (v)
NIST's Fire Dynamics Simulator (FDS).
Approach
This program is based on the guidelines of the ASTM E1355, "Evaluating the Predictive
Capability of Deterministic Fire Models," for verification and validation of the selected fire
models. The guide provides four areas of evaluation:
* Defining the model and scenarios for which the evaluation is to be conducted,
* Assessing the appropriateness of the theoretical basis and assumptions used in the model,
* Assessing the mathematical and numerical robustness of the model, and
* Validating a model by quantifying the accuracy of the model results in predicting the course
of events for specific fire scenarios.
Traditionally, a V&V study reports the comparison of model results with experimental data, and
therefore, the V&V of the fire model is for the specific fire scenarios of the test series. While
V&V studies for the selected fire models exist, it is necessary to ensure that technical issues
specific to the use of these fire models in NPP applications are investigated. The approach
below was followed to fulfill this objective.
1. A set of fire scenarios were developed. These fire scenarios establish the "ranges of
conditions" for which fire models will be applied in NPPs.
2. The next step summarizes the same attributes or "range of conditions" of the "fire
scenarios" in test series available for fire model benchmarking and validation exercises.
3. Once the above two pieces of information were available, the validation test series, or
tests within a series, that represent the "range of conditions" was mapped for the fire
scenarios developed in Step 1. The range of uncertainties in the output variable of
interest as predicted by the model for a specific "range of conditions" or "fire scenario"
are calculated and reported.
The scope of this V&V study is limited to the capabilities of the selected fire models. There are
potential fire scenarios in NPP fire modeling applications that do not fall within the capabilities
of these fire models and therefore are not covered by this V&V study.
Results
The results of this study are presented in the form of relative differences between fire model
predictions and experimental data for fire modeling attributes important to NPP fire modeling
applications, e.g., plume temperature. The relative differences sometimes show agreement, but
may also show both under-prediction and over-prediction. These relative differences are
affected by the capabilities of the models, the availability of accurate applicable experimental
data, and the experimental uncertainty of this data. The relative differences were used, in
combination with some engineering judgment as to the appropriateness of the model and the
xx
agreement between model and experiment, to produce a graded characterization of the fire
model's capability to predict attributes important to NPP fire modeling applications.
This report does not provide relative differences for all known fire scenarios in NPP applications.
This incompleteness is due to a combination of model capability and lack of relevant
experimental data. The first can be addressed by improving the fire models while the second
needs more applicable fire experiments.
EPRI Perspective
The use of fire models to support fire protection decision-making requires that their limitations
and confidence in their predictive capability is well understood. While this report makes
considerable progress towards that goal, it also points to ranges of accuracies in the predictive
capability of these fire models that could limit their use in fire modeling applications. Use of
these fire models present challenges that should be addressed if the fire protection community is
to realize the full benefit of fire modeling and performance-based fire protection. This requires
both short term and long term solutions. In the short term a methodology will be to educate the
users on how the results of this work may affect known applications of fire modeling. This may
be accomplished through pilot application of the findings of this report and documentation of the
insights as they may influence decision-making. Note that the intent is not to describe how a
decision is to be made, but rather to offer insights as to where and how these results may, or may
not be used as the technical basis for a decision. In the long term, additional work on improving
the models and performing additional experiments should be considered.
Keywords
Fire
Fire Modeling
Risk-informed regulation
Performance-based
Fire safety
Fire protection
Fire Probabilistic Risk Assessment (PRA)
Verification and Validation (V&V)
Fire Hazard Analysis (FHA)
Nuclear Power Plant
Fire Probabilistic Safety Assessment
(PSA)
xxi
PREFACE
This report is presented in seven volumes. Volume 1, the Main Report, provides general
background information, programmatic and technical overviews, and project insights and
conclusions. Volumes 2 through 6 provide detailed discussions of the verification and validation
(V&V) of the following five fire models:
Volume 2
Fire Dynamics Tools (FDTS)
Volume 3
Fire-Induced Vulnerability Evaluation, Revision 1 (FIVE-Revl)
Volume 4
Consolidated Model of Fire Growth and Smoke Transport (CFAST)
Volume 5
MAGIC
Volume 6
Fire Dynamics Simulator (FDS)
Finally, Volume 7 quantifies the uncertainty of the experiments used in the V&V study of these
five fire models.
xxiii
FOREWORD
Fire modeling and fire dynamics calculations are used in a number of fire hazards analysis (FHA) studies and
documents, including fire risk analysis (FRA) calculations; compliance with, and exemptions to the regulatory
requirements for fire protection in 10 CFR Part 50; "Specific Exemptions"; the Significance Determination
Process (SDP) used in the inspection program conducted by the U.S. Nuclear Regulatory Commission (NRC);
and, most recently, the risk-informed performance-based (RI/PB) voluntary fire protection licensing basis
established under 10 CFR 50.48(c). The RI/PB method is based on the National Fire Protection Association
(NFPA) Standard 805, "Performance-Based Standard for Fire Protection for Light-Water Reactor Generating
Plants."
The seven volumes of this NUREG-series report provide technical documentation concerning the predictive
capabilities of a specific set of fire dynamics calculation tools and fire models for the analysis of fire hazards in
nuclear power plant (NPP) scenarios. Under a joint memorandum of understanding (MOU), the NRC Office of
Nuclear Regulatory Research (RES) and the Electric Power Research Institute (EPRI) agreed to develop this
technical document for NPP application of these fire modeling tools. The objectives of this agreement include
creating a library of typical NPP fire scenarios and providing information on the ability of specific fire models to
predict the consequences of those typical NPP fire scenarios. To meet these objectives, RES and EPRI initiated
this collaborative project to provide an evaluation, in the form of verification and validation (V&V), for a set of five
commonly available fire modeling tools.
The road map for this project was derived from NFPA 805 and the American Society for Testing and Materials
(ASTM) Standard E1355-04, "Evaluating the Predictive Capability of Deterministic Fire Models." These
industry standards form the methodology and process used to perform this study. Technical review of fire
models is also necessary to ensure that those using the models can accurately assess the adequacy of the scientific and
technical bases for the models, select models that are appropriate for a desired use, and understand the levels
of confidence that can be attributed to the results predicted by the models. This work was performed using
state-of-the-art fire dynamics calculation methods/models and the most applicable fire test data. Future
improvements in the fire dynamics calculation methods/models and additional fire test data may impact the results
presented in the seven volumes of this report.
This document does not constitute regulatoryrequirements,and RES participationin this study neither
constitutes nor implies regulatoryapprovalof applicationsbased on the analysis containedin this text. The
analyses documented in this report represent the combined efforts of individuals from RES and EPRI, both of
which provided specialists in the use of fire models and other FHA tools. The results from this combined
effort do not constitute either a regulatory position or regulatory guidance. Rather, these results are intended
to provide technical analysis, and they may also help to identify areas where further research and analysis are
needed.
Carl J. Paperiello, Director
Office of Nuclear Regulatory Research
U.S. Nuclear Regulatory Commission
xxv
ACKNOWLEDGMENTS
The work documented in this report benefited from contributions and considerable technical
support from several organizations.
The verification and validation (V&V) studies for FDTS (Volume 2), CFAST (Volume 4), and
EDS (Volume 6) were conducted in collaboration with the U.S. Department of Commerce,
National Institute of Standards and Technology (NIST), Building and Fire Research Laboratory
(BFRL). Since the inception of this project in 1999, the NRC has collaborated with NIST
through an interagency memorandum of understanding (MOU) and conducted research to provide
the necessary technical data and tools to support the use of fire models in nuclear power plant
fire hazard analysis (FHA).
We appreciate the efforts of Doug Carpenter and Rob Schmidt of Combustion Science
Engineers, Inc. for their comments and contribution to Volume 2.
In addition, we acknowledge and appreciate the extensive contributions of Electricit6 de France
(EdF) in preparing Volume 5 for MAGIC.
We also appreciate the efforts of organizations participating in the International Collaborative
Fire Model Project (ICFMP) to Evaluate Fire Models for Nuclear Power Plant Applications,
which provided experimental data, problem specifications, and insights and peer comment for
the international fire model benchmarking and validation exercises, and jointly prepared the
panel reports used and referred in this study. We specifically appreciate the efforts of the
Building Research Establishment (BRE) and the Nuclear Installations Inspectorate in the United
Kingdom, which provided leadership for ICFMP Benchmark Exercise (BE) #2, as well as
Gesellschaft fuer Anlagen-und Reaktorsicherheit (GRS) and Institut fuer Baustoffe, Massivbau
und Brandschutz (iBMB) in Germany, which provided leadership and valuable experimental data
for ICFMP BE #4 and BE #5. In particular, ICFMP BE #2 was led by Stewart Miles at BRE;
ICFMP BE #4 was led by Walter Klein-Hessling and Marina Rowekamp at GRS, and R.
Dobbernack and Olaf Riese at iBMB; and ICFMP BE #5 was led by Olaf Riese and D. Hosser at
iBMB, and Marina Rowekamp at GRS. We acknowledge and sincerely appreciate all of their
efforts.
We greatly appreciate Paula Garrity, Technical Editor for the Office of Nuclear Regulatory
Research, and Linda Stevenson, agency Publication Specialist, for providing editorial and
publishing support for this report. We also greatly appreciate Dariusz Szwarc, Nuclear Safety
Professional Development Program participant, for his assistance finalizing this report.
xxvii
LIST OF ACRONYMS
AGA
American Gas Association
AHJ
Authority Having Jurisdiction
ASME
American Society of Mechanical Engineers
ASTM
American Society for Testing and Materials
BE
Benchmark Exercise
BFRL
Building and Fire Research Laboratory
BRE
Building Research Establishment
CFAST
Consolidated Fire Growth and Smoke Transport Model
CFR
Code of FederalRegulations
EdF
Electricit6 de France
EPRI
Electric Power Research Institute
FDS
Fire Dynamics Simulator
FDTs
Fire Dynamics Tools (NUREG-.1805)
FHA
Fire Hazard Analysis
FIVE-Revl
Fire-Induced Vulnerability Evaluation, Revision 1
FM-SNL
Factory Mutual & Sandia National Laboratories
FPA
Foote, Pagni, and Alvares
FRA
Fire Risk Analysis
GRS
Gesellschaft fuer Anlagen-und Reaktorsicherheit (Germany)
HRR
Heat Release Rate
IAFSS
International Association of Fire Safety Science
iBMB
Institut fur Baustoffe, Massivbau und Brandschutz
ICFMP
International Collaborative Fire Model Project
IEEE
Institute of Electrical and Electronics Engineers
MCC
Motor Control Center
xxix
MQH
McCaffrey, Quintiere, and Harkleroad
MOU
Memorandum of Understanding
NBS
National Bureau of Standards (now NIST)
NFPA
National Fire Protection Association
NIST
National Institute of Standards and Technology
NPP
Nuclear Power Plant
NRC
U.S. Nuclear Regulatory Commission
NRR
Office of Nuclear Reactor Regulation (NRC)
RES
Office of Nuclear Regulatory Research (NRC)
RI/PB
Risk-Informed, Performance-Based
SDP
Significance Determination Process
SFPE
Society of Fire Protection Engineers
V&V
Verification & Validation
xxx
1
INTRODUCTION
As the use of fire modeling tools increases in support of day-to-day nuclear power plant
applications including fire risk studies, the importance of verification and validation (V&V)
studies for these tools also increases. V&V studies afford fire modeling analysts confidence
in applying analytical tools by quantifying and discussing the performance of the given model
in predicting the fire conditions measured in a particular experiment. The underlying assumptions,
capabilities, and limitations of the model are discussed and evaluated as part of the V&V study.
The main objective of this study is to document a V&V study for the MAGIC zone model,
in accordance with ASTM E1355, StandardGuidefor Evaluatingthe Predictive Capability
ofDeterministicFireModels [Ref. 1]. MAGIC is a zone model developed and maintained by
Electricit6 de France (EdF), which officially released the latest version of the model in 2005. The
MAGIC software calculates fire-generated conditions single or multi-compartment geometries as a
function of time [Refs. 2, 3, and 4].
The MAGIC software is a classical thermal model for fire simulations in zones able to process
communicative multi-compartment problems. Each compartment is divided into two volumes,
which are assumed to be homogeneous. The solution of the mass and energy balances
accumulated on each zone, together with the ideal gas law and equation of heat conduction
into the walls, results in the environmental conditions generated by the fire.
Consistent with ASTM E1355, this document is structured as follows:
*
Chapter 2 provides qualitative background information about MAGIC and the V&V process.
*
Chapter 3 presents a technical description of MAGIC, which includes the underlying physics
and chemistry inherent in the model. The description includes assumptions and approximations,
an assessment of whether the open literature provides sufficient scientific evidence to justify
the approaches and assumptions used, and an assessment of empirical or reference data used
for constant or default values in the context of the model. MAGIC's source code and
technical description are EDF proprietary material; consequently, this report provides only
a technical summary of this material.
*
Chapter 4 documents the mathematical and numerical robustness of MAGIC, which involves
verifying that the implementation of the model matches the stated documentation.
* Chapter 5 presents a sensitivity analysis, for which the researchers defined a base case scenario
and varied selected input parameters in order to explore MAGIC's capabilities for modeling
typical characteristics of NPP fire scenarios.
1-1
Introduction
* Chapter 6 presents the results of the validation study in the form of relative differences
classified by fire modeling parameter. The following parameters were selected for validation
purposes:
*
Hot gas layer temperature and height
*
Ceiling jet temperature
*
Plume temperature
*
Flame height
*
Oxygen concentration
*
Smoke concentration
*
Room pressure
*
Target surface temperature and incident radiant and total heat flux
* Wall surface temperature and incident total heat flux
*
Appendix A presents the technical details supporting the calculated relative differences
discussed in Chapter 6 and provides graphical comparisons of experimental measurements
and modeling results.
*
1-2
Appendix B presents MAGIC input files.
2
MODEL DEFINITION
This chapter provides qualitative background information about MAGIC and the V&V process,
as suggested by ASTM E1355.
2.1 Name and Version of the Model
This V&V study focused on the latest version (V4.1.Ib) of the MAGIC zone model, which EdF
released in November 2005.
2.2 Type of Model
MAGIC is a two-zone fire model that predicts the environmental conditions resulting from a fire
prescribed by the user within a compartmented structure. Essentially, the space to be modeled is
subdivided into two control volumes that represent upper and lower layers. The fundamental
equations of conservation of energy and mass are solved in each control volume as the fire heat
release rate develops over time. The thermal conditions of the control volumes are the boundary
conditions for localized heat transfer problems solved at the surfaces and targets in the room.
2.3 Model Developers
MAGIC was developed and is maintained by Electricit6 de France (EdF).
2.4 Relevant Publications
MAGIC is supported by three EdF publications, including (1) the technical manual, which provides
a mathematical description of the model [Ref. 2]; (2) the user's manual, which details how to use
the graphical interface [Ref. 3]; and (3) the validation studies, which compare MAGIC's results
with experimental measurements [Ref. 4]. These three proprietary publications are available
through EPRI to EPRI members.
2.5 Governing Equations and Assumptions
MAGIC solves the conservation equations for mass and energy. The model does not explicitly
solve the momentum equation, except for use of the Bernoulli equation for the flow velocity
at room openings. These three equations and the ideal gas law are solved to obtain firegenerated conditions in the selected control volumes.
MAGIC assumes that the room is divided in two zones (upper and lower control volumes),
in which the equations described above are solved. The upper control volume, referred to
in this report as the hot gas layer, is assumed to have uniform density and, therefore, temperature.
The same assumption applies to the lower control volume (also known as the lower layer).
2-1
Model Definition
Chapter 3 of this report and Ref. 2 provide a complete technical description of MAGIC algorithms and
sub-models.
2.6 Input Data Required To Run the Model
In general, the following data is necessary to develop the input file for MAGIC. The required
inputs for each individual analysis may vary, and depend on the characteristics and objectives
of the fire scenario under analysis.
(1) The following parameters describe the compartment geometry and ventilation conditions:
*
Compartment geometry (length, width, and height): The compartment (or each compartment in
a multi-room scenario) is assumed to have a rectangular floor base and flat ceiling.
*
Floor, ceiling, and wall material properties (density, specific heat, and thermal conductivity):
Depending on the selected material, this information may be available in the MAGIC database.
Natural ventilation (height and width of doors; height, width, and elevation of windows;
time to open/close doors and windows during a fire simulation; and leakage paths).
* Mechanical ventilation (injection and extraction rates, vent elevations, and time
to start/stop the system).
*
(2) The following parameters describe the characteristics of the fire:
* Fuel type and fire heat release rate profile. The heat release rate profile is specified using
the heat of combustion and the mass loss rate of the fuel.
* Fire location (elevation, near a wall, near a corner, or center of room).
*
Footprint area of the fire: circular (e.g., pool fires specified by the diameter)
or rectangular (e.g., bounded pool fires, electrical cabinets specified by length and width)
*
Fuel mass, irradiated fraction, and stochiometric fuel-oxygen ratio.
(3) Two sets of parameters (thermo-physical properties and location) describe targets.
Thermo-physical properties include the density, specific heat, and thermal conductivity
of the material. Location refers to where the target is with respect to the fire (expressed with
three-dimensional coordinates).
(4) The inputs for sprinklers and detectors are the device's location with respect to the fire
and its response characteristics, which include activation temperature and response time index.
The MAGIC user's guide [Ref. 3] provides a complete description of the input parameters
required to run MAGIC.
2.7 Property Data
Various equations associated with the MAGIC model require the following property data:
* For walls: density, thermal conductivity, and specific heat
* For targets: density, thermal conductivity, and specific heat
* For fuels: heat of combustion, mass loss rate, stochiometric fuel-oxygen ratio, specific area,
and radiated fraction
2-2
Model Definition
These properties may be available in fire protection engineering handbooks or the MAGIC database.
However, depending on the application, properties for specific materials may not be readily available.
2.8 Model Results
MAGIC has an extensive library of output values. Once a given simulation is completed,
MAGIC generates an output file with all of the solution variables. Through a "post-processor"
interface, the user selects the relevant output variables for the analysis. Typical outputs include
(but are not limited to) the following examples:
*
environmental conditions in the room (such as hot gas layer temperature, oxygen
concentration, and smoke concentration)
*
heat transfer-related outputs to wall and targets (such as incident, convective, radiated, and
total heat fluxes)
*
fire intensity and flame height
*
flow velocities through vents and openings.
2-3
-------
3
THEORETICAL BASIS FOR MAGIC
3.1 Introduction
This chapter provides a brief technical summary of the MAGIC zone model to address
the ASTM E1355 requirement to "verify the appropriateness of the theoretical basis
and assumptions used in the model." However, given the proprietary nature of the software,
readers should refer to Ref. 2 for a complete technical description.
MAGIC is a "standalone" computer program for simulating fire conditions inside a compartment.
Technical details, as well as a user's guide and validation studies of this computer program,
can be found in Refs. 2, 3, and 4.
3.2 Theoretical Basis for MAGIC
MAGIC is a classical two-zone fire model. That is, a room is divided into upper and lower zones
(or layers). The upper layer (also referred to as the hot gas or smoke layer) accumulates hot gases
generated in the combustion zone and primarily transported by the fire plume. The lower layer
primarily consists of fresh air and has its own energy and mass balance.
Perhaps the most important characteristic of the two-zone model formulation is that each zone
is assumed to have homogeneous properties. The gas density (and, consequently, the temperature),
oxygen concentration, and concentration of unburned gases are assumed to remain constant
throughout each layer. These properties change only as a function of time.
Resulting fire conditions are obtained by solving equations for conservation of mass, species,
and energy, together with the ideal gas law. The species equation yields the concentration
of unburned fuel and oxygen in each layer. The compartment pressure, layer temperature,
and layer heights are obtained from the mass and energy equation. Finally, the gas densities
are calculated using the ideal gas law.
Specific sub-models are used to characterize the various physical processes:
* pyrolysis of the fuel(s) on fire
*
combustion in gaseous phase, governed by the properties of emitted products and the air
supply attributable to the plume flow
* smoke production and unburned products, for which the properties depend on the fuels
*
the fire plume over each source in various configurations (circular or linear fire, at different
heights, above the openings)
*
heat exchange by convection and radiation between the flame, air, hot gas layer, walls,
and environment
3-1
TheoreticalBasis For MAGIC
* natural flows through the openings (vertical and horizontal), which allow the compartments
to communicate with each other and the outside
*
forced or natural ventilation
*
thermal behavior and reaction of critical elements to determine their malfunction or ignition
*
thermal behavior and combustion of electric cables
* water spray from sprinklers
From a geometric point of view, MAGIC works on a set of rectangular rooms with flat ceilings,
with their edges parallel to the reference axes. These rooms communicate with each other and
the outside through horizontal or vertical openings.
MAGIC provides the following general results:
* temperatures of hot and cold zones
*
concentrations of oxygen and unburned gases
*
smoke migration into each room
*
the mass flow rates of air and smoke through the openings and vents
*
the pressures at the floor level of each room
*
the temperatures at the surface of the walls
*
the thermal fluxes (radiative and total) exchanged by the targets placed by the user
Figure 3-1 summarizes MAGIC's modeling features.
Figure 3-1: Pictorial Representation of MAGIC's Features
3-2
TheoreticalBasis For MAGIC
3.2.1 Combustion
The standard combustion model in MAGIC assumes a perfect oxidation reaction; that is, the fire
will burn at the specified heat release rate if oxygen is available. MAGIC tracks the amount
of oxygen in the fuel (in the case of a pre-mixed fuel), oxygen entrained by the fire, unburned
fuel in the environment, and the predefined fuel source in order to determine whether complete
combustion will occur. The chemical aspects of combustion are not considered. If the input
of oxygen quantity into the plume is at least equal to the quantity necessary to burn all of
the gaseous fuels in the plume, combustion is considered to be complete and controlled by
the fuel flow rate. If not, the combustion is incomplete and controlled by the available oxygen.
The user can also specify a low oxygen limit (LOL).
3.2.2 Hot Gas Layer Temperature and Height
Hot gas layer temperature and height result from balance equation of energy and mass for the
defined control volume. Properties are assumed to be homogeneous in the volume except in the
specific regions of the plume and ceiling-jet. Mass balance takes into account the fire plume flow
from the lower layer, and air supplied or exhaust through vents or openings. Energy balance
takes into consideration convection and radiation to the room surfaces (walls ceiling and floor)
and to the lower layer. The radiation properties of the layer are obtained from its opacity (based
on smoke concentration resulting from the mass balance). Oxygen and un-burnt gas
concentrations also result from the mass and energy balances in the hot gas layer volume. Similar
conservation equations are applied to the lower layer.
3.2.3 Walls, Ceiling and Floor
Walls, ceiling and floor are represented using one-dimensional finite difference meshing of
conduction. Two separate calculations are made: one for the section of wall in the upper layer
and the ceiling and a similar one for the lower layer and the floor. Boundary conditions for wall
inside a room use convection and a detailed radiation exchange. As a default, heat transfer
coefficient and wall emissivity are fixed to 15 W/m 2/K and 0.9 respectively. The heat transfer
coefficient can also be correlated to the temperature and the estimated velocity in the layer, as an
option. This study is based on default values.
Each wall can be constructed with multiple successive layers of a homogenous material; however,
the characteristics of each material are assumed to remain constant. The initial temperature condition
at both sides of the wall is the ambient temperature. The boundary conditions are calculated as
the simulation goes on and are based on heat exchange between surfaces and gas layers.
3.2.4 Flame Height, Fire Plume & Ceiling Jets
The fire plume in MAGIC is modeled using McCaffrey's semi-empirical correlations for fire
plume entrainment [Ref. 5] McCaffrey's correlation for temperature and velocity in the flame
region, Heskestad's correlation [Ref. 7] for temperature and velocity in the plume region. The
software incorporates the effects of the smoke layer on fire plume temperature [Ref. 6], and
simulates ceiling jets using the model developed by Cooper [Ref. 6] to account for hot gas layer
effects. As such, MAGIC models both confined and unconfined ceiling jets and considers the
adiabatic ceiling jet correlation and exchanges to walls from the layers' properties. In addition,
3-3
Theoretical Basis For MAGIC
MAGIC accounts for fires located along a wall or in corners, and it estimates flame height using
Heskestad's correlation [Ref. 7].
3.2.5 Natural & Mechanical Ventilation
The model for flows through horizontal openings is based on the formulation proposed by Cooper
[Ref. 8]. This model addresses the issue of one- or two-way flow at the opening using
experimental results. It is important to note that this model has been developed from
experimental conditions in which the horizontal opening was not directly above the fire source.
The model does not apply to configurationsin which the fire plume directly influences theflow.
MAGIC uses the Bernoulli equation to model flows through vertical openings with a
corresponding orifice flow coefficient. Flows are assumed to be perpendicular to the surface of
the opening.
The ventilation model used in MAGIC is based on the KIRCHOFF equation, and is represented
by a fan between ducts. In each duct, regular and singular pressure differences are considered.
Upstream and downstream nozzles make the link between rooms and vent systems. In the case
of no fan, the model calculates the mass flow through the ducts considering pressure differences.
3.2.6 Radiation
Radiation modeling is relatively complex in MAGIC. The gas layer is treated as a semitransparent gas. Radiation exchanges between surfaces (walls and openings), flames, and gas
layers are considered. One system is built for the upper layer and another for the lower layer.
Those systems exchange through the layer interface. View factors are re-evaluated for every
iteration due to the layer interface height variations.
3.2.7 Targets
Two kinds of targets are implemented in MAGIC. The basic target is equivalent to a flux meter
(with controlled surface temperature), and the thermal target is equivalent to a one-dimensional
homogeneous material. Fire sources, gas layers, walls, and openings generate the incident heat fluxes.
Convective and radiative fluxes are considered, and the heat exchange is calculated in detail.
For example, MAGIC considers direct radiation flux from sources located in adjacent rooms, and
correlates the convective exchange to local temperature and gas velocity. The target can be
located in the plume or ceiling jet, and MAGIC calculates the target temperature using
a one-dimensional finite difference conduction model.
3.2.8 Electrical Cables
Electrical cables are numerous in NPPs, and they serve as both fuel and targets in NPP fire scenarios.
Figure 3-2 summarizes the modeling of electrical cables in MAGIC.
3-4
TheoreticalBasis For MAGIC
Figure 3-2: Simplified Inner Structure of an Armored Electrical Cable
The cable is composed of successive layers of materials that are thermally described by the user.
Each layer includes a certain amount of discretization points, limited to 40 per layer. The code
automatically implements a node at the center of the cable, and the heat transfer inside the cable
is considered to follow an axial symmetry.
In the calculation, an electrical cable is divided in 20-cm segments along its length. The total
number of segments should not exceed 50. For each segment, MAGIC calculates the thermal
exchanges with the outside and the thermal heating.
The maximum surface temperature encountered on all the segments is the criterion to start
the cable ignition, from a piloted ignition threshold value or (if needed) a pyrolysis output
(introduced by the user). After the ignition, the cable behaves as a classical fuel, and its thermal
behavior is no longer modeled (that is, the surface temperature retains its last value).
An important consideration in this validation study is the treatment of multi-conductor cables. In
MAGIC, multi conductor cables were modelled as single conductor as follows:
n
*
The cross sectional area of the equivalent single conductor is
E
A, where Ai is the area of
1=l
each individual conductor and n is the total number of conductors in the cable.
*
The thickness of the jacket remains the same
* The thickness of the insulation is given by (cable thickness - jacket thickness - equivalent
conductor radius).
Figure 3.3 illustrates this process.
3-5
Theoretical Basis For MAGIC
4
I
Figure 3-3: Modeling multi-conductor cables in MAGIC
3.2.9 SprinklerSuppression
Modeling of sprinkler suppression is divided into three phases:
(1) Sprinkler activation determines the instant when the device is activated. Specifically,
the sprinkler is triggered when the temperature of the gas contained in the sprinkler bulb
reaches its activation temperature, which generally varies from 70 to 150'C depending on the
sprinkler. The model for sprinkler activation developed by Heskestad [Ref. 9] is implemented
in MAGIC.
(2) Cooling of the hot gas layer by the water spray is achieved through the interaction between
water droplets and the hot gas layer, which results in several physical and thermal
phenomena. The spray comprises a multitude of drops that have different speeds, diameters,
and directions. The thermal exchanges between the hot gases and the drops increase the
temperature of the drops and lead to partial or total evaporation and cooling of hot gases.
(3) Fire extinction takes into account the interaction between water drops and the fire.
Specifically, the power of the fire is stabilized when the spray is in contact with the fire.
This modeling is conservative and has been adopted in MAGIC.
3.3 Concluding Remarks
This chapter provided an overview of the modeling features of MAGIC. A complete technical
description is available in Reference 2.
MAGIC is based a combination of on macroscopic conservation equation and empirical
correlations for specific phenomena. This combination between fundamental principles and
experimental observations leads to a sound quantitative approach for its intended domain of
application. In addition to the validity of semi-empirical sub-models that may be used
independently, confidence in the predictive capabilities of the code is mainly obtained through
validation exercises, were the sub-models work together in order to provide consistent results for
the different relevant outputs in a specific scenario.
3-6
Theoretical Basis ForMAGIC
Although MAGIC can be used for general fire modeling applications, it has been intended since
the beginning for nuclear power plant applications. For this reason its validation file and some of
its sub-models, e.g., electrical cables, are specially adapted to this field.
It is necessary to stress the importance of the input parameters. Databases in MAGIC give
consistent information to the user, who can also customize with preferential materials. MAGIC
also provides some checks for the validity of the input values. However, typical fire modeling
studies usually involve uncertain inputs. In those cases, the analyst's expertise and experience in
the field is important for developing valid input files and obtaining consistent conclusions from
the model results.
3-7
4
MATHEMATICAL AND NUMERICAL ROBUSTNESS
4.1 Introduction
This chapter documents the mathematical and numerical robustness of MAGIC, which involves
verifying that the implementation of the model matches the stated documentation. Specifically,
ASTM E1355 requires the following analyses to address the mathematical and numerical
robustness of models:
*
Analytical tests involve testing the correct functioning of the model. In other words, these tests
use the code to solve a problem with a known mathematical solution. However, there are
relatively few situations for which analytical solutions are known.
* Code checking refers to verifying the computer code on a structural basis. This verification
can be achieved manually or by using a code-checking program to detect irregularities
and inconsistencies within the computer code.
* Numerical tests investigate the magnitude of the residuals from the solution of a numerically
solved system of equations (as an indicator of numerical accuracy) and the reduction in residuals
(as an indicator of numerical convergence).
4.2 Mathematical and Numerical Robustness Analyses for MAGIC
MAGIC consists of a user interface and a mathematical source code models. This section covers
only the second module in detail. Section 4.2.4 describes a classical quality assurance policy
for the user interface.
4.2.1 Comparison with AnalyticalSolutions
General analytical solutions do not exist for fire problems - even for the simplest cases.
Nonetheless, it is possible to test specific aspects of the model in typical situations. Some studies
have been performed to control the correct behavior of the following sub-models of MAGIC:
* conduction into the wall: comparison to other models and analytic solutions
*
target and cable thermal behavior: consistency of the behavior in typical situations
* plumes model: comparison with the theoretical model
* vent and opening: comparison to other zone and field models
* room pressure: comparison with pressure estimated by the perfect gas law and simplified
energy equation
4-1
Mathematicaland Numerical Robustness
4.2.2 Code Checking and Code Quality
In general, MAGIC is structured as shown in Figure 4-1. The code reads data in a case file
(*.cas, which can be accessed using any word processing software or the MAGIC interface
itself), and initializes all of the variables for the problem. To solve the system of differential
equations, the model divides time into successive intervals, and then solves the equations
recursively from instant to (where the variables are known) and by using a recurrence formula
linking instant tn to tn+i.
All the solution variables at instant t. are obtained through Array Y (temperatures, pressure,
concentrations, and gas characteristics). The 26 constituent equations related to a given room
are numbered from Y(1) to Y(26).
A subroutine calculates all of the fluxes (Y') derived from the physical model implementation
between tn and tn+l . This Array, Y', is transmitted to the solver for the calculation of tn+1.
The ordinary differential equation (ODE) solver is based on a trial-and-error process to estimate
the solution variables at tn+1.
4-2
Mathematical and NumericalRobustness
Figure 4-1: Simplified Functional Breakdown of the MAGIC Code
4-3
Mathematicaland Numerical Robustness
The equations of conduction inside the walls and cables, the sprinkler spray, and the equations
of transport in the ventilation system are solved independently during the calculation cycle
of transfers. All temperatures are updated at each calculation step.
The ODE system is solved using backward differential formulas (Gear method). The solver uses
a specific algorithm based on the BROYDEN approach [Refs. 10 and 11]. This method is interesting
because there are several distinct time scales in the resolution of the ODE. Indeed, the fire combustion
is the fastest phenomenon encountered, and the transport time scale is much slower than the fire
reaction. This is why a numerical resolution (such as Gear) with trial-and-error enables the model
to dynamically adjust the resolution time step. If the problem presents dramatic physical changes
in relatively short periods of time, the time step decreases; however, in the opposite case,
the time step increases.
In addition, the BROYDEN approach is a quasi-Newton method to process the system of
nonlinear algebraic equations. In this method, the Jacobean matrix is replaced by a series
of "approached" matrices converging toward the exact matrix at the solution point. First,
the "approached" matrix is decomposed in a product of two matrices - LU with L as the lower
triangular matrix, and U as the upper triangular matrix with 1 on the diagonal (CROUT method).
The resolution makes a first iteration with the Newton method, and three iterations with
the BROYDEN method. This solution enables the model to improve the convergence
when the problem presents dramatic physical changes in relatively short periods of time.
This method avoids recalculation of the Jacobean matrix at each iteration (thereby saving time).
The source code itself is tested with the following methods:
* First, to control robustness, the code may be compiled in several different platforms and
software applications. The MAGIC code has been compiled under Microsoft Windows 2000
and Windows XP, with a variety of compilers, including Absoft Pro Fortran, Visual
FORTRAN, and G77. In addition, a global update of the FORTRAN sources was performed
in 2004 [Ref. 12], and aspects such as code documentation, variable glossary, and source
cleanup were addressed.
* In terms of code quality, two tools have been used to control the language:
> FORSTUDY from Cobalt Blue
>
PLUSFORT from Polyhedron Software
These tools confirm the consistency of variables and constants (undefined and incorrectly or
redundantly declared) and use of good FORTRAN syntax.
The software quality assurance system provides a process to fix detected anomalies concerning
the interface of the code. Maintenance of MAGIC is based on observation forms, which identify
problems. Then, a modification form describes the problem analysis and proposed solutions.
Finally, a correction form explains the chosen solution and implementation features. The project
manager decides on the implementation of the correction in future versions.
4-4
Mathematicaland Numerical Robustness
4.2.3 Numerical Tests
For each new code version, a set of tests is used to ensure that the calculation is correct. These tests
come from previous case studies. The convergence and speed of the calculation is the first step
of control. Main results from the original study are then compared, and significant differences
are analyzed.
Specific tests are performed in the maintenance process when new models are implemented into
the code, or when existing models are corrected or improved. Those tests are not systematically
conducted for new versions, but they are available in case problems arise with the model under study.
4.2.4 User Interface
The method used is a classical V-cycle development with tests, as illustrated in Figure 4-2.
Figure 4-2: V-Cycle Representation
The code meets the corresponding specifications at each step of the cycle. The following
reference documents are available:
* Conception documentation [Ref. 131 presents the general "architecture" and input/output files,
and summarizes all of the class, function, and subroutine codes in the interface. The document
includes a short description of objectives and parameters for each.
* Reception test framework [Ref. 14] validates all of the interface functions and enables the user
to verify consistency between the specification and software.
*
User reference guide [Ref. 15] presents the various interface menus and details their uses.
* Tutorials [Ref. 16] allow self-teaching through step-by-step exercises.
4.3 MAGIC Improvements as a Result of the V&V Process
Some improvements were made in MAGIC as a result of the V&V process. This highlights the
importance of the verification and validation process including a rigorous comparison of code
predictions with experimental observations. MAGIC version 4.1. lb includes the latest the
4-5
Mathematicaland Numerical Robustness
improvements resulting from the V&V. Specifically, the following MAGIC features were
corrected during the V&V process:
*
Improvement of the soot mass balance within the plume
*
Improvement of the correlation for temperature in the flame region
*
Correction of a problem in the flame length calculation when lower than the layer interface.
4.4 Concluding remarks
MAGIC has been developed to allow quick and robust calculations of typical fire conditions in
single and multi-compartment building, on a standard PC platform. Calculations will be very
quick (a few seconds) for simple scenarios, e.g., single room with opening and vents.
Configurations with several communicating rooms (up to 24) can be managed by the code.
Calculation times however are correlated to the complexity of the problem. The number of
communicating rooms maybe the most influencing parameter. The use of the cable model can
also have a significant "cost" on calculation time.
The development and maintenance of MAGIC is performed by EdF R&D. On average, a
MAGIC revision is released once a year.
4-6
5
MODEL SENSITIVITY
This chapter discusses sensitivity analysis, which ASTM E1355 defines as a study of how
changes in model parameters affect the results. In other words, sensitivity refers to the rate of
change of the model output with respect to inpult variations. The purpose of this sensitivity analysis is
twofold:
1. Test MAGIC predictive capabilities with a range of different inputs to check for consistency in the
results, and
2. Compare different modeling strategies in MAGIC in support of the validation study described later
in Chapter 6 and Appendix A. Specifically, two modeling strategies selected for validation
includes:
a. The use of thenmic-target (slab) sub-model in MAGIC for predicting cable surface
temperature VS the use of the cable sub-model.
b. The use of thenmic-target (slab) sub-model in MAGIC for predicting temperature and heat
fluxes to room surfaces VS the use of the wall temperature sub-model.
5.1 Definition of Base Case Scenario for Sensitivity Analysis
Conducting a sensitivity analysis requires the definition of a base case scenario. Variations in
the output of the model are measured with respect to the base case scenario.
The base case scenario for this study was analyzed in Benchmark Exercise #1 as part of an
ongoing International Collaborative Fire Modeling Project (ICFMP) [Ref. 17]. This section
summarizes the technical description of the scenario. (Note that only Part 1 of the benchmark
exercise was selected as the base case.)
*
Room
*
*
*
*
*
*
*
*
Length: 15.2 m (50 ft)
Width: 9.1m(30ft)
Height: 4.6 m (15 ft)
Walls: 0.15 m thick concrete (6 in)
Door: 2.4 m x 2.4 m (62 ft2 )
Mechanical ventilation: 5 air changes per hour
Vent size: 0.5 m2 (5.4 ft2 )
Vent elevation: 2.4 m (7.9 ft)
* Target
* Cable Tray A: 0.6 m wide, 0.08 m deep
* Elevation (cable tray A): 2.3 m, 0.9 m off the right wall of the room
5-1
Model Sensitivity
* Cable Tray B: 0.6 m wide, 0.08 m deep
* Elevation (Cable Tray B): 2.3 m, along the left wall of the room
* Material properties for concrete
* Specific heat: 1,000 J/kg-K
* Thermal conductivity: 1.75 W/m-K
* Density: 2200 kg/m3
* Emissivity: 0.95
* Material properties for cables (targets): See Table 5-1.
Table 5-1: Material Properties for Cables
Thermal cond
[kW/mK]
Density [kgtm3l Cp [kJ/Kg-kI
Waterial
1
IXPE
PVC
I
0.00021
0.000147
1375
1380
_
1.566
1.469
Ambient conditions
*
*
*
*
*
Temperature: 27 C (81 F)
Relative humidity: 50%
Pressure: 101,300 Pa
Elevation: 0
Wind speed: 0
* Fire (heat release rate): The fire heat release rate was assumed to have a t2 growth profile.
The fire reaches its peak intensity in 600 seconds. Two peak intensities (1.0 MW and 5.0 MW)
were selected for this sensitivity analysis, in order to explore different MAGIC features
and capabilities. For example, fire intensity capable of consuming all of the oxygen
in the enclosure allows the sensitivity analysis to explore situations that exercise MAGIC's
extinction model. Figure 5-1 illustrates the two heat release rate profiles.
Heat Release Rate Profiles
MM.,
----
5000
__
4000
33000 - -X
Xd
.-1(11
I
0
.
0
200
400
-
600
800
1000
1200
Tine [secJ
5.0 MW
1.0 MW -
Figure 5-1: Selected Heat Release Rate Profiles
5-2
Model Sensitivity
5.2 Sensitivity Analysis
A total of 16 MAGIC simulations were conducted for this sensitivity analysis. A key input
parameter was modified in each run in order to explore MAGIC's capabilities and sensitivities
with respect to input parameters. Note, however, that MAGIC requires numerous input
parameters, and this study did not analyze all of those input parameters and their combinations.
Instead, the researchers selected key parameters relevant to typical commercial NPP fire scenarios.
Table 5-2 summarizes the 16 MAGIC fire simulations selected for sensitivity analysis.
Table 5-2: Summary of MAGIC Simulations Selected for Sensitivity Analysis
Rate [kW]
Natural
Ventilation
[m2 i
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1000
1000
1000
1000
1000
1000
1000
1000
5000
5000
5000
5000
5000
5000
5000
0.015
0.015
0.015
0.015
5.76
5.76
5.76
5.76
0.015
0.015
0.015
0.015
5.76
5.76
5.76
0
0.88
0
0.88
0
0
0
0
0
0.88
0
0.88
0
0
0
0
0
10
10
0
0
0
0
0
0
10
10
0
0
0
Heptane
Heptane
Heptane
Heptane
Heptane
Toluene
Heptane
Heptane
Heptane
Heptane
Heptane
Heptane
Heptane
Toluene
Heptane
Floor
Floor
Floor
Floor
Floor
Floor
0.5H
Floor
Floor
Floor
Floor
Floor
Floor
Floor
0.5H
Center
Center
Center
Center
Center
Center
Center
0.25W
Center
Center
Center
Center
Center
Center
Center
16
5000
5.76
0
0
Heptane
Floor
0.25W
Case
Heat Release
Mech.
Lower
Ventilation Oxygen
[m 3/sl
Limit [%]
Fuel Type
Vertical
Fire
Position
Horizontal
Fire
Position
The first eight simulations were conducted with the assumption of a peak fire intensity of 1.0 MW.
The researchers varied parameters affecting the size of the ventilation openings, the mechanical
ventilation system, the fuel type, and the fire location, in order to explore their effects on selected
results. Simulations 9-16 are identical to the first eight, but with a peak fire intensity of 5.0 MW.
Targets of both PVC and XPL material were specified in the computational domain. Therefore,
sensitivities to thermo-physical properties of targets can be explored in each of the analyzed cases.
Figure 5-2 provides a pictorial representation of the fire scenario selected for sensitivity analysis,
as defined in MAGIC.
5-3
Model Sensitivity
Wall Target
PVC Target
Mechanical
ventilation
Figure 5-2: Problem Specification in MAGIC
The sensitivity analysis and its results are classified by relevant fire-modeling attributes selected
for this V&V study, as presented in the following sections.
5.2.1 Hot Gas Layer Temperature and Height
The hot gas layer temperature is perhaps the single most important output of a zone model,
since it is the direct result of the energy and mass balance in the upper control volume.
In general, the hot gas layer temperature is affected by the fire intensity, natural and mechanical
ventilation characteristics, and material properties of the room. This study analyzed the effect
of the first three groups of inputs on the hot gas layer temperature. The properties of the wall
were not evaluated, since most NPP fire scenarios involve concrete rooms.
Figure 5-3 summarizes the hot gas layer temperature profiles for selected cases in the sensitivity
analysis. The first group of profiles (Cases 1, 2, and 5) was associated with a heat release rate
of 1.0 MW, and the predicted hot gas layer temperature was just below 140 'C. Notice that
the temperature profile is similar to the heat release rate. That is, once the fire reaches steady-state
at 600 seconds, the temperature profile is almost steady.
The second group of profiles (Cases 9, 10 and 13) reached temperatures just below 350 'C.
Notice that the profiles for Cases 9 and 10 show decay after 600 seconds, which is attributed to
a reduction in the heat release rate as a result of low oxygen concentration. Recall that only
air leakages were assumed in these two simulations. Notice that Case 13 was not affected
by the amount of oxygen because the door was open and fresh air was constantly moving
into the enclosure.
5-4
Model Sensitivity
In summary, although it is generally obvious that the heat release rate affects the hot gas layer
temperature, higher fire intensities consume additional oxygen, which may prevent the fire
from burning at its specified heat release rate.
The hot gas layer height output is directly associated with the hot gas layer temperature, as it is
also a direct output of the energy and mass balance in the upper control volume. This output result
is also generally affected by the same input parameters as the hot gas layer temperature.
Figure 5-3 also illustrates a selected set of hot gas layer heights. Notice two distinctive sets of
results. First, the profiles for Cases 1, 2, 9, and 10 reached the floor of the room. Those cases
consist of fire simulations that assume a small leakage area below the door (closed door
simulation). In Cases 5 and 13, the hot gas layer did not reach the floor, because the door was
assumed to be open. As expected, the layer interface in Case 13 leveled lower than the one in
Case 5, because of the higher heat release rate.
Fbt Gas Layer Terrperature
WbtGas Layer Hight
160
5
4.5
-
100
I-
f
50
)^
Ca(se 1
40
AO
cae 2
O
20
40
-Case
r
10
5
0.5
0
20
0
I
The (rrn)
HlotGas Layer Terrperattxe
5
10
15
250
I
tHotGas Layer Fbght
-_Case 9
Awe 9
4
300 -
20
Tere (nin)
5
4.5
400
350 -_
Ase 5
2.5
2
5
15
Id
1.5
O
0
3 35
35
Case1
Case 2
1,
'4
-.
-Caseloo
-UCse
Case 10
-Case
13
200
13
_
okk
lo
2.5
2
\
150If_
1.5
100
0.5
0
50
0
0
5
10
15
20
a
5
Time (fin)
10
is
20
Trhe (nin)
Figure 5-3: Hot Gas Layer Temperature Profiles
5.2.2 Ceiling Jet Temperature
Two sensors were specified in MAGIC's computational domain to record the gas temperature
at the specified locations. Figure 5-4 illustrates selected ceiling jet temperature profiles.
5-5
Model Sensitivity
Ceing Jet Temperature
Ceing Jet Temnperature
300,
900
-Casel.R=2m
250
r
_-asseS.R=2m
- CaseSR=4m
200
U
800 _1Ca2e9,R=2m
0
Case 9, A= 4m
-CaselR=4m
70
12
I-onm
le\
Case 13,R=2m
X
U
150
E
1!
100
50s
0
0
5
10
15
3
2X
Temn
(sin)
Tine (mrit)
Cirng Jet Tenrperature
350 -
_
Case 7, R=2 m-...
300 -
-
Case 7,
400
U
250
E
2X00
1!
CeifroJet Tenyerature
200
800
100
700
400
=4 m
_2m-
U
300
150
100
50
0
0
5
10
15
20
0
5
The (WM)
10
15
20
Tlre (sin)
Figure 5-4: Ceiling Jet Temperature Profiles
MAGIC performed as expected. First, for each case, the temperature at a larger radial distance,
R, was lower than at a shorter R. Second, the ceiling jet temperature was higher than the
predicted hot gas layer temperatures for the respective cases. For example, at a relatively large radial
distance from the fire, R = 4 m, the ceiling jet temperature was just above 150 'C in Case 5.
Recall that the predicted hot gas layer temperature for Case 5 was below 150 0C. Another
interesting observation is that the ceiling jet temperatures are higher in the closed room
simulation (Case 1), compared to the open room simulation (Case 5). This behavior was also
observed in the corresponding simulations with a 5-MW heat release rate. Consider, for example,
the ceiling jet temperature profiles for Case 9. In this case, with an input heat release rate of 5.0
MW, the peak ceiling jet temperatures were above 600 'C. The decaying nature of the heat release
rate profile (resulting from an oxygen-limited environment) is also reflected, similar to the one
observed for hot gas layer temperature.
Finally, Cases 7 and 15 are also relevant to the ceiling jet temperature. In this case, the input
parameter of interest is the fire elevation (as opposed to the horizontal radial distance and heat
release rate). For a fire located 2.3 m above the floor, the ceiling jet temperatures were above
250 'C. Case 5, which had identical conditions but with a floor base fire, resulted in temperatures
more than 50 'C lower. Figure 5-4 illustrates the temperature profiles for the ceiling jet in Cases 7
and 15.
5-6
Model Sensitivity
5.2.3 Plume Temperature
Three plume temperature sensors were specified in MAGIC's computational domain,
as illustrated in Figure 5-3. The input parameters included in the sensitivity analysis were
the elevation of the sensor above the fire and fire intensity. MAGIC again performed as expected.
Specifically, plume temperatures were lower as the elevation above the fire increased,
temperatures were higher for higher heat release rate profiles, and temperatures were higher than
the corresponding hot gas layer temperatures for evaluated cases. Figures 5-5 illustrates
the plume temperature profiles for Cases 1 and 9, respectively.
Two important observations can be made regarding Figure 5-5. First, the plume temperature for
the lowest sensor (z = 2.5 m) in Case 1 reached values above 700 'C. This is a clear in
indication that the sensor is just outside the steady flame region of the fire. In Case 9, however,
where the fire intensity was 5.0 MW, all peak plume temperatures were above 1,000 'C. These
values should be interpreted as sensors immersed in flames.
In the case of a fire elevated 2.3 m from the floor, MAGIC predicted plume temperatures on the
order of thousands of degrees for the lowest two sensors. Peak flame temperatures are generally
on the order of 1,500 'C.
5-7
Model Sensitivity
urmeTenp
Rum Temp
c00
-Ca-se9,H=a5m
Case 1, H = 2.5 m
Case 1 H 35mM
700
500
2)000-
-CBaseH=3.5m
-Case
Case1,H=45
1500 -
~
9, H=4.5 m
,
I-
400
1000
370
200
500
_
100
0
0
0
5
10
15
,
0
20
5
10
15
20
15
2(
Tr- (mrn)
Trre (rrun)
1!0
Rare Tenp
Aiie Tenp
400
Case 5, H=Z:5 nr
350
30030
250370
700
-
_
1400
C
ase 5, I. = 35 r
Case5 H -4.5n
,g
se
13,H=2Sm
Case 13, H = 3.5 m
1200
-
Case13,H=4.5m
1000
I-
600
,5o
aAJ
400.
20050
0
0
10
5
15
20
5
0
Twre(mdn)
10
Turn(nin)
I9
Anie Tenp
PRim Tenp
250D
16C0
Case 15.H = Z5 m
1400
-
2000
1200
_
'
I/
100C
6C0.Cas
-.
4C0
-.- (se7,H=3.5rr
se7,
H=2.5Fr
500
O-ne7,H=4.5 rr_
200
0
0
0
5
10
15
20
0
5
10
Tim (min)
Tn- (m'n)
Figure 5-5: Plume Temperature Profiles
5-8
C-se15, H =35 m
Case 15H = 4.5 m
150D
15
20
Model Sensitivity
5.2.4 Flame Height
The flame height results illustrated in Figure 5-6 suggest two observations. First, the flame
presents a linear growth during the t2 growth period of the fire. Flame height are constant during
the steady burning period. Second, notice that MAGIC predicted flame heights above the ceiling
height of 4.6 m in Cases 9 and 13.
3
Pam Fight
Parm Heiht
6
Case9
-Co
i
2
5
ggase
13
__0
10
12
fs.1.5
Cse 1
1a,
1
-Ca~se
52
0.5
0,
0
0
5
15
10
r
20
0
5
10
15
20
e (n)
(X)
Figure 5-6: Flame Height Profiles
5.2.5 Oxygen Concentration
Two important aspects of modeling oxygen concentrations in commercial NPP scenarios
are the amount of oxygen available for combustion in the room and the lower oxygen limit
(LOL). These two aspects are, of course, closely related. In terms of the oxygen available for
combustion, the fire consumes whatever oxygen is available. As long as there is oxygen above
the LOL, the fire will burn at its specified heat release rate. The larger the heat release rate, the
larger the amount of the consumed oxygen. Natural and mechanical ventilation conditions will
affect the amount of oxygen available. The LOL is a user input, and the most conservative value
is 0 percent. That is, the fire will bum with an intensity governed by the amount of oxygen or
fuel until all the oxygen in the room has been consumed.
Figure 5-7 illustrates the oxygen concentration profile for Cases 5 and 13, which are simulations
with one open door. Notice that the concentration was well above 10 percent. As expected, Case
13 showed a lower oxygen concentration because of the higher heat release rate.
The mechanical ventilation effects in oxygen concentration profiles can be observed in Figure 5-7.
Notice that there is more oxygen in Cases 2 and 10, in which the mechanical ventilation system
(both injection and extraction) was operating.
However, the effects of low oxygen concentration on the heat release rate can be observed
in simulations with closed doors (specifically Cases 9 and 11). Figure 5-7 compiles the results.
The heat release rate for these cases can be react in the right y-axis. Notice that in Case 9,
where the LOL is 0 percent, the heat release rate begins to decay when the oxygen concentration
is 0 percent. In Case 11, where the LOL is 10 percent, the heat release rate begins to decay
5-9
Model Sensitivity
at 550 seconds, when the oxygen concentration is 10 percent. It is interesting to note that the two
oxygen concentration profiles are identical up to 10 percent. At that point, the fire in Case 9
maintains its original intensity and, therefore, consumes more oxygen than the fire in Case 11.
Heat Pelease Pate
Oxygen Concentration
I.
100
/
0.2
6000.15
'
Ot_
A
3e
Case21
-Case
-._Case
400 -
-/
200-
5
200
>-Case
-
Case 2
5
0.1
0
5
10
1s
0
2C
5
10
T-m (rin)
6000
_ Case 9
5000 - -Case lo
-Ose
13
4000 Case
02
0.15-
A
-.-
7ll
3000
ssel)
°
Case 10
-Case 13
-Oasell
0.05
2(
heatRelease Rate
Oxygen Olcentraton
025
0.1
15
Trne (rin)
1000_
0,
0
0
5
10
15
20
0
5
10
is
20
rTe (frin)
Trne (nirn)
Figure 5-7: Oxygen Concentration Profiles
5.2.6 Smoke Concentration
The relevant output for smoke concentration in MAGIC is average extinction coefficient, k,
with units of 1/m. Average specific area is also available [Ref. 17], which can be also a way to
input soot yield using a conversion factor of 7600 (ye = k/0.0076) [Ref.18]. The average
extinction coefficient can be converted to concentration in units of mg/m 3 or visibility in units of
m with relatively simple algebraic manipulations. For the purpose of NPP applications, visibility
would be the most relevant output. Recall from Ref. 1 that the average extinction coefficient
correlates linearly with visibility, based on the equation S = 3/k for a light-reflecting object, or S
= 8/k for a light-emitting object, where S is the visibility distance in m.
In this sensitivity analysis, Case studies 5, 6, 13 and 14 are relevant to visibility. In those cases,
the fuel was varied from heptane to toluene in order to explore the effects on the average
extinction coefficient. The MAGIC input governing the average extinction coefficient is
the specific area, s, which has units of m2 /kg. The specific area for heptane is 106.4 m2 /kg, while
the value for toluene is 1482 m2 /kg.
5-10
Model Sensitivity
Figure 5-8 summarizes the average extinction coefficient results. Cases 5 and 6 are associated
with a 1.0-MW fire, while Cases 13 and 14 are associated with a 5.0-MW fire. As expected,
the highest extinction coefficient resulted from the toluene fuel burning at an intensity of 5.0 MW.
Shake Concentration
1zi~~o
16001400 -
-Case
1200
-
-
~1000
eg
_-
t
8
.
5
3500
1
'
2500_
_ _ _ _
_ _ _ _
-.-
Case 14
nbtncnrto
Caset -13At.
Cae I
/____
Concentraton
S
ase
ase63DO
_Case9
2000
_ _
1500
600
400
1000
200
_10,500
o
0
0
5
10
15
200
5
Tre (finn)
10
ma
15
(nm)
Figure 5-8: Smoke! concentration profiles
5.2.7 Room Pressure
In addition to the variations in heat release rate, the room openings varied from a leakage path
of 0.015 m2 to a 5.76-M2 open door in order to explore the impact on room pressure. Figure 5-9
illustrates the pressure profiles for Cases 1, 5, 9, and 13, which were simulations with closed
(only leakage paths) and open doors for the two heat release rates selected for the study.
Given the differences in magnitude, profiles for Cases 5 and 13 should be read on the right y-axis.
In open door simulations (Cases 5 and 13) the pressure at the floor was negative, indicating that
fresh air was moving into the enclosure. Recall that the hot gas layer in these simulations did not
reach the floor. The region below this hot gas layer interface is associated with the negative
pressure profiles in Figure 5-9. In terms of sensitivity to heat release rate, the 5.0-MW fire (Case
13) resulted in higher negative pressure, indicating that air would move into the room at higher
velocities than in the case of the 1.0-MW fire. It is interesting to note that the pressurization
levels are on the order of Pascals.
By contrast, for rooms with only leakage paths, the pressure profiles were positive (for the most part)
and on the order of thousands of Pascals. This is an indication that flows are moving out of the
room through the leakage paths. In addition, notice that Case 9 had a negative pressure spike
after 600 seconds. As shown in Figure 5-9, this is the time when the heat release rate suddenly
decays as a result of an oxygen-limited environment. This pressure spike is attributable to
sudden change in heat release rate. At this point, the heat lost to the boundaries is greater than
the heat generated by the fire. After this spike, air begins to move into the enclosure through
the leakage paths, and the fire is able to bum with an intensity governed by the amount of air
drawn into the room. This spike was not observed in Case 1 because the fire had enough oxygen
to burn at its specified intensity.
5-11
Model Sensitivity
Frn Pressure
Poaonmessure
0.5
8000 6000
-Ae1
4-.
4000
4000
ae 2000
Jf
_
0.5I1
_;
6
-2
-2000
o
-2.5
-.
0- -3
-4000
-3.5
5
5
-1
-1.5
L
0
5
t
0B
-
Case 9 --
10 t
15
2
.55
-8000i|5_1e3.
1m (sin)
Trore(rn)
Figure 5-9: Room Pressure Profiles
5.2.8 Target Temperature and Heat Flux
Of primary interest in NPP applications are the effects of the cables' thermo-physical properties
on the predicted surface temperature. As illustrated in Figure 5-2, this analysis included two
types of cables (XPE and PVC). The two cables have different material properties. The effects
of the material properties were explored by comparing surface temperature results for Cases 1
and 9. The only difference between these two cases was the fire heat release rate. As depicted in
Figure 5-11, the selected material properties did not have a significant impact on the surface
temperature profile. Notice that the profiles are almost identical for the XPE and PVC targets in
both cases. However, the damage or ignition temperature was an important distinction.
Another important aspect of evaluating target response in NPP fire scenarios is the difference
between gas temperatures at the location of the target and the surface temperature of that target.
MAGIC provides both results as part of its output library, and Figure 5-11 illustrates this comparison.
In Case 9, the gas temperature was higher than the surface temperature for the first 800 seconds
of the simulation, and the highest temperature difference was just above 100 OC. The temperatures
then converged when the fire was well into its decay stage. By contrast, the gas temperature
was always higher than the surface temperature throughout the simulation in Case 1
and the temperature difference was around 50 'C.
MAGIC offers two modeling alternatives for predicting cable temperature as a thermal target
or a cable. This section compares the two alternatives. The fundamental difference between the
two alternatives is that a cable is treated as a cylinder, while a thermal target is treated as a slab.
This shape difference requires the following computational distinctions:
* numerical resolution one-dimensional plane for targets and one-dimensional cylinder for cables
* convective heat exchange coefficient on a plane or cylindrical surface
* configuration factors for the radiative flux calculation
In the following example, a cable is compared to a target. The target is considered similar
because the thickness of the target (1) is equal to the radius of the cable, and (2) is calculated
conserving the same surface-to-volume ratio. Figure 5-10 illustrates the cable and target.
5-12
Model Sensitivity
/Z
e
D
I
L
Target
Cable
Surface Area
Volume
7DL
D2
v-
4
L
DL
DLe
Figure 5-10: Equivalence between Cable and Targets
. Figure 5-11 above also
4
includes temperature profiles with targets with thicknesses D/4 and D/2. Targets with thickness
D/4 resulted with the highest surface temperature.
The surface-to-volume ratio gives a thickness target value of e =
5-13
Model Sensitivity
XP. Target Terrperature - Case 1
XPt Target Terrerature - Case 9
160
2.
140
Gas Teryp
120
Target V4
Target d1t2
CaleMoe
lo
12010-
A
/|
2-
350
-Gas Terrp
300
-
Target d'4
250
200
t-
50
40
120
20
50
5
10
15
20
0
5
Tira(rrin)
. Gas Terrp
120-
Target V4
100
f/t
10
15
20
15
2D
Tra (rri)
PVCTarget Tefrrerature - Case 1
160 140
A
/E
60
0
15
-
Tart 2
CableModel
PVCTarget Terrgerature - Case9
400
-
-Ga
350
30
TargetV2
O
-
I-
60
--
Terrp
Target 64
250
TargetW2
2502D-
Cabe Moe
150
40
102
20
5O
0
0
5
10
15
20
0
5
IThe
(rrn)
10
Tire (nri)
Figure 5-11: Target temperature profiles
In terms of heat flux, the MAGIC output options "Total heat flux" and "Incident heat flux" are
relevant in this study. The former is the total radiative and convective heat flux contributions to
the target. The later is total radiative heat flux received by the target. The thickness of the target
does not affect these output options.
As illustrated in Figure 5-2, PVC and XPE targets were located 3.55 m away from the fire,
and the elevation of both targets was 2.3 m. In MAGIC, these targets serve as sensors and record
thermal conditions in their specified location. Each case study exhibited the identical predicted
heat flux to each target type. That is, given the symmetrical arrangement of targets relative to
the fire source, both XPE and PVC targets receive the same radiated heat flux in each case. As
expected, the total heat flux is higher than the Incident heat flux due to the contribution of the
convective heat transfer.
Finally, the fuel type appears to have some effect on thermal radiation levels. According to
the results, the simulations conducted with heptane fires produced higher heat fluxes than the
corresponding simulations conducted with toluene fires (Cases 6 & 14), although the magnitude
differences were less than 1 kW/m2 . Figures 5-12 and 5-13 compile the graphical results.
5-14
Model Sensitivity
XFRTarget Heat Rux - Case 1
PVCTarget Flw RFux
- Case 1
3
2.5
- Total W
-hcideent HFF
2
_
I
0.5
07
.
0
5
10
15
20
5
The (fmn)
1.5-
10
15
20
Timr(min)
PVC Target lal RIux- Case 5
XFL Target Heat Rux - (CseS
3-a-
Total FF
25
WicidntFF
_
2-
25
Y
Total FF
1.5
1.5
m
I
-J:
1
0.5
0.51
0
2
0
0
0D
5
10
15
0
20
5
Tkne(rnin)
X
10
15
20
15
20
Tia (mdn)
P/C Target HWatFux - Case 6
XR- Target Flet Fks - Case 6
2.5
ZS
2 _Total FF
_ Total FF
'
I--h
ncident I-F
V
Er.
y3
I.
1
,
1.5
1
0.51
051
0
0
5
10
Tir
15
20
0
10
5
(nin)
r-
(min)
Figure 5-12: Target heat flux profiles, 1MW fire.
5-15
Model Sensitivity
XR Target Feat Rux - Case 9
FVCTarget Flaw Rux - Case 9
14
14
12
-i-Total
10
-_-
W-
cident
-+-Total -FF
12
I-F
10 - -bcidentrd-F
8
6
6
I
-
4
4
2
2
0
0
0
5
10
15
3
20
Tre (nin)
PVC Target Heat Rux - Case 13
XaL Target l-aI FRux
- Case 13
12
12
0F
15
10
5
0
Tire (min)
1.-Total
I0
Total 220
FF.
-hwident FF,,
6
~
/
6
IX
I
4
2
0
5
10
15
5
0
20
15
10
20
Tim (rrnh)
Tim (nin)
PVCTarget Flag FRux- Case 14
12
10 10
,
Total FF__
_
hcident tfw
-
8
z
I7
UI
6
4
2
0
5
10
Time(trn)
15
20
0
0
10
Twy-(ninh)
Figure 5-13: Target heat flux profiles, 5MW fire.
5-16
15
20
Model Sensitivity
5.2.9 Wall Temperature
Figures 5-14 and 5-15 summarize a comparison between the use on MAGIC output options
"Wall Temperature", and "Surface Temperature of the target". The output "Wall temperature"
results from a one dimensional finite difference conduction calculation in to the walls. The
internal boundary condition is the thermal properties of the gas in contact with the particular wall
surface (upper or lower gas layer). The "Surface temperature of the target" output option results
from a conduction calculation into a slab of similar thickness as the wall.
This comparison is important because the validation study described in Chapter 6 for wall
temperature was developed using the later option. That is, virtual sensors in MAGIC, e.g.,
targets, were specified in the same location as the wall thermocouples in the experimental series.
The targets were specified with the same thermo-physical properties and thickness as the walls.
The only difference in the specification is the emissivity. The targets had an emmisivity of 0.95
and MAGIC does not require emissivity as a wall property input.
Results suggests that the use of the "Target" feature in MAGIC to predict wall surface
temperature can produce higher temperatures with the exception of the floor surface in open door
tests. In the cases ran with a 1 MW fire, the temperature difference between both modeling
strategies is approximately 10 'C. In cases ran with a 5 MW fire, the temperature difference is
between 40 'C and 60 'C.
5-17
Model Sensitivity
WMilTerrperature - Case 1
70 -a-Wag
60
2.
50
I-
40
Well Tenqperature
- Case S
60
Terrp
Target Terrp
-Wal Terrp
0E
TargetTerrp
40
E
I-
30;
30
20
20
10
10
0
0
0
5
10
15
20I
0
10
5
20
15
Trre (rmh)
T-re (rrn)
Rccr Terrperature - Case 1
loor Tenperature - Case 5
70
_
eo
6 _ Wal Terrp
Wal Terrp
50
_
Target Tenp_
40
60
-
-2
Target Terrp
30
20
20
10
10
0
10
5
15
0
20I
re,-
Cling Terrperature - Case 1
70
70
-Wall Terrp
60
_Target Terrp
60 5
1
Wa
Terrp
ZTargetTeang
50
40
t--
I;.
30
_
30
20
104-
U10
,
O-,
0
5
10
Tre (min)
15
20
0
5
10
Tur (nin)
Figure 5-14: Target temperature vs. wall temperature, 1 MW fire
5-18
20
15
(rin)
Cakirg Tenperature - Case 5
so
b
10
5
Too~(rin)
15
_
20
Model Sensitivity
Wai Ternperature - Case 13
WainTerrperature - Case 9
200
160
140
-_
120 - -
_
160 -
Wal Tenp
-
Wal Terrp
160-
-TargetTenp
2.
100
120 - -
Target Teitp
120
100
60
40
200
0
5
10
15
0
20
5
10
Tare (rrn)
_
Wall Teinp
180
80
160 -a-Wal
f
140 100 -
Target Terrp
a-
15
20
Tert
Target Terrp
120
100
20
20
0
140
0
20
Floor Terrperature.-Case 13
Foor Terrperature - Case 9
140
120 -
15
Talre(irrm)
5
10
15
/_~
0
20
_-TargetTerrp
5
_10
Tftre (rrn)
mrre(rnn)
so
Cainrg Teirperature - Case 9
CRag Tenrperature - Case 13
60
160
2C00
140 - _Wat TOWr
20
160
0
40
::
120 - _ Target Terrp
50
52
Tur
(ri
120
100
I-
60
/
20
cu
_
0
5
10
Torne(rrn)
15
20
0
5
10
15
20
Tena(crn)
Figure 5-15: Target temperature vs. wall temperature, 5MW fire
5-19
Model Sensitivity
5.3 Concluding Remarks
This chapter illustrates the effect of the most important parameters in fire modeling with the code
MAGIC. The set of models included in MAGIC are intended to translate the impact of those
parameters on the fire-generated conditions in a compartment. It is therefore important to
understand the effects the input parameters have in the predicted fire conditions considering that
the simulation results are simplifications and idealizations of real fire induced temperatures and
flows.
It is difficult to generalize which input parameters are more important that others since it depend
on specific applications, and most (if not all) of the parameters are mathematically related. As
illustrated in the chapter, different parameters are important for different sub-models. In most
applications, the fire modeling analyst will need to determine which outputs are relevant for the
scenario under evaluation, which parameters will affect those outputs, and how variations in
those parameters will impact the conclusions made from the simulation results.
5-20
6
MODEL VALIDATION
This chapter summarizes the results of a validation study conducted for the zone model MAGIC,
in which its predictions are compared with measurements collected from six sets of large-scale
fire experiments. A brief description of each set of experiments is given here. Further details can
be found in Volume 7 and in the individual test reports.
ICFMP BE #2: Benchmark Exercise #2 consists of 8 experiments, representing 3 sets of
conditions, to study the movement of smoke in a large hall with a sloped ceiling. The results of
the experiments were contributed to the International Collaborative Fire Model Project (ICFMP)
for use in evaluating model predictions of fires in larger volumes representative of turbine halls
in NPPs. The tests were conducted inside the VPTT Fire Test Hall, which has dimensions of 19 m
high by 27 m long by 14 m wide (62 ft x 88.5 ft x 46 ft). Each case involved a single heptane
pool fire, ranging from 2 MW to 4 MW.
ICFMP BE #3: Benchmark Exercise #3, conducted as part of the International Collaborative Fire
Model Project (ICFMP) and sponsored by the US NRC, consists of 15 large-scale tests
performed at NIST in June, 2003. The fire sizes range from 350 kW to 2.2 MW in a
compartment with dimensions 21.7 m x 7.1 m x 3.8 m (71 ft x 23 ft x 12.5 ft), designed to
represent a variety of spaces in a NPP containing power and control cables. The walls and ceiling
are covered with two layers of 25 mm thick marinate boards, while the floor is covered with two
layers of 25 mm thick gypsum boards. The room has one 2 m x 2 m (6.6 ft x 6.6 ft) door and a
mechanical air injection and extraction system. Ventilation conditions and fire size and location
are varied, and the numerous experimental measurements include gas and surface temperatures,
heat fluxes, and gas velocities.
ICFMP BE #4: Benchmark Exercise #4 consists of kerosene pool fire experiments conducted at
the Institut fur Baustoffe, Massivbau und Brandschutz (iBMB) of the Braunschweig University
of Technology in Germany. The results of two experiments were contributed to the International
Collaborative Fire Model Project (ICFMP). These fire experiments involve relatively large fires
in a relatively small (3.6 m x 3.6 m x 5.7 m high, 11.8 ft x 11.8 ft x 18.7 ft) concrete enclosure.
Only one of the two experiments was selected for the present V&V study (Test 1).
ICFMP BE #5: Benchmark Exercise #5 consists of fire experiments conducted with realistically
routed cable trays in the same test compartment as BE #4. Only one test (Test 4) was selected for
the present evaluation, and only the first 20 min during which time an ethanol pool fire preheats
the compartment.
FM/SNL Series: The Factory Mutual & Sandia National Laboratories (FM/SNL) Test Series is a
series of 25 fire tests conducted for the NRC by Factory Mutual Research Corporation (FMRC),
under the direction of Sandia National Laboratories (SNL). The primary purpose of these tests
was to provide data with which to validate computer models for various types of NPP
compartments. The experiments were conducted in an enclosure measuring 60 ft long x 40 ft
wide x 20 ft high (18 m x 12 m x 6 in), constructed at the FMRC fire test facility in Rhode
6-1
Model Validation
Island. All of the tests involved forced ventilation to simulate typical NPP installation practices.
The fires consist of a simple gas bumer, a heptane pool, a methanol pool, or a polymethylmethacrylate (PMMA) solid fire. Four of these tests were conducted with a full-scale control
room mockup in place. Parameters varied during testing are the heat release rate, enclosure
ventilation rate, and fire location. Only three of these tests have been used in the present
evaluation (Tests 4, 5 and 21). Test 21 involves the full-scale mock-up. All are gas burner fires.
NBS Multi-Room Series: The National Bureau of Standards (NBS, now the National Institute of
Standards and Technology, NIST) Multi-Compartment Test Series consists of 45 fire tests
representing 9 different sets of conditions, with multiple replicates of each set, which were
conducted in a three-room suite. The suite consists of two relatively small rooms, connected via
a relatively long corridor. The fire source, a gas burner, is located against the rear wall of one of
the small compartments. Fire tests of 100, 300 and 500 kW were conducted, but for the current
V&V study, only three 100 kW fire experiments have been used (Test 100A, 1000, and 100Z).
Technical details of the calculations, including output of the model and comparison with
experimental data are provided in Appendix A. The results are organized by quantity as follows:
*
Section 6.1: Hot gas layer temperature and height
*
Section 6.2: Ceiling jet temperature
*
Section 6.3: Plume temperature
* Section 6.4: Flame height
* Section 6.5: Oxygen concentration
* Section 6.6: Smoke concentration
* Section 6.7: Room pressure
*
Section 6.8: Target temperature and heat fluxes
*
Section 6.9: Wall temperature and heat fluxes
The model predictions are compared to the experimental measurements in terms of the relative
difference between the maximum (or where appropriate, minimum) values of each time history:
&M -AE
A.E
-
(M p - M.)- (EP - E)
-E 0
AM is the difference between the peak value of the model prediction, Mp, and its original value,
Mo. AE is the difference between the experimental measurement, Ep, and its original value, ED.
A positive value of the relative difference indicates that the model has over-predicted the severity
of the fire; for example, a higher temperature, lower oxygen concentration, higher smoke
concentration, etc.
6-2
Model Validation
Each section in this chapter contains a scatter plot that summarizes the relative difference results
for all of the predictions and measurements of the quantity under consideration. The details of
the calculations, the input assumptions, and the time histories of the predicted and measured
output are included in Appendix A. Only a brief discussion of the results is included in this
chapter. At the end of each section, a color rating is assigned to each of the output category,
indicating, in a very broad sense, how well the model treats that particular quantity. Colors are
assigned based on the following criteria. Once the user determines the validation results reported
here are applicable (see Volume 1), the user must determine the predictive capability of the fire
models. The following two criteria are used to characterize the predictive capability of the
model:
Criterion 1 - Are the physics of the model appropriatefor the calculation being made? This
criterion reflects an evaluation of the underlying physics described by the model and the physics
of the fire scenario. Generally the scope of this study is limited to the fire scenarios that are
within the stated capability of the selected fire models, e.g., this study does not address the fire
scenarios that involve flame spread within single and multiple cable trays.
Criterion 2 - Are there calculatedrelative differences outside the experimental uncertainty? This
criterion is used as an indication of the accuracy of the model prediction. Since fire experiments
are used as a way of establishing confidence in model prediction, the confidence can only be as
good as our experiments. Therefore, if model predictions fall within the ranges of experimental
uncertainties, the predictions are determined to be accurate. However, one should recognize that
the experimental uncertainties vary with the experiment and the attribute being measured (see
volume 7 of this report). These ranges could be as much as +50% for the experiments and
attributes we used in this validation study. This leads to some judgment about the relative
difference proximity relative to the range of uncertainty, as the uncertainty ranges are not
necessarily all-inclusive or definitive.
The predictive capability of the model is characterized as follows based on the above criteria.
: If both criteria are satisfied, i.e., the model physics are appropriate for the calculation
being made and the calculated relative differences are within or very near experimental
uncertainty, then the V&V team concluded that the fire model prediction is accurate for the
ranges of experiments in this study, and as described in tables 2-4 and 2-5. A grade of GREEN
indicates the model can be used with confidence to calculate the specific attribute. The user
should recognize, however, that the accuracy of the model prediction is still somewhat uncertain
and for some attributes, such as smoke concentration and room pressure, these uncertainties may
be rather large.
I
: If the first criterion is satisfied and the calculated relative differences are outside
the experimental uncertainty but indicate a consistent pattern of model over-prediction or underprediction, then the model predictive capability is characterized as YELLOW+ for overprediction, and YELLOW- for under-prediction. The model prediction for the specific attribute
may be useful within the ranges of experiments in this study, and as described in tables 2-4 and
2-5, but the users should take care and use caution when interpreting the results of the model. A
complete understanding of model assumptions and scenario applicability to these V&V results is
necessary. Generally, the model may be used if the grade is YELLOW+ when the user ensures
that model over-prediction reflects conservatism. The user must exercise caution when using
models with capabilities described as YELLOW+.
6-3
Model Validation
YELLO: If the first criterion is satisfied and the calculated relative differences are outside
experimental uncertainty with no clear pattern of over- or under-prediction, then the model
predictive capability is characterized as YELLOW. Caution should be exercised when using a
fire model for predicting these attributes. In this case, the user is referred to the details related to
the experimental conditions and validation results documented in volumes 2 through 6. The user
is advised to review and understand the model assumptions and inputs, as well as the conditions
and results to determine and justify the appropriateness of the model prediction to the fire
scenario for which it is being used.
M : If the first criterion is not met, then the particular fire model capability should not be used.
No color: This V&V study did not investigate this capability. This may be due to one or more
reasons that include unavailability of appropriate data or lack of model, sub-model, or output.
As suggested in the criteria above, there is a level of engineering judgment in the classification
of fire model predictive capabilities. Specifically, engineering judgment is exercised in the
following two areas:
1. Evaluation of the modeling capabilities of the particular tool if the model physics are
appropriate.
2. Evaluation of the magnitude of relative differences when compared to the experimental
uncertainty. Judgment in this area impacts the determination of Green versus Yellow colors.
In general, a Green or Yellow classification suggests that the V&V team determined that the
model physics are appropriate for the calculation been made, within assumptions. The difference
between the colors is due to the magnitude of the calculated relative differences. Judgment
considerations include general experimental conditions, experimental data quality, and the
characterization of the experimental uncertainty.
6.1 Hot Gas Layer Temperature and Height
The single most important prediction a fire model can make is the temperature of the hot gas
layer. After all, the impact of the fire is often assessed not only a function of the heat release
rate, but also as a function of the compartment temperature. A good prediction of the height of
the hot gas layer is largely a consequence of a good prediction of its temperature because smoke
and heat are largely transported together and most numerical models describe the transport of
both with the same type of algorithm. Following is a summary of the accuracy assessment for
the hot gas layer predictions of the six test series:
ICFMP BE #2: MAGIC under-predicts the hot gas layer temperature by less than 10 % for all
three cases. This falls within the range of experimental uncertainty. In addition, MAGIC underpredicts also by less than 10 % the hot gas layer height in all three cases. That is, the MAGIC
height prediction is above the measured one. A graphical comparison of the MAGIC predictions
and the experimental observations for these three cases is presented in Figure A-2. The scatter
plot in Figure 6-1 illustrates the relative differences between the measured and predicted peak
hot gas layer temperatures and heights.
ICFMIP BE #3: MAGIC predicts the hot gas layer temperature and height to within
experimental uncertainty for all 15 tests. It should be noted that the discrepancies in the hot gas
6-4
Model Validation
layer height depicted in Figures A-4, and A-5 (which refers to closed door tests) is due to the
data reduction method used to determine the experimental layer interface. This method is not
applicable for tests in which a single gas layer develops. Notice that MAGIC predicts that the
hot gas layer eventually reaches the floor generating a single gas layer in the room. That
prediction is consistent with visual observations during the experiments. Due to the
inconsistency between model results and the reduced experimental data, no relative differences
were calculated for closed-door tests.
The collection of graphical comparisons between MAGIC predictions for hot gas layer
temperatures and heights for ICFMP BE #3 is presented in Figures A-4 to A-7. The relative
differences calculated for peak values are summarized in Table A-2 and Figure 6-1.
ICFMP BE #4: MAGIC predicts the hot gas layer temperature within experimental uncertainty
for the single test (Test 1). However, there is some discrepancy in the shapes of the curves for
the hot gas layer height (see Figure A-i 1). This discrepancy is associated with a relative
difference of 25%, which is outside the range of experimental uncertainty. A possible
explanation for the discrepancy in the layer height is the fact that the room was almost engulfed
in flames, which may not be consistent with the fundamental assumption in MAGIC of two
distinct gas layers. The relative differences for layer temperature and height are plotted also in
Figure 6-1.
ICFMP BE #5: MAGIC predicts the hot gas layer temperature and height to within
experimental uncertainty for the single test (Test 4). The graphical comparison between
experimental measurements and model predictions, illustrated in Figure A-Il suggests very good
agreement between the profiles. The calculated relative differences for peak hot gas layer
temperature and height are listed in Table A-4.
FMISNL: MAGIC predicts the hot gas layer temperature to within experimental uncertainty for
Tests 4, 5 and 21. In the case of the hot gas layer height, there are inconsistencies in the
comparison of hot experimental measurements and model predictions. As discussed earlier for
the case of closed door tests in ICFNP BE#3, the data reduction method for determining hot gas
layer height is not applicable for closed door tests. Consequently, the graphical comparisons
presented in Figure A-13 do not show good agreement between model predictions and
experimental measurements. For that same reason, no relative differences were calculated for
hot gas layer height in this test series.
NBS Multi-Room: MAGIC predictions in this test series are for the most part outside the
experimental uncertainty for both the fire room and adjacent compartments. The largest relative
differences are associated with adjacent compartments to the fire room. As depicted in Figures
A-1S to A-17, and in Figure 6-1, MAGIC over predicted the hot gas layer temperature in all
cases. The hot gas layer height shows under-predictions of up to around 50% in some of the
rooms.
Summary: Hot Gas Layer Temperature and Height
*
The research team considers the MAGIC model for calculating hot gas layer temperatures to
be appropriate for its intended applications.
6-5
Model Validation
*
The MAGIC predictions of the hot gas layer temperature and height are, with the exception
of the selected tests from the NIBS test series, within experimental uncertainty of 15%.
The scatterplot in Figure 6-1 summarizes the relative differences calculated for hot gas layer
temperatures and height. As explained earlier, no relative differences were calculated for
close door tests.
Validation results suggest that MAGIC is certainly suited for prediction hot gas layer
temperatures and heights in scenarios where this study is applicable. Because most of the
validation results are within experimental uncertainty, and MAGIC is over predicting hot gas
layer temperatures in the selected tests from the NIBS test series, a color assignment of green
is assigned for the room of fire origin and a yellow + for adjacent rooms. In the case of hot
gas layer height, a green classification is assigned for the room of fire origin and a yellow for
adjacent rooms.
*
*
100%
Hat Gas Layer Terrp and Height
80%
60% U
40% 20% -
--
0
--
~~
~
o
ko
00
~
--
-
~
~
~
--
~
o
0
- - -
- -
-M-
-20%-40% -60%o-_
Terp (PRoom of fire origin)
*
Temp (Adj rooam)
*
-80% --.------
Height (Rtom of fire origin)
EU Height
x
Height (Adj rooms)
-100% -_._
- CJ co
y
-;
N~
Ir.
CY
co
6^^N 6
0
-
co
co
-7^
-
'-_ co
-
A
Terrp (Adj rooms)
- - - - EJ Teryp
0) Lo
-
w
m w
m w
m w
m um w
m m m WW X
WWW mWmWm
It
4^7
in
CO
7
;
'It
688
COui OM
Xwm
m
~
~
z
0
N
- -
-
U- Z
Z
Z
Figure 6-1: Scatter plot of relative differences for hot gas layer temperature and height in
ICFMP BE #2,3,4,5,and the selected FMWSNL and NBS Tests. Experimental uncertainties
are 13% (HGL temp) and 9% (HGL height).
6.2 Ceiling Jet Temperature
The ceiling jet algorithm in MAGIC consists primarily in the model proposed by Cooper [Ref 66]. It is recommended that analysts review MAGIC's technical reference (Ref 6-2) for specific
details about the implementation of the ceiling jet algorithm. Typical of ceiling jet correlations,
it applies only to the flow of hot gases under a flat ceiling. Only two of the six test series
(ICFMP BE #3 and FM/SNL) involved a ceiling jet formed over a relatively wide, flat ceiling.
ICFMP BE #3: MAGIC predicts the ceiling jet temperature to within experimental uncertainty
with the exception of three tests, Test 10, 13, & 16 as illustrated in Figure 6-2. Interestingly,
Test 10 is a replicate of Test 4, which was predicted to within experimental uncertainty. The
ventilation system was on during these two tests, and the inconsistent results may be attributed to
it. It is difficult to draw conclusions about over predictions in Test 15 and 16. Figure 6-2 also
suggests that the relative differences for opened door tests are smaller (near 0%) than those in
closed door tests. Furthermore, only two under-predictions, -7% in Test 1 and -1% in Test 14
were calculated.
6-6
Model Validation
The graphical comparisons between experimental measurements and MAGIC predictions for
ceiling jet temperature are grouped in Figures A-18 and A-19. Table A-9 lists the calculated
relative differences.
FM/SNL: MAGIC predicts the ceiling jet temperature at two locations in Test 4, 5 and 21 to
within experimental uncertainty. The graphical comparisons are provided in Figure A-20. The
calculated relative differences are listed in Table A-10, and plotted in Figure 6-2.
Summary: Ceiling Jet Temperature
* With three exceptions, corresponding to closed-door tests in ICFMP BE#3, MAGIC ceiling
jet predictions are within experimental uncertainty. .
* The MAGIC ceiling jet sub-model is well suited for the range of scenarios validated in this
study. In general, this validation applies to ceiling jet flows under flat unobstructed ceilings
and a rfH up to 1.7 (r is the horizontal radial distance and H is the distance between the fire
source and the ceiling). Notice that the MAGIC technical manual (Ref 6-2) suggest that "the
model is valid up to r/H = 3. In MAGIC they are applied up to r/H =10, to avoid a
discontinuity in gas temperature. In cases where the ceiling-jet exceeds r/H = 10, it is
assumed that the gas temperature beyond r/H = 10 equals the temperature at r/H = 10. These
hypotheses are conservative even if they have not been explored experimentally".
* Based on the model robustness and the fact that most relative differences are within
experimental uncertainty of 15% and no outliers were observed, a green classification
assigned.
100%
CeDing Jet Tenperature
80%/
60%
40%
o
20%
..................................................
.............................
..............
0200
A.
-40%
-60%
-60%.80%/
°o Tree 7-10
* 11/98H
A 11198H
-100%
cm
ww
co
LU
0
w
7
-)
M
r')
Iii WLUw r
N
C'U
wm
_J Xf0
^^te0tJ
w C')
L
LLU.
Z
Z
2
Figure 6-2: Scatter plot of relative differences for ceiling Jet temperatures in ICFMP BE #3,
and the selected FM/SNL Tests. Experimental uncertainty Is 16%.
6-7
Model Validation
6.3 Plume Temperature
As with the ceiling jet, MAGIC has a specific plume sub-model. This validation refers primarily
to the implementation of the McCaffrey plume temperature correlation and the correction for
plume flows above the hot gas layer interface. The line plume model in MAGIC was not
evaluated in this study. Data from ICFMP BE #2 and the FM/SNL test series have been used to
assess the accuracy of plume temperature predictions.
ICFMP BE #2: MAGIC predictions of plume temperature are within the experimental
uncertainty of 15%. Figure A-22 provides the graphical comparisons between model predictions
and experimental measurements. The calculated relative differences are listed in Table A-1I1 and
plotted in Figure 6-3.
FM/SNL: MAGIC predicts the plume temperatures in Test 4 and 5 to within experimental
uncertainty. See Figure A-23 and Table A-12 for the graphical comparisons and the calculated
relative differences.
Summary: Plume Temperature
* The axisymmetric plume temperature model in MAGIC is well suited for applications similar
to the ones evaluated in this study.
* Since all the relative differences are within experimental uncertainty and the experimental
and predicted temperature profiles show good agreement, a classification of green is assigned
for the axisymmetric plume model in MAGIC.
100%
Ruris Tenperature
80%
60%
40%20%
0%/
A,
io/ ..9 ................................................................................
A............
-20%
o TG.1
-40%
TG.2
-60% -
vA28/98H
-80%
-100%X, -
_
l
X.
_
T
N
X
0
t
X
-
C
X
O)
D
)
UC
Ut
m mm
mm m
m
uL
uL
CD
-
*
0
U)
O
LU L
N
o
co
c
X
m
L
U-
zZ
Z
Figure 6-3: Scatter plot of relative differences for plume temperatures In ICFMP BE #2, and
the selected FWSNL Tests. Experimental uncertainty is 14%.
6-8
Model Validation
6.4 Flame Height
Flame height is recorded by visual observations, photographs or video footage. Videos from the
ICFMP BE # 3 test series and photographs from BE #2 are available. It is difficult to precisely
measure the flame height, but the photos and videos allow one to make estimates accurate to
within a pan diameter.
The MAGIC model for flame height consists in the Heskestad's flame height correlation. See
reference 6-2 for technical details.
ICFMP BE #2: The height of the visible flame in the photographs of BE #2 has been estimated
to be between 2.4 and 3 pan diameters (3.8 m to 4.8 m). From Figure A-24, which reports
MAGIC flame height predictions, flame heights are between 3 and 7 m.
ICFMP BE #3: MAGIC appears to predict the flame height correctly in this test series, at least
to the accuracy of visual observations and a few photographs taken before the hot gas layer
obscures the upper part of the fire. The experiments were not designed to measure the flame
height other than through visual observation. Flame height pictures and MAGIC predictions can
be found in Figures A-26 to A-28. Notice for example that Figure A-26 suggests flames with
heights similar to the height of the door (2 m). MAGIC predictions peak above 2 m in all cases.
Summary: Flame Height
* MAGIC appears to provide flame height predictions consistent with the ones observed in
available photographs for BE#2. MAGIC flame height predictions for BE#3 are also
consistent with observations made from available photographs.
* This evaluation does not suggest that MAGIC is under-predicting flame height. Therefore,
based on the consistency with visual evidence, a green classification is assigned.
6-9
Model Validation
6.5 Oxygen Concentration
The oxygen concentration in MAGIC results directly from the conservation of mass equation in
both the upper and lower layer. The evaluation results are based on oxygen concentrations
calculated in the upper layer. It should be stressed that this study is limited to well-ventilated
fires.
ICFMIP BE #3: The relative differences associated with MAGIC predictions of upper layer
oxygen concentration range from approximately -30% to 25%. Some of these relative
differences are outside the range of experimental uncertainty of 9%. As suggested in Figure 64,
there appears to be a pattern of negative relative differences associated with open door tests not
observed in the close door tests. In all these cases, the measured oxygen concentration was
above 15%. In terms of the close door tests, all the relative differences are within experimental
uncertainty with the exception of tests 4 and 10, which consisted of mechanically ventilated
room. Recall that negative relative differences indicate that MAGIC predicted higher oxygen
concentrations that those measured in the experiments. Figures A-29 and A-30 illustrate the
experimental and model oxygen concentration profiles.
ICFM[P BE #5: MAGIC prediction of the upper layer oxygen concentration in Test 4 of this test
series is above the experimental uncertainty of 9%.
Summary: Oxygen Concentration
* The MAGIC model is capable of making oxygen concentration predictions, assuming that
the basic stoichiometry of the combustion reaction is known. Recall that this study is limited
to well-ventilated compartment fires only.
* Relative differences in BE # 3 are comparable to experimental uncertainty in the case of
close room tests. In the case of open door tests, they are below experimental uncertainty
range in BE#3. The relative difference associated with the single comparison in BE#5 is also
outside the range of experimental uncertainty.
* Based on the above discussion, a classification of yellow is assigned for the oxygen
concentration predictions in MAGIC.
6-10
Model Validation
100%
Oxygen ODncentration
80%
60%/
40%/
20%
0
-
-
~~~-0%
>-av
- o-0 -
-
- --
--
-20% ---
0
---
----
-o-
7OO
020
---
---
---
40
a
0
---
---
o
o
---
-_- -----
_-_--_-__-_-__-_-
-40%
-60%
-80%
-100%
_
c
e
-
.
N
C%
V
0
CO
CD
t-
M
L£0
,
V
C
-
V
to
UULZ
0
N
Z'
Z
Figure 6-4: Scatter plot of relative differences for oxygen concentration in ICFMP BE#3 and
BE#5. Experimental uncertainty is 16%.
6.6 Smoke Concentration
Only ICFMP BE #3 has been used to assess predictions of smoke concentration. For these tests,
the smoke yield was specified as one of the test parameters. MAGIC consistently over predicted
smoke concentrations from approximately 30%b to 500% in the closed door tests. For open door
tests, MAGIC predictions are within the experimental uncertainty of 33%.
The graphical comparisons for smoke concentration are summarized in Figures A-32 and A-33.
The relative differences are listed in Table A-15 and plotted in Figure 6-5.
Summary: Smoke Concentration
* MAGIC is capable of transporting smoke throughout a compartment, assuming that the
production rate is known and that its transport properties are comparable to gaseous exhaust
products.
* MAGIC over predicts the smoke concentration in close door tests. The predictions for open
door tests are within experimental uncertainty.
*
No firm conclusions can be drawn explaining why the drastic differences in predictions
between open and close door tests. Therefore, a yellow classification is assigned since there
is no clear indication that MAGIC would always result in conservative estimates.
6-11
Model Validation
600%
Smoke Concentration
500
-
0
400% 0
300%
0
0
200%
°
0
0
0
100% 0
---
---
---
---
--
------- ----
---
,
00/6
,
,o
A-
0 ,X
-100%..
.
CU
X
G'
.
r
V)
c
CY
l)
6
cm
UJ
mUJ Co CD
m OD
Ch
0O
h '
m
X) c0 D
-7
7 ;
-
X
am
Chw j
u
ua
u)
0)
"' W
m m
WWm m m
0)
0
C*
-
w
u j
C
".t
LO
5
-
O
c
L.Z
N
C 2Z
Figure 6-5: Scatter plot of relative differences for smoke concentration In ICFMP BE#3.
Experimental uncertainty is 33%.
6.7 Compartment Pressure
Comparisons between measurement and prediction of compartment pressure for BE #3 are
shown in Figure A-34 and A-35. For those tests in which the door to the compartment is open,
the over-pressures are only a few Pascals, whereas when the door is closed, the over-pressures
are several hundred Pascals.
The relative differences were calculated as follows:
*
For closed-door rooms, the relative difference refers to the positive peak at the early stages of
the fire. Positive relative differences indicate that MAGIC over-predicted the measured
peak.
*
For opened door rooms, the relative difference refers to the negative magnitudes of the
pressure, typically at the late stages of the test. Positive relative differences suggest that
MAGIC calculated a more negative difference than the experimental measurement.
Relative differences are listed in Table A-16.
Visual examination of experimental data and model results plots (see Figure A-34) strongly
suggest that tests with open doors, where leakages are not critical because of the large door
opening, MAGIC captures both the magnitude and the profile of the pressure. These figures
describe a negative pressure profile at the floor of the room, indicating that fresh air is moving
into the enclosure.
In close door tests, MAGIC is able to capture the both peaks and pressure profiles (see Figure A35). It is important to mention that fan tests were conducted before some of the tests resulting in
relatively well known leakage areas. Furthermore, notice that MAGIC captures the positive and
negative pressure peaks. These peaks are an indication of a positively pressurize room in the
early stages of the test, and a negatively pressurize room when the fuel supply is discontinued
and heat loses to the boundaries are higher than the fire hear release rate.
6-12
Model Validation
In general, the predicted pressures are of comparable magnitude to the measured pressures, and
in most cases differences can be explained using the reported uncertainties in the leakage area
and the fact that the leakage area changed from test to test because of the thermal stress on the
compartment walls. The one notable exception is Test 16. This experiment was performed with
the door closed and the ventilation on, and there is considerable uncertainty in the magnitude of
both the supply and exhaust flow rates.
The relative differences are plotted in Figure 6-6. Notice that only the relative difference
associated with Test 16 is outside the experimental uncertainty ranges of 50% and 75% for tests
with ventilation system off or on respectively.
Summary: Compartment Pressure
*
The basic mass and energy conservation equations solved by MAGIC ensure reliable
predictions of compartment pressure. It should be stressed that compartment pressure
predictions are extremely sensitive to the leakage area and forced ventilation. In the MAGIC
runs, leakage area listed in the experimental descriptions was divided by the orifice flow
coefficient, 0.68, so that it is reflected in the model as the actual opening area.
*
The MAGIC pressure predictions for BE #3 are within experimental uncertainty, with an
exception that may be related to the behavior of a ventilation fan.
*
A green classification is assigned for compartment pressure predictions in MAGIC assuming
that room leakages are known in closed-door scenarios.
150%
Room Pressure
|U
-
100% .
-
-
-
-
-
-
-
-
-6 -
-
-
-
-
-
-
-
-
-
*
-
Forced vent
E- ND forced vent
-
-
-
-
-
500/%
0
01/0
...........
-50%
.......
.
O
-100%/
N
to
CN) -X.
u.)W r
N
X
V
r
ER
a
M
CD
I'-
CO
C)
LCu
Ln
U
LO
L
__
Ln
t CO
Lu
LU
Su
N
<
%
LLLL
_
_
_
z
Z
Z
Figure 6-6: Scatter plot of relative differences for room pressure in ICFMP BE#3.
Experimental uncertainties are 40% (no forced vent) and 80% (forced vent).
6-13
Model Validation
6.8 Radiation, Total Heat Flux and Target Temperature
Target temperature and heat flux data are available from ICFMP BE #3, #4 and #5. In BE #3,
the targets are various types of cables in various configurations - horizontal, vertical, in trays or
free-hanging. In BE #4, the targets are three rectangular slabs of different materials instrumented
with heat flux gauges and thermocouples. In BE #5, the targets are again cables, in this case
bundled power and control cables in a vertical ladder.
ICFMP BE #3: There are nearly 200 comparisons of heat flux and surface temperature on four
different cables that are graphed in the Section A.8.1. Consequently, it is difficult to make
sweeping generalizations about the accuracy of MAGIC. The section is classified by target. For
each target, the graphical comparison of experimental measurements and MAGIC predictions are
presented for Target temperature, Cable temperature, radiation and total heat flux. At best, one
can scan the figures and the associated tables to get a sense of the overall performance. The
experimental uncertainty is about 20 % and 14% for heat flux and surface temperature
respectively. The following important aspects of this evaluation should be considered:
* MAGIC provides the capability of modeling cable temperature as Targets, where the material
is simulated as a slab, or a cable itself, where the material is simulated as a cylinder with
concentrical layers of conductor, insulation and jacket. This evaluation includes both
alternatives. When cables are modeled as targets (slab), the thickness of the target was
selected as d/4 were d is the diameter of the cable (see discussion on cable modeling and
sensitivity analysis in Chapters 3 & 5).
* The measured radiative heat flux is compared with MAGIC output "Incident heat flux"
which is the sum of all radiated heat fluxes to a target. The total heat flux measurements is
compared with MAGIC output "Total heat flux, flux meter", which simulates a typical watercooled heat flux meter.
Figures 6-7 to 6-10 show the relative differences for target and cable temperature as well as
radiative and total heat fluxes for targets B-TS-14, D-TS-12, F-TS-20, and G-TS-33. The
following observations are relevant:
* It can be concluded that the majority of the comparisons resulted within experimental
uncertainty or over predictions for surface temperature.
* There is in general more scatter in heat flux predictions than in surface temperature
predictions.
* In the case of temperature, there is almost no difference in modeling cables as targets or
cables provided that thermo-physical properties are the same and the thickness of the target is
a quarter of the diameter of the cable.
* Relative differences for heat flux suggest under and over predictions.
* Specific conclusions can be made on a case-by-case basis. For example, temperatures for GTS-33 are for the most part over-predicted.
ICFMP BE #4: MAGIC over-predicts both the heat flux and surface temperature of three "slab"
targets located about 1 m from the fire. The trend is consistent, but it cannot be explained solely
in terms of experimental uncertainty. The technical details supporting relative differences are
included in Figure A-68 and Table A-21.
6-14
Model Validation
ICFMP BE #5: MAGIC predicts both temperature and total heat flux to targets in BE#5 test 4
approximately to within experimental uncertainty. The technical details supporting relative
differences are included in Figure A-69 and Table A-22.
Summary: Target Heat Flux and Surface Temperature
* MAGIC is capable of predicting the radiative and total heat flux to targets, assuming
know thermo-physical properties. MAGIC is also capable of predicting the surface
temperature of a target.
*
Based on the scatter plots of relative differences, the following classification is assigned:
-
Yellow for surface temperature: of the target
-
Yellow for total heat flux, and
-
Yellow for radiated heat flux
100%/6
Target Teirperature
X
X
X
X
75%
0
a
A
A,
AL
*A
25%
X
x
Steel
*
+
ancrete
L cOncretp
A
X
B
X
-
Cable B
Cable D
Cabe F
Cable G
*
*
>
x
I
--
a-U
*
X
A
I
a
.X--4-
I .
-
xA
A
*
_
0%
*
-
TC1-3
a
TO3-3
TrCO1-5
TC 3-5
oX
-
-25%
-50%
A
a
-75%
X
TC 1-7
TCO 3-7
-100%
dl
d,:;
LI
t
l^ R!
,z ^ ^
RI
6
R
°bLI
9
le
19
E W W
1
9
ii if
i
Figure 6-7: Scatter plot of relative differences for target temperature in ICFMP BE#3.
Experimental uncertainty Is 14%.
6-15
Model Validation
100%
Cable Terrperature (Cablb
75%
X
X
del hWMinGIC
X .
X
X
50%/6
X
25%
*-
Al
0%/0
*
X
X
A.
,,
X
,
.*......
.......
,......
.*............. g.
... ... .
_
~
_
x
-25%
..........................
-50%
A
-75%
* Cable D
*Cable B
-100%
-
N
w
w
w
_-
r
NJ CD
w
w
w
w
0
t
C CO
LUw
w
m m m m m mm m
L
Lm
LU
N
Co 0
1
U)
LU LL
w w
Lm
m m m
M
LU
A Cable F
U)
D
-_o
LU
LU LU
w5
2 m
L
'
X Cable G
t
0 N
'-
cn;20 Z
-
-
-
U.LL.
L
Z
Z
Z
Figure 6-8: Scatter plot of relative differences for target temperature in ICFMP BE#3.
Experimental uncertainty is 14%.
150°h
-100°%
_C~
N
N
0o ,
_ CoCo
_o
_
CoCoN N
N M OMCoCoCo CO
QM0
6
<
LUI LU LU LU LU Ul LU LU Co Co C
wm
m mmmm
6
6o
<C
Co LUJ Ul
LU
LU
LU LU
02
mmm
202
_
6 -*
.
.
LU Co Co Co
C
L
U ILU m
LU
mm
-,
,
1 U)
gg
j
L
LU CO Co
mm
N
Z
O
-
-
_
,
N
o
LLz ZZ
Figure 6-9: Scatter plot of relative differences for radiant heat flux in ICFMP BE#3.
Experimental uncertainty Is 20%.
6-16
-
Model Validation
200O%
Total Heat Rux
+
150%0
*Gauge2,CableF
* Gauge 4, CabLe B_
100%
* ;.
50%
0%
----.I.
I.
A
,., i.,
+Steel, WS2
WS3
----.. -CoIcrete,
*,
*
X,
I
...I I ,,I
AGauge 8 CableD
XGaugegCableG
S
#x
f -0
a
I-.--
*++
* *
*
,
.
a ......
'
k
'
,
I
!
A
*~
-50%
4|6
0 -GasConcrete,WS4
,
L
WS2
&VVS
0
CWS3
1A3SA
U
-100%
0- _
_
N
N
N
A)
T
N
)C'
)
O0
co
)
7.
0(LOCD
.
`I<
"
-v
-
0s M[CO]MAM1w1CO m
M1mm
W W W
W
m WmM Mm M m
'
<LOON
'I_
Q
-
Z mZM m o
o
Qcn
w0)mm
U..
LL
z
Figure 6-10: Scatter plot of relative differences for total heat flux in ICFMP BE#3.
Experimental uncertainty is 20%.
6-17
Model Validation
6.9 Wall Heat Flux and Surface Temperature
Heat flux and wall surface temperature measurements are available from ICFMP BE #3, plus
wall surface temperature measurements are available from BE #4 and BE #5. As with target heat
flux and surface temperature above, there are numerous comparisons.
It should be noted that the wall temperatures and heat fluxes in MAGIC were calculated locating
targets in the walls. The targets are characterized by the thermno-physical properties and
thickness of the wall. The targets were located in the same model location as the experimental
instruments in the test room. Consequently, this evaluation does not include the MAGIC output
option "Wall Temperature", or "Wall heat flux" available in the Wall output category.
Experimental measurements were compared with MAGIC's output option "Total absorbed heat
flux".
ICFMP BE #3: It cannot be generalized that MAGIC predicts wall temperatures and heat fluxes
within the experimental uncertainty of 14% and 20 % respectively. For the most part, walls are
over predicted and ceiling and floor are under predicted. As noted by the corresponding markers
in Figures 6-11 and 6-12, most of the relative differences for the ceiling and floor temperature
and heat flux are negative. The over predictions for the wall and floor can be up to
approximately 100% with very few exceptions.
The graphical comparison of experimental and predicted temperature and heat flux profiles is
presented in Figures A-70 to A-85 and Tables A-24 to A-27.
ICFMP BE #4: MAGIC predicted two wall surface temperatures to within the experimental
uncertainty of 20%. The two points are presumably very close to the fire because the
temperatures are 600 'C to 700 'C (see Figure A-86) above ambient. The relative differences are
-11% and 10% as listed in Table A-28.
ICFMP BE #5: MAGIC predictions of wall temperature are comparable to experimental
uncertainty with a significant outlier of more than 800%. At this point, there is no explanation
for such an outlier.
Summary: Wall Heat Flux and Surface Temperature
* MAGIC has the capability of predicting the radiative and total heat flux to walls.
MAGIC is also capable of predicting the surface temperature of a wall, assuming that its
composition is fairly uniform and its thermal properties are well-characterized.
* MAGIC predictions of heat flux and surface temperature are generally over predictions
for walls with few comparisons below the lower limit of experimental uncertainty.
Ceiling and floor were consistently under predicted. Based on these results, a yellow
classification is assigned.
6-18
Model Validation
300%/
Wall Tenrperature
250% -
A
Outlier of 868% hi BEi
*006
Long Wall
Short Wall
200%
X
150% - A Ceiling
100%10%-_X
Roor
XXX
50%/6 -
x
0
*
X
X
FM
... A-........
-50%-
X
S
-
A
R...
- .sv=v
X
A
AX
A
A
-100%
7
N
NNw
L
w
-
C
u
w
w
w
c~
)
w
w
')
C')
I
LO
100
U)
*
z
nwco
' w iw
Co
'
'
e
U
CO')C')
ww)
N'
Z
,
-
Z
Z
Figure 6-11: Scatter plot of relative differences for target temperature in ICFMP BE#3,
4, and 5. Experimental uncertainty Is 14%.
1500/a
t
Heat Flux to Wails
a,
* Long Wail
X
.
.
.
*
50%s
.
i
.
A
0°/0
........ .'
..
..
A Cefling
.
*
Xv 11RV.
Iz
-
"*'
'. ...
..
..
.
..
.
5
U
X
O
U
x. AA
. .
.
.. .... I I.
A
-50%
-100%
* Short Wal
.
..
.
x-
A
_
N
rw rw
C
rw
_
1-1 C
rw rw
C
rw rt
VO
wr
0
^;
M
;6
L
Il
V
6' 6- 6
6
rw
COnPn wL LM:
6
C
r
''
a)
U)
6L
rz
CD -_
WL
t
W~
t
_
g
g-J
,
88 o
s
B
Figure 6-12: Scatter plot of relative differences for heat flux in ICFMP BE# 3, 4 and 5.
Experimental uncertainty is 20%
6-19
Model Validation
6.10 Summary
This chapter presents a summary of numerous comparisons of the MAGIC model with a range of
experimental results conducted as part of this V&V effort. Thirteen quantities were selected for
comparison and a color rating assigned to each of the output categories, indicating, in a very
broad sense, how well the model treats that particular quantity:
* Hot Gas Layer (HGL) Temperature and Height: Green
*
Ceiling Jet Temperature: Green
*
Plume Temperature: Green
*
Flame Height: Green
*
Oxygen: Yellow
* Smoke Concentration: Yellow
* Compartment Pressure: Green
* Radiation Heat Flux, Total Heat Flux, and Target Temperature: Yellow
* Wall Heat Flux and Surface Temperature: Yellow
Five of the quantities were assigned a green rating indicating that the research team concluded
the physics of the model accurately represent the experimental conditions and the calculated
relative differences comparing the model and the experimental are consistent with the combined
experimental and input uncertainty. A few notes on the comparisons are appropriate:
* The MAGIC predictions of the HGL temperature and height are, with a few exceptions,
within or close to experimental uncertainty.
* MAGIC predictions for ceiling jet and plume temperatures are comparable to experimental
uncertainty. In the case of the ceiling jet, results suggest a higher scatter among relative
difference in BE #3 closed-door tests than in BE #3 opened-door tests. At this point, no
specific explanation for this behavior is available. In the case of plume temperature, all
relative differences were within experimental uncertainty.
* MAGIC predicts the flame height consistent with visual observations of flame height for the
experiments. This is not surprising since MAGIC simply uses a well-characterized
experimental correlation to calculate flame height.
* Compartment pressure: MAGIC predicted compartment pressure to within experimental
uncertainty.
Four of the quantities were assigned a yellow rating indicating the user should exercise caution
when using the model to evaluate that quantity. This typically indicates limitations in the use of
the model. A few notes on the comparisons are appropriate:
* Predictions of smoke concentration by MAGIC are typically over-predicted. Predicted
concentrations for open-door tests are within experimental uncertainties, but those for closeddoor tests are far higher.
6-20
Model Validation
*
Most cable surface temperatures are predicted within or above experimental uncertainties.
Very few under predictions were observed. However, this is not the case for total and radiant
heat fluxes. Relative differences are both under and over predicted. Total heat flux to targets
is typically predicted to within about 30 %/,and often under-predicted. Care should be taken
in the prediction of localized conditions such as target temperature and heat flux due to
inherent limitations in all zone fire models.
*
Oxygen concentrations were consistently under predicted at about 30% for opened door tests
in BE#3. However, these under predictions resulted from oxygen concentration
comparisons above 15%, which are above concentrations suggesting fire extinction. In the
case of close door tests, MAGIC results are comparable to experimental uncertainty.
* Predictions of compartment surface temperature and heat flux are for the most part over
predicted for walls. Consistent under predictions are observed for the ceiling temperature
and heat flux. Finally, the floor surface presents both over and under predictions.
Differences between the model and the experiments were evident in these studies. Some of the
differences can be explained by limitations of the model as well as of the experiments. Like all
predictive models, the best predictions come with a clear understanding of the limitations of the
model and of the inputs provided to do the calculations.
6-21
7
REFERENCES
1. "ASTM Standard Guide for Evaluating the Predictive Capability of Deterministic Fire Models,"
ASTM E1355-04, American Society for Testing and Materials, West Conshohocken, PA 2004.
2. L. Gay, C. Epiard, B. Gautier "MAGIC Software version4.1.1: Mathematicalmoder' EDF
HI82/04/024/B November 2005
3. L. Gay "User guide of the MAGIC Software V4.1.1 " EDF H182/04123/A April 2005
4. L. Gay, J. Frezabeu, B. Gautier "Qualificationfile offire code MAGIC version 4.1.1. ". HI82/04/022/B November 2005
5. B. McCaffrey "Flame Height" SFPE Handbook of Fire Protection Engineering,
NFPA, 1995.
2 nd
edition,
6. Cooper, L., "Fire Plume-Generated Ceiling Jet Characteristics and Convective Heat Transfer
to Ceiling and Wall Surfaces in a Two-Layer Fire Environment: Uniform Temperature,
Ceiling and Walls," Fire Science & Technology, Vol 13, No. 1 & No. 2 (1-17), 1993.
Also NISTIR 4705, Nov 1991.
7. G. Heskestad "Fire Plume, Flame Height, and Air Entrainment " SFPE Handbook of Fire
Protection Engineering, 3"d edition, NFPA, 2002.
8. L. Y. Cooper, "Combined Buoyancy and Pressure-Driven Flow Through a Horizontal Vent,"
NISTIR 5384.
9. G. Heskestad "Quantificationof thermal responsiveness of automaticsprinklers including
conduction effects" in Fire Safety Journal, vol. 14, 1988.
10. R. Gear "Difsub for solution of OrdinaryDifferential Equations': Collected algorithms from
CACM, Algorithm 407, 1970
11. W. H. Press, S.A. Teukolsky, W.H. Vetterling, B.P. Flannery " Numerical Recipes in
Fortran ", Cambridge University Press, 1992
12. A. Benmamoun "Rapport d'analyse et des modifications du code FORTRAN" SYSAM - SE0310AB - 2004
13. J. Casal T. Noguer "MAGIC Version 4 Dossierde conception globale - Mise ajour" Doc.
ILM-technologie September 2004
14. L. Gay, J. Frezabeu "Cahierde recette de l 'interface utilisateurdu logiciel MAGIC en
version 4.1.1" EDF HI-82/04/021/P November 2004
15. L.Gay, J. Casal, T. Noguer "Manuel de refeirence de l'interface utilisateurdu logiciel
MAGIC en version 4" HI-82/04/025/P September 2004
7-1
References
16. NUREG-1758, "Evaluation of Fire Models for Nuclear Power Plant Applications:
Cable Tray Fires," U.S. Nuclear Regulatory Commission, Washington, DC, June 2002.
17. Barakatm M., "Interaction Rayonnment-Particules.Cas de Fummes Generees par Differents
Types de Combustibles," Thesis from Poitiers University, France, 1994.
18. Mulholland, G., "Smoke Production and Properties" SFPEHandbook of Fire Protection
Engineering,3Pd Edition, Chapter 2-13, National Fire Protection Association, Quincy, MA 2002.
7-2
A
TECHNICAL DETAILS FOR THE MAGIC VALIDATION
STUDY
Appendix A provides comparisons of FDS predictions and experimental measurements for the
six series of fire experiments under consideration. Each section to follow contains an assessment
of the model predictions for the following quantities:
A. 1
Hot Gas Layer Temperature and Height
A.2
Ceiling Jet Temperature
A.3
Plume Temperature
A.4
Flame Height
A.5
Oxygen Concentration
A.6
Smoke Concentration
A.7
Compartment Pressure
A.8
Target Heat Flux and Surface Temperature
A.9
Wall Heat Flux and Surface Temperature
The model predictions are compared to the experimental measurements in terms of the relative
difference between the maximum (or where appropriate, minimum) values of each time history:
AM -AE
AE
(p - M.)- (Ep,- E.)
(E -E
AM is the difference between the peak value of the model prediction, Mp, and its original value,
Mo. AE is the difference between the experimental measurement, Ep, and its original value, Eo.
A positive value of the relative difference indicates that the model has over-predicted the severity
of the fire; for example, a higher temperature, lower oxygen concentration, higher smoke
concentration, etc.
Finally, all of the calculations performed in the evaluation were open; that is, the heat release rate
of the fire was a specified model input, and the results of the experiments were provided to the
analysts.
A-1
Technical Detailsfor the MAGIC Validation Study
A.1 Hot Gas Layer Temperature and Height
Relative differences for hot gas layer temperature were calculated using experimental data from
ICFvP benchmark exercises 2, 3, 4, and 5, the FM/SNL test series, and the NBS multicompartment fire test series. In the case of hot gas layer temperature, positive relative
differences are an indication that the MAGIC predictions are higher than the experimental
observations. In contrast, in the case of hot gas layer height, positive relative differences suggest
that the MAGIC prediction is below the measured layer height.
A. 1.1 1CFMP BE #2
The HGL temperature and depth were calculated from the averaged gas temperatures from three
vertical thermocouple arrays using the standard reduction method. There were 10 thermocouples
in each vertical array, spaced 2 m (6.6 ft) apart in the lower two-thirds of the hall, and 1 m (3.3
ft) apart near the ceiling. Figure A-I presents a snapshot from one of the simulations.
K
E395U09
388.29
381.50
m 374.70
367.90
361.11
:354.31
347.52
a340.72
o 333.92
" 327.13
a320.33
n 313-54
o306.74
V299.95
2 293.15
Figure A-1: Cut-away view of the MAGIC simulation of ICFMP BE #2, Case 2.
The comparison between measured hot gas layer temperatures and heights for ICFMP BE #2
Cases 1, 3 and 3 are presented in Figure A-2.
A-2
Technical Detailsfor the MAGIC Validation Study
80
Fbt Gas Layer TerL ee.
BE2, Case 1
70
I/
60
20
18
16
14 12
-
50
E
9
40
EM: 10
I)
L 30i-
/'
20
Tjupper
_.
10
-
MAgKZ Layer riterface heihtg
8
L
MAGIC Upper ayer
terqperat1ure
-
Llbt Gas Layer Height
BE 2, Case 1
42
-
0
0
0
0
2
4
TirrL (rri
6
8
2
4
6
Tirne (min)
10
|Fbt Gas Layer Ternperatu!!.
0-
-
I-HotGas Layer Height
BE2, Case 2
16
1Z
80
I
--
Height
IIDz
E
a
4a'
MAGIC Upper layer
-
I-
terrperature
20
20
T~upper
--
0
2
4
Uae (nvr)
00
6
h
6~
E
I
40MAGIC Upper layer
-
20 -I-.
2
te,-,-erature
T_upper
.
4
Tirre (ienr)
.
6
H
7
8
8
\
12
k
BE2,Case 3
MAGIC Layer intertace height
Height
-
-
I.-
6
HrFlot
Gas Layer Height
I
1
s0o60
4
Trrue (nrui)
100| BE2.-Case 3
0 0
2
8
120 -Ft Gas Layer Terrperature
I'N
MAGIC Layer riterface height
-
12
60
e
10
2;20 .I
12v,
iuu
8
8
ll.
4
0
0
2
4
Tom (rrin)
6
8
Figure A-2: Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #2.
Table A-1 summarizes the relative differences calculated for the hot gas layer temperature and
height. MAGIC slightly under predicts the temperature. At the same time, measured hot gas
layer heights are consistently higher than the MAGIC prediction by a relatively small margin.
A-3
Technical Detailsfor the MAGIC Validation Study
Table A-1: Relative differences of hot gas layer temperature and height in ICFMP BE# 2
Test
ICFMP 2-1
ICFMP 2-2
ICFMP 2-3
Hot Gas Layer Heiaht
Hot Gas Layer TemDerature
Relative
Relative
Differenc
Differenc
AE (C) AM (0C)
e
AE (m) AM (m)
e
54.8
50.8
-7%
-14.57
-13.66
-6%
-14.77
-14.35
-3%
86.3
81.6
-5%
82.6
81.3
-2%
-13.86
-12.55
-9%
A. 1.2 ICFMP BE # 3
BE #3 consists of 15 liquid spray fire tests with different heat release rate, pan locations, and
ventilation conditions. The basic geometry as modeled in MAGIC is shown in Error!
Reference source not found.. Gas temperatures were measured using seven floor-to-ceiling
thermocouple arrays (or "trees") distributed throughout the compartment. The average hot gas
layer temperature and height were calculated using thermocouple Trees 1, 2, 3, 5, 6 and 7. Tree
4 was not used because one of its thermocouples (4-9) malfunctioned during most of the
experiments.
Figure A-3: Snapshot of the MAGIC simulation of ICFMP BE #3, Test 3.
In the closed-door tests, the HGL layer descended all the way to the floor. However, the
reduction method, used on both the measured and predicted temperatures, does not account for
the formation of a single layer, and therefore does not indicate that the layer dropped all the way
to the floor. Notice that in the MAGIC simulations, the hot gas layer is predicted to reach the
floor.
A-4
Technical Detailsfor the MAGIC Validation Study
It is important to indicate also that the HGL reduction method produces spurious results in the
first few minutes of each test because no clear layer has yet formed.
The comparison between MAGIC simulations and measured hot gas layer temperatures and
heights are compiled in Error! Reference source not found. to Figure A-7.
A-5
Technical Detailsfor the MAGIC Validation Study
4,0
.i
Hot Gas Layer Height
KtvIBE#3,TesI -
f
3.5
*.
3.
6
-a-
,9 2.5
Layer W-ight above floor (rr
-MAGECRre-PForn:
j
2,0
.0'
Layer
interface height
E
0,5
0.n
0
5
10
15
Trre (rrn)
20
25
30
0
5
10
0
5
10
iS
Tirre (rin)
20
25
30
20
25
30
120
s
0
80
E
60
.IP
I0
40
20
0
0
5
10
15
Trnw(rrin)
20
25
30
300
q'1.
Gas Layer H-ight
FbHot
3,5 -
250
3,0-.
o
15
T-rn (nun)
200
VBE #3, Test2
f
2,5-
U 150
.
.0'
Laert et cehet
2,0
% 100-
\Ns
1
110"O
r
0.5
0.0
0
2
4
6
Trne (rni)
a
10
12
0
.
2
4
6
Trne (rrn)
8
10
12
Hot Gas Layer Terrperature
OW BE #3, Test 8
250
enn
0
U
.0
100-
x
I I
iv
I-.-
50
MPAGIC
Fire-Roomr
Lpper by
;te' re
.,r
-
I
0
2
upper
4
6
TrI (rInn)
6
10
12
0
2
4
6
Tern (rnin)
8
10
12
Figure A-4. Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #3, closed door tests.
A-6
Technical Detailsfor the MAGIC Validation Study
C
o
150
a
El100
e50
15
10
Tirm (rrin)
250
200
6
1t50
REE 100
I
a)
0
10
5
15
15
10
Tom (fin)
Tre (Win)
C
V
I.
0
350
2
4
6
Tam (rrh)
8
10
12
0
2
4
6
T-m (rrh)
8
10
12
.HFt Gas Layer Tenperature
ItW BE t3. Test 16
300
250 1
-
a
C
200
f
1/
150
F100
5
50-
-
f
0
.Tjpper
WG1C Rre-Rcom: Ltper ayer
-
ten1perature
2
4
6
Time (rrn)
8
10
12
0
2
4
6
Tme (rrn)
8
10
12
Figure A-5. Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #3, closed door tests.
A-7
Technical Detailsfor the MAGIC Validation Study
200
4,0
180
160
a
A
1401120
_bt Gas Layer Terrperature _
l
BE #3,Test 17
H-t Gas Layertlaight
-CFPBE #3, Teat17
3,5 1.1
2,0
E
I-
An
-N
--
s0 _
2to
-
T-upper
MAGIC Rre-Paom: Layer
interface height
_
1.0
-
-
2O
MAGIC Fire-Peom: Lipper layer
\~
_
U., I.
terrperature
0
UO
0
Wight
-
-
11,
I
2
4
6
Tnme(mim)
8
10
12
0
.
4
2
6
Time(rin)
8
12
10
Open Door Tests
4,0
Wt Gas Layer Temperature
ICFW BE #3, Test 3
Hot Gas Layer ibght
-CAP BE #3, Test 3
3,5
200 .
I
3,0 -
6
El
i!
II
150
/10-0'10-
_
100
F
--
-
2,0
I
1
ri-upper
-MAGICre-Pm: Layer
interface height
-'
0.5
MAGICFire-Room: Upper layer
So
n
Weight
2.5
0,1
temperature
-
0
5
15
10
Trne (rmn)
20
25
0
10
15
Tmre (min)
5
20
25
40btGas Layer Wight
D:PBE#3, Test 9
3,5
3,0
0
2,
I.
V-,
j
2.0
I-
I
1.5-
a
t
r
-
,
^
.Uj1
a
_
\
\
MAG61C
Rre-Room:
Layer nterlace height
0.5
0,0
0
5
15
10
Time (rin)
20
25
.
0
5
.
.
10
15
Tire (rmi)
.
20
25
Figure A-6. Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #3, open door tests.
A-8
Technical Detailsfor the MAGIC Validation Study
H-t Gas Layer Terrperature
I
5BE#3, Test 5
200-
2
150 .
E
100
Aw
V.
6'(
50-
-
f
I
T-upper
-O_-
MAGIC Rre-Room: Lper layer
terrperature
vn'
0
5
10
15
Ture (nin)
20
25
30
Trr (m^i)
250
3,5
3.0
i
2!
150
E
100
50
v
2,5
-
2,0
!
1,5
50
0
0
5
10
Is
Tune (min)
20
25
0
4
250
5
10
15
Tim (rin)
25
20
.0
Irt Gas Layer Ieight
CFPBE#3, Test 15
3,5
200-
c
150
e
100
2.!
inefc _egd
5
I-
I1,C!
I
.I
U0,.5
I
\
--
0.0
0
5
10
Tir
15
20
25
0
(ri)
5
10
15
20
25
20
25
rT (fimn)
0
.
150
1
100
6E
I-
0
5
10
Tir
15
(rim)
20
25
0
10
15
Trm (frui)
Figure A-7: Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #3, open door tests.
A-9
Technical Detailsfor the MAGIC Validation Study
Table A-2: Relative differences of hot gas layer temperature and height in ICFMP BE# 3
Test
HOt uas Layer
HOt Ga Layer I em perature
Relative
AM (00) Difference
AE (0C)
AE (m)
AM (i)
leight
Relative
Difference
N/Al
N/A_
ICFMP 3-7
122.9
116.8
120.3
117.3
-2%
0%
ICFMP 3-2
ICFMP 3-8
ICFMP 3-4
229.2
217.7
204.3
219.2
218.3
210.8
-4%
0%
3%
ICFMP 3-10
197.8
209.4
6%
NAl
ICFMP 3-13
ICFMP 3-16
290.5
268.4
298.9
278.6
3%
4%
N/Al
N/A'
ICFMP 3-17
ICFMP 3-3
ICFMP 3-9
ICFMP 3-5
135.3
129.2
-5%
207.3
204.0
175.5
207.1
204.5
176.5
0%
0%
1%
-3.26
-3.23
-2.98
-3.82
-3.82
-3.82
ICFMP 3-14
208.2
205.8
-1%
-3.29
-3.82
16%
3-15
ICFMP 3-18
210.6
193.4
-3.75
-3.82
20%
17%
ICFMP 3-1
jCFMP
205.3
204.6
-3%
6%
_
N/A'
N/A
_
-3.13
-3.26
17%
18%
28%
1. Relative difference not applicable for closed door compartment fire experiments since the
data reduction method does not account for the formation of a single layer, which is the case
when the hot gas layer reaches the floor.
A-10
Technical Detailsfor the MAGIC Validation Study
A. 1.3 ICFMPBE #4
ICFMP BE # 4 consisted of two experiments, of which one was chosen for validation, Test 1.
Compared to the other experiments, this fire was relatively large in a relatively small
compartment. Thus, its HGL temperature is considerably higher than the other fire tests under
study. As shown in Figure A-8, the compartment geometry is fairly simple, consisting of a
rectangular shape room.
*1022.50
973.88
X925.26
a876.63
> 828.01
779.38
*730.76
_682.14
633.51
584.89
* 536.27
*487.64
R439.02
2390g.40
a 341.77
293.15
Figure A-8: Snapshot of the MAGIC simulation of ICFMP BE #4, Test 1.
Figure A-9 includes the comparison between experimental and predicted hot gas layer
temperature and height. The relative differences calculated for this experiment are listed in
Table A-3.
6
IH2t Gas
Layer TerfperatuWe6,
- BE4 TeSt
I
W-
1.
5
,,
AIM0
.5F
X
e-
-
-
1
o
~
OI
u
10
20
Trnw(inF)
30
40
MAGIC Fire:
Layer
minerface
t_
MAGIClire:Lpperlayertenperatu
I
0O-
~Aver ~S.
-.-oLayer height
3
z
IO
.
BE4. Test 1
i 1
4
8 500-
IrK. Gas
DI
4
0
--
_
/
10
20
30
40
Tirre (nn)
Figure A-9: Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #4, Test 1.
A-1l
Technical Detailsfor the MAGIC Validation Study
Table A-3: Relative differences of hot gas layer temperature and height in ICFMP BE# 4
Test
ICFMP 4-1
Hot Gas Layer Temperature
Relative
AE (0C)
AM (0C) Difference
700.1
741.4
6%
Hot Gas Layer Heiaht
Relative
AE (m)
AM (m) Difference
-4.20
-5.27
25%
A. 1.4: ICFMP BE #5
BE #5 was performed in the same fire test facility as BE #4. Figure A-10 displays the overall
geometry of the compartment, as idealized by MAGIC. Only one of the experiments from this
test series was used in the evaluation, Test 4, and only the first 20 min of the test, during the
"pre-heating" stage when only the ethanol pool fire was active. The burner was lit after that
point, and the cables began to burn.
K
* 531.15
* 515.29
a 499.42
E 483.55
' 467.69
451.82
435.95
420.09
U 404.22
388.35
* 372.48
N 356.62
340.75
a 324.88
* 309.02
a 293.15
Figure A-10: Snapshot of the MAGIC simulation of ICFMP BE #5, Test 4.
Figure A-1I summarizes the comparison between the experimental and predicted hot gas layer
and height during the first 20 minutes of simulation. The corresponding relative differences are
listed in Table A-4.
A-12
Technical Detailsforthe MAGIC Validation Study
G
e
Ir
E
0
5
10
15
20
25
30
rwm (mrn)
0
5
10
15
Tse (min)
20
25
30
Figure A-1i1: Hot Gas Layer (HGL) Temperature and Height, ICFMP BE #5, Test 4.
Table A-4: Relative differences of hot gas; layer temperature and height in ICFMP BE# 5
Test
ICFMP 5-4
Hot Gas Layer Temperature
Relative
AE (0C)
AM (0C) _Difference
185.7
182.8
-2%
Hot Gas Layer Height
Relative
AE (m) AM (m) Difference
-4.70
-4.72
1%
A. 1.5 FM/SNL Test Series
Tests 4, 5, and 21 from the FM-SNL test series were selected for corn arison.
K
34S.36
342.81
339.26
335.72
332.17
328-62
325.08
321.53
317.98
314.43
a 310.89
300.24
296.70
293.15
Figure A-12 provides a pictorial representation of the experimental geometry as idealized in
MAGIC. The experimental hot gas layer temperature and height were calculated using the
standard method. The thermocouple arrays that are referred to as Sectors 1, 2 and 3 were
averaged (with an equal weighting for each) for Tests 4 and 5. For Test 21, only Sectors 1 and 3
were used, as Sector 2 fell within the smoke plume.
A-13
Technical Detailsfor the MAGIC Validation Study
K
*346.35
*342.81
a339.26
a
335.72
332.17
328.62
325.08
"321.53
E317.98
In 314.43
310.89
a307.34
a303.79
a 30024
E 296.70
a293.15
Figure A-12: Snapshot from the MAGIC simulation of FMISNL Test 5.
Figure A-13 summarizes the graphical comparison of hot gas layer temperatures and heights for
tests 4, 5, and 21. The relative differences are included in Table A-5.
A-14
Technical Detailsfor the MAGIC Validation Study
An.
Fbt Gas Layer Teqperature
S N " Test 4
70
60
a
.\
-
50 -
-Y
2
E
40 -
z
30 20 -X
_
0
5
w
10
15
Tlne (min)
20
2!5
I0
5
10
15
Tme (rrin)
^
T-
6A
:
20
25
.HFtGas Layer Fbight
FM'S
Test 5
FlbtGas Layer Tenperatur
60
iiS!
teirperature
_.- T unner
7
70
S
MAGIC Localeu: Lpper layer
-
I
10
51
50
40-
JE4
30
.
20
2
MAGICocaueu: L7e
.0
0
5
10
Tkne (rrin)
15
MAGIC Localfeu: Layer
-4-
Iyer ternerature
_r
10
3
interface
inght height
1*
20I
0
5
10
Tk- (rrin)
15
20
90
Fbt Gas Layer TerrperW
FM1NVS Test2/l
seo
70
iC 60
4
e 50
fb
40
30
a
4-
ff
20
-MAGI
-
10 _
3
Local~eu: pe
byer
l
tperature
Tupper
A
0
10
20
Ta (n)
30
40
0
5
10
15
Trne (min)
20
Figure A-13: Hot Gas Layer (HGL) Temperature and Height, FM/SNL Series.
Table A-5: Relative differences of hot gas layer temperature and height in FM/SNL
Test
FM/SNL 4
FM/SNL 5
FM/SNL 21
-
Hot Gas Layer Temperature
Relative
0
AE ( C)
AM (OC) Difference_
59.2
51.0
-14%
46.6
40.4
-13%
66.0
59.6 _ -10%
Hot Gas Layer Height
Relative
AE (m) AM (m) Difference
-3.40
-5.50
N/A
-3.23
-5.41
N/A
-3.43 1 -5.79
N/A
A-15
TechnicalDetailsfor the MAGIC Validation Study
A. 1.6 The NBS Multi-Room Test Series
This series of experiments consisted of two relatively small rooms connected by a long corridor.
The fire was located in one of the rooms. Eight vertical arrays of thermocouples were positioned
throughout the test space: one in the burn room, one near the door of the burn room, three in the
corridor, one in the exit to the outside at the far end of the corridor, one near the door of the other
or "target" room, and one inside the target room. Four of the eight arrays were selected for
comparison with model prediction: the array in the burn room, the array in the middle of the
corridor, the array at the far end of the corridor, and the array in the target room. In Tests 100A
and 1000, the target room was closed, in which case the array in the exit doorway was used.
The standard reduction method was not used to compute the experimental HGL temperature or
height for this test series. Rather, the test director reduced the layer information individually for
the eight thermocouple arrays using an alternative method (Peacock 1991).
K
*528.45
*606.10
*583.74
E 561.39
i539.04
516.68
-494.33
5 471.98
a 449.62
* 427.27
X404.92
f 382.56
360.21
337.86
315.50
293.15
Figure A-14: Snapshot from the MAGIC simulation of NBS Multi-Room Test 100Z
The following three figures (Figure A-15 to Figure A-17) compiled the graphical comparison
between experimental measurements and modeling results for hot gas layer temperature and
height for the three selected experiments. Recall that the target room was closed in the first two
experiments (A and 0). Consequently, no relative difference was calculated for the target room
in those two experiments. The relative differences are listed in Table A-6 to Table A-8.
A-16
Technical Detailsfor the MAGIC Validation Study
400
HotGas Layer Height, Burn Rboor
HotGasLaver Terr,. Burn Rroo
350
2,0 -
300
)
250
E
200
150
'
NS. M'J100A_2
1,5 -
1,0-
t-
O
v
_P
U
BR
-
MAGCReRon: Upperbayertenperature
0,5 -
'
MAGIC FireFborn: Layer
0,0 -
0
5
10
15
20
0
5
10
Time (min)
20
2.5
400 .
Hot Gas Layer Terp, Burn FbornoOrw ay
-BS,
MV100A_2
350
15
Time (min)
_
25HtGas Layer HbigKitBumrn
Roonm
Iborw ay
_
2.0
NBS,MVtODA_2
3CO
250
I-
1.5
200
I0
150
too
A
50
-MAGE
-
t
1,0
I
L -
-
GT DO
0,5
ConeotToRreRoorn: Lpper bayer
-M_ tilAGtCConneeftrolireRoom: Layer interface
0,0~
0
5
10
15
0
2C
5
Trem(rinn)
1s
20
Tkre (rrin)
3,00
1
20 HotGasLayer T
*
10
2
80
.FtGas Layer Height.Corrirtor
2.50 -NBS.MV100A_2
so
C
2.00 -
n
I-
40
-U-P4.5
20
0
I -MAGnC
0.50
U-PB
0.00 ' r!
0
_, _ _
0
5
10
Tire (rrh)
Is
M111:11.9m
1.00-
-MAGC
Corridor: LUpper
layer terrperature
71-U-P
1a
UP 38
20
Ceetbr: .Wrkfla.feoh.gl
-_HGT 3E
+ HG4TEXI
_,
5
10
15
20
The (rni)
Figure A-15: Hot Gas Layer (HGL) Temperature and Height, NBS Multiroom, Test 1O0A.
A-17
Technical Detailsfor the MAGIC Validation Study
Table A-6: Relative differences of hot gas layer temperature and height in NBS Tests
NBS A
NBS A
NBS A
NBS A
NBS A
NBS A
A-18
Hot Gas Layer Height
Hot Gas Layer Temperature
Relative
Relative
AM (0C) Difference AE (m) AM (m) Difference
AE (0C)
-1%
-1.17
-1.18
29%
318.4
247.8
Burn Room
-47%
-1.18
-2.24
9%
93.5
85.9
orridor 4.5
-7%
-1.18
-1.27
21%
93.5
77.5
orridor 18
-47%
-1.18
-2.22
31%
93.5
71.1
orridor 38
-50%
-1.18
-2.37
41%
93.5
66.4
Corridor EXI
Target
Target room closed
Room
Technical Detailsforthe MAGIC Validation Study
SD0
MV10002
400 -NLS.
E
I-
2,5
Hot Gas Layer Temp.Burn Rrorn
HotGas Layer HF ht. Burn Room
NBS.MW100D_2
2,0
i9
3t0
r
2C00 P
a,
P
d
100
1.0
10,
LUPBR
--
-MAGICFeRrot: LUperbyer
temperature
0
5
0O
1D
0
2D
15
MAGICReRwrn: Layer intertaerheig
-
5
Thie (min)
10
15
2C
Time (min)
I2
2.5
450 -Hot Gas Layer Temp.BurnRoomDIorw ay
NS, MV1000-2
400
.Ht Gas Layer Height BurnRoom DNorway
NBSM
tIW10X,2
2,0
W>v
G
3D00
250n
200
150
1.0
I
/e
100
50 I
0,5
-
0 .
-GbLer
MAGIC C(>9CrbwctoreRorno: Llpper b~yer
.
0
5
10
2CI
15
0
5
t
*ML
~ai-
2
l
O
-
Upper hy ert7wH
E
1,50
-I
r
IW
I
I HGT4.5
MAGIC
EOrrkIDr: Layer hiterface heght
tHGT18
GT 38
-
2.w -
lel t.4.5
MA
Cor
UP 18
2C
HotGas Layer Heoot G~ior
2.50- NBS.MV100w_2
_
100-
20
15
300.
WAqMVflM 9
so
10
Tieh (mn)
140 Hot GasLayer Temp.(brridor
120-
,,
-
Twre(mnn)
B
r
7
ae
.
0,0 I
.
L
K.HESI
,
=
0,50
Pz3
.P.E,
S'_
0,00
vw,
0
5
10
Tne (mn)
15
20
0
5
10
15
20
Tom(mnn)
Figure A-16: Hot Gas Layer (HGL) Temperature and Height, NBS Multiroom, Test 1000.
A-19
Technical Detailsfor the MAGIC Validation Study
Table A-7: Relative differences of hot gas layer temperature and height in NBS Tests
Hot Gas Layer Temperature
Hot Gas Layer Height
Relative
Relative
AM (m) Difference
AE (CC) AM (0C) Difference AE (m)
-1.19
-25%
-1.59
29%
310.3
399.3
NBS 0 Burn Room
NBS 0
orridor 4.5
94.5
97.3
3%
-1.55
-1.06
-32%
94.7
97.3
3%
-2.12
-2.44
15%
NBS 0
orridor 18
8%
-2.14
-2.44
14%
NBS 0
orridor 38
89.8
97.3
NBS 0 Corridor EXI
65.5
97.3
48%
-2.23
-2.44
10%
NBS 0 Target
Target room closed
Room
A-20
Technical Detailsfor the MAGIC Validation Study
400 - Hot Gas Layer Terrp. tBurn Fborn
1 8
Fbt Gas Layer Height,
1 6 -NBeS, MV102Z_2
350O - Nasb, MV juU_z
urn Poorn
1.4 a.
250
c
20O
op
150
Pr
-
100
UP
-
50
1.
0,6-
BR
0,4
MAGC FireFlorn: LUper layer
ternperature
5
0
1,2
10
_
HGTt3F
0,2
2a3
15
-
GtC Fireborn: Layer interface heiht
-
0
5
Hot Gas Layer Height, Burn PRiom Docrway
HotGas Layer Terrip.Burn RbornDorway
NBS.tV100Z_2
2.0
NBS, W1tOZ_2
250
200
j
I
!
150
100
Lip
50
I
%_
-MAGICnneefforeR|or
uipertlyer
t.0
_r
0,5 .
-
-
0
AAGICO n
0,-.
D
5
Time(rdn)
20
_
-
UP 38
-
5
-LIP
V
2.00
Al
1,50
i
IIf
1.00
-MAGIC Crriktor: Layer intertace heiug!(F1.
HGT 18
_
T 38
.
H&TrE
0.50
ULPEt
10
20
NBS,W10OZ_2
II
W-P4.5
15
Hot Gas Layer Height, Carrktor
2.50
I
0.00
0
10
nei
3,00 .
80 II
40
eooe:tayer
Tire (rfin)
120~.U~00
__
t btGasLayerTeuprridor
I4
tO 8_0S, MU10Z$2
a
C)
20
2.5T
400,
a
15
Time (min)
Tame (min)
350
10
iS
__
0
20
S
Tim (fin)
10
15
2CI
Th- (rnin)
2.50
Ft
zoo
N
Gas Layer HeWht. Target Fborn lbaw ay
S' MVIOOZ_2
1.50
I
I9
0,50
m-tMGIA~0C
=
ectToTargetloo: Layer intertace
hecght
l
0.-00 4
0
10
Tim (rfi)
15
20
0
5
10
15
20
Tim (fih)
A-21
TechnicalDetailsfor the MAGIC Validation Study
60 V ;'
l.
HDt Gas L~ayer Tenrp, Target Fbom
HotGas LayerHjgtht. Target Room
250 -NBS, MV1OOZ-2
:?
E
2,C0 -
I6
30 t '
I
____IC
- ------
Im
,w
10
8
0.501
MAGbCTarget~aorn: Upperbyer terrperature
0
5
10
15
_-
| -M
0,00I
0
20
AGICTargetborn: Layer hlterface height
5
10
15
20
Terne
(in)
The (fin)
Figure A-17. Hot Gas Layer (HGL) Temperature and Height, NBS Multiroom, Test 100Z.
Table A-8: Relative differences of hot gas layer temperature and height in NBS Tests
NBS
NBS
NBS
NBS
NBS
Z
Z
Z
Z
Z
NBSZ
____Rom
A-22
Hot Gas Layer Temperature
Relative
AE (°C) AM (0C) Difference
Burn Room
283.9
311.9
10%
orridor 4.5
68.7
92.9
35%
orridor 18
61.1
92.9
52%
orridor 38
58.2
92.9
60%
orridor EXI
51.2
92.9
82%
arget
31.7
137.8
119%
Hot Gas Layer Height
Relative
AM (m) Difference
AE (m)
-1.67
-1.17
-30%
-1.24
-1.05
-15%
-1.70
-1.05
-38%
-1.76
-1.05
-40%
-1.36
-1.05
-23%
-2.06
-.
6
114%
Technical Detailsfor the MAGIC Validation Study
A.2 Ceiling Jet Temperature
MAGIC has an explicit ceiling jet temperature model based on the model developed by Cooper
[Ref. 6]. The model also accounts for the hot gas layer effects using Cooper's method. In
general, a target is specified in the computational domain. If the target is exposed to ceiling jet,
the "Target/Gas Temperature" output option provides the ceiling jet temperature.
Experimental measurements for this category are available from ICFMP BE #3 and the FM/SNL
series only.
Positive relative differences are an indication that the MAGIC prediction is higher than the
experimental observation.
A.2.1 ICFMP BE # 3
The thermocouple nearest the ceiling in Tree 7, located towards the back of the compartment,
was chosen as a surrogate for the ceiling jet temperature. The 15 graphical comparisons of
experimental measurements and model results are grouped in Figure A-18 and Figure A-19. The
relative differences are listed in Table A-9.
A-23
Technical Detailsfor the MAGIC Validation Study
it 60
If 3D Ceing Jet Tenperature
oCFNWBE#3, Test I
Ceing Jet Teurperature
to O
BE#3, Test7
II
_
60
13
6
a
I:
it
I
It
30
E-
I
E
30
30
-
/t
0
0
MAGtCTree7-10: Gas tenperature J
0
10
5
15
Twrs(rrin)
20
25
30
250 .
5
10
15
Tumr(mhn)
20
25
30
0
2
4
6
TjT (rin)
8
10
12
2.
200 -
I
I
0 I
0
Ceing Jet Tenperature
BE#3, Test 2
CR
300
a
Tree 7-10
-MAGICTree7-10: Gas terperature
20
S
150
i1o0
50
Tree7-t0: Ga. tenprature
I E-AGI
0
0
2
4
6
Trm (nrin)
8
10
12
350
Ceging Jet Te ryprature
3CO - IPW BE J3, Test 4
300 +
2.50
6
2
Ceing Jet Terrperature
WFWBE#3, Test 10
a
R
200
v
150iI-
100Y
M! Cre 1
0
10.
_
re7-10
50 2t _t-AGICTree7-10: GaIsterrperte
f
.
I
5
0
ee-0
-a
0
15
10
~
5
Trmr (rrin)
:Btrreau
71
-a-Te
a t~r~r
10
15
Trom(nun)
J
S .C
450-Oe
Jet Trrerature
40 wBE#3, Test 13
400
U
350
z
1
,2
200
T
eI:
gt
150
I
-..-Tree7-10
too
50s-oMAGOTree7-tO Gas terperature
0
2
4
6
Tre (rrm)
8
10
12
0
5
10
15
Trrm (rfn)
20
25
30
Figure A-18: Near-ceiling gas (ceiling jet) temperatures, ICFMP BE #3, closed door tests.
A-24
Technical Detailsfor the MAGIC Validation Study
I
250
et Terrperature
0
tgPBE#t3, Test 17
2C0A
a
e
150
_
/
Open Door Tests
100
0
50
/
0
0
Tre 7
<
.
4
.
2
0
10
6
Timne(rrim)
: Ga
10
8
12
MU0,
Ceiing Jet Terriperature
a
BE#3. Tes
=
250
U
200
e
150
,
100
_
I
/C
e
1/
-.
50
Tree 7-10
|
At3CTree7-10
G
Gas terrperature
0
10
15
rue (rrin)
5
0
20
25
0
--~JJ
10
15
Tkre (rrin)
5
25
20
Cteing Jet Terrperature
250
-CFWBE #3.Test14
200
-
a
I
I
,II
100
1
MAGI Tree7-10: Gas temperature
0
10
5
is
Tire (rrin)
20
25
30
0
5
10
15
Time (rnn)
20
25
350,
Ceiing Jet Terrperature
250 S1
|
300 -
8E #3,Test 1
Ceing jet Ternperature
IFWBE#3, Test 18
250 -
210
200
e
a.
I
10010
50
}
Tre71
i
150 I
t-
100
-
|MAGIC
Tree7P-10: Gas terrperatre
Tree 7-10
WGkIC Tree7-10: Gas tenperature
50- V-
oi
0
5
10
15
Tare (rnr)
20
25
0
6
10
15
Twre (nrr)
20
25
30
Figure A-19: Near-ceiling gas (ceiling jet) temperatures, ICFMP BE #3, open door tests
A-25
Technical Detailsfor the MAGIC Validation Study
Table A-9: Relative differences for ceiling jet temperature in ICFMP BE #3
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
ICFMP
A-26
3-1
3-7
3-2
3-8
3-4
3-10
3-13
3-16
3-17
3-3
3-9
3-5
3-14
3-15
3-18
Instrument
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
Tree 7-10
AE (0C)
154.9
139.3
270.6
246.9
228.9
217.5
330.5
277.7
155.9
240.7
234.6
207.7
240.8
243.7
235.1
AM (OC)
140.1
136.8
269.4
268.3
261.0
259.1
400.9
376.6
172.3
234.6
231.4
204.7
232.5
231.4
233.2
Relative
Differenc
e
-10%
-2%
0%
9%
14%
19%
21%
36%
11%
-3%
-1%
-1%
-3%
-5%
-1%
Technical Detailsfor the MAGIC Validation Study
A.2.2 The FM/SNL Test Series
The near-ceiling thermocouples in Sectors 1 and 3 were chosen as surrogates for the ceiling jet
temperature. Figure A-20 compiles the graphical comparisons between experimental
measurements for ceiling jet temperature and the MAGIC predictions. The corresponding
relative differences are listed in Table A-10.
Celrig Jet Terrp
FW'SMW
Test 4 - Sector 1
a
i!
U
80
a
9n
60
i
Gas ten~perature
-tAAGICSE~t98:
-
20
e
-..-- l/(098H
0
15
10
5
0
0
5
10tt
90
15
10
Trh (rnn)
rre (nin)
80
Ceifing Jet Terrp
Caeilg Jet Tenwp
F8- S. Test 5 - Sector
70
70
60
R`SN.Test5 - Sector1
a 50
E
9- 40
I-.2
E
0
40
30
30
20
SE_98: Gas etrlfprature
MAGt
20 -
-o
MAGIC
~Sa:98: Gas terWperatr
-.- /(0.981j
10 _
0
0
10
5
15
Ture(nm)
E
8D1/0.8i
120
100
0o
Gerrag Jet Terrp
R
Oehrng Jet Tenrp
Test 21 - Sectrar 3
RWSN.
Test 21- Sector I
1
1000AA1
9-
soz
U
0
60
a
0
I-
I
A
O1
-M
tC
S
Gas tenperature t
AGIC S'ESC3_98:
204 W-M
1_98:Gas termperature
1/(0.98 1I
0
5
10
15
T-m (rinr)
20
25
30
0
5
10
15
20
25
30
rTh (rrh)
Figure A-20: Near-ceiling gas (ceiling jet) temperatures, FWSNL Series, Sectors 1 and 3.
A-27
Technical Detailsfor the MAGIC Validation Study
Table A-10: Relative differences for ceiling jet temperatures in FMISNL Tests
Instrument
FMSL4
1/98H
1/98H
FM/SNL14
FM/SNL 5
1/98H
11/98H
FM/SNL 21
1/98H
I 11/98H
A-28
AE (0C)
82.8
66.1
73.7
52.6
75.9
77.2
AM (OC)
78.8
58.1
66.8
47.5
74.2
74.2
Relative
Differenc
e
-5%
-12%
-9%
-10%
-2%
-4%
Technical Detailsforthe MAGIC Validation Study
A.3 Plume Temperature
Plume temperature measurements are available from ICFMIP BE #2 and the FM/SNL series. For
all the other series of experiments, the temperature was not measured above the fire (BE#3), the
fire plume leaned because of the flow pattern within the compartment (BE#4), or the fire was set
up against a wall (NBS). Only for BE #2 and the FMISNL series were the plumes relatively free
from perturbations.
Once a target is specified in the calculation domain, MAGIC identifies if it is located within the
fire plume region. Plume temperature results are captured in the 'Target/Gas Temperature"
output option. The output would include hot gas layer effects in the plume temperature if the
target is located inside the fire plume and above the hot gas layer interface. Positive accuracies
are an indication that MAGIC predictions are higher than the experimental observations.
A.3.1 ICFMPBE # 2
BE #2 consisted of liquid fuel pan fires conducted in the middle of a large fire test hall. Plume
temperatures were measured at two heights above the fire, 6 m (19.7 ft) and 12 m (39.4 ft). The
flames extended to about 4 m (13.12 ft) above the fire pan. (Figure A-21). The suspended
rectangle contains an array of thermocouples designed to locate the plume centerline. Notice
that the smoke plume does not always rise straight up because of air currents within the large test
hall.
11
I ,1,;
Figure A-21: Fire plumes in ICFMP BE #2. Courtesy Simo Hostikka, VTT Building and
Transport, Espoo, Finland.
A-29
Technical Detailsfor the MAGIC Validation Study
200
Pure Terrp
BE2, Case 1
180
Rune Terrp
300 _ BE2, Case 2
250 .
120
E
200
100
i0
I
60
80
-
MAGICGas enw ature
10
Iefrvature
50
40 i/-MAG.IGas
o0
0
0
.
-Ml
TG.2
MAGIC Gas ieiperature
-
MAGIC Gas terperature
-
TG.2
0
2
4
Toe (rrh)
6
8
0
10
2
4
1THm
(min)
6
8
300
imume
Terip
8E2,Case3
250
200
150
E
100
Gas tem~eature
-MtAGIC
MAGIC Gas tefyperature
50
rTG2
0
0
2
4
Time (min)
6
8
Figure A-22: Near-ceiling gas temperatures, FM/SNL Series, Sectors 1 and 3.
Figure A-22 above illustrates the graphical comparison between experimental plume
temperatures and MAGIC prediction for the three cases in ICFMP BE #2. The corresponding
relative differences are listed in Table A-l1.
Table A-11: Relative differences for plume temperature in ICFMP BE #2
ICFMP 2-1
ICFMP 2-2
ICFMP 2-3
Instrume
nt _E (°C) AM (°C)
TG.1
166
161
TG.2
79
87
258
TG.1
288
TG.2
128
141
10%
TG.1
252
229
_9%
128
132
3%
_TG.2
A-30
Relative
Differenc
e
-3%
10%
-11%
Technical Detailsforthe MAGIC Validation Study
A .3.2 The FMISNL Test Series
In Tests 4 and 5, thermocouples were positioned near the ceiling directly (5.9 m, 19.4 ft) over the
fire pan. In Test 21, the fire pan was inside a cabinet. For that reason, no plume temperature
comparison has been made. Figure A-23 presents the graphical comparisons and Table A-12
lists the corresponding relative differences.
140
Rurm Terrp
120
a
Pkjrns Teryp
Tt.N
FI -
120
_ FM/Test4_
100
100
__
a
80
I-
2EO
40
:
40
_
0 I
-A
50
jf-MAICSTTION3_98: Gas teffperalurri
20
.p,
-. 220.8I
10
0
15
0
5
10
15
Tim (rrdn)
Tffe (rrin)
Figure A-23: Near-plume temperatures, FMISNL Series, Sectors 13.
Table A-12: Relative differences of plume temperature in FWSNL tests
FM/SNL 4
FM/SNL 5_
Instrume
AE (0C) AM (0C)
nt
28/98H
116
115
28/98H
94 1 101
Relative
Differenc
e
0%
8%
A-31
Technical Detailsfor the MAGIC Validation Study
A.4 Flame Height
Flame height is recorded by visual observations, photographs or video footage. Videos from the
ICFMP BE # 3 test series and photographs from BE #2 are available. It is difficult to precisely
measure the flame height, but the photos and videos allow one to make estimates accurate to
within a pan diameter.
A.4.1 ICFMPBE #2
Shown in Figure A-24 are MAGIC predictions for flame height. Figure A-25 contains
photographs of the actual fire. The height of the visible flame in the photographs has been
estimated to be between 2.4 and 3 pan diameters (3.8 m to 4.8 m, 12.5 to 15.7 ft). The height of
the simulated fire fluctuates from 5 m (16.4 ft) to 6 m (19.7 ft) during the peak heat release rate
phase.
6
6Rarne
7
7RFare
FbgtM
5? Z, Case 1-
Fleht
Laise
ct
45
-
4
t
0
L-MAGr
2
2
_____-II_2
i
Rareheight
4
2
___.
_
MGC Rr
Vhe
2
-
6
2
Tre (nin)
Tae (rrin)
7
Parne Fbgt
BE2. Case 3-
-
65
I
I
0
2
halt
-MAGVCF
4
6
TUne(muin)
Figure A-24: Flame heights for ICFMP BE #2
A-32
4
_
.
6
_
Technical Detailsfor the MAGIC Validation Study
Figure A-25: Photographs of heptane pan fires, ICFMP BE #2, Case 2. Courtesy, Simo
Hostikka, VTT Building and Transport, Espoo, Finland.
A-33
Technical Detailsforthe MAGIC Validation Study
A.4.2 ICFMP BE #3
No measurements were made of the flame height during BE #3, but numerous photographs were
taken. Figure A-26 is one of these photographs. These photographs provide at least a qualitative
assessment of the MAGIC flame height prediction. Recall that the size of the door is 2.0 m (6.6
ft) high. Inspection of the picture suggests that the flame height, at least is some of its
oscillations, can be more than 2.0 m (6.6 ft) high. MAGIC however appears to be overpredicting flame heights, as most of the predictions are over 3 m high.
Figure A-26: Photograph and simulation of ICFMP BE #3, Test 3, as seen through the 2 m
by 2 m doorway. Photo courtesy of Francisco Joglar, SAIC.
A-34
Technical Detailsfor the MAGIC Validation Study
1I
2,5
E 1,5
I
I
'
1,0
U-
5
0
20
25
30
20
25
30
AGCre-Fan: Fa
height
10
15
Tre (rrin)
E
9
Eb
5
0
10
15
The (min)
20
25
0
30
5
10
1S
Too (itn)
3,5
Pa
N-
3,'0
Fbight
PF BE #3, TeSt4 -
2.5
v-
-
2.0
15% -,-
I
__
I
MAG Fke-PSn: Piano height
-
0.5
00.1.
0
_
,
.
5
,
|
.
10
E
.
.
1S
Two (nin)
D
r
.
20
'
.
25
30
IE
/
2,05
I2.0
I
-i-
0
2
4
r
6
Tre (mrn)
8
h0,0
10
12
0
2
4
6
Tirn
a
10
12
(nin)
Figure A-27: Near-ceiling gas temperatures, ICFMP BE #3, closed door tests.
A-35
Technical Detailsfor the MAGIC Validation Study
3,5
Ranfe Height
tBE#3,Test
F
3,0 -
1!
2,s
E
I
17_
Open Door Tests
/
1,5
1,0
I
/
as
~ehih
MG~r-an
,
o.o
0
2
4
I1
6
Tne (^in)
8
10
12
-E
E
I
E
I
0
5
10
15
Tim (rin)
20
25
30
0
5
10
15
Tew^Inf)
20
I
II
30
FnWe
Hsi#Tht
PBE #J, Test 14
|lW
7n .
25
I
1,5I.0 I
-MGAZF-e-Fhn
0,5
0
5
10
15
Trne(mu)
20
25
30
0
10
5
15
Time(mn)
Fnu height
20
30
25
3,a0e Height
2,5 _
CFPBE#3, Test18
2.0
E
I
1,5
ID
I0
1,0-
I
0.5
0,0
0
5
10
15
TRW(n*u)
20
25
30
I
.
-
MAGIC
re-Pen: Pnw height
t-
-v,
Ti
(Mi)
Figure A-28: Flame heights, ICFMP BE #3, open door tests
A-36
<
ID
Technical Detailsfor the MAGIC Validation Study
A.5 Oxygen Concentration
Oxygen concentration data is available for accuracy calculations in ICFMP benchmark exercises
number 3, and 5. For the calculations, measured values in the experiments are compared with
the upper layer oxygen concentration,which is an output available in MAGIC. Relative
differences are calculated comparing the lowest concentration measured in the experiments with
the lowest concentration predicted by MAGIC. Positive relative differences indicate that
MAGIC is predicting a lower concentration than the one measured in the experiments.
A.5.1 ICFMP BE #3
In experiments with close room doors in the ICFMP BE # 3 test series, the fuel supply was
discontinued when the oxygen concentration was measured around 12 to 15 %. In these tests,
relative differences are measured at the lowest concentration before the experiment was
terminated.
The graphical comparisons for closed and open door tests and relative differences are provided in
Figure A-29, Figure A-30, and Table A-13 respectively.
A-37
Technical Detailsfor the MAGIC Validation Study
Oxygen Concentration
00.I
_W
Oxygen Concentration
LMICP BE #3, Test 7
BE#3, Test 1
0n2n
0.20=
S
'§
0,15-
_
0,15
_
L0
E
? 0,10
ninI-_02-1
0,05
-
MAGICFIre-Paorn: Upiperlayer oxygen
concentration
-
0.00
-02-1
0,05
0.00
.
.
0
15
10
Tfre (rnn)
5
MAGIC Fire-arnom: Lpoer layer
oxygen concentration
-
.
20
0
25
5
20
10
15
Tune(min)
25
rOxygen Concentration
BE#3. Test 8
.20
,,_ICFW
0,20
S n.1
S
-_
E
,
0.10
I_02-1
0.05
-MAGICFire-Rocrn: L4per layer
oxygen concentration
0.00
0
2
4
6
Tim (rrin)
8
10
0
12
4
6
Torr (rrin)
8
10
12
0,25
Oxygen Concentration
Test 4
00N._BE#3,
2
Oxygen Concentration
'CFMPBE#3.
Test 10
0,20-
0,20
I
0,15 ,
-
0,15
IL
J
_.10 I
I.
1_02-11
0.10
1_02-1
0.05
0,05
-
0
S
10
MAGIC Rre-RPom: Lpoer layer oxygen
concentration
-
MAGIC Fre-Pcomn: tpper layer oxygen
concentration
5
0
15
Tiiie (rin)
I5
10
Tmre(rrn)
A09
Oxygen Concentration
lICFWBE#3.Test16
0L20
I
0.15
0,10
0.05 '
000
2
4
6
Tie (rin)
Figure A-29:
A-38
8
02
10
12
MAGIC
Rre-Room: L~,pr layer oxygen
concentration
.
0
2
.
4
6
The (rrin)
concentration, ICFMP BE #3, closed door tests.
8
10
12
Technical Detailsfor the MAGIC Validation Study
Cou
Oxygen Concentration
WBE #3, Test 17
0,20
S
U-
Open Door Tests
*0 0,10
0,5|MAGIC Fire-Floom:. 4ger layer
0,00
.
2
0
.
4
.
6
Tine (nin)
_
.
8
_
10
12
0.25
0,25
Oxygen Concentration
ICFWBE#3, Test 3
_CM
0,20
0,20
9
nil
.,d
?
Oxygen Concentration
RBE #3, Test 9
vs .v
o
0,15 i-
0,10
?
0,10
-02-1
0,05
.
r
_r
0,05
-
MAGIC Fre-FRom: LQper layer oxygen
concentration
.
5
10
15
20
Taue(mnn)
0,00w_
0
Fire-Room: Lpper tayer oxygen
concentration
-MAGIC
0,00
0
25
5
10
15
Trie (nin)
20
25
025
Oxygen Concentration
_ MI+P
BE1#3,Test S
0,20
I
5 0,15
F0,10
_02-1
0,05-
MAGIC Fire-Fboom:Upper layer oxygen
concentration
-
I
.
.
.
.
0
5
10
re
15
(nmn)
.
.
20
25
0
30
10
15
Trne (rrmn)
20
25
10
15
Time(mn)
20
25
0,25
Oxygen Concentration
1R BE3, Test 15
_.
_2
0,20
.E
I
U.
0.10
I
0.05 -
e02-1oyge
| MAGIC Fire-FIloom: L4)per layer oxye
Owion
0
,
____
5
.
_..
_
10
15
Trte (nin)
20
25
0
Figure A-30: 02, ICFMIP BE #3, open door tests.
A-39
Technical Detailsfor the MAGIC Validation Study
Table A-13: Relative differences of oxygen concentration in ICFMP BE #3 tests
AE
ICFMP 3-1
ICFMP 3-7
ICFMP 3-2
ICFMP 3-8
ICFMP 3-4
ICFMP 3-10
ICFMP 3-13
ICFMP 3-16
ICFMP 3-17
ICFMP 3-3
ICFMP 3-9
ICFMP 3-5
ICFMP 3-14
ICFMP 3-15
ICFMP 3-18
A-40
-0.065
-0.064
-0.092
-0.096
-0.079
-0.079
-0.101
-0.091
-0.033
-0.052
-0.054
-0.030
-0.055
-0.052
-0.051
AM
-0.070
-0.068
-0.100
-0.098
-0.065
-0.065
-0.110
-0.081
-0.028
-0.039
-0.038
-0.025
-0.038
-0.038
-0.037
Relative
Differenc
e
8%
6%
8%
2%
-17%
-18%
9%
-11%
-16%
-24%
-30%
-17%
-31%
-27%
-27%
Technical Detailsfor the MAGIC ValidationStudy
A.5.1 ICFMP BE #5
Figure A-31 and Table A-14 present the graphical comparison and relative difference for oxygen
concentration in ICFMP BE#5 Test 4.
S
.0
cV
By
F)
0
5
10
15
Tim (rrn)
20
25
30
Figure A-31: 02, ICFMP BE #5, Test 1
Table A-14: Relative differences of oxygen concentration in ICFMP BE #3 tests
AE
|ICFMP 5-4
|
-0.028
|
AM
-0.035
Relative
Differenc
e
24%
A-41
Technical Detailsfor the MAGIC Validation Study
A.6 Smoke Concentration
Data for smoke concentration is only available in ICFMP BE # 3. Positive accuracies are an
indication of MAGIC prediction higher smoke concentrations than the ones measured in the
experiments. Depending on the application, this may not mean a conservative result.
The units used for accuracy calculations are mg/rM3 . Notice that MAGIC output is average
extinction coefficient, k, with units of 1/m. As a result, the direct output from the model was
converted to mg/m3 using the following equation:
v = Yk.
/km
where v is the concentration in mg/m 3 , and km is a constant with value 0.0076 m2 /mg (Ref. 18).
A. 6.1 ICFMP BE #3
Figure A-32and Figure A-33 contain comparisons of measured and predicted smoke
concentration at one measuring station in the upper layer for closed and open door tests.
MAGIC consistently under predicts the smoke concentration with the exception of Test 17,
which consisted of a Toluene fuel. . This trend is reflected in the relative differences listed in
Table A-15.
A-42
Technical Detailsfor the MAGIC Validation Study
50
350,
Srmke Concentration
DO.ICFWBE #3,TestI
Smke Concentration
^FPBE
#3, Test 7 300 -1
/l
;2(
2.50
z
c-
1^ 250
i21
i
200
0
150
50
0!
0
Sntke
f_
Conc.
if
MAGIC Aire Poom: Soke
Concentration
-
Co
E
100
z.-.-
--
-MAGIC
Fire ,born: Smoke
-Conrcentration
/
.-
U,
50
I
50-
--------- r--
x
0.. -0
5
10
15
20
25
0
30
5
10
rne (rrin)
SuO
_ Sn-ke Concentration
450
FP BE #3, Test
E
350 300
.
250
-
v
Sn-^ke OAnc.
-MAGIC Fire Ibom: Snmke
Concentration
Concentration
I40tABE#3, Test 8
350 -- _
Smoke t
300. -
MAGIC Fre Paorn: Smnke
Concentration
e9-e
a
ej 200
e
30
400 -
400
t
25
_C
450 -Srnoke
2
20
15
Time(min)
.---
r^
nc..
I
7
200
150 -{
100-
,50
to.
0
0
2
4
6
Trn (min)
8
10
12
0
250
250,
-
200
SnI(okeConcentration
IAF BE#3, Test 4
Srr'ke Corn.
-MAGIC Rre Pon:
Smilke Concentration
150 )
oa
_.Q
10
8
12
-
a
Simke Concentrataon
Smok1e
Con_.
200 -MAGIC Fire eorn: Soe
Concentration
_
I
,
_A
BE#3,Test 10
150.
,
100
0
0
6
Tare (rrin)
E
<
100 )
4
2
10
5
15
/
15
10
5
0
Tin (min)
Time(min)
250
Sn ^keConcentration
|ICP(BEtl3,Test16
2003
400
V' 150
e
~
I?
* 300
r3
100-I
-z
1 50 -
200
100-
-4-
Simlke Cone
-MAGIC Fire Iborn:Simile
Concentration
0
0
2
4
6
rn- (nin)
8
10
12
0
2
4
6
Tie (nin)
8
10
12
Figure A-32: Smoke concentration In ICFMP BE #3, closed door tests.
A-43
Technical Detailsfor the MAGIC Validation Study
o 1000
Open Door Tests
0
540
0
4
2
6
Tsm (nin)
140
10
8
12
Smnke Concefntration
3;F9'BE U. Test3
jA)
0
I
100
/
80
1
v
0
e
60
eI
An
.
Conc.
-MAGICRre Room: Srrtn
A~ncentration
I-tSirnke
/
220J
<
OxS
0
20
15
10
TIM (min)
5
25
140-
SlInke Concrentraksn
K;FWBE #3, Test 14
120
100
aI
Ij
80~
60
I
Sam
keQnc.
/.
M --AGJC F'ie moaon:&i2c
20
0
5
15
10
20
25
30
0
T- (nin)
5
10
15
Trm (rin)
Ccgcentrato_
20
25
V,
I
0
5
10
15
T-r, (min)
20
25
10
15
Trre(min)
Figure A-33: Smoke concentration in ICFMP BE #3, open door tests.
A-44
20
25
Technical Detailsfor the MAGIC Validation Study
Table A-15: Relative differences of smoke concentration in ICFMP BE #3 tests
ICFMP 3-1
ICFMP 3-7
ICFMP 3-2
ICFMP 3-8
ICFMP 3-4
ICFMP 3-10
ICFMP 3-13
ICFMP 3-16
ICFMP 3-17
ICFMP 3-3
ICFMP 3-9
ICFMP 3-5
ICFMP 3-14
ICFMP 3-15
ICFMP 3-18
AE (mg/n 3 )
AM (mg/M3 )
41.50
55.05
128.00
99.53
79.90
70.75,
223.51
139.07
353.09
118.03
117.00
87.34
91.30
123.71
110.18
258.62
250.39
305.08
359.64
196.32
194.81
282.84
219.10
1453.17
127.06
126.03
91.36
126.62
126.65
126.08
Relative
Differenc
e
523%
355%
138%
261%
146%
175%
27%
58%
312%
8%
8%
5%
39%
2%
14%
A-45
Technical Detailsfor the MAGIC Validation Study
A.7 Room Pressure
Experimental measurements for room pressure are available from the ICFMP BE #3 test series
only. The pressure within the compartment was measured at a single point, near the floor. In the
simulations of the closed door tests, the compartment is assumed to leak via a small opening near
the ceiling. In order to reflect the actual leakage area in the model, the measured area was
divided by 0.68, which is the orifice flow coefficient used in MAGIC for all flows through
vertical openings.
A.7.1 ICFMP BE #3
Visual examination of experimental data and model results plots strongly suggest that tests with
open doors, where leakages are not critical because of the large door opening, MAGIC captures
both the magnitude and the profile of the pressure. These figures describe a negative pressure
profile at the floor of the room, indicating that fresh air is moving into the enclosure.
Similarly, in close door tests, MAGIC is able to capture the both peaks and pressure profiles. It
is important to mention that fan tests were conducted before some of the tests resulting in
relatively well known leakage areas. Furthermore, notice that MAGIC captures the positive and
negative pressure peaks. These peaks are an indication of a positively pressurize room in the
early stages of the test, and a negatively pressurize room when the fuel supply is discontinued
and heat loses to the boundaries are higher than the fire hear release rate.
Comparisons between measurement and prediction are shown in Figure A-34 and Figure A-35.
For tests in which the door to the compartment is open, the over-pressures are only a few Pascals
at the early stages of the fire, whereas when the door is closed, the over-pressures can be up to
several hundred Pascals. The calculated relative differences are listed in Table A-16.
The relative differences were calculated as follows:
* For closed-door rooms, the relative difference refers to the positive peak at the early stages of
the fire. Positive relative differences indicate that MAGIC over-predicted the measured
peak.
* For opened door rooms, the relative difference refers to the negative magnitudes of the
pressure, typically at the late stages of the test. Positive relative differences suggest that
MAGIC calculated a more negative difference than the experimental measurement.
A-46
Technical Detailsfor the MAGIC Validation Study
60
~Room Pressure
P BE*3, Test 7
.
40
20
IL
A
k
0~
e
I
50
rL
0. -20
_
Time(mn)
.
Tee (rrn)
300
300
Room Pressure
7FtP BE #3, Test 2
Room Pressure
tBhPBEl3.Testa-
200 -
200
-100-
26
8
t
100
e
0-
.100 1
-gm
2
4
6
8
l1l
!
2
6
a.
10r
-300-
-500-
8
ill
-500
-
MAGIC re-Room: Pressure at th
room floor
-6W0
Time(rrh)
rTme(rrn)
200
RoomMossure
100
BE #3,Test 4
CP
5
10
-100
-400
6
-200
-400
IMAGC Rre-Room: Pressure at
Ttheroomfloor
-7U0-
-300
4
-300
-400-
rL
2
.10
rL
L1- _OwPP
_-
.60t0
/
I
0.
MAGC Rre-Room: Pressure
at the roomrfloor
~
The (irn)
Time(mir)
e
0.
400
300
300
Room Pressure
t
PBE#3. Test 13
.\
100
10
.-100
-100
2
&
i -200 -300
-400
-500
-w
RoomPressure
CPBE#3, Test 16
200
4
8
10
1
2
4
6
~
2
6 I-200
- 300
]-
-400
MAGC Rio-Room:
the room floor
-600
4
]ssure
f
a
_[
-600 -
-MAGIC Rre-Rom: Prssure
at the room ftloor
-70011
Tie (rrn)
Tim (wrn)
Figure A-34: Compartment pressure in ICFMP BE #3, closed door tests.
A-47
Technical Detailsfor the MAGIC Validation Study
a2
e
Open Door Tests
rL
Time (in)
A
2,.
Room Pressure
2
_
_
_
_
2-
PBE#3, Test 3
_
-_-ConrpP
1-
I1-
20
O5
5
e.
,e -2
a.
-IC
MAGICFire-Room:
Pressure at the room toor
0-
5
__
-
-4
-2
-2
- -6
AGIC Rre-oom: Pressure at the
room ittrr
-3
.q
-8
Time(rmn)
Tme (rmir)
4
Room Pressure
1 RWBE#3, Test 5
2
15
a-
20
25
-2
a
2
CL
-4
-.- CoripP
-6 _-MAGIC Rre-Pom: Pressure at the room tflor
Tine (rmin)
Time(rin)
0
Time(rnin)
Tim (rin)
Figure A-35: Compartment pressure in ICFMP BE #3, open door tests.
A-48
BE#3, Test 9
Technical Detailsfor the MAGIC Validation Study
Table A-16: Relative differences for compartment pressure in ICFMP BE #3 tests
ICFMP 3-1
ICFMP 3-7
ICFMP 3-2
ICFMP 3-8
ICFMP 3-4
ICFMP 3-10
ICFMP 3-13
ICFMP 3-16
ICFMP 3-17
ICFMP 3-3
ICFMP 3-9
ICFMP 3-5
ICFMP 3-14
ICFMP 3-15
ICFMP 3-18
AE (Pa)
57.6
45.9
290.0
189.3
56.6
49.3
231.5
80.6
194.9
-1.9
-2.0
-1.8
-2.1
-2.4
-2.0
AM (Pa)
35.7
21.0
241.2
195.7
57.6
34.3
313.6
162.5
144.8
-2.7
-2.7
-2.5
-2.7
-2.8
-2.8
Relative
Differenc
e
-38%
-54%
-17%
3%
2%
-30%
35%
102%
-26%
41%
36%
40%
31%
17%
41%
A.8 Target Temperature and Heat FlIx
Target temperature and heat flux data are available from ICFMP BE #3, #4 and #5. In BE #3,
the targets are various types of cables in various configurations - horizontal, vertical, in trays or
free-hanging. In BE #4, the targets are three rectangular slabs of different materials instrumented
with heat flux gauges and thermocouples. In BE #5, the targets are again cables, in this case
bundled power and control cables in a vertical ladder.
Cable targets in MAGIC can be represented as Cables or as Thermal Targets. This section
provides graphical comparisons and relative differences for both options. Targets are virtual
sensors in the computational domain characterized by a thickness and themo-physical properties.
On the other hand, cables are specified as cylinders of some length, with multiple concentric
layers of different materials to account for jacket, insulation and conductor.
For radiated heat flux, the Target/Heat Flux.lncident heat flux output option in MAGIC was
selected for comparison with experimental results. This is the sum of all external radiative heat
fluxes impacting the target. In the case of total heat flux, the Target/Heat Flux/ Total heat flux
"fluxmeter" output option in MAGIC was selected for comparison.
The total flux gauges used in the experiment corresponds to the total heat flux with a target
calibrated using 750 C (167 0F) cooling water. In MAGIC, the target measuring total heat flux is
based on ambient temperature (-20 'C, 68 0F).
A-49
Technical Detailsfor the MAGIC Validation Study
A.8.1 ICFMP BE #3
For each of the four cable targets considered, measurements of the local gas temperature, surface
temperature, radiative heat flux, and total heat flux are available. The following pages display
comparisons of these quantities for Control Cable B, Horizontal Cable Tray D, Power Cable F
and Vertical Cable Tray G.
The superposition of gas temperature, heat flux and surface temperature in the Figures on the
following pages provides information about how cables heat up in fires. In MAGIC, cables can
be modeled using the Targets option or the Source/Cable option. This study evaluates both. The
target option is listed in the graphical comparisons as "Surface Temperature of the Target". The
Cable/Source option is listed as "Maximum Surface Temperature".
Favorable or unfavorable predictions of cable surface temperatures can usually be explained in
terms of comparable errors in the prediction of the thermal environment in the vicinity of the
cable. Regardless of the complexity of the target, the model must be able to predict the thermal
insult to it.
The following Figures and Tables provide the graphical comparisons and calculated relative
differences. Results are classified by cable. That is, for a selected cable, the surface and gas
temperature and the total and radiated heat fluxes for the 15 tests are grouped together. The
tables for relative differences for those 15 tests follow each group of graphical comparisons.
Relative differences were calculated for surface temperature, radiative heat flux and total heat
flux.
A-SO
Technical Detailsfor the MAGIC Validation Study
Target B-TS-14 Temrperature
140- tCFiPBE#3 Test1
120 2
.~
100
A
---
R
60
.2
B-TS-14
....---
L.
MAGICB-TS-14:Gas
a
l-
te
0
s~teibrperature
20 tt
T
_
0
5
10
Catur
rMAt
B-TS14: Surface
_ree
ure o the target
15
Trre (rrin)
20
25
0
30
5
10
15
TkM (rrin)
20
25
30
0
120
100
ar
E
60
a)
0
5
10
15
20
25
30
Tirre (rnin)
rJU
rarget B-TS-14 rerrperature
140 CFPBE#3,Test7
--
Target B-TS-14 tLeat ttux
ICRPIBE#3, Test7
7.0
6
OO120
-0
5.0
MAGIC FG3&4;
-ncident heat ttux
Cable Total Rux 4
4.0
.
9
.MAB-T
B-14 14Ga
40 *
Yterrpeature
MAGIC B-TS-14: Surface
tenerature of the target
20
-'--
3,0.
Cable Ra* Gauge 3
MAGIC R33&4: Total
heat flux 'tktxrreter'
..
1,0
0.0
0
5
0
5
10
15
Ttrne (rnir)
20
25
30
0
5
10
15
Twne(rrin)
20
25
30
a
E
I.-
10
15
Thre (rni)
20
25
30
Figure A-36: Thermal environment near Cable B, ICFMP BE #3,Tests 1and 7.
A-51
Technical Detailsfor the MAGIC Validation Study
12,0
Target B-TS-14 Tenperature
tCFWBE#3. Test 2
200 -
,..-
10,0 -_
I
Cable PladGauge 3
*
8.0 ._
h
.r7.,
-MAGIC
100
6,0 _-
*B-Ts-14
t
50
C.2
...
MAGC B-TS-14: Gas
Surface
teerature
I
0
.f
BE #3, Test 2
CtF
'
2C0
(
Target B-TS-14 FLeatRux
,.. --- A.
x
40
MAGIC PG3&4: Tota
heat fIux fltuxmeter'
PG3&4:
x1identtheattlluxu
Cable Total Rux 4
,
_
.-
2,0
tenpe4rature of the target
0,0
0
2
-
6
TIM (rnn)
4
8
10
Cable B-TS-14 Terrperature
2 FWP BE #3, Test 2
0
12
2
4
6
Trne (nin)
8
10
12
_
200-
2
1501-
a
11
wn
E
C
_B-Ts-1t4
[--
so -
I
__.
I
0
0
2
4
MAGIC B-Ta- 14: wnum
surfarce terrperature
6
Ttme(nin)
8
10
12
250
2
1bu
,
iii
R
E-
6,0
iL
100
I
4.0
2,0
0
2
4
6
Trne (rrin)
0
2
4
6
Time(nin)
8
10
12
0
2
4
6
Tym (nin)
8
10
6
0
E
E
I!
8
10
12
Figure A-37: Thermal environment near Cable B, ICFMP BE #3, Tests 2 and 8.
A-52
12
Technical Detailsfor the MAGIC Validation Study
204O
YU I
Target B-To-14 Terrperature_
-
8,0
20 MPBE #3,Test 4
7.0
--
.0
it
. -
5,0 U-
4,0
Ii
3,0
B-Ts-14
E100
*X
50
"
I MAGIC
W
B-TS-14: Gas
tenperature
e-MAGICB-TS:o
thacrt
Cable Total lex 4,
Target B-TS-14 FteatFlux
OP BE #3, Test 4
0
5
1E
10
5
f
RG3&4: WOWden
MAGtIC
heat fbux
:
0,0
.
0
MAGIC RG3&4: Total
heat ilux 'iuxrreter'
2,0 I
1,0
tempterature of the target
-.- Tree 4-8
O
250
.
60-.
8,0
6
Cable Rad Gauge 3
-.
2
4
Tna (rrin)
6
8
Tinre(min)
10
14
12
Cable B-TS-14 Tenaperature
BE#3. Test 4
ttF0PFW
200_.
a
150
E
100:
15
-Z 0
-
B-Ts-14
-
MAGIC B-Ts- 14: Vaximum
surfare termperature
50-
0
5
0
10
15
Ttre (mn)
9.0
---
8.0
*
_
Cable FladGauge 3
I
I MAGIC RG3&4: Total heat
hIiuxnreter'
s
MAGIC RG3&4: tncirent
6.0
heat luxs
Cable Total Fux 4
5.0 I -a^
.
7,0
._
a
I-
u.
UI
I,
2.0
1,0
I
I Tz,,2
~Target B-TS-14 t-eat Rux
ICM BE#3, Test 10
0.0
0
Trie (mn)
2
4
6
8
Trh (rin)
10
12
14
>
2!
a
t
I-
0
5
10
15
The (mn)
Figure A-38: Thermal environment near Cable B, ICFMP BE #3,Tests 4 and 10.
A-53
Technical Detailsforthe MAGIC Validation Study
350
300
250
°
200
:2
I
100
I
50
0
0
2
4
8
6
Trre (min)
300
10
2
4
6
Tire (rrm)
8
10
12
Cable B-TS-14 Ternperature
I
PBE#3, Test 13
I
__n I
6
0
12
1501wv
I.-
*/
50.
-
-M
f
B-Ts-14
AGIC-Ts-14: Maxirnum
surface termerature
0
2
0
4
6
Tme (mrin)
450
8
10
12
__:T- ..
*B-Ts-14
M*GICB-TS.14:Gas
ternerature
00IO B-TS-l14: Surface
'-temvperature of the target
350
0
300 *
I:
250
,;t-
.
14,0
4....
.
_
4-8
?
10,0
_
8,0
.*
..
MAGIC 33&4: Total
heat flux Ihjxreter'
-MAZGCFFC34:
..
Irciderit heat flux
.
Cable Total Fux 4
A
.
200a
I
150
Cable Rad Gauge 3
*
12,0
-
1100
/
011111
TargetB-TS-14 Terrperature
I`
B~PE
#3, Test 16
U-n .
w l
0
2
4
6
Trre (rinn)
.
8
10
Target B-TS214 F
0,0 Fs
CFWBE#3, Testl16
CD
2
4
6
Tirre (irn)
llet
8
10
200
150 .-
I-
12
2,0 1
Cable B TS-14Temruerature
ICF'8BE#3. Test 16
300
G
-
.
,11
100 I
_
~a
B-Ts-14
50-
_
i-hAGVCB.Ts-14: Maximum
surface tetqperature
.
0
2
4
6
8
10
12
Tire (rri )
Figure A-39: Thernal environment near Cable B, ICFMP BE #3, Tests 13 and 16
A-54
12
Technical Detailsfor the MAGIC Validation Study
250
a
E
150
-
6,0
*-M,.G.
B TS-14:Gas
terroerature
-MAGIB-TS-14: Surfare
,.4errperature, of the target
_
Vree 4-8
200
e
flu I
B-Ts-14
-
w
I~
100
5,0
-
4,0
-
3,0 _,
I.
;
2,0
50
g
g
*
CabbeRladGauge 3
.....
-
MAGIC F3&4: Total
heat flux lluxrrter'
MAGIC IC.3&4:
Inckient heat flux
Cable Total Flux 4
Target B-TS-14 FbeatFlux
CF
BE#3,Test17
Il
1,0
Target S-14 Terperature
CFt'P BE#3. Test 17
.
0
0
2
4
6
Twre (rrin)
0
2
4
6
Tom (rnt )
8
10
12
0
2
4
6
Tir (rrtn)
8
10
12
120
100
a
80
E
60
E
8
10
12
Figure A-40: Thermal environment near Cable B, ICFMP BE #3, Test 17.
A-55
Technical Detailsfor the MAGIC Validation Study
U,80
300
Target B-TS-14 Temperature
tICF
250
Target B-TS-14 Heat Fux
BE 113.Test 3
tF
7,0
BE#3,Test3
b,U ,
iv
g00 -
150
B-Ts-14
4
t°
_
50
E
U-
100*
if
X
.MaGICB-TS-14:
.....
Gas
terrperature
-MIAtCB-TS-14: Surface
tenferature of the target
_Tree
4-8
50.
_
- I
3,0
_
Y"
M
zo
ef
1,0-
r
^-
hkthent
0
15
10
TrNr (rrin)
5
0
25
20
>Cable
3
Rad Gauge3
*G*-... MAGtC FG3&4T Total
heat fi
luxrmeter _
MAGIC FtG3&4:
heat fhxi_
Cable Total Flux 4
,
0.0
O
g
5
10
15
Tine (rrin)
20
25
300,
Cable B-TS-14 Temperature
tr BE #3, Test 3
250
& 200
a
_
N
//'
150-
~w
100lB-Ts.14
f
50
GIC B-Ts-14: Maxkmm|
_
surface temfperature
|-MPI
0
0
5
10
15
Tkre (rin)
20
25
30
-"A.
Target B-TS-14 Temperature
250
BE #3, Test 9
4 t^
loz~
I
I
A-eoA,
;
150
-
100
B-Ts-14
-
I
V}I
50
-, --
.. C
GtCB-TS-14:
-W. Gas
I
teu¶veratureI
MC.GKC B-TS-14: Surface
| enpe ature of the target
- -T r
a
0
0
5
15
10
Tim (rm)
20
25
5
10
15
Tre (rrin)
20
a
0
5
10
15
Trne (rrn)
20
25
Figure A-41: Thermal environment near Cable B, ICFMP BE #3, Tests 3 and 9.
A-56
25
Technical Detailsfor the MAGIC Validation Study
12,0
250
Cable FladGauge 3
.
10,0
____
8,0
86,0,
El
4.0 T
Ia
0,0.
5
10
0
5
10
E
15
irTne(rrnn)
20
15
20
30
25
_s
-
a
-
2,0
0
MAGIC FRG3&4:
Total heat flux
fluxmeter'
Fr.3&4: Incidentheatflux
-MAGIC
'
Target B-TS-14 H-eatRux
0
5
10
1
BE#3, Test 5
CFt
n
15
Tirn (rrn*)
20
25
30
a
E
I-
30
25
Tom (rnin)
7,0
Target B-TS-14 FleatRux
6,0 -atWBE#3, Test 14
W
e
150
5,0
4,0
100
u
3.0
I
2.0
/'z
50
1,0
0.0
00
250
5
10
15
Trne (rrnr)
20
25
-.-
Cable Rad Gauge 3
-
MAGIC RG3&4: Total
heat flux ¶lkxrreteri
-
0
5
Tine (rrhi)
10
MAGIC
Fr3&4: itcident
heat flux
-- e- Cable Total RLux4
1--
Cable B-TS-4UTerrperature
CMPBE 43. Test 14
200-
o
k 1001
tI
>
D
150
/f
'
--
B-Ta-14
50-MAGIC B-Ts-14: Maximu
surface tenperature
/
0
5
10
15
Tenm(nin)
20
25
30
Figure A-42: Thermal environment near Cable B, ICFMP BE #3, Tests 5 and 14.
A-57
Technical Detailsfor the MAGIC Validation Study
900
Target B-TS-14 Terperature
ICF'BE #3, Test 15
DOW,"
750600 4F
fAJI-
I
E
I
1
0
0
B-Ts-14
.
0
---
I....
MAGIC B-TS-14: Gas
temperature
eMAICB-TS-14: Surlare
tenperature of the target
Tree 4-8
5
0
25
Trim(nin)
10
15
Tim (min)
20
25
I-
E
a
0
5
10
15
Tfrne (min)
20
25
30
--18,0
18.0 .......16,0 14.0 -
a
MAGIC G3&4: Totat
heat tlux ftnwmeter
MAGIC P3S4:
hicident heat thjx
-Cable Total Rux 4
12,0
LI
Cabl ReadGauge 3
10.0
U-
E
It
I!
8.0
6.0
2,0
2.0
fnn
0
5
10
15
Twre (rrm)
20
_
250 I
25
0
-FM
5
B-1S-14
Heat
TargetBE#3,
Test
18FkJX]
10
15
Trne (nin)
20
-
200
a
C
GableB-TS-14 Tenperature-
150
FW BEE#3 Test 18
100
- _B-Ts-14
50
-
M-GIC B-Ts-14: Mixrum
surface temrperature
u
0
5
10
15
Tim (min)
20
25
30
Figure A-43: Thermal environment near Cable B, ICFMP BE #3, Tests 15 and 18.
A-58
25
Technical Detailsfor the MAGIC Validation Study
Table A-17: Relative differences for surface temperature to cable B
Control
Cable
Test 1
Test 7
Test 2
Test 8
Test 4
Test 10
Test 13
Test 16
Test 17
Test 3
Test 9
Test 5
Test 14
Test 15
Test 18
Target Surface Temp, B- Cable Surface Temp, BTS-14
TS-14
Relativ
Relative
0
e Diff
AE (C) AM (C0)
AE ( C) AM (0C) Diff
5%
112
106
8%
115
106
-1%
108
87
23%
111
90
6%
187
176
16%
204
176
1%
186
183
11%
203
183
29%
192
149
35%
201
149
31%
189
144
38%
198
144
25%
228
186
42%
264
186
38%
222
160
52%
244
160
80
83
-11%
202
226
-9%
206
226
-12%
199
228
-.10%
204
228
16%
175
150
21%
182
150
0%
199
199
1%
201
199
-20%
333
416
-24%
315
416
-15%
200
236
15%
200
236
Table A-18: Relative differences for radiative and total heat flux to cable B
Radiant Heat Flux Gauge
_
3
AE
Control (kW/m2 AM
(kW/m2 )
)
Cable
1.3
1.1
Test 1
3.6
2.9
Test 7
3.6
4.4
Test 2
3.4
2.9
Test 8
3.0
3.9
Test 4
1.2
1.2
Test 10
3.6
2.9
Test 13
3.5
4.3
Test 16
3.4
2.7
Test 17
7.0
4.8
Test 3
3.3
2.8
Test 9
7.4
46.5
Test 5
5.9
4.1
Test 14
1.5
1.3
Test 15
3.8
5.2
18
Test
Total Heat Flux, Gauge 4
AE
Relative
Diff (kW/m2 )
1.85
_12%
1.84
.26%
5.26
-18%
5.58
I 7%
5.52
-22%
4.91
2%
8.26
23%
8.37
-19%
2.36
25%
7.10
47%
6.58
15%
6.86
-84%
3.82
42%
57.72
15%
7.61
-28%
AM
(kW/m2 )
3.06
2.99
6.92
6.85
6.57
6.51
11.48
10.03
3.43
6.74
6.56
5.67
6.37
11.72
6.86
Relative
Diff
65%
62%
31%
23%
19%
33%
39%
20%
45%
-5%
0%
-17%
65%
-80%
-10%
A-59
Technical Detailsfor the MAGIC Validation Study
_
Target D-TS-12 Terrperature
1C
IC
ai
E
....--. MAGIC D-TS- 12: Gas
temperature
E R so
IL
DTS-12: Surface.
terrperature of the target
-MAGIC
4
I-
e-
20
Tree 3-9
n
0
5
10
15
rTie (frn)
20
0
5
10
15
Tro (rrin)
20
25
30
0
5
10
15
20
30
25
Trr~e(rrin)
i2
a
E
C
I-
30
25
3.5
180
Target D-TS-12Terrperature
3FWBE#3, Test 7
_
160
3,0
140 -
~
a. 120 -
2.0
T
__
10D-
9q
a-
41rTel .
.....
60
-'
-
__
N
A__
_
12: a
M
MAGIC D-TS-12:
.2
1.5
I
1.0
Gas
-
-a-Tree 3-9
20
0
5
10
20
15
rene (rir)
30
25
Cable D-TS-12 Tenpzerature
tCUWBE#3 Test77
140
e
100
2
_
/
120
a
__
temperature
MAGIC D-TS-12: Surtace
of t target
temperature
3
-
40
_
__
a
E
//-tGIC
C2
pf
40
20
D-TS-12: Mabxknurn
surface lemfperatur
D-DTs-12
0
5
10
20
is
The (rrin)
25
30
Figure A-44: Thermal environment near Cable D, ICFMP BE #3, Tests 1 and 7.
A-60
Technical Detailsfor the MAGIC ValidationStudy
9.0
.Target D-TS-12 Heat Rux
.P
BE 113,
Test 2
8.0
6,0 -
U
C
II
_"
5,0 -
_
An.
3,02,0 ---
Caflie Red Gauge 7
1.0
0,0
..
,
0
T
(rrin)
.
2
-
M.AGC RO7&M: Total
heat flux fluxreter'
-AGIC r7&8:
khcident heat ftux
Cabl Total
tPux 6
300
a)
0
4
2
6
8
10
12
Taoe (rmn)
6
a)
E-
C
I
0
2
0
2
4
6
Tane (fmn)
8
10
12
e
4
8
6
Ttne (min)
10
12
Figure A-45: Thermal environment near Cable D, ICFMP BE #3, Tests 2 and 8.
A-61
Technical Detailsfor the MAGIC Validation Study
300
250
6S 200
<
E'
150
L.2
E100
x
50
0
0
15
10
S
5
Trm (mn)
10
15
Tre (mrn)
300
-MAGIC0TS-12: MaxiTum
surface aenerature
250
a
a
E
200
lDU
100
/~
550
Cable TS-35 Terrerature
fP8E#3,Test2
f
.'
0
0
5
15
20
Trne (min)
10
25
30
10
250
9
able Flad Gauge 7
MAGicFr7&8:Totawheattkux
1tlixreter.
8 .......
7
a
6
Gble
ledGuge
-.-
~
FWGIC RG7&8: kicident heat
-
-
i!
EU.
a
.
4
I
I
15
10
5
TiM (rmn)
w
.iie
o
0
{t
I3
I
0
_10
_
Target D-TS-12 leat Rux N
CRAPBE #3. Test 10
5
10
Tre (mn)
3-MAGICD-TS-12: Maxhmm
250
surface temlperature
D-Ts-12
-.-
I
200
2Cto6
9S
a
I-
-
100
GableD-TS-12
Tenferature
KFPWBE#3. Test 10
so
50
o
5
10
20
15
Terr (mrrn)
25
30
Figure A-46: Thermal environment near Cable D, ICFMP BE #3, Tests 4 and 10.
A-62
15
Technical Detailsforthe MAGIC Validation Study
a
E
I-E
I
2
0
4
8
6
Trne (nin)
10
12
0
2
4
6
Time (rrin)
10
8
12
-MAGICD-TS-12: Maxkmm
250
-
surface tenperature
_D-Ts-12
200
5
E
.
101
ll
f50
O0
/
Cal -T 2Ten~perature
.Ct3BEf 3, Test 13
_,n
15
10
5
Tme (nin)
20
25
son
RIG7&8:
Total
*
III....I WGIC
heat flux 'ltuxmeter'
KvAGICRG7&8:
ki[
cident heat ttux
Cable Tota Rux 8
14 0 12,0 e
w
E
a
E
11Q0
0f
'^
,
8,0 -
X
Ia
--/
40
'
uu_
0
2
4
6
Tim (.-in)
8
10
12
e\L
H
12
D.TSe
tC3E^3E #3, Test
c
.,
..
0
2
4
,
6
Tne (rrin)
.
.
8
10
.
..
I
12
00
F WBB#A
3. Teat 16
hZt
0
5
10
15
20
25
30
Time (rin)
Figure A-47: Thermal environment near Cable D, ICFMP BE #3, Tests 13 and 16
A-63
Technical Detailsfor the MAGIC Validation Study
8,0
Cable Pad Gauge 7
7,0
FG7&8: Total
heat flux Iluxrnter
WGIC Fr.7&8:
hIcident heat flux
Cable Total Flux 6
6.0
&
E
E
U-
0~
I-
5,0
4,0
4,0
3.0
' :
-
.MLAGC
Target D-TS-12 Feat Fux
I
2.0
7
i
0.0
0
2
4
6
Tno (rrn)
8
10
12
0
2
4
6
Trr (rtn)
8
120
/\
100 &
Cable D-TS-12 Terrperature
aNP BE#3, Test 17
80
00
S
0
40-
20
-
-
0
2
4
MAGICD-TS-12: Maxmrn
surface tenperature
I
6
Twne(rin)
8
10
12
Figure A-48: Thermal environment near Cable D, ICFMP BE #3, Test 17.
A-64
10
12
Technical Detailsfor the MAGIC Validation Study
50
Target D-TS-12Terrperature
aM BE #3 Test 3
50
3.
,
21
E
S
e
1
50
..
w0
-M
M.. GICD-TS-12: Gas
-
DTS12: Surface
terrerature of the target
I
AGC
50
D-s1
0
20
15
10
Twm (rrin)
5
25
2,
T
0
5
10
15
20
25
30
Tffw (rdn)
E
a
2.
81P
te
I-
z
I
0
5
15
10
Tre (ffi)
20
25
0
5
10
15
20
25
mare (rnn)
I-
E
0
5
1S
10
rim (nfi)
20
25
Figure A-49: Thermal environment near Cable D, ICFMP BE #3, Tests 3 and 9.
A-65
Technical Detailsfor the MAGIC Validation Study
250
Target D-TS-12 Peat Rux
6.0 - - tP8W E #3, Test S
--
a
5,0
4,0-
a
rL
100
50
te
3,0-
e
2.0-
.. ..
MAGIC FR0&8: Total
1,030
t
flu{xfltuxrreter
MAGIC RG7&8: Incident
heat tfux
nn0
5
10
15
tirme(rrni)
25
20
30
0
5
10
15
Tom (rinn)
20
t .
25
300
Cable D-TS-12 Terrperature
tP
BE UY3,Test
5
250
£
200
a! 150
a
E
.0
I"-
0-
14
-
Ni.
100
-MGlC D-TS-12: Nbxrnurn
suracetererature
50.
0
0
5
10
15
20
Trom (rrin)
25
30
8,0
Target D-TS-12 Terrperature,
-
250
7,0
BE#3, Test14
6,0
6
9.
E
I-
150
t v/
100
t
...........
fygr -
M
.
* tM.
C D-TS-12: Gas
-
5,0
,;
4,0
U.
a
U
D.1S-12:Surface
terprature of the target
3.0
A
D-Ts-12
-
n
0
15
10
Tirn (rin)
5
20
25
300
Cable D-TS-12Terrperature
ICFPBE#3, Test 14
250
o
200
a
re
.. I
1W
f
I
-
I
-1D-T-12
50 OF
0
-MAGICD-TS-12: MwXffIurn
surface terrperature
5
10
15
20
rrme (rin)
25
30
Figure A-50: Thermal environment near Cable D, ICFMP BE #3, Tests 5 and 14.
A-66
IV I
30
Technical Detailsfor the MAGIC Validation Study
Target D-TS-12 Terperature
tF BE #3, Test 15
3S0
250
9
0
-
}
200
...... . MAGICD-TS-12:Gas _
tenverature
- M
AGICDTS-12:Surface
terrperature of the target
-- Tree 3-9
1I'50
E
I-
.
0
.
10
5
x
f
--
. -\ 11
-
-
,I
D-Ts-12
---
0
CL
-
.. 15
20
0
5
Trie (inn)
25
ime(rrin)
10 -
Cable RadGauge 7
MAGICFR37&8: Total
heat flux ltuxrreter'
MAGIC FG7&8:
h ent heat flux
CabeTotalFlux 6
F
a
a
E
I-
0
5
10
15
20
Trm (frin)
25
30
S
E
a
E
C)
I-
IL
I
0
15
10
Trim (rin)
5
20
25
10
15
Tie (rrin)
20
25
0
e
E
1-
0
5
10
15
20
Twm (rin)
25
30
Figure A-51: Thermal environment near Cable D, ICFMP BE #3, Tests 15 and 18.
A-67
--------
Technical Detailsfor the MAGIC Validation Study
Table A-19: Relative differences for surface temperature to cable D
Target Surface Temp, D-
Cable Surface Temp, D-
TS-12
TS-12
Control
Cable AE (C)
Test 1
Test 7
Test 2
Test 8
Test 4
Test 10
Test 13
Test 16
AM (C)
Relative
Diff AE 00
111
204
203
201
198
263
243
27%
62%
36%
77%
50%
52%
56%
Test 17
Test 3
Test 9
Test 5
Test 14
Test 15
rest 18
Relativ
e Diff
112
115
87
126
150
113
132
173
156
AM (C)
87
126
150
113
132
173
156
108
187
186
192
189
229
222
-4%
48%
4%
70%
43%
32%
42%
_201
199
174
202
251
199
-4%
-8%
9%
14%
8%
-8%
___
210
220
132
178
243
217
204
201
178
203
247
199
-3%
-9%
34%
14%
1%
-8%
210
220
132
178
243
217
Table A-20: Relative differences for radiative and total heat flux to cable D
. Radiant Heat Flux Gauge 7 Total Heat Flux, Gauge 8
Relative
AM
AE
Relative
AM
AE
Control
Diff
Cable (kW/m 2 ) (kW/m 2 ) Diff (kW/m2 ) (kW/m2)
Test 1
1.44
1.59
10%
Test 7
4.16
3.64
3.75
3.43
3.22
1.54
3.59
3.61
3.39
7.02
3.33
-12%
Test 2
Test 8
Test 4
Test 10
Test 13
Test 16
Test 17
Test 3
Test 9
Test
Test 18_
5%.
-33%
14%
1%
-31%
16%
7%
0%
3.95
14
Test 5
Test 15
A-68
3.26
4.78
1.35
3.55
5.26
2.91
6.58
3.32
4.83
1.52
5.90
1.49
3.43
22%
-2%
3.07
2.52
9.83
8.51
7.23
6.71
11.22
11.67
3.29
9.45
9.06
2.99
6.93
6.86
6.60
6.54
11.50
10.08
3.43
6.85
6.68
19%
-29%
-19%
-9%
-3%
3%
-14%
4%
-28%
-26%
6.07
6.45
6%
8.52
20.87
5.87
8.39
-31%
-60%
7.83
6.50
-17%
Technical Detailsfor the MAGIC Validation Study
II60
w
Target F-TST20 Terrperature
1,40 tfXMV BE #3, Test 1
20__
12
&i
it
U-
iS
gz
30
a-
40o
*
:!5
f-Ts-20
......
* -* MAGIlCF-TS-20: Gas
terrrperature
MAGC F-TS-20: Surface
terrperature of the target
-"Tree
5-6
*
2
20
O
0
5
10
15
Trne (rrfi)
20
25
30
0
5
Trre (rin)
10
15
20
25
30
Cable F-TS-20 (Bottoar Terrperature
12 M BE#3, Test 2
100 I
e
fO-
80
RD
40-
-.-
20
-MAGIC
F-Ts-20: Maxkrum surlare terrperature
0
a
I
5
10
15
Tirm (rrin)
20
25
30
Target F-TS-20 Ternperature
KCRIABE #3, Test 7
140
rS
FTs 20
O0
Wu
-
-
-
,
I
N.
-
*
F-Ts-20
L2
MAGIC F-TS-20: Gas
lterrerature
-MGC
F-TS-20: Surface
terrperature of the target
Tree 5-6
I.f
I I.
40
20
-.
0
0
.
5
10
15
Trnm(rfin)
20
25
I
II
30
,
0
5
Trrge(nin)
,heat
C Cable FkedGauge 1
-MAGIC
10
-
MAGIC R3I&2: Total
ftlx 'tkjxmter'
RG1o2:
tncident heat ftux
Cable Total Flux2
.......
I
T.2
0
5
10
15
Tne (fin)
20
25
30
Figure A-52: Thermal environment near Cable F, ICFMP BE #3, Tests 1 and 7.
A-69
Technical Detailsfor the MAGIC Validation Study
250
200
,
150
C
U.
E1 00
18
*
50
-
Cable Rad Gauge 1
-- MAGCRG1&2:Total
-0
2
4
6
TIM (rnn)
8
15
Trme(rin)
20
10
12
0
2
Tune (inn)
4
-.
-
.-
heat flux %luweter'
GMA
RG1&2:
hcident heatflux
Cable Total Rux 2
200
180
t60
140
-
o
120
~-18114060
80
0
5
10
25
30
2Dn
Tw
arget F-S-20 Terrperature
200
ltFMPBE 43, Test 8
k 1
E1 100
-
_
x
F-s2
I-
J
I
MAGIC t-TS-20: Gas
50
MAGIC1TS-20: Surface
terrperature of the target
- -Tree
O
.
2
0
4
6
Tire (frfn)
D
-
O
Dli
10
Cable
Rd
Gauge 1
.MAGtCRG1&
5-6
8
-
12
0
e
2
rTne (fnm)
4~
. .......
-rrGe
-w--
heat flux Iixereter'
(ri1
&2:
Ikldent heat flux
Cable Total FRux2
2t00
&
T
180
Cabte #3
180 -
CRR EtE13,Test 8
140 -
Terrperature
~
120 -
t
18 10D -
/
Fm I-
I-
ez
60
40
-
20.1
MAGIC F-Ts-20: MUxirurn
surface terrperature
u1
0
2
4
6
8
10
12
Tire (rrn)
Figure A-53: Thermal environment near Cable F, ICFMP BE #3, Tests 2 and 8.
A-70
Technical Detailsfor the MAGIC Validation Study
10,0
9.0
-
8,0
*
Cable FladGauge 1
*
7,0 --
a
6,0
ii
u-
MAGC FG1&2: Total heatllkx
'flexreter
MAGIC RG1&2: hinident heat
.
Cable Total Rux 2
I.
I,:
U.
I
E
4,0
ta
ww
--
2.0
-_
1.
-
5
10
15
-
-I-
KCPBE#3, Test 4
PO'
0,0
0
"
3,0\
T
Tine (rirn)
0
Tim (rdn)
10
15
a
an
I!
0
5
10
15
Tmn (rin)
20
25
30
9,0
-_
Cable Red Gauge 1
6.0
- M
ats 5.0
- -
MAGIC Fr1&2: Total heat
flex tluxvrter'
AGICRF1I&2: hrciient
heat flex
Cable Total RFux2
2.
-
8,0
-
7.0
a
1
a
u' 4,0
I-
*'
I
_XY
...
_
.
- IY.
1'
A
I
2.0
0501
0
5
10
15
Trre (rin)
O
0,0FMP3BEa#,
et1
5
10
15
Trne (in)
Cable F-TS-20 Terrprature
lCFRBE#3. Test 10
a
e
1!
100F-Ts-20
50
Fl,
0
0
surlae terrperature
5
10
15
Twne(rin)
20
25
30
Figure A-54: Thermal environment near Cable F, ICFMP BE #3, Tests 4 and 10.
A-71
Technical Detailsfor the MAGIC Validation Study
1
u6,0
Cable red Gauge 1
--
14.0 12,0
..
g
10.0]
,
U-
8,0
:.
! .1
6,0
.
A,
,.
M....
WG Fr1&2: Total
heat flux fluxneter'
UAGC F31&2:
' kcident heat flux
_ . Cable Total Rux 2
arget F-TS-20 -leatRux
WLT
BE #3, Test 13
4,0
2,0
0,0
0
2
4
6
Trn (rin)
8
12
10
J
C0
2
4
6
The (rri in)
8
10
12
U
F-
0
15
10
Twre (rrin)
5
20
25
600
500
S 400
i
1!
300
1
U-
I,
200
0
4
2
6
Tie (rlin)
8
10
12
0
2
4
6
Time (trin)
8
10
250Cable F-TS20 Terrperature
Al
200 -
a
ISO6-
i
100-
1C
BE#3, Test 16
A
-_- F-Ts-20
50 -
__
0
5
GfZIF-Ts-20: Maxinum
surface temrperature
10
15
Twm (rrin)
20
25
30
Figure A-55: Thermal environment near Cable F, ICFMP BE #3, Tests 13 and 16
A-72
12
Technical Detailsforthe MAGIC ValidationStudy
300W
u_,_____
-.....
250
-
-o-
200
a
E
u.
/
t
50
MGIC F-TS-20: Gas
tenperature
WGICF-TS-20: Surface
temperature of the target
Tree5-6
-
I
Target F-TS-20 Terrperature1
-
fCFNPBE#3, Test 17
0
0
2
4
0
a-
6
Twne(rrin)
8
10
12
0
2
4
6
Tune (rin)
8
10
12
e
120
Cable F-TS-20 Terrperature
#3, Test 17
100 -lCFWBE
80 60
/
-
IAnY
FpTa-20:t
surf ace tenperature
-MAGIC
20-
0
2
4
6
Temr(rrin)
8
10
12
Figure A-56: Thermal environment near Cable F, ICFMP BE #3, Test 17.
A-73
Technical Detailsfor the MAGIC Validation Study
E
a
E-
I
-.
O-IF
-r.
OI
0
5
10
15
Tire (rin)
20
25
Cable Red Gauge 1
-.
It....... .,MG, FG1&2: Total
heat flux tiuxireter
1^
1
0
5
Tire (min)
GrORG1&2:
-
Incident heattfux
Cable Total tRux2
250
Cable F-TS-20 Tenperature
tC3BE#3,Test3
2cX3
_!
"1
,A
2.
e
a
t
icx
100
S-
50
-
F-Ta-20
-
WGIC F-Ts-20: lanxi.nu
surface temperature
0+1
0
5
10
15
Tfm (rrin)
20
25
30
250
Target i=TS-20 Terrperature_
200-tVPtE#3 Test9
a
15v -
405/
10
E
a)
-
i-Ts-20
-r
50
.
0
I
r i-TS20: Surface
T=,m 6ratureof the targe
f...
O-
U-
MAGICi TS-20: Gas
Jt
{
10
15
Trne (rin)
.
2t5
20
-
Cable Rad Gauge 1
D.
MAGIC RG1&2:Total
heattfiu taurmeter
GC RG1&2:
.......
.
_,I
5
DI-
0 56
10
10
hIncident
heatPux
flex
CableTotal
2
Tire (rin)
250
20D0a
e
Cable FTS-20 Terrperature
DM'BE#3. Test1401125
9
150 -
'
DS
-
50-
-
0
Oi
5
10
F-Ts-20
M-rGIC F-Ts-20: iruxirnr
surface terierature
15
Trne (rrin)
20
25
30
Figure A-57: Thermal environment near Cable F, ICFMP BE #3, Tests 3 and 9.
A-74
Technical Detailsfor the MAGIC Validation Study
Target F-TS-20 Terrperature
lCFW BE #3, TestS5
U
a
41a
I-
x
50 f
. .te:Mrature
MAGICFqTS20: Surtare
of the target
_
0
5
10
15
Time(rrn)
20
25
0
30
5
10
15
Trie (rrn)
20
25
30
250 .
Cable F-TS-20 Terrterature
ICFF BEt3. Test 5
S
1bu
-
I
-00,01011r-
50-
0
5
--
F-Ts-20
-
MAGIC F-Ts-20: MCaxiflJ
surface tevnperature
10
15
Tirre (rrin)
20
25
30
6'
E
f-
U
3,0
0
I
2,0 -__
1 0 -
~~~.
Vble ad Gaugte 1
<e
_........ AGIC FGI&2: Total
hea flu. MAfluxter&'
0,0
0
15
10
The (ran)
5
20
25
0
5
Trrn (rrn)
10
-
hickient heat flux
Cable Total Flux 2
Cable F-TS-20 Tenrperature
E #3, Test 14
lCF'
I
Bt
200
r6
e
ai
150 -
s
-a-F-Ts-20
50
-
0
5
10
PA~GIC
F-Ts-20: Maxirrur1
surf ace levrrerature,
15
Tire (rrhn)
20
25
30
Figure A-58: Thermal environment near Cable F, ICFMP BE #3, Tests 5 and 14.
A-75
Technical Detailsfor the MAGIC Validation Study
25,0
a
0
U
a-
-
700
-
600
F-Ts-20
M- GIC F-Ts-20:.Ixir r
surface terrperature
a
400 -
e
Ipo
30020D
Y100 -able
'
r-TS320 Ter1erature
U=W BE U3, Test 15
5
0
10
15
Tkne (frin)
20
25
30
Target F-TS-20 Terrperature
ICFUBE#3,Test18
250
0
a
S
02
150 r
/
F-Ts-20
U.
100
/
If*.. I
I MAGIC F-TS-20: Gas
-soJ/MI
I
TS-20: Surface
of the target
5terrperature
-
- Tree 5-6
On
0
15
10
Trmn(rrrn
5
20
25
5
10
15
Tme (rrrn)
20
I-
0
5
10
15
Tibm (rrin)
20
25
30
Figure A-59: Thermal environment near Cable F, ICFMP BE #3, Tests 15 and 18.
A-76
25
Technical Detailsfor the MAGIC Validation Study
Table A-21: Relative differences for surface temperature to cable F
Surface Temp, F-TS-20
_Target
AE (0C)
Cable
83
Test 1
90
Test 7
129
Test 2
131
Test 8
149
Test 4
Test 10_ 150
143
Test 13
168
Test 16
Test 14
Test 15
Test
18
Relative
Relative
Power
Test 17
Test 3
Test 9
Test 5
Cable Surface Temp, F-TS-20
195
195
175
171
669
232
AM (0C)
Diff
109
109
183
181
189
194
219
227
32%
21%
42%
38%
27%
29%
52%
35%
202
200
178
197
245
205
3%
3%
2%
15%
-63%
-11%
AE (0C)AM (00)
95
83
101
87
150
129
148
131
158
149
174
150
188
143
199
168
195
195
175
171
669
232
182
180
157
179
231
184
Diff
15%
12%
16%
13%
6%
16%
31%
19%
-7%
-8%
-10%
5%
-66%
-20%
Table A-22: Relative differences for radiative and total heat flux to cable F
Power
Cable
rest 1
Test 7
Test 2
Test 8
Test 4
Test 10
Test 13
Test 16
Test 17
Test 3
Test 9
Test 5
Test 14
Test 15
rest
18
Radiant Heat Flux Gauge 1 Total Heat Flux, Gauge 2
Relative
AM
Relative AE
AM
AE
Diff
(kW/m2 ) (kW/m2 )
(kW/m2 ) (kW/m2 ) _ Diff
90%
3.05
1.60
44%
1.24
0.87
95%
2.98
1.51
81%
3.60
1.99
44%
6.89
4.77
15%
3.40
2.95
38%
6.82
4.93
65%
3.33
2.02
29%
6.49
5.02
2%
2.69
2.65
48%
6.43
4.36
48%
1.22
0.82
56%
11.40
7.28
83%
3.54
1.93
61%
9.90
6.13
%
_71
3.29
1.93
83%
3.43
1.85
20%
3.26
2.73
17%
6.51
5.55
139%
6.91
2.90
25%
6.33
5.08
50%
3.17
2.12
-18%
5.34
6.45
-48%
9.59
18.29
77%
6.26
3.46
108%
5.73
2.76
-56%
10.44
23.94
169%
1.49
0.88
-21%
6.94
8.74
-25%
3.87
5.18
A-77
Technical Detailsfor the MAGIC Validation Study
*
Vertscal Cable Ts-33
.MAGCTS-33: Gas tefrperature
250
-....-
200 -
Surface
-NAGtCTS-33:
tenperature of the target
,
Tree 2-5
E
a.
-
150
LL
E
100
I0
f
0
5
1.0
Target TS-33 Terrperature
rF BE#3, Test 1
_
50
10
15
Twr (win)
20
25
0,0
0
30
5
10
15
rem (min)
*Cabte
Rad Gauge I10
20
25
30
140 ._.Vertical Cabte Ts-33
-M
120
GICTS-33: Max
sWurface
ternperat
-p
--
100 -_.
80
60
F-
40
20
Cable TS-33 Temperature
2 WBE#3, Test 1
Vn
0
5
10
15
Tirn (rrrn)
20
25
30
250
200
e
..... .. MAGIC RB1&10: Total
heat flux tuxrreter
3.0 -
-MAGICFPG9&10:
kcident heat flux
150
x
?1
4,0 -
100
tr
140
Target TS33 Weat ux
l
BE #3, Test 7
0,04
0
5
10
0
5
10
15
Tne (rrin)
20
25
30
0
5
10
15
Tfne (rrtn)
20
25
e
1!
15
Teta (rrin)
20
25
30
Figure A-60: Thermal environment near Cable G, ICFMP BE #3, Tests 1 and 7.
A-78
30
Technical Detailsfor the MAGIC Validation Study
12.0
1
-w
Cable Pad Gauge 10
Vertical Cable Ts-33
350
10,0
.MAGICTS-33:Gas
.----300
terrurature
a
250
-MAGICTS-33:Surface
terrperature of the target
Tree 2-5
E
100
heat flux Iluxrreter'
MAGIC RG,9&10:
kicident heat f ux
Cable Total Rlux 9
8,0
6,0
I.
H
,
-
I
2,0
TS-33 Tenperature
UMP BE #3 Test 2
ITarget
50;
A40
.S
UTest 2r
lCFBE
0,0
12
10
8
6
Trne (rin)
4
2
0
Tagel=TS~-33 Heat Rux.
.
^
0
2
4
8
6
Ture (mnn)
10
12
te
250
Cable TS-33 Tenperature
FbP BE #3, Test 2
200-
rD 150--
S@
a
9
..1w
/
I
|_--VertFical Cable Ts-33
MACM TS-33: Wxi.urn
50
o
.
0
5
L350
-
surface tenperature
I
20
15
Trre (nin)
10
30
25
Vertical Cabe Ts-33
350
-.*- ... *MAGICTS-33: Gas
terrprature
Surface
-WUGtCTS-33:
tenperature of the target
300
a
250
iiS
200
E
150
II-a
Tree 2-5
-
-
*
tL
-
0
I
100
- Target TS-33 Tenperature
1F BE#3, Test 8
50
-;
0
4
2
6
Tirre (min)
8
10
12
0
2
4
6
TkT* (min)
8
10
12
250
Cable TS-33 Te.nperature
BE#3. Test 8
CRP
R
a
2
150
100
_
A-
50
-
|~::
Vertical Cable Ts-33
Gt TS-33: Mbxirrumj
surface tenperature
-
11±-
0
2
4
6
Tirne (nin)
8
10
12
Figure A-61: Thermal environment near Cable G, ICFMP BE #3, Tests 2 and 8.
A-79
Technical Detailsfor the MAGIC Validation Study
350
-.
Vertcai Cable Ts-33
300
2501
2
200
ajn
150
* f.MAGICTS-33: Gas
..
terrerlure
MAGTS-33: Surface
terTerature of the taQt,
:
__ -: _ W.Tree2-5
-
s
E
e
F*n
I.
I
so
I
0
5
15i
10
10
15
Tmrr (nit)
Tom (rrin)
250
Cable TS-33 Terrperature
/\,
200
rD
1rn
I
-
BE#3, Teat
4
I
D0
5ic
5
_a Vertical Cable Ts-33
TS-33: Mwxirrum
-h-AGIC
M
0
5
.
surf ace terrperature
-
10
15
Tare (rrin)
20
25
30
Target TS-33 Heat F
6 -ICfWBE#3, Test 10
i-
hux
-
5
101
0
I0 Cable Pre Gauge 10(
/N-*
.
0
10
5
15
Tnm (rin)
0
AG3C FKM910: Totat
~heat
f lux 'f luxrreter'
MAGIC P10~i: hkidkntr_
B
t
1
&
.
heat fhux_
5
10
Tmw (rrrn)
250
Cable TS-33 Terrperature
IWBE#3, Test 10
200
2.
a
e
C
Vi
100
f
-Vedcat OCabte
Ta-33
,Isurface
-MAGICT;33: A~xn
tenperature
0
10
05
15
(rrh)
Tiarae
20
25
30
Figure A-62: Thermal environment near Cable G, ICFMP BE #3, Tests 4 and 10.
A-80
III
15
Technical Detailsfor the MAGIC Validation Study
a)
a
S
-
a-
0
2
4
6
Trne (nin)
8
10
0
12
2
4
Tr
6
(min)
8
10
12
300
Cable TS-33 Tenperature
AP RF jfn Trut
2'50
1=
1'30
50
t
1.A
1(
Vertical Cable Ts-33
f _
f50
-
Ov
A4CGc TfS33: Maixirin
surface terrperature
n-I
0
5
10
15
Trrm (nin)
20
25
20,0
600
-.
500
i3
_.
Vertical Cable Ta-33
MGIC
A TS-33: Gas .tenuerature
MA^GICTS-33: Surface
tenperature of the targe
-..Tree 2-5
4C0
ID
.
14,0
12,0
200
100
-
U.
8,0
I
6,0
-
#3.
BE Test 16
0lCR.P
0
2
4
6
Tniri(mrin)
8
I
>S"U'S
>TrkS3
4,0
are 3TTerreratur8ff
TS-3rgtT
2,un
frJs
0.0
10
12
PAGICFr39&10: Totad
heat f lux lfluxn eter'
I~MAG@C
19&10:
tncident heat tlu
Ca
ble Total Flux 9
~-
.v
10,0
k
S
I-
.. I . I
16,0
S
Cable Rad Gauge 10
-----
18,0
-
eaINJ
_
; _
ICFBE#3, Test 16
-
.
0
2
4
6
Tir (nyin)
8
10
12
300
Cable TS-33 Terrperature
250
W BE #3, Test 16
A
200
a
150
E-
I
50
--
Vertical Cabi
-MAGICTS-33: lMaxmrun
0
.
0
5
*.
surface temperature
10
15
Tire (rrin)
20
25
30
Figure A-63: Thermal environment near Cable G, ICFMP BE #3, Tests 13 and 16
A-81
Technical Detailsfor the MAGIC Validation Study
300
Vertica" Cable Ts-33
*
2501
200
MAGIC TS-33: Gas
terrperature
MAGICTS-33: Surface
terrperatureeo the target
Te2A.
.....
-_
150 -_
E
43
=l
.F
_'£-
nn
550-
4,0
=
3.0
I
Target TS-33 TerrperatureCF 8E#3, Test 17
,
§!
2,0
1.0
0
0.0
0
4
2
6
Tirre (rrin)
8
10
12
0
2
4
6
Trre (mnn)
8
120,
1Was
/
A
\
Cabb TS-33 Terrperature
ItwP8E#3, Test 17
--
80-
e
w
_
An -
/
w
/1
40
-
20
W.GCTS-33: hbximm
surtace te.Tperature
.
0
2
4
6
Trr (rrin)
8
.
10
12
Figure A-64: Thermal environment near Cable G, ICFMP BE #3, Test 17.
A-82
10
12
Technical Detailsfor the MAGIC Validation Study
Target TS-33 Terrperature
CA BE c03,Test 3_ fi
200
2.O
5,0
F
150
a
100
x
/
g~
I-
50
_
Vertical Cable Ts-33
a
~
WAGIC TS-33: Gas
I
M ITS-33: Surface
= 1temerature
of the target
Tree
7 2-5
Y
1-
4,0
U.
3,02,0 1,02
o
0,0
5
0
10
15
Tnre (rirn)
25
20
5
10
15
Tim (rrtn)
20
25
2a
f!
I-
0
10
5
15
Tim (rrin)
.4
20
25
30
S
a4:0
I'I
15
10
Tire (mn)
5
0
zooF-
20
25
0
5
10
15
Tir (rnn)
20
25
BEY3, Test 9
200^
2.
E.
a
E
_
100
|-Vertical Cable Ts-33
50 <j
0
I
_ _
M
A
G'C TS-33: llbxrum
surface temrperature
-
0
5
10
15
Tme (rr)
20
25
Figure A-65: Thermal environment near Cable G, ICFMP BE #3, Tests 3 and 9.
A-83
Technical Detailsfor the MAGIC Validation Study
3W0
v
50,
§
4,0
,,
3,0
I
2,0
a
aT
a
1,0
0
5
0
5
10
E
15
Tre (nin)
20
25
30
20
25
30
0
10
5
15
TVM (min)
20
30
25
a
10
15
Time (rrin)
3C0
Target TS-33 Tenperature
PBE 3, Test 14
250
O
200
i
150
....
MAGCTS-33: Gas
.
nterperature
-MMACTS-33:
Surtace
terTrerature of the target
50
,e
8,0
I
a
4.0
-.-
zo
0.0
0
2
a
200
1
150
ISO
100
5
15
10
Trm (rrin)
5
20
25
Cable Rad Gauge 10
-M4GC
FG9&10: Total
heat Iux
kmluxrreter'
-MAGC RG&10: hcidenl
heat flux
-*Cable Total Rux 9
2U0
-ti-Tree 2-5
_^
_N
Vertical Cable Ts-33
-ti-
100
Target TS-33 Fat RuxI
_
I CWBE #3, Test 14
10,0
0
5
10
15
Tnm (nrin)
20
50
0
5
10
15
Tirre (rrn)
20
25
30
Figure A-66: Thermal environment near Cable G, ICFMP BE #3, Tests 5 and 14.
A-84
2
5
Technical Detailsforthe MAGIC Validation Study
7.0
250UI
Target TS-33 Terrperature
ICFPBE#3, Test 15
/;
200
6,0
5,04_
e
150
E
100
-7d
Cable Ts-33
*Vertical
U-
....... MAGITS-33: Gas
i
Cable Red Gauge 10
G
I:
M-GMCTS 33: SurfaceI|
terrerature
2_5 of the target
-w-Tre
I/I
50
0
i
1,0
1,0
20
10
15
TIme (rni)
25
heat flux Iuxrreter'
G CR91
cident heat flux
Cable Total Fux 9
M
. I.of
0,0
0I
5
FAGC FG9&10: Tota
I
2o
5
0
15
10
Tim (rmin)
25
20
0
250
.
I-
5
0
10
15
Tffm (nri)
25
20
30
9,0 -
Cable Rad Gauge 10
-
U
8,0......MAGIC G&10: Total
fluxRt3&10:
lluxrreer'
_heat
7,0
W~GIC
E
6,0 _
a
F
3.
§
8O
1L-
E-
I
hchent heat fhux_
_
,
Cable Total Rux 9
50_K
5,0
4.0
3,0
2.0
Tarot TS-33 Heat Rux
1,0
0.0
0
10
5
Trre (nin)
15
20
25
0
5
10
#3, Test 18
ICW PBE
20
15
25
Tne (rdn)
6
2
E
F°
0
5
10
20
15
Tie (rrme)
25
30
Figure A-67: Thermal environment near Cable G, ICFMP BE #3, Tests 15 and 18.
A-85
Technical Detailsfor the MAGIC Validation Study
Table A-23: Relative differences for surface temperature to cable G
Control
Cable
Test 1
Test 7
Test 2
Test 8
Test 4
Test 10
Test 13
Test 16
Test 17
Test 3
Test 9
Test 5
Test 14
Test 15
Test 18
Target Surface Temp, G- Cable Surface Temp, GTS-33
TS-33
Relativ
Relative
AE (0C) AM (0C) Diff AE (0) AM (0C) e Diff
74%
64
111
114
79%
64
37%
90
107
111
42%
78
72%
184
91%
107
204
107
182
70%
107
90%
203
107
189
51%
61%
125
201
125
25%
148
186
34%
148
198
66%
222
98%
133
264
133
216
27%
44%
169
169
243
169
165
161
270
160
106
206
204
183
230
200
200
22%
23%
14%
-15%
25%
89%
169
165
161
270
160
106
200
198
172
204
197
197
19%
19%
7%
-24%
23%
87%
Table A-24: Relative differences for radiative and total heat flux to cable G
Control
Cable
Test 1
Test 7
Test 2
Test 8
Test 4
Test 10
Test 13
Test 16
Test 17
Test 3
Test 9
Test 5
Test 14
Lest 15
Test 18
A-86
Radiant Heat Flux Gauge
10
AE
AM
Relative
2
2
Diff
(kW/m ) (kW/m )
1.41
-6%
1.51
-39%
5.97
3.64
3.71
-31%
5.36
3.41
-43%
6.00
-42%
5.45
3.16
1.33
-10%
1.47
3.59
-40%
6.03
3.58
-31%
5.15
3.37
-38%
5.42
-30%
7.01
10.06
6.12
-42%
10.50
3.28
-12%
3.73
-51%
5.87
11.96
-11%
2.42
2.15
8%
3.06
2.85
Total Heat Flux, Gauge 9
AE
AM
Relative
2
2
(kW/M ) (kW/m ) Diff
1.89
2.99
56%
5.98
6.42
6.20
12.18
12.23
3.07
6.45
6.37
6.69
10.90
5.12
4.45
6.86
6.58
6.51
11.50
10.05
3.43
6.82
6.65
5.80
8.39
6.36
6.13
15%
2%
5%
-6%
-18%
11%
6%
4%
-13%
-23%
24%
38%
Technical Detailsfor the MAGIC Validation Study
A.8.2 ICFMPBE #4
Targets in BE #4, Test 1 were three material probes made of concrete, aerated concrete and steel.
Sensor M29 represents the aerated concrete material while Sensors M33 and M34 represent the
concrete and steel materials respectively.
MAGIC appears to over predict both total heat flux and surface temperature to the targets. The
graphical comparisons for heat flux and surface temperature and the resulting relative differences
are presented in Figure A-68 and Table A-17respectively.
A-87
Technical Detailsfor the MAGIC Validation Study
a
a
E
0
I
10
20
rne (rrin)
40
30
10
20
Tfme(min)
30
40
30
40
80
a
.a
U..
E
-I
0
20
Tim (rrin)
10
30
40
0
10
20
Tie (rdn)
0
10
20
Tfie (eim)
8C0
700 Steel Pate Surface Terrerature
00 BE 4, Test 1
\
c
a
500 -
/
400
3O
I
200
100
0I
0
.
10
M3~404.Steef:
Surface
Werat're of "h target
20
Tire (rin)
I
40
30
40
30
Figure A-68: Heat Flux and Surface Temperatures of Target Slabs, ICFMP BE #4, Test 1.
Table A-25: Relative differences for surface temperature, and total heat flux to targets
AE
ICFMP 4-1
A-88
( 0C)
teel, M34
356
oncrete, M33
as Concrete,
M29
308
489
Relative
AM (°C)
Diff
684
92%
608
7
J728
AM
27.18
(kw/m2)
75.65
Relative
Diff
178%
97%
46.56
75.65
63%
49%
132.41
75.65
133%
AE
(kW/m2 )
Technical Detailsfor the MAGIC Validation Study
A.8.3 ICFMP BE # 5
A vertical cable tray was positioned near a wall opposite the fire. Heat flux gauges were inserted
in between two bundles of cables, one containing power cables, the other, control. On the
following pages are plots of the gas temperature, heat flux and cable surface temperatures at
three vertical locations along the tray.
Figure A-69 compiles the graphical comparisons for total heat flux and surface temperature.
Table A-26 and Table A-27 list the corresponding relative differences.
A-89
Technical Detailsfor the MAGIC Validation Study
C
4
t
3
Is2
0
5
10
15
Tfne (rnn)
20
25
0
30
5
10
15
20
25
20
25
30
TIM (rrhn)
z
a
U0
i
0
5
10
15
Tme (rrin)
20
25
0
30
5
10
15
Tim (rni)
30
YJ
30 - Cabe Surface Terrperature
- BE 5, Test 4_
60
£
a
.
I,
ic
£
e'
Fe
60
r
40
0
|
0
5
10
15
Tre (nin)
20
25
30
_-
3-7-F
MAGIC TCO1-7
TMO 1-7
TOO 3-7
O1
0
;
K
--T
.
5
'I ....
10
15
Tim (nin)
TM
, 3,
20
25
Figure A-69: Thermal environment near Vertical Cable Tray, ICFMP BE #5, Test 4.
Table A-26: Relative differences for surface temperature
Surface
Temperature Instrument
.
TCO 1-3
TCO 3-3
IP
TCO 1-5
ICF
A-90
5
TCO 3-5
TCO
1-7
TCO
3-7
AE (°C)
141.1
144.3
147.8 222.5
182.6
180.2
AM (0C)
165.1
165.2
159.2
159.3
157.6
157.7
Relative
Difference
17%
14%
8%
-28%
-14%
-13%
30
Technical Detailsfor the MAGIC Validation Study
Table A-27: Relative differences for total heat flux
Total Heat
AE
Relative
(kW/m 2 )
(kW/m 2 )
Diff
Flux
Instrument
ICFMP 5-4
WS2
WS3
141
144
161
174
14%
21%
WS4
148
158
7%
A-91
Technical Detailsfor the MAGIC Validation Study
A.9 Compartment Wall Temperature and Heat Flux
Heat fluxes and surfaces temperatures at compartment walls, floor and ceiling are available from
ICFMP BE #3 and #5. This category is similar to that of the previous section, Heat Flux and
Surface Temperature of Targets, only here the focus is on compartment walls, ceiling and floors.
MAGIC offers two alternatives for wall temperature and heat flux results. The first alternative
results from a heat balance at the surface of the wall in both the upper and lower layer. The
second option is to locate a target characterized by the wall properties. The second option is
preferred for validation purposes since the target can be placed in the same location as the
experimental sensors.
A.9. 1 ICFMP BE #3
Thirty-six heat flux gauges were positioned at various locations on all four walls of the
compartment, plus the ceiling and floor. Comparisons between measured and predicted heat
fluxes and surface temperatures are shown on the following pages for a selected number of
locations. Over half of the measurement points were in roughly the same relative location to the
fire and hence the measurements and predictions were similar. For this reason, data for the east
and north walls are shown because the data from the south and west walls are comparable. Data
from the south wall is used in cases where the corresponding instrument on the north wall failed,
or in cases where the fire was positioned close to the south wall.
The heat flux gauges used on the compartment walls measured the net, not total, heat flux. In
MAGIC, this measured heat flux is compared with the Target/Heat Flux/ Total Absorbed Heat
Flux output option.
The following graphical comparisons are grouped per room surface (long wall, short wall,
ceiling or floor). The term long wall refers to either the north or south wall. The term short wall
refers to either the east or west wall. Two sensors have been selected for comparison for each
surface. Comparisons include both surface temperature and heat flux. The corresponding
relative differences are provided after the graphical comparison for each room surface.
A-92
Technical Detailsfor the MAGIC Validation Study
250
200
-
TC South U-4-2
-
TC
tNorth U-1-2
4.0 - North Wall Heat Flux
35-ICFMiP BE #3, Test 1
*
MAGIC South-U4 Surface
temp1erature of the target
2 150
100
-
2
2,0
_
s
2S
2,5
terperature of the target
I
0MAGIC
0
10
heat fluxTotal
absorbedrSouth-LA4
0.0
a0
20
MAGIC
b
erth-de flu
absorberJ heat f lux
.......
n 0;
North Wal Terrperature
;FW BE #3, Test 1
0
North U1
i
1
50
South UJ-4
3,0
10
0
Trne (rrin)
20
3tD
Tvre (rfin)
3.0
South U-4
2,5 -
6
2,0 -
I
1.5
MAGIC South-U4 Total
.
U-
e:
0.5
10
Tire (nin)
kCrFwBE#3. Test 7
0
30
20
|-350-f4
30- ---
1
TC South U-4-2
5 0 11
I
TO North U-1-2
250- 20
-Sn
200-
@-
||
F
300 --
2
10
Ttre (nin)
7,0
40U -
South U-4
MAGIC North-Ul
-
Total
absorbed heat flux
3,0
,
I
3C
MAGIC South-U4: Total
0
MAGICtNorth-Ue
ere
tSurace
t
terrperature of the target
___
_
20
North U-I
5.0.j
AGIC South-U4 Surface
tenperature of the target
absorbed heat flux
MAGCNorth-UI Total
absorbed heat flux
North Wall Heat Flux
e
nn~
0
I
-w--Nbrth U-I
\1
PAI
100
N hP Wall#,Teat Flux
1,0 -
50
o
0,0.
10
5
0
0
5
.n
-
TC North U-4-2
350
6
I
300
-W--TCNorth U1-2
250
200
-
1SO
MAGIC NMrth-U4 Surface
terrperatureof the target
North-U
-MAGIC Surfac
-....
tenperatureof
11j-
North U-i
11
MAGIC NorthU4A Total
absorbed heat flux
North-U1 Total
absorbed
4.0
3,0
I-
201
100 -
.
00119="A
I
1,0]
ICFWP BE #3, Test 8
uu
0
10
Toe (inn)
Tne (min)
404,
'K
IIABE#,Teat 2
tFM' BE #3, Test 2
5
Tte (rrin)
10
-
.MAGIC
..
K
,]
- --
~
~
North Wall Heat Flux
S
Test8
ICFLPRlE
-
Qk
I
I
10
0
Tirme(rrin)
Figure A-70: Long wall heat flux and surface temperature, ICFMP BE #3,closed door tests.
A-93
Technical Detailsfor the MAGIC Validation Study
400
TC North U-4-2
350
250
ni
200
5,0 -
TC North U-1-2
-
300
j
.
lRfl
,u,
-
North-U4: Surface
terrperature of the target
MAGIC North-U1 Srace
terrperature of the target
-MAGIC
I
4,0
i
3,0-
!
_
Y
North U-1
-
2,0
~T
5:
*
I
CL150
100
North U-4
_
i
?
-_ *
A
1.0
_
North Wall Heat Rux
FCFvP BE #3, Test 4
le
ICFvPBE#3, Test 4
MAGIC North-K4: Total
absorbed heat flux
I MAGIC North-Ul: Total
absorbed heat flux
0,0
0
5
5
0
-,
N
.
.
15
10
10
15
Tirme (rrin)
Tirre (rin)
TC North W4-2
250 200 -
a,
C
a,
E-
•-..
TC North U-1-2
WAGIC
North-U4: Surface
ternrerature of the target
MAGIC North-Ut: Surface
terrperature of the target
15t
I
_
inn
50 -
NTrature
StOU=PBEs73,Test 10
0
10
5
0
15
5
10
15
Trne (rrin)
Tre (nin)
10.0
T
9,0
8,0+
7.0-
t'7,0+-
a
3
I,
2E,
- MAGIC North-U3: Total
6,0-
2t
absorbed heat flux
- MAGC WesT-U1 :Total
4,0 +
1
absorbed heat flux
3'0t-
1t
-North Walt F-eat Flux -ICW BE#3, Test 13-
2,0
1.0
0,0
5
500- I
-|*
450
400
6
15
10
Thne (rrin)
|
350 -
-
Ei 250 |4
p 200 1Ign
100 I
5n
I
0
-K
2
MAGIC Norh-U: Surface
terrperature of the target
MGIC North-U1 ....
:ESurface
:ft
fure of the target
8
15
20
25
tr
|
7,0 6,0 -
-
North U-1
MAGC North-U4: Total
absorbed heat flux
MAG- WesT-U1: Total
sobdheat flux
5,0
12
0,0.
0
I
I
3,0
2.01
1,0
0
10
North U1-4
-
8.0 |
TCNorth U-1-2
6
Tre (rnn)
10
9.0
Nr
PNth Wa lnbr
-W BE #3. Test 16
4
5
Tmef(min)
IX
0!-,;>O
_
0
25
TC North U-4-2
--
* 300 -
20
/1
1/
2
North Wallt eat Rux
k
_4
IF t J
#3,
Test 16
-
6
8
10
12
Tmre (Crnn)
Figure A-71: Long wall heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-94
-
Technical Detailsfor the MAGIC Validation Study
5..0
,5
4.,
North U-4
-*.-
4.,,0 -*-1
i
!
~-M
, _
3.,
5
2. 5
.
.
w
0.5
0.0
4
6
Trme (rrin)
8
10
absorbed heat flux
MAGIC North-U1 Total
absorbed heat flux
.....
__
_ -
I
-
2
MAGIC North-U4: Total
2'
I1.
,u
0
North U-t
ls_
_-
0
0
12
hNorth Waft Fleat Flux
BE #3 Teat 17
10
5
Tarn (inn)
Open ]Door Tests
400
.-. *
TC North W-4-2
300
-
TC Norttl U-1-2
U
250
-
j
200
50.,
North U-4
350
MAGIC North
: Surface
tenrprrature of the target
.. MAGIC Norh
t
the targe
3 0 .
....
2,0
'
hrw
fh
1,0
t /North
0
CFP BE #3, Test 3
r
-.
n
5
10
15
Twrn (nin)
_
Wat Hea Flu
KIFMP BE #3, Test 3
O
0
absorbed heat flux
II
~North Wall TenfperatureI
50
MAGIC North-U.4: Total
-
..*.*.... MAGIC North-Utl: Total
inn
f
North UL1
4.0 - -
20
0
25
5
10
15
Tine (rin)
5,0
.
4.5
3,0
.......
North U-4
MAGIC North-U4: Total
absorbed heat flux
MAGIC North-U1 Total
absorbed heat flux
2.5
U-
25
tNorth U-1
.
4,0
3.5
U
20
2,0
I!
North Wal Heat Rux
KTW BE #3, Test 9
0,5
50
0
0
5
10
15
Tom (rrin)
20
25
0
5
10
15
Tkre (rrn)
20
25
Figure A-72: Long wall heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-95
Technical Detailsfor the MAGIC Validation Study
5,0
*
North UL4
4,5
4
4.0
3,5.
W
3,0
a.
.------
2,5
IT
MAGIC North- L4: Total
absorbed heat flux
MAGIC North-U1 :Total
absorbed heat flux
-
0
20
10
30
0,5 -
North Wall Heat Flux
ICFP BE #3, Test 5
0,0 I r
0
10
Tffe (nir)
500
_
_
450
_
_
6,0
_
.
3CI
Norrt U-4
e
5,0 -
4C00-
TC North
350 30 -
MAGIC Nwrth-U4: Surface
U-1-2
North U-1
4,0 -
temperature of the target
MAGICNo
the target
200 -eof
20
Twre (n-fn)
TC North U-4-2
20
250 -__
l'
North U- I
T!!_
3.0i
'
MAGIC North-3: Total
absorbed heat flux
*------ MAGIC WesT-Ul :Total
_bsorbed
heat f lux
At
KNrorth
2,0 -
150
100
North Wal Terrperature
LI-w t Yu, test 14
so
AI
0,0
0
5
10
15
Trna (rrh)
20
0
255
wal Heat Fux1
I C vIBE #3, Test 14
10
15
20
25
Twm (,Ain)
5
8.0
450 -tW
400 -e
a
4.0I
TC South U-1-2=
South U-3
I
TCtSouth U-3-2
I
sSouth
350
X
5C0
3C0I _GISouth-Ul : Surface
30-tesfpreature, of the target
250 ---------. MAGIC South-U.3: Surfam_
200 -
so
r
t
50p
s
_
T
H
7,0
42.0
6, | tMAGIC SouthU1.2: Tota
r5.0
tabsoraedheatfaux
z;0--
[
4,0;
s
U- I
n
MAGIC North-U6 :ToWt
absorbed heat f tux
:<
||
e
Nlsrth Wall Terrperature
MP BE #3, Test 15
1.0 >
ICvPBE#3.Test 15
0,0
5
0
10
15
20
0
25
5
10
15
Twre (nin)
Tne (mn)
20
25
6,01
-
5,0 1
i
South U- I
-"---South U-4
4,0
*
3.0-
MAGICSouth-Ul :Total
absorbed heat flux
AGIC SouthU4: Total
absorbed heat flux
-M
.4.-
I zo-
L
1,0
1:s 1
0
5
10
Two (nin)
15
20
25
North Wall Heat Flux 1
ICFRP BE #3, Test 18
-/'
oo
2
0
5
10
15
Tom (nin)
20
25
Figure A-73: Long wall heat flux and surface temperature, ICFMP BE #3, open door tests.
A-96
Technical Detailsfor the MAGIC Validation Study
Table A-28: Relative differences for temperature and total heat flux corresponding to
the long wall
Wall Total Heat Flux
Wall Tem erature
AM Relative
Relative AE
Instrumen
Long
2
0
AE ( C) AM (°C) Diff (kW/m ) (kW/m2) Diff
all
49%
1.1
0.7
65%
89
54
Ni
25%
1.2
1.0
31%
89
68
S4
est 1
29%
1.0
0.8
63%
87
53
Ni
-5%
1.1
1.1
23%
87
70
S4
Test 7
17%
2.8
2.4
65%
158
96
N1
-1%
2.8
2.8
32%
158
120
S4
Test 2
12%
2.8
2.5
66%
157
95
N1
83%
2.8
1.5
19%
157
120
N4
Test 8
20%
2.4
2.0
64%
159
97
Ni
59%
2.3
1.5
9%
1519
124
N4
Test 4
20%
2.4
2.0
68%
158
94
N1
53%
2.3
1.5
-3%
158
124
N4
Test 10
91%
209
110
Ni
5%
210
151
N4
Test 13
83%
195
107
N1
-10%
196
150
N4
Test 16
47%
2.1
1.5
79%
7()
39
Ni
125%
2.1
0.9
-13%
71
54
N4
Test 17
18%
2.0
1.7
47%
168
114
N1
40%
2.0
1.4
-1%
170
169
N4
Test 3
17%
2.0
1.7
47%
165
113
N1
Test 9
Test 5
Test 14
Test 15
est 18
N4
N1
N4
N1
N4
S1
S3
Si
S4
167
94
150
114
146
115
308
109
312
168
143
147
167
176
167
182
166
191
-6%
1.4
2.0
40%
52%
-5%
46%
-31%
34%
-17%
41%
1.5
1.7
1.8
2.7
1.9
5.9
1.8
1.8
2.0
2.0
2.6
2.0
3.1
2.0
20%
20%
16%
-4%
9%
-46%
14%
-39%
A-97
Technical Detailsfor the MAGIC Validation Study
250
I
TC East U-1-2
\ _N--
200
.*MAGC EAST-U1: Surface
terperature of the target
MAGIC EAST-U2: Surface
terrperature of the targe
., .-
a* t50
TC East
U-2-2
-
150
0
East Wal Terrperature
tCrbP BE #3, Test 1
-
3t
20
10
o
10
0
20
30
TurI (nln)
Time (min)
c-
250
East U-3
TC East U-2-2_
200
2.0
TC;East UJ3-2
-
*------ MAGIC EAST-U)3: Surface
_terroerature of the target
tsAGtC EAST-U.2: Surface
_
terrperature oftetr_
2 150
-1-.
I
I 1.0
I
0.5
~
H-eat Fu
East Wagn
CRAP BE #3,Tet
BE #3, Test 7
0,0
10
5
-M
MAGIC EAST-tJ2: Total
absorbed heat ftux
MAGtC EAST-U3: Total
absorbed heat ftux
East Wall Temperature
ItCvF
0
East U-2
s-
1.t5
| 100
S!
-
15
TIM (nin)
20
25
-
0
3C
20
10
D
Trme (nin)
6,0
350
300
_
IC East U-2-2
-|
300
TC East U- 1-2
--
2. 250
EAST-Ul:
-MAGIC
-Surface
terrerature od the target
t4GtCAEAST-U2: Surface
1tenperature of the target
200
I-
East U-1
East tJ-2
5,0 -
4,0-.........
.
MAGIC EAST-12: Total
absorbed heat flux
MAG-C EAST-UI :Total
absorbed heat flux
-
3,0-
150
100
50 -
_0
,
g
'
1 0
0.0
East Watt Terrperature
ICFMPBE #3, Test 2
,0
- 7
East Watt Hreat ti-5u
[
Tes2_.
tCt;MBE#3,
0
0,0
2
0
4
6
Tore (min)
8
10
12
5
Time (rin)
0
Rn,
vv
400-
East U-1|
TIC East UI-2-2=
30-
5,0 i
250 -. .
.MAGIC EAST-Ul :Surface
teriperature of the target
tAST-U2: Surface
EC
-ilAG
150
temperatu
3.0
,2
I
of re
East Wag Tenperature
.CFWP BE#3, Test 8
500
5
Thim (mni)
absorbed heat f tux
MAGIC EAST-Ut:Total
absorbed heal
4-
I
100-
0
East U2|
TICEast UJ-1-2
300 - -
200 - -
10
10
East WaHtIeat
1.E
tW
lux
BE#3, Test 8
nfl
vV
0
2
4
6
rem (min)
8
10
12
Figure A-74: Short wall heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-98
Technical Detailsfor the MAGIC Validation Study
5,0
400-
4.0
300°
4
East U- 1
nf-
East U-2
4,5
350-
-i
'3,5
250-
3,0
i
MAGIC EAST-U2: Total
absorbed heat flux
MAGIC EAST-UI: Total
abs
_
50
-
2,0
1,5
iso
100-
1,0
East Wall Heat Flux
ICFW BE #3, Test 4
0,5
0,0
5
10
15
0
rTue (min)
5,.,0
20
ht2-2
te*TCEast
!250 ........
' j
100 -1
200
e
.
.
.
East U-3
-4---
4,5
350
6'
East U-2
||
TC East U'3-2
4,0 -
-
II
MAGICEAST-U3: Surface
eraturedo the tarceI
AIEAST-U2: Sur..r
terrperature of the targe
3,5- 1-
-MAGIC
EAST-2: Total
absorbed heat flux
EAST-U3: Total
absor
atflux
~
15
10
5
Tre (rrin)
3,0
$
2.52.0
' 50 -
;
-MAGIC
y
-
-
-
,
1.5
|1,OD
J
50 -
1,0
East Tenperature
-
1,fi
0,0
0
10
5
15
East, -e0atRux
ICRAP BE #3, Test 10
I
0.5
ICFBE#3, Test 10
Turn (rrin)
Thue (rrn)
4501
i
-
|
350 6
12.0
,
400.
TC East UJ-4-2
|
110,0
TC East U-1-2
300.
8.0
EAST-Ul:
.MAGIC Surface ,
terrperature of the target L
j 250 .
T A1MAGIC EAST-U4: Surface
tenperature of the target
p200 150 f
1
100
:._
_
0
5
10
15
I
20
East U- I
-i|--
East ULI-4
MAGIC CELtGC-5: Total
absorbed heat flux
MAGIC EAST-U2: Total
absorbed heat flux
.
6,0 -
_
u I
-4-
I-I
-
_,fEast Wall Te,~er.-e"INIIIIIIII
M-'BE #3, Test 13
50
15
10
5
0
1
4,0
2I0
I
2.0
I
0.0 .- -
3
2E
0
East Wat HIeat Flux
B3N1'aE#3,
Test 13-
I
10
15
Tkme (nin)
5
Thu (rrxn
20
25
SCO
4500
450
a
TC East U-2-2
-.-.-
~
400 |
-(
|
C East U-1-2
H
3501
LO
MAGIC
-CO
EAST-U1 : Surface
terrperature of the target
300
250
MAGIC EAST-12: Surface
-
200
re of the tar
F- 150
I
£,-'
1001
X,00WEast Wall TermertueraiJ"
WR
0
-
IC
BE#3, Test 16
5
Tre (rrin)
10
0
2
4
6
Trhu (rin)
8
10
12
Figure A-75: Short wall heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-99
Technical Detailsfor the MAGIC Validation Study
OU -
a
1,5
MAGIC EAST-.3: Total
absorbed heat flux
MAGIC EAST-W2: Total
absorbed heat flux
....
2,5 20 I
I-
East U-3
a
3.5 3.0 -
.
-
-
Ad
East Wag -bat Flux
iAf
1,0
U-2
-East
4
4,5
4.0 -
hIFW
BE #3, Test 17 _
0,5
0
5
0.01~
0
10
10
5
rTre (rnn)
Tine (rrfn)
Open Door Test
6'
0
0
5
*
350
e
25
0
5
10
15
20
25
TRm (mu)
TC East Uf-3-2
M LAG IC EA ST- -:3Su rfa ce
terrperature of the target
MAGIC EAST-U4: Surface
tewperature of the tare_
*.I* *
250
20
TC East U4-2
_
300
6'
10
15
Tire (ffn)
II 200
I
I.
150 I
1000
so I
Eat
aWTrrerature
K3FMBE #3. Test 9
I
O
0
5
15
10
Ture (rnm)
20
25
0
5
10
TD
15
20
25
(rdn)
Figure A-76: Short wall heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-100
Technical Detailsfor the MAGIC Validation Study
400
400 |
TC East U-2-2
3.5
350
X*-
TC East U-1-2
3,0
300
MAGI EAST-U1: Surface
terrperature of the target
EAST-U2: Surface
terrperature of the target
.......
250
I
;2,5
eMMAGC
200
i 2,0
1.5
150
1.0
100
5t Wa1 TerrleraturTW
ICFW' BE #3, Test 5
0
1
0,5
0,0
0
10
-
30
20
0
20
10
Tire (rrin)
30
Tekn (nin)
4004,5
-
East U-2
350 4.0
East
-
U-3
3006
3.5 . *
3,0
250I
200-
MAGICFLOOFPUJ8:Total
absorbed heat flux
MAGIC EAST-U3: Total
absorbed heat flux
2,5
MM:
2,0-
1.5
100Walt Peat Fu
05East
50
0,5
Test14
rEFUPBE#3
0,0
0
5
15
10
Tine (rrin
20
25
5
0
10
15
Tre (rrin)
20
25
5.0
4,5
7
4.0
3,5
I~
1
3.0
^
2,5-
!
2,0-
1,5-
0
5
15
10
Tkne (rrin
20
25
400
4,
2
0
5
15
10
Trre (rrin)
20
25
0
5
10
15
Ture (rrin)
20
25
Figure A-77: Short wall heat flux and surface temperature, ICFMP BE #3, open door tests.
A-101
Technical Detailsfor the MAGIC Validation Study
Table A-29: Relative differences for temperature and total heat flux corresponding to
the short wall
hort
all
est 1
est 7
est 2
Test 8
Test 4
est1O
I
El
E2
E3
E2
El
E2
El
E2
El
E2
E3
E2
El
est 13
Test 16
est 17
est 3
Test 9
est 5
Test 14
Test 15
Test 18
A-102
E4
El
E2
E3
E2
El
E2
E3
E4
El
E2
E3
E2
E3
E2
E3
E2
Wall Total Heat Flux
Wall Tempera ture
AM Relative
Relative AE
2
Diff (kW/M ) (kW/m2) Diff
(°C) AM (°C)
Amen
AE
64%
1.1
0.7
61%
89
55
17%
1.0
0.9
28%
91
71
77%
1.2
0.7
59%
87
53
6%
1.0
1.0
26%
88
70
20%
2.8
2.3
44%
158
110
-2%
2.8
2.9
29%
162
125
12%
2.8
2.5
44%
157
109
-3%
2.8
2.9
29%
160
125
26%
2.4
1.9
50%
159
106
12%
2.4
2.2
34%
162
121
36%
2.4
1.8
51%
158
102
14%
2.4
2.1
38%
161
117
65%
208
127
.
290%
214
55
58%
195
123
42%
201
141
40%
2.2
1.6
22%
69
52
14%
2.2
1.9
21%
74
61
27%
2.1
1.6
92%
168
87
1%
2.1
2.0
16%
170
146
49%
2.1
1.4
90%
165
83
_124%
167
75
80%
2.2
1.2
99%
142
71
4%
1.8
1.7
23%
144
118
70%
2.1
1.2
94%
167
90
-2%
2.0
2.1
14%
169
148
46%
2.1
1.4
73%
167
84
-7%
2.0
2.2
11%
168
151
58%
2.1
1.3
76%
166
87
-9%
2.0
2.2
9%
167
153
Technical Detailsforthe MAGIC Validation Study
3,5
350, -
*-
Ceiling U-4
C
TCCeiling U1-1-2
3.0
300250-
,
200-
TC Ceiling C-4-2
-
-
5
--- *-
Ceiling U-1
2.5
150 -
MAGIC CELNGC-1: Surface
terperapure of
*MAGiCF4:Surface 1j
kt
,_u _'Ure of the target
--
100
2.0
50-
1 n-
_
MAGIC CEI G(- I :Total
bsorbed heat flux
KCFP BE #3. Test 1
00
0
313
20
10
...
Po
CeUiingHeat Flux
Ceiling Tenrperature
ICFP BE #3, Test I
0
CEZNGC-4. Total
heat flux
.2 1.5
U.
_
o'
2-0-GIi
_absorbed
15
10
5
Tirre (rrin)
|
_
30
25
20
Tire (min)
25.-
400
.350
*
Ceing U-4
*
TC CeidingU-I-2
2,0 -
TC Ceiling C-4-2
"
-
Celiing U-I
300
9-
MAGICCEItlIGC1: Surface
ternperature of the target
rf ace
IAG -
250
I.
200 -.
I
1.5 -
iL
1,0-
150
ISO0
MAGIC CELNG3-4: Total
absorbed heat flux
MAGIC CELNGC-1 :Total
absorbed heat flux
AS
------
_i
100-
50~f'
B#f3.-
Ceiing I-leat Flux
0W BE #3. Test 7
/i
Test7
0,0
2.
E
C
10
0
3CI
20
500
_
_
*
-
450o
400
-
350
-AGIC
Tim (min)
_
_
_
10,0 1
9,0*
TC Ce4ing C-4-2
8.0-
|
"
Ceiing 1I
7,06,0
|
-
AGIC CEJGC-4: Total
absorbed heat flux
MAGICCE1LlC-1 :Total
[
absorbed heat fkv*
CEL
rfac
v
r
e
.$
****^;IZCCn tGC4:Su ce
terpe rature of the target
250
5,06
E
200
I-
Te rrperature
;_S9~Ceiling
100
0
I
0
0
2
2
4
4
8
6
Tire (rrin)
6
Tim (nin)
8
10
10
-M-----
f~tf
7 7
1,0
BE e3 Test 2
50
40
I
3,0
2,0
150
100
Ceiling U-4
-
TC Ceiling UI-1-2
tc
300
30
20
10
0
Trme (min)
12
12
0
n
Ceiin Yeat Flux
1CFMP BE#3 Test 2
z
.
nrz. F0
5
Tim (rnin)
2
4
6
Timr (rrin)
10
8
10
12
Figure A-78: Ceiling heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-103
Technical Detailsfor the MAGIC Validation Study
400W
-
TC Ceiling Ut-1-2
-
TC Ceiing C-4-2
350
300
0
250
MAGIC CELNGr-1 Surface
terreratureofthe taret
200
terr
per a
40
L.
I
3n0
Ceoling U-1
-0--
5.0 -_
'
Celing U-4
-
~
6,0
-_
-
MAGIC CELNG(-4 Tota
absorbed heat flux
MAGIC CELWIC, 1: Tota
absorbed heat flux
I .....
tu
Ia
100
50so
n
Celing Temperature
D
ICF,
M0 W
Ceitung Heat Flux
lr'
ePC
BE# Tes A
BE #3, Test 4
0.0
5
0
10
15
w
I
I
---
5
0
15
10
Tifre (minn)
Tire (nnin)
400
3TCCeiling U-1-2
300
s
errerature
250
Surface
.*-MAGICCEU'3G-4:
strtace
U
rerature of the target
ueiling G4-2
v
200
am5
u.
150
I-
so
1'00
_Gling
Terrperature
50 31BE#3, Test 10
5
0
15
10
5
0
Tmre (rrin)
l
-
14,0
TC Celing C-5-2
e
500-
12.0
e
F
2.
FTC
/
Celing C-7-2
MAGICCEILNGC-5: Surface
terrperature of the target
-MAGIC CELNGt>1: Toztal
300 i
300 7 3i-
f
200
|-uCeiing
U-5
10.0
400-
11
absorbed heat f lux
-
E
U
8,0
i
77
6.0
2.0 -
Ceiling Termperature
ICF
FN B3E#*3Test 13
0
25
I
400 -
10
15
Tire (nnin)
|
-
Celing U-5
MAGIC CELNG1>7: Total
6.0
absorbed heat flux
MAGIC CBE'LJGC-5: Total
absorbed heat fhux
4.0
I
200-
1.0
0,0
2
4
6
Trre (nnin)
8
10
12
.
Ceiling Heat Flux
3,0
|
__7
2,0
00
I
25
7.0t
5.0t
1 300-
.
20
Ceiling U-7
-.---
9.0
600
0
BE O3.Test 13
5
8.0 |
.
Ceoing IHat Rux
/CFW
0
700-
i
|
0,0
20
15
10
Tme (min)
5
MAGIC CE1LNtC-7: Total
absorbed heat flux
CELN3C-7:
/
4,U
100
15
10
Twre (rnin)
tfX
Iz
__.
A_
IeY
0
IICWPBE #3. Test 16
2
4
6
Tire (nnin)
8
10
12
Figure A-79: Ceiling heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-104
Technical Detailsfor the MAGIC Validation Study
5,0
400
TC Ceding C-4-2
-
'
250
AGCCELNG~C-4: Surface
of the target
* GCCELNG~t:3C1
Surface
'erture ot the target
150
I
-
3,5
1
3,0
I;
J
I-^
;1
i 2,0
1.5
1
I
-1
..
_
,f
fI
1,0
Ceiling U-4
-
MAGICCEILNGI> 1:
Tota absorbed heat flux
MAGIC CEIMIC-G4:
-
Total absorbed heat flux.
-
fnperature
200
i
4,0
TCCeing U-1-2
-
300
a
Ceiling BL1
4.5
350
|
Cling Hat Rlux
Ii
ICFW Br #3, Test 17 _
U0,5
o I
0
Ceiling Tenperature
lCRFP BE #3. Test 17
v
0,0
10
5
5
Tirne (min)
0
Tvre (nin)
.
10
Open Door Tests to Follow
7.0
BOO,
TC Ceding
500 -
,
'1--2
6.0
TC Ceiling C4-2
5,0
a
Surface
400 -W.CAGCCELN301:
terrperature of the target
CELMC4: Surface .............. ,"
3D... 1. . I W MGIC
300-
!
4,0
2
3,0
:
2.0
I
200
Terrperature CRA BE#3, Test 3
{Ceilng
100
0
10
15
Tire (nin)
5
5
0
25
20
10
15
20
25
Tire (mrin)
5,0
4,5
4,0
Ar 3,5
absorbed heat flux
25
1502,0
100
-
50
......
1.
2,
1,0
0
MAGC CELI4GC-2 Surface
terrperature of the target
MAGC CELFJC4: Surface
oEhetage
Ceding Hoat Rux
t
0.5praur
lF
BE #3, Teat 9
0,0
0
0
5
15
10
Tore (rnii)
20
25
0
5
15
10
Tere (Amin)
20
25
Figure A-80: Ceiling heat flux and surface temperature, ICFMP BE #3, open door tests.
A-105
Technical Detailsfor the MAGIC Validation Study
500
-
450 400 -
TC Ceing U-1-2
5,0
TIC Ceiling C-5-2
4,5
4,0 -
a
*
300 ........
9
-.
200 -
MAGIC CELNG-1: Surface
terrperature of the target
MAGIC CELNGIC-5: Surface
-
2.0
;
1,5
1,0
0,5
Ceiling Tenperature
BE
#3,Test 5
SO~ICF1
0
350-
CeiTingTenrperature
ICF1 BE#3, Test 14
30
_
_
MAGIC CELI.GC-5:
Total absorbed heat flux
---- MAGIC CELNGC-1:
Total absorbed heat flux
09101000"mgmwLl --------I
Ej130
20
0
10
Celing eat Flux
ICFI BE #3. Test 5
Trne (rrin)
400,
Ceiling U-1
-
0,0
20
10
-
30 4
150sojF
Ceiling U-5
W~ 3,5
250-
100 -
.-
Tr.e (nin)
_
W
300
6
250
;
200-
t4
Sa
9
I
Ceiling
r-5-2
e.
C
5
Surface
CAGIC
terperature of the target
10 t
100
-
50
TC
--
-iTC
I
0
0
5
MAGIC CELWGGC-1:Surface
teperature of the target
15
10
ITi.e (mrin)
0
25
20
350
2,0
Ceing Tenperature
PBE #3, Test 15
300- t
1,8
250 -_
_
Mef!~
_
1
0 .
O
.
.
5
15
10
Thu (rrin)
0.2
Ceiling U-4
I
MAGIC South-U6: Total
j
-
absorbed heat flux Totall
MAGICELGC-3
absorbed heat flux
04
_
0,0
.
.
-z
0,6
AGICCELING-4: Surface
tenperature of the target
------- MAGIC CEL'AGC-1: Surface
terrperature of the target
i
P
5100i
Cetig=U-i
10
8
150l-_
_
I\
12
I
ICWBE#l3,Tes 15
f
I
r1.4-
20u
w _o
25
20
___
1,6
0
15
10
Time (rnin)
5
20
0
25
5
10
Tro
15
20
25
(rrin)
7.0
|500-
S 300
6.0
1-e--T Ceilng U-1-2
-
6 400
TICCeling C-4-2
-510
MAGIC CELaGC-4: Surface
tenrerature of the target
I MAGIC CEINGC-1: .Surface
terrperature of the target
94
1'
200 -
a 3.0
I
100I
,0
4.0-
t
I
2.0
Ceiling Terrperature
ICFW BE #3, Test 18I
0
0
5
15
10
TMb (nrn)
20
25
0
5
15
10
Trte (rrin)
20
25
Figure A-81: Ceiling heat flux and surface temperature, ICFMP BE #3, open door tests.
A-106
Technical Detailsfor the MAGIC Validation Study
Table A-30: Relative differences for temperature and total heat flux corresponding to
the ceiling
Ceiling Total H at Flux
Ternpe ature
AM Relativ
Relative AE
Instrumen
2
AE (0C) AM (°C) Diff (kW/m )(kW/m 2 ) Dff
Ceiling
3%
1.0
1.0
10%
89
81
C1
-17%
1.1
1.3
-49%
89
176
C4
lest 1
-1%
1.0
1.0
8%
86
80
C1
-55%
87
191
C4
Test7
-21%
2.8
3.5
7%
158
148
C1
-42%
2.8
4.8
-49%
159
308
C4
Test 2
-28%
2.8
3.9
6%
157
148
C1
-50%
2.7
5.5
-52%
157
325
C4
Test 8
-19%
2.4
2.9
8%
158
147
C1
-52%
2.3
4.9
-12%
159
180
C4
Test 4
-10%
2.3
2.6
14%
158
138
C1
-29%
158
221
C4
est 10
209 501%
35
C7
-58%
210
500
C5
Test13
15%
196
171
C7
-53%
197
419
C5
Test 16
4%
71
69
C1
-69%
71
230
C4
Test 17
-9%
2.0
2.2
8%
167
155
C1
-36%
2.8
4.5
-37%
1131
287
C4
Test 3
-2%
2.0
2.0
1166 260%
46
C2
-32%
2.8
4.1
-38%
1-79
290
C4
Test 9
-13%
1.8
2.0
13%
142
125
C1
-61%
1.8
4.7
-12%
146
166
C5
Test 5
-11%
2.0
2.2
5%
166
158
C1
12%
277
248
C5
est 14
6%
166
157
C1
-33%
192
287
C4
Test 15
-16%
2.0
2.3
14%
165
145
C1
-33%__
167
250
04
est 18
I_Ceilin,
A-107
Technical Detailsfor the MAGIC Validation Study
4,0
TC Floor U-4-2
-
250-
TC Floor U-1-2
-_-
-
Floor Heat flux
3,5 --KFl BE #3, Test1
Foor U-I
--
Roor U-4
MAGIC FLOOR-U1: Tota
absorbed heat flux
Tota
MAGIC FLOORC)U4:
3.0
U
200 -
a
150-
,
100
FLOORU4: Surface
_-MAGIC
termperature of the target
MAGIC FLOOR-U1: Surface.
-.
terrperature of the target
-
.......
3! .
on .
Bzu
absorbed heat flux
'.
.
~1,5
O'
50
0.0
0
10
20
3aI
Time (min)
300
CFU BE #3. Test 7
- *Floor
0
IF 200
MAGIC FLOOR-U4: Surface
terrperature of the target
150 -....... *.MAGICFLOORU1:Surface
tenperature of the target
inn
I.
13.0
25
30
U-1
Floor U-4
-
FLOORUI: Total
absorbed heatflux
MAGC FLOORU4: Total
12.0
/
50
20
15
Tame(mrn)
Rrvor Hleat Rlux
TC Floor U-1-2
-e
10
5.0
TC Floor U-4-2
250
5
0
_ u
_Floor Terperature
10
0
0
20
I
BE #3. Test 7
3B3
I- amI
10
0
Tne (inn)
Tie (mrin)
20
hattFlux
-abForbed
|*
TC
350
12,0-
Floor U4-2
.FRoor
2.0
_
TC Floor U-1-2
|-
MAWFLtGFIOR-U4: Surface
terrperature of the target
MAGIC FLOORUI :Surface
30
Li-1I
-
oor UI-4
10 0 -Fl
300
250
0
0.0
8,0 -MAGIC
S
<
~ ~
-
~
s
so
3!6 0 -
4A
100
terp__et
-
I4,0
lu
.
NMaCrLOOS4:
RLOORUI: Total
absorbed heat flux
Total
at liux
Fl.oor Heat FRux :
(ICFMPBE
#3 Test
I
BE #3. Test 2
ICF
I I
0
2
4
6
Tme (rin)
10
8
12
5
Trne (min)
0
_
12,0
10,0-
|
|
Floor U-1
|-
|
Floor U-4
a.
e
I
_A d
50 so
I
/10.e
_r
a
CIFW
BE #3, Test
°
0
5
Tme (rin)
10
81
10
FLOORUt :Total
absorbed heat tlux
GIC FLOOR-U4: Total
heatfkt
d'>
- VVteT
"
oo
JL
0
2
Floor HeatRux
KICFMBE#3 Testa8
4
6
8
10
12
Te (rmn)
Figure A-82: Floor heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-108
Technical Detailsfor the MAGIC Validation Study
10,0
9,0
Roor U-1
-
8,0 .
Floor U-4
--
? 7,0
MAGIC FLOORU1I :Total
sorbed heat flux
FLOORU4: Total
.bed atflux
1 6,0
i 5,0
L
4,0
3,0
-W
2,0
.I
0
5
Roor Rieat Flux
ICFbPBE #3, Test 4
15
10
0
.
I.
1,0
0,0
Tiw (mn)
10
15
Tim (rrn)
101
I1
9
e
Roor U-1
F
8 -
-
Roor U-4
7
-
MAGIC RFOOR-U: Total
absorbed heat flux
MAGIC FLOORFU4: Total
orbed heat flux
6 -
<*
. &~,so
5
4
e
3
2
-
--
-
1'
0
10
5
n
15
Tffm (rin)
12
350
5 Tfe(rruin)
ea
10
15
7.0,
*TC
FRoor UJ-2-2
. Roor Peat Rlux
8.0 .P BEW
#3, Test 13
300-
TO Roor U- 1 2
250 -
U
RI91
,L
0
200 -_
LD t50
/
~
Roor U-1
F
-
tenmoerature of the target
X**
^
.
5.0
-MAGIC RLOOFI U2: Surflarce
1
MAGIC R COR-U1.: Surface
terrperature of the target
L
3,0-
.11
t
2.0 -
ifoTenl1orat
5
Roor tJ-2
-*MAGIC R.OORLU3: Total
absorbed heat fhux
.
I
/
100 -
"
4,0
.... I AI FLORL3 os
1,0 0.0
10
115
Tare (nmn)
5
0
20
25
5
0
10
15
Tfm (rim)
20
25
400
350
300
I
5.0
O 250
1 4,0
U 200
9 3.0
1 150
!
0
2
4
6
TMn (rrh)
8
10
12
2.0
0
2
4
6
Trne (rrh)
8
10
12
Figure A-83: Floor heat flux and surface temperature, ICFMP BE #3, closed door tests.
A-109
Technical Detailsfor the MAGIC Validation Study
140
120
6
-
(U
80
r
60
IT
TC Rloor U-2-2
-
TC Floor UJ-1t-2
FL-OOR U12:Surface
-MAGIC
tenperature of the target
I - -x-AGIC
FLOOP~UI : Surface
-ttur of the targjet
100
a
r
=
E__
1_
40
-
L-
I
Tenperature
Roor
l
CFNWBE #3. Test 17
20
04,
2
0
4
6
TiRme
(mn)
8
10
12
0
5
Time (Imn)
10
Open Door Tests to Follow
250
6
|
@TC
---
w
200
6,0
-5.0
IMAGIC
RFOO-12: Surface
tenperature of the target
MAGtC ROOR-U1 Su
4,
* -.---
150
Floor U-2-2
TC Floor U-1-2
tenpertur
lenperature 1
Floor
Fl_oor Tenperature
tCFW1
BE #3.Test 3J
sW
0
0
5
10
15
20
25
TIre (nin)
-0 *
250
--
p23.0-
-
12.0
-
-
-----
-
Poor U-2
-
FAAGIC 0FLOORU1:Total
absorbed heat flux
tlMAGIC FLOOfPLJ2: Total
absorbed heat thx
'1
1,0 -
SS
3
FborFHeat Fhx
IXUP BE #3, Test 3
0,04
0
5
10
15
Time (min)
20
f
2
0
5
10
15
Tine (imn)
20
25
TC Floor U-2-2
TCFRoor U-1-2
-
slAGIC FLOOR-U2: Surface
tenperature of the target
200 - a
-F-----
150
1 4.0
-RPoor U- I
MAGIC RLOOR-IUJ: Surface_
tenperature Of
100
50
'
Floor Tenperature
ICRIP BE #3. Test9 I
0 +.
0
_;
5
10
Time (min)
15
20
25
Figure A-84: Floor heat flux and surface temperature, ICFMP BE #3, open door tests.
A-1 10
Technical Detailsfor the MAGIC Validation Study
12,0
T
.
Nw
---
450
400
-
TC Floor U-4-2
-"
TC Floor U-1-2
Z
MAGtC RFLOORL4: Surface
.0. .
-a 350
300
m
-
-
I
Floor U-4
8,0-
g
8.0 -
terrperature of the target
.
Suurl
ce
enperare
_
....
i 250
i 200-
C
F- 150
* ....
-
I bsorbecd heat flux
I
IL
0Z0-
-
WN00
5
MAGtCFLOOR-U1:
Total absorbed heat flux
MAGIC FLOO-U4:
6,04,0-
1
50V
0
Floor U-1
10,0-
10
15
20
25
Floor Tenperature
lCFPBE #3, Test 5
30
20
0
10
Floor Heat Flux
ITW BE #3, Test 5
Terne(rfin)
30
Trre (rmi)
300 1
250 -
-
TC Floor U-2-2
L_-
TC Floor U 1--2
-
MAGIC FLOOR-U2: Surface
terrperature of the target
MAGIC FLOOR-U1
Surface
terrperature of the target_.
-
f
150
I
100 -
S
_
50
r
1,4
t
1,2
i 1,0
Ci 0,8
_
0,4
Floor Terrperature
ICFW BE #3, Test 14
02
0.0
0
5
400
10
15
Tim (rrin)
*
250
....
200
15
20
y
25
MAGIC ROOAU2: Surface
terrperature of the target
-
m
10
Trn (rrin)
TO Rloor U 1-2
300-i
4i
ia
25
TC Floor UJ2-2
35030
)
20
*. *MAGIC RZOOR-U1: Surface
terrperature of the target
I
I
150
100
Floor Terrperature
50
50-
k_5
0
5
10
sR~~
15
T
20
25
0
Thue (rin)
5
10
15
Trn (rrh)
20
25
350
-.- *TC
Floor U1-2-2
300
||
-
I
TC Floor U1-1-2
250-1
200 -
I
150
NAG CR.LOORl-12 Surface
terrperature of the target
...:Surface
MAGC FLOOR-Ul
lteperature of the target
100
;5
-
wFlor TeIrperature
BE #3 Test 18|
,nP
0
5
10
Trm (nin)
15
20
25
0
5
10
15
Tim (rmi)
20
25
Figure A-85: Floor heat flux and surface temiperature, ICFMP BE #3, open door tests.
A-ill
Technical Detailsfor the MAGIC Validation Study
Table A-31: Relative differences for temperature and total heat flux corresponding to
the floor
Floor Total Heat Flux
Floor Temrerature
AM Relative
Relative AE
Instrumen
2
Diff (kW/m (kW/m2 ) Diff
AE (0 C) AM (°C)
Floor
18%
0.7
0.6
40%
53
38
F1
43%
2.3
1.6
-15%
66
77
F4
Test 1
9%
0.6
0.6
43%
52
36
F1
34%
2.3
1.7
-19%
63
78
F4
Test 7
6%
1.9
1.8
34%
99
74
F1
-27%
4.7
6.4
-16%
131
156
F4
Test 2
2%
1.9
1.9
37%
98
71
F1
-23%
4.8
6.2
-11%
132
148
F4
Test 8
7%
1.7
1.6
55%
118
76
F1
-22%
4.7
5.9
-5%
144
152
F4
Test 4
17%
1.7
1.5
68%
120
71
F1
-17%
4.7
5.7
-8%
145
158
F4
Test 10
48%
132
89
F1
77%
130
73
F2
est 13
49%
119
80
F1
-43%
118
206
F2
Test 16
96%
1.7
0.9
80%
44
24
F1
1%
1.5
1.5
-56%
51
117
F2
Test 17
8%
1.2
1.2
110%
112
54
F1
-34%
1.5
2.3
-27%
135
186
F2
Test 3
5%
1.2
1.2
106%
110
53
F1
-21%
1.5
1.9
41%
132
94
F2
Test 9
16%
1.0
0.9
118%
91
42
F1
-52%
4.9
10.0
36%
232
171
F4
Test 5
9%
1.2
1.1
113%
111
52
F1
-3%
1.3
1.3
149%
117
47
F2
Test 14
6%
1.2
1.2
113%
112
52
F1
-76%
1.8
7.5
10%
155
140
F2
Test 15
10%
1.2
1.1
118%
108
50
F1
-4%
1.3
1.3
115%
117
55
F2
Test 18
A-112
Technical Detailsfor the MAGIC Validation Study
A.9.2 ICFMP BE #4
Three thermocouples were mounted on the back wall of the compartment. Because the fire
leaned towards the back wall, the temperatures measured by the thermocouples are considerably
hotter than most of the other wall surface points considered in this report.
700 Well Surface Te vprature
6COBE4,
600
t> Test
700
6
Test 1
500
Soo
400
400
E
Wall Surface Terrperatur
BE4, e
B
E 300
300
200
M19: Surface
,MAGIC
100
terperatureof the target
200
MAGICMO0: Surface
100
t;erature of the target
0
0
0
10
20
Tn-e (rrin)
30
40
0
10
20
Trre (nn)
30
40
Figure A-86: Back wall surface temperatures, ICFMP BE #4,
Table A-32: Relative differences for wall temperature
ICEMP 4-1
Instrumeni
M19
M20
E (0C)
596
724
(0C)
AM
656
656
Relative
Diff
10%
-9%
A-1 13
Technical Detailsforthe MAGIC Validation Study
A.9.3 ICFMP BE #5
Wall surface temperatures were measured in two locations during the BE #5 test series. The
thermocouples labeled TW 1-x (Wall Chain 1) were against the back wall; those labeled TW 2-x
(Wall Chain 2) were behind the vertical cable tray. Seven thermocouples were in each chain,
spaced 80 cm apart. In Figure A-87, the lowest (1), middle (4), and highest (7) locations are
used for comparison.
l
140
Wall Surface Terrperature BE5, Test 4
v
120
-
a
MAGICTW 1-1
1_
100-
.TWI
e
2
60
MAG 1W 2-1
K G
....
E
I-
20E0
==
5
10
15
Tme (irdn)
20
25
30
0
5
10
15
20
25
Tir (rrin)
a
E
I-
0
5
10
15
Time(ein)
20
25
30
Figure A-87: Back and side wall surface temperatures, ICFMP BE #5, Test 4.
Table A-33: Relative differences for wall temperature
ICFMP 5
A-1 14
Instrumen
TW 1-1
TW 2-1
TW 1-4
2-4
TW 1-7
TW 2-7
AE (0C)
79
12
118
96
121
100
AM (°C)
51
113
134
132
131
135
Relative
Diff
-35%
868%
13%
38%
9%
36%
30
B
MAGIC INPUT FILES
Appendix B includes the MAGIC input files used for the simulations in this V&V study. These
file will only be available electronically due to their size and formatting.
B-1
U.S. NUCLEAR REGULATORY COMMISSION
NRC FORM 335
1. REPORT NUMBER
(9-2004)
(Assigned by NRC, Add Vol., Supp., Rev.,
and Addendum Numbers, if any.)
NRCMD 3.7
BIBLIOGRAPHIC DATA SHEEr
NUREG-1 824
(See instructions on the reverse)
3. DATE REPORT PUBLISHED
2. TITLE AND SUBTITLE
Venficaton &Nadation of Selected Fire Models For Nuclear Power Plant Applications
Volume 5: MAGIC Draft Report for Comment
YEAR
MONTH
200
Janua
4. FIN OR GRANT NUMBER
S.AUTHOR(S)
6. TYPE OF REPORT
FS=Ofco Joglar-Biloch (EPRI/SAIC), Jason Driesbach (NRC), Kendra Hill (NRC), Bijan Najafi
(EPRI/SAIC), Kevin McGrattan (NIST), Richard Peacock (NISTl, Anthony Hamins (NIST)
Technical
7. PERIOD COVERED (incluswe Oates)
8.
PERFORMING QRGANIZATION - NAME AND ADDRESS (If NRC, prod
prnide name anfliv adrsSs.)
Divesion.
Office or Region,.US. Nudear Regulatory Commission, and mading address; if contractor,
Electric Power Research Institute (EPRI), 3412 Hillview Avenue, Palo Alto, CA 94303
U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research (RES), Washington, DC 20555-0001
Science Applications International Corp (SAIC), 4920 El Camino Real, Los Altos, CA 94022
National Institute of Starlidards arid Technology, 100 Bureau Drive, Stop 8600, Gaithersburg, MD 20899
9. SPOQNSORING ORGANIZATI1iN -NAME AND ADDRESS
arnd~mAT#Vadd"Ss~
I
{effNRC,
Wvs 'Same as aJovye'; If contractor, provide NRC Division, Offic or Region, U.S. Nudear Regulatoty Commission,
i
Electric Power Research Institute (EPRI), 3412 Hillview Avenue, Palo Alto, CA 94303
U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research (RES), Washington, DC 20555-0001
10.
SUPPLEMENTARY NOTES
II
f
11. ABSTRACT (200 words or less)
There is a movenent to introduce risk- and performance-based analyses into fire protection engineering practice, both
domestically and woridwide. This movement exists in the general fire protection community, as well as the nuclear power plant
(NPP) fire protection cbomrmunit*.
In 2002, the National tire Protection Association (NFPA) developed NFPA 805, Performance-Based Standard for Fire
Protection for Light-Wpter Reator Electric Generating Plants, 2001 Edition. In July 2004, the U.S. Nuclear Regulatory
ilended i Sfire protection requirements in litle 10, Section 50.48, of the Code of Federal Regulations (10
Commission (NRC}
l
t
i~ting relactor licensees to voluntarily adopt fire protection requirements contained in NFPA 805 as an
CFR 50.48) to permt
alternative to the exist
deterinistic
ng
fire protection requirements. In addition, the nuclear fire protection community wants to
(RI/PB) approaches and insights to support fire protection decision-making in general.
use risk-informed, $rfwrnlance-based
i
RI/PB fire protection is the availability of verified and validated fire models that can reliably
One key tool needed 4upport
predict the conseqb riCes of fires. Section 2.4.1.2 of NFPA 805 requires that only fire models acceptable to the Authority
Having Jursdictiorj A,>-J) shall be used in fire modeling calculations. Further, Sections 2.4.1.2.2 and 2.4.1.2.3 of NFPA 805
state that fire model s all onI be applied within the limitations of the given model, and shall be verified and validated.
This report is the firt pffort to document the verification and validation (V&V) of five fire models that are commonly used in NPP
applications. The poject was performed in accordance with the guidelines that the American Society for Testing and Materials
(ASTM) set forth in Standard El 355-04, "Evaluating the Predictive Capability of Deterministic Fire Models." The results of this
V&V are reported in the form of ranges of accuracies for the fire model predictions.
12. KEY WORDS/DESCRIPTORS (List words orphiases that wfll assist researcers inioating the report)
Fire, Fire Modeling, Verification and Validation (V&V)
Performance-based , 0isk-informed Regulation, Fire Hazard Analysis (FHA),
Fire safety, Fire Protqction, Nuclear Power Plant
Fire Probabiste
Assessmnent (PRA), Fire Probabilistic Safety Assessment (PSA)
13. AVAILABILITY STATEMENT
^E
unlimited
4
14. SECURITY CLASSIFICATON
(This Page)
fiik
unclassified
(This Repolt)
unclassified
15. NUMBER OF PAGES
16. PRICE
NRC FORM 335 (9-2004)
rm-
I-
UN ---
TUtu-e
I
I
il
,.1
*I
Federal Recycling Program