UNCLASSIFIED

AD NUMBER

AD906699

NEW LIMITATION CHANGE

TOApproved for public release, distributionunlimited

FROMDistribution authorized to U.S. Gov't.agencies only; Test and Evaluation; NOV1972. Other requests shall be referred toArmy Mobility Research and DevelopmentLab., Fort Eustis, VA.

AUTHORITY

USAAMRDL ltr, 30 Mar 1984

THIS PAGE IS UNCLASSIFIED

USAAMROL TECHNICAL REPORT 72-52

INVESTIGATION AND EVALUATION OF

NONFLAMMABLE, FIRE-RETARDANT MATERIALS

By

S. AtllhH. Duccigross D

November 1912 JAN

EUSTIS DIRECTORATE B

U. S. ARMY AIR MOBILITY RESEARCH AND DEVELOPMENT LABORATORYFORT EUSTIS, VIRGINIA

CONTRACT DAAJ02-71-C-0042ARTHUR D. LITTLE, INC.

CAMBRIDGE. MASSALHUSETTS

Distribution limited to U. S. Governmentagencies only;, test and evaluation;November 1972. Other requests for thisdocument must be referred to the EustisDirectorate, U. S. Army Air MobilityResearch and Development Laboratory,

Fort Eustis, VA 23604.

r

!p

DISCLAIMERS

The findings in this report are not to be construed as an officialDepartment of the Army position unless so designated by other authorizeddocuments.

When Government drawings, specifications, or other data are used for anypurpose other than in connection with a definitely related Governmentprocurement operation, the U.S. Government thereby incurs no responsi-bility nor any obligation whatsoever; and the fact that the Governmentmay have formulated, furnished, or in any way supplied the said drawings,specifications, or other data is not to be regarded by implication orotherwise as in any manner licensing the holder or any other perscl orcorporation, or conveying any rights or permission, to manufacture, use,or sell any patented invention that may in any way be related thereto.

Trade names cited in this report do not constitute an official endorse-ment or approval of the use of such commercial hardware or software.

DISPOSITION INSTRU.*TIONS

Destroy this report when no longer needed. Do not return it to theoriginator.

DEPARTMENT OF THE ARMYU. S. ARMY AIR MOBILITY RESEARCH & DEVELOPMENT LABORATORY

EUSTIS DIRECTORATEFORT EUSTIS. VIRGINIA 23604

This report was prepared by Arthur D. Little, Inc., under the termsof Contract DAAJ02-71-C-0042. The technical monitor for thisprogram was Mr. H. W. Holland of the Safety and SurvivabilityDivision.

The purpose of this effo' was to evaluate the feasibility ofcontaining or restricting in-flight or postcrash fire in anattempt to allow the crew and passengers to escape or remainwithin a livable environment until the fire can be extinguishedand rescue accomplished.

During program, an evaluation was conducted of currently

availay Lonflammable, fire-retardant materials. After comple-tion , this evaluation, the most promising of these materialswere selected for application on two helicopters, which were thensubjected to full-scale fire tests. This report presents theresults of the evaluation of the various materials and also thepreparation for and outcome of the full-scai- fire tests.

The conclusions and recommendations contained in this report areconcurred in by this Directorate.

Project l162203A529Contract DAAJ02-71-C-0042

USAAMRDL Tec1hnLc" Rep..rt 72-52

November 1-72

INVESTIGATION AND EVALUATION OF

NONFLAMMABLE, F-RE-RETARDANT MATERIALS

Final Report

ADL- 73588

By

S. AtalahH. Buccigross

Prepared by

Arthur D. Little, Inc.Cambridge, Massachusetts

for

EUSTIS DIRECTORATE

U.S. ARMY AIR MOBILITY RESEARCH AND DEVELOPMEN' .ABORATORYFORT EUSTIS, VIRGINIA

Distribution limited to U.S. Government agenciesonly; test and evaluation; November 1972, Otherrequests for this document must be referred to the[ustis Directorate, U.S. Army Air Mobility Research

and Development Laboratory, Fort lustis, VA 23004.

ABSTRACT

The objective of this study was to evaluate the feasibility ofcontaining or restricting in-flight or postcrash helicopterfires to allow the crew and passengers to escape or remain withina livable environment until the fire could be extinguished orthe L'rning fuel consumed.

A -omprehensive survey was made of present materials technology.A number of materials and composites were selected and tested ina specially designed furnace capable of providing a thermal fluxequivalent to that encountered In JP-4 fires. Final candidatew6ll systems were compared for protection effectiveness, costand weight penalty. Various combinations of isocyanurate foams,sodium silicate hydrate panels, a mineral insulation, and intumes-cen't mastic paints were then applied to the walls of two crash-

damaged helicopters (UH-lD and CH-47) and exposed to full-scale fires simulating in-flight and postcrash fires. The heli-copters were fully instrumented to measure temperature, heatf-ux, smoke density, and toxic gases.

The results of the in-flight simulation tests indicated thatit should be possibie to protect the habitable compartmentagainst a fire occurring in an adjacent compartment resultingfrom a fuel or hydraulic oil line leak. Sodium silicate hydratepanels placed on the fire side appeared to give the bestperformance

Interior tenipcrature and heat fluxes were above tolerable levelsfor humans during the post.rash fire tests in both helicopters.Smoke and particulates were also judged to be too high for human

tolerance Penetrations in the CH-47 walls occurred where isocy-anurate foam could not be applied because of the presence ofwiring, air ducts and hydraulic oil tubes.

The UH-lD walls did not lend themselves to foaming because of theabsence of ribs and formers. The sodium silicate hydrate panelsused to protect the interior walls partially collapsed because ofthe absence of structural support.

It was concluded that total wall protection of existing helicopters

against postcrash fires is not feasible and should not be pursued

any further because of cost, unreliability, and lack of assurance

that the walls will maintain their integrity in a crash.

iii 1W 103 BLANUCT MIJW

FOREWORD

This study was conducted by Arthur D. Little, Inc., under ContractDAAJ02-71-C-0042, Project 1F162203A529, with the Eustis Directorate,U.S. Army Air Mobility Research and Development Laboratory, FortEustis, Virginia.

The authors wish '.o acknowledge the valuable assistance providedat Arthur D. Little, Inc., by J. Oberholtzer and J. Valentine(toxic product analysis), R. Lindstrom (material q(.ection),A. Camus (instrumentation), J. Hagopian (fire tests), andH. Survilas (photographer). In addition, the authors are gratefulto the manufacturers who provided free samples of their materialsto be tested and evaluated, and to the Fire Department of LaurenceG. Hanscom Air Force Base for allowing the use of its site for thelarge-scale fire tests and for providing valuable standby assistance.

v| ii ii | iTIMM

TABLE OF CONTENTS

Page

ABSTRACT ......... ........................ ... iii

FOREWORD ......... .......................... v

* LIST OF ILLUSTRATIONS ........ ................. ix

LIST OF TABLES ..................... xii

INTRODUCTION......................1

LABORATORY EVALUATION PROGRAM .... ............... 3

Test Furnace Facility ........ ................ 3

Results of Laboratory Test Program .... ......... 9

SELECTION AND APPLICATION OF PROTECTIVE WALLS ..... 20

Selection Criteria ...... ................. ... 20

Compatibility With the Helicopter ... .......... ... 21

Wall Systems Used in the CH-47 ..... ........... 21

Wall Systems Used in the UH-ID ........... ....... 23

FIRE TEST PROGRAM ...... ................... ... 25

General Description of Tests ...... ........... 25

Sampling and Measurement of Toxic Gases and Smoke . 27

Thermal Instrumentation .... ............... .... 38

Photographic and Video Coverage .. ........... .... 39

Data Acquisition ......... .................. 43

Power Supply ....... .................... ... 43

RESULTS ......... ........................ ... 44

In-Flight Fire Simulation Tests ..... ... .. 44

CH-47 Postcrash Fire ...... ................ ... 49

11)-ID Posterash Fire ...... ................ ... 65

vii "I )D A W 1

'"wao PA m~ww nu

Page

CONCLUSIONS AND RECOMENDATIONS ... ............. 74

APPENDIX, TEST DATA ...... .................. ... 75

DISTRIBUTION .......... ..................... . 83

v

viii

LIST OF ILLUSTRATIONS

Figure Page

1 NASA-Ames T3 Thermal Test Facility ........ . 4

2 Helicopter Simulation Box ..... ........... 5

3 Details of Corner of Simulation Box . . . ... 6

4 Overall View of the ADL Test Furnace ... ...... 7

5 Test Enclosure Above Furnace Opening ... ...... 8

6 Photograph of Interior Plexiglas of Window After9 Mintes of Fire Exposure ............. ... 19

7 Photograph of Side Exposed to Fire ......... ... 19

8 Second In-Flight Fire Simulation CompartmentWith Partial Cover Down ... ............ ... 26

9 First In-Flight Fire Test in Progress ..... 26

10 CH-47 Before Postcrae'i Fire Test .... ........ 28

11 CH-47 Fire in Progress Soon After Ignition . . 28

12 CH.-47 Fire at Maximum Intensity .. ...... ... 29

13 CH-47 Cockpit Near Conclusion of Test .... 29

14 UH-ID Before Postcrash Fire Test .... ....... 30

15 UH-ID Postcrash Fire After 30 Seconds .... . 30

16 Fully Developed UH-lD Postcrash Fire ....... ... 31

17 UH-ID Fire Near Conclusion ....... ........ 31

18 Exterior Sampling System for CO2, CO, and 02 . . 34

19 Interior Sampling System for Reactive Gases . . 36

20 Thermocouple and Heat Flux Meter Distributionin the UH-lD for the In-Flight Tests ...... ... 40

21 Thermocouple and Heat Flux Meter Distribution

in the CH-47 ...... .................. ... 41

22 Thermocouple and Heat Flux Meter Distribution

in the UH-ID for Postcrash Test . ........ ... 42

23 Ttaperatures Recorded During the UH-ID FirstIn-Flight Fire Test .... .............. ... 45

ix

Figure Page

24 Temperatures Recorded During the UH-ID SecondIn-Flight Fire Test . . . . . . . . . . .......... 46

25 Heat Flux Data Within the UH-ID Cabin During theFirst In-Flight Fire Test . . . ............... 47

26 Heat Flux Recorded in the UH-lD During the SecondIn-Flight Simulation Test .. . . . .. . .......... 48

27 Temperatures Recorded in the CH-47 Cabin During thePostcrash Fire Test . . . ...... .......... 50

28 Temperatures Recorded in the CH-47 Cabin During thePostcrash Fire Test . . . ................. 51

29 Temperatures Recorded by Thermocouples Exposed tothe CH-47 Postcrash Fire ....... ............... .. 52

30 Total Heat Flux Recorded During the CH-47 PostcrashFire Test . . . ....................... 53

31 Cargo Door Where Explosion Occurred .... ............ .55

32 Cargo Floor Space With Aluminum Ribbed Panel Usedas Cover . . . . . . . . . . . . . . . . . . . . . . . .. . 55

33 Sequence of Explosions in CH-47 Postcrash Fire ......... 56

34 Light Transmission at the 4-Ft and 1-Ft Levels Duringthe CH-47 Postcrash Fire Test ..... ............... .58

35 CH-47 Interior After Fire Looking Toward CockpitDoor ... .. ........................... 61

36 CH-47 Interior Typical Wall Construction Before Fire . . . 62

37 CH-47 Interior Typical Wall Construction After Fire . . 62

38 Right Side of CH-47 After Fire ..... .............. .63

39 Left Side of CH-47 After Fire ............... 63

40 Unribbed Section Where 1CU Foam Could Not Support

Itself . . . . . . . . . . . . . . . . . . . . . . . . .. 64

41 Typical Cellular Structure of Charred ICU Foam ........ .. 64

x

Figure Page

42 Temperatures Recorded in the UH-ID Cabin During thePostcrash Fire Test .. . .*........... 66

43 Temperatures Recorded within the UH-lD Cabin Duringthe Postcrash Fire Test . . . . ........... . . . . 67

44 Temperatures Recorded by Thermocouples Outside theUH-lD Cabin . . . . . . . . . . . . . . . . . . . . . 68

45 Total Heat Flux Measured by Brass Sphere Calorimeterin the UH-lD Postcrash Fire Test . . . . .......... 69

46 Light Transmission in the UH-lD Cabin During thePostcrash Fire Test . . . . . . ............... 72

47 Upwind Side of UH-lD After Postcrash Fire Test ........ .. 73

48 Lee Side of UH-lD After Test Showing Absence ofAluminum Walls and Sagging Fire Plate Walls ........ 73

xi

LIST OF TABLES

Table Pa__

I Furnace Test Results for Paints and Coatings . ..... 10

II Furnace Test Results for Inorganic Insulations . .... 13

III Furnace Test Results for Organic Fclms . . . . . 14

IV NASA-Ames ICU Foam Formulation. . . . . . ........ 15

V Furnace Test Results on Combinations of Materials . . . 16

VI Furnace Test Results on Window Materials ............ 17

VII Cost and Weight Penalty of Selected Wall Systems ...... . 22

VIII Toxic Gases and Particulates Generated in theCH-47 Fire ..... .............. ...... ... 60

IX Toxic Gases and Particulates Generated in theUH-ID Fire . . . . . . . . ................. 71

X Furnace Tests - Uncoated (Shiny) Aluminum .......... ... 75

XI Results of Furnace Tests on Paints and Coatings ..... 76

XII Results of Furnace Tests on Organic, Inorganicand Composite Insulations ..... ................... 78

XIII Results of Furnace Tests on Box Panels SimulatingPresence of Wiring and Tubing ..... .............. ... 80

XIV Results of Window Materials Test Program ........ 81

xii

INTRODUCTION

Although it has been possible to reduce the number of postcrashfires of U.S. Army aircraft by the construction of crashworthyfuel s-ostems, it will be several years before these systems areinstalled in all U.S. Army aircraft. Even then, it is doubtfulthat all crash fires will be eliminated.

The crew and passengers aboard such aircraft have often survived

the crash to find themselves trapped in a fire that engulfstheir *abin. In other cases, localized fires in engine compart-ments (r cargo areas have penetrated the walls of the habitablecabin t) render it untenable.

Recent a.vances in cryogenic and high-temperature (reentry)

technology have brought forth a wide variety of materials withinteresting physical and chemical properties. A number of

materials are now available which combine the properties offlame retardancy, mechancal strength and insulation. It wouldseem that one should be able to prevent postcrash or in-flightdeaths from fires by the appropriate selection and application ofone of these new materials to the aircraft fuselage or interiorpartitions. Unfortunately, flame retardancy is usually Impartedto materials by the addition of chemicals which inhibit flamepropagation. These chemicals, as well as the materials them-selves, generate irritating smoke and toxic gaseous productswhen they are heated to high temperatures. The identiry ot the

toxic gases, their concentrations, and their rates of generationare unpredictable because they are a complex function of therate of heating and the temperatures to which the material naivbe exposed. Thus, the selection oI a protective material for

aircraft fuselages should be accompanied by realistic tests inwhich smoke and toxic gases are monitored.

The objective of this study was to evaluate the feasibility ofcontaining or restricting in-flight or postcrash fires to allowthe crew and passengers of a helicopter to escape or remainwithin a livable environment until the fire could be extinguishedand rescue accomplished.

To achieve this objective, it was desired to review data onvarious flame-retardant materials and to select seveial materialsthat appeared to have the greatest potenLial for maintaining alivable envizonment within an aircraft exposed to in-flight orpostcrash fire. The thermal properties, weight, cost and

1

installation feasibility were to be examined, and a test procedurewas to be developed to evaluate and compare the behavior andeffectiveness of the selected materials (and combinations thereof)when they were applied to helicopter fuselage skin and to interiorpartition materials. The results of the experimental programwere to be used to recommend the most promising fire-retardantmaterial (or combination of materials) for protecting aircrafthabitable compartments. Finally, UH-lD and CH-47 helicopterswere to be examined to determine appropriate methods for applyingthe selected fire protection materials and the areas that shouldbe protected. Two crash-damaged helicopters were then to be soprotected, instrumented and subjected to full-scale simulationsof in-flight and postcrash fires.

2

.1

LABORATORY EVALUATION PROGRAM

The objective of the laboratory test program was to screen a largenumber of candidate wall materials by exposiug aluminum panelsprotected by tihe test materials to a thermal flux equivalent to thatencountered in a large-scale JP-4 fire. Various window materialsand vent plugging techniques were also to be evaluated.

TEST FURNACE FACILITY

The desired thermal flux level (31,000-35,000 Btu/hr sq ft) wasobtained in a furnace, the design of which was quite similar to NASA-Ames' T-3 Thermal Test Facility (Figure 1) but which was scaled-upso that 16 inch x 16 inch panels could be exposed. This was done to

reduce end effects, particularly for thick samples. A stainlesssteel enclosure was constructed which was used to hold most panels inplace above the horizontal furnace opening. Sketches of the testenclosure are shown in Figures 2 and 3. Figures 4 and 5 are photo-graphs of the Arthur D. Little (ADL) furnace and the test enclosure

in position. Thermocouples, gas-sampling tubes and a smoke densitymeter were used to evaluate conditions within the enclosure. Acalibrated (NBS traceable) multichannel millivolt recorder was usedto record the data.

Chromel-Alumel thermocouples were utilized to measure the enclosureair temperature and the interior surface temperature of the panelat two or three locations.

The smoke detector consisted of a light source and a photo cell.This was calibrated using standard filters having various absorptioncoefficients.

Gas samples were drawn either into evacuated glass sample bottles forchromotographic analysis or into Kitagawa tubes. These tubes detect

the presence and measure thb "oncentration of specific gases by thelength of the chemical er'.,r iange in the packed tube.

The cond .rior surface of the panel was noted

visually cest through a mirror attached to kn openingin the to, 'f the enclosure. An electric timer was started when

the panel w . nlaced above the furnace and stopped when burn-through was oL rved or when the test was terminated.

Tests on tht :andidate materials were conducted by placing

the window in a s .1 frame on the furnace opening so that theflames impinged directly on the window. The time for the windowto burn through was noted.

3

L000

0)0 '2

d' m N

W 4- 4- 4

(E- w\

0 0

-('4 r(~~~V~*\K M: ULJL

**~t'c', L)

imr~hr-2Z

m

~~ZY Z)w

(Y) /

15 I NSIDE -

151" OUTSIDE4-" STAINLESS

CHIMNEY FOR GAS -3" DIA 8STEELSAMPLING ANDTHERMOCOUPLES --- ~

71

~--SMOKE DETECTOR(TYP OPPOSITE SI DE)

14

4 --- ?-~-~

JTP

STD &NT

3 PRX EIH:7L

Figue 2 Helcoper SmultionB17

-- " STAINLESS STEEL BOX

WELD (TYP)

PROTECTIVE MATERIAL 3"

LALUMINUM TEST (TYP)PANEL (.032) 2 4

-- 16UNC STUD8 NUT

Figure 3. Details of Corner of Simulation Box.

6

Figure 4. Overall View of the ADL TeC5L Furnace.

7

U

w

a-4

U

(I)

H

-4

f

8

RESULTS OF LABORATORY TEST 'ROGRAM

The interior surface of ; .elicopter may be divided into four areas,each of which requires pArticular protection techniques. Theseareas consist of the exterior aluminum wall, windows, ventilationopenings, and interior partitions.

Wall Areas

Much of the wall area in a helicopter is free of wiring and Lubing so thatfire protection materials can be placed directly on the wall withoutinterfering with other mechanisms. However, some areas do containhydraulic, mechanical or electrical components, and the selected in-sulation system should allow access to these pieces of hardware.

The materials that were evaluated for wall procection can be con-veniently classified under the following categories:

* Paints and coatings

* Inorganic or mineral insulations

* Organic foams

* Fire-protecting panels and composites

Tables I, II, III, and V summarize pertinent results of testson promising materials in these categories. Tables X, XI, XII, XIII,and XIV summarize the results for all materials tested.

Data of particular concern in the test program were the time for theflame to burn through the panel, the presence of toxic or noxiousfumes in the enclosure, and the maximum temperature of the airin the enclosure.

Intumescent Paints and Coatings

A large number of intumescent paints and coatings, which are appliedin thicknesses from 3 mil to 1/8-inch or more, are conmerciallyavailable for fire protection. Upon heating, these materials in-tumesce or char, forming a protective insulating layer. Table I

summarizes the furnace test results for these materials. The dataindicate that no intumescent paint or coating could provide thedesired fire protection. While the paints did intumesce readily andprovided an insulating char, the fire gases eroded the char veryquickly, exposing the bare aluminum ?anel to the fire. When such paintswere used on the inside of a panel, they were not effective because thepanel burned through and the char could not support itself. Further-more, most of them produced noxious fumes in the enclosure. Thereis no apparent difference in the performance of the very expenive

9

(0

4, 4

1> 0 to0 (0( 0

4O,4 4 od -W 41 -H" 00 0 00 &0~ 0 M0 00 (0 -

a4 84 0 C o 0 ~ 4- '041( 41 '0.3

0 I 0 (00 0J, (H t 4m. 4 13 m- u0 -r -

c4 >0 00 -4 . .

-4

> 0

(04

H3 tn 03, C1 ar 0 ( N " 0

0 ~ ... t C2 0 " -. 12 02 00fl 14 1-4 N4 C1 1- 4 (Y) (

V) w

10 0 4 41 0 0 C ( 0 0 C( 0H 0 H CO oj4 V0 0 C0 C

(0 03 n4-1:1 (0 14 ) (

w0.4. 0) 0 C' ~

0. -4 H N '-4 H H 0

ccX 4 0 0 0n 00 0

C" 04(04 . -4

0.0040 .1 -.1 04 0 133 0.0.0 -0N 0 0 0 "~ 0

<z w4 ~ 0 H U4 4Q3 w0' rA 0 - ( 4c

1-( -4 ,1*00

.144

4-4- (0i 0c 0

to' OC, r- 03 r 0) Q) 10 40 ~4 (11-4~ '1 0 0 05 V) W' NC:~(0. C: & * i- .04 (

> 4 > 4 0 j 04 0

100

paints or the less expensive types, although there are claimeddifferences in weatherability and stability. The one coating thatappeared to bc potentially useful, North American Rockwell's Larodyne,wos applied in 1/8-inch thickness to the aluminum panel and theLarodyne side of the panel was exposed to the fire. Although it took11 min 39 sec for the panel to burn through the first test, thisexceptional perfor'ance could not be duplicated. Subsequent panelsdid not perform as well because the char crumbled quickly. Itappeared that if Larodyne is applied over large areas, stressescould develop during fire exposure which would fracture the coatingand allow the rapid penetration of fire. Furthermore, Larodyne atpresent does not pass military specifications for exterior coatingsand is fairly expensive.

Inorganic Insulations

Certain high-temperature inorganic (mineral) insulations arecommercially available which have the potential of affordingprotection to the occupants of the helicopter without generatingany toxic gases.

Two materials from Johns-Manville called Dynaflex and Microquartzwere evaluated. Both of these materials afford good fire protec-tion. However, being structurally weak (similar to fiberglassblankets), they need to be held in place and they offer little aidin filling cracks, etc. In combination with other materials, theyproved to be quite useful. Table II describes the results of testson Dynaflex and Microquartz felt panels.

An interesting fire protective panel developed by Badische Anilin &Soda-Fabrik (BASF) called Brandschutzplatte (fire plate) wasevaluated. This is a 1/16-inch-thick fiberglass reinforced sodiumsilicate hydrate. When heated, the water of hydration was drivenout, expanding the panel to 1/2-inch thickness. This panel providedexcellent protection when tested on the furnace. After 7 minutes,the air temperature on the opposite side of the enclosure was only40*C and after 12 minutes of exposure, there was no burnthrough ofthe fire plate (see Table II).

Organic Foams

A variety of fire-retardant organic foams, particularly polyurethane

and polyisocyanurate foams, are available for foaming in place aswell as i. 9refoamed panel form. A large number of organic foamswere evaluated including Pyrell polyurethane foams from Scott,polyisocyanurate (ICU) foams SS-0011/0012 from Witco, an ICU foamfrom Uniroyal, and an ICU foam developed by NASA-Ames. Table III

11

summarizes the results of these tests. Excellent fire protection(up to 10 minutes) was provided with 2 inches of NASA-Ames foam.Such foams can easily be applied by spraying or pouring andprovide a convenient means of insulating an aircraft. Also, thefoams are self-supporting and need little or no reinforcement tohold them in place. They appear t(, provide many of the desirablerequirements of a wall thermal protection system.

The formulation for the NASA-Ames ICU foam is shown in Table IV.

Fire Protecting Panels and Composites

Table V shows the results for some of the most effective compositesthat were evaluated. Combinations which provided optimum performanceconsistent with low smoke and low toxic or noxious gas levels aswell as economy involved the BASF fire plate in combination with eitherinorganic insulation or Witco or NASA-Ames polyisocyanurate (ICU) foam.

Window Areas

A number of readily available window materials were evaluated by

exposing them directly to the flame in the test furnace. It wasquickly determined that helicopter windows are a weak point in thestructure and that the protection of windows from fire would requirea full-scale development study.

Table VI summarizes pertinent results on window materials. Thistable also shows the results of attempts made to examine varioustechniques and approaches that may be used to provide adequate thermalprotection. Working jointly with BASF, a very promising panelconsisting of 1/8-inch acrylic - 1/8-inch transparent sodium silicatehydrate - 1/8-inch acrylic was developed which withstood the testfurnace for 9 minutes. The surface temperature at the end of thetest was about 15C. Figures 6 and 7 are photographs of the twosides of the window after the 9-minute exposure to 35,000 Btu/hrsq ft heat flux.

Ventilatiou Openings

Since a helicopter is an air-breathing aircraft, ventilation openingsare provided which draw air from the exterior. In the case of a fire,such openings could provide a path for fumes and vapors to enter thepassenger compartment. Two systems which could plug the vents whenthey are exposed to heat were evaluated. One of these involvescciting the vent interior wall or a screen inserted in the vent withan intumescent paint or mastic. When exposed to high temperatures,the coating intumesces and effectively plugs the hole. In anothercase, pieces of BASF fire plate were inserted in the tube in such away that when the fire plate expanded, it completely filled the tube.Both of these approaches work quite well since the tube itself is notexposed to the erosive eftects of the flame.

12

(0 0O0 p0 p.J.4 U)4 -H -H 0 ~- -4 WDw. 41 p44 040 V)C4~4Cd co 0

rn1 .,CDD 4J . D 0C

-0 44 4) (v ,4

u 0 p

0 ca

-4 41 0H

CcD 41 C

4.41

0 ED CD

Hn 1.4 m

~~4 CD1

0L** 0r 0

411C-4

4-1 14 0 0 0

F- 1-U0 w -

0

13.

$04~ 0- P

C"H 40H -4 > *mp

~ 0

C4.0Ci0>~4 W 'd44 *

41 ... '04( 0 C 424..m0 0-M- 0d r. E u4- 1 4 H .41 m a"H Q

C ~ ~ W 0d-C aJ--~ 4- 2 00 (0 Qm 1 w4H4C E 0~ -41 0

UU

-a) 0 -r-,4 .00 10 VN 00 ej Ln V)

-~ 44

u m0-4 44a)

C4 ) (1) 0 w4 V) J)(

(3) u. - ~00 H (

00

0 . 0 CDw 0 0

-44 00 .1

(44

-. 44

44) u m.40 0 0 0

-4 I

4-~~ u- r-00r)

Q0 0 1

4- C w

14 C -.1004

144

TABLE IV. NASA-AMES ICU FOAM FORNULATION

Components Parts by Weight

Niax Polyol 34-45* 100

Potassium Fluoroborate 52 Mixture A

Zinc Oxide 52

Freon li** 100Mixture B

DMP-30 70M

Mondur M.R.tt 400 / Mixture CSilicone Fluid L-5340"

8 )

Mixture A is milled to 5 N.S. fineness then mixed with B.Foaming occurs when A-B mixturu is added to C.

Union Carbide Corporatiou

** E. I. duPont de Nemours & Co. (Inc.)

t Rohm & 11ais Corporation

t! Mobay Ciemical Corporation

15

44~

C)0 a) 0W) ~ W O ) 0 0()

oC) 004 440 44- 0~ >) 044( 44 w 44 44 CC 0w 4

1-44 g)( 4j000 0)- 0 ~ 4

(U 0r 0) cu 40C 0 0 ,- aj (n9 40C) C ) U) 4) .C 0) v r.C Q) C)

-! )C4 Q)t 0)5 co1 m) 4 - Q4~ ) -H 44.4c

0 Wc U u 4 0 4-5 0

00 0 x x 0 0 x 00 0 > z44 '1 9 a

.,4 .0k. v1H r d ,-4 1-. -, 'n--H 10

It~ V) 1- ))V

m'0 0 Q~

4A

P-1

0 J

In 0 0

0 0

z H V

00

- 0 16

TABLE VI. FURNACE TEST RESULT5 ON WINDOW MATERIALS

Material Description Burnthrough Time (Min:Sec

Plexiglas GM, 1/8" (Rohm and Haas) 1:08,0:54

Plexiglas SE-3, 1/8" 0:48

Plexiglas GM, 1/4" 1:48

Plexiglas GM, 1/2" 2:48

Plexiglas SE-3, 1/2" 2:57

Plexiglas GM, 3/4" 5:38

Abcite-coated Lucite 1/8" (duPont) 1:12

Plexiglas GM, 1/8" treated 0:58

with Kellogg Urethane Varnish

Plexiglas SE-3, 1/8" treated with 0:54

Kellogg Urethane Varnish

Merlon Polycarbonate, 1/8" (G.E.) 0:46

Merlon Polycarbonate E-297, 1/8" 0:48

Two 1/8" Merlon Polycarbonate sheets 1:13 (ist layerwith 1/2" air space fell in

0:44)

Two 1/8" Merlon Polycarbonate sheets, 1:42 (1st layer1/2" air space filled with Gillette fell inshaving foam 1:06)

Two 1/8" Acrylic with 1/8" BASF (Removed afterspecial transparent plate between 9 min. No

burnthrough of

the top Plexi-

glas layer)

17

Interior Partitions

Since present windows are the weakest points in a helicopter, thecrew cockpit would be a highly vulnerable area in a postcrash fire.Not only is the probability of the window breaking high, but thelarge surface area of windows exposed to the fire presents a seriousfire hazard to the crew.

The only solution to this problem appeared to be an emergency doorthat would be accordion-folded against the ceiling of the CH-47 orthe sidewalls of the UH-lD and clamped shut in case of postcrashfire after the crew had climbed into the passenger compartment.Panels consisting of an aluminum sheet (.03-inch thick) on thecockpit side, a 1-inch layer of Dynaflex and 1/16-inch BASF firepla:e on the passenger compartment side provided adequate protectionwhen tested on the furnace.

18

Figure 6. Photograph of Interior Plexiglas ofWindow After 9 Minutes of Fire Exposure.

Figure 7. Photograph of Side Exposed to Fire(Exterior Plexiglas Layer Had Been Burnt).

19

SELECTION AND APPLICATION OF PROTECTIVE WALLS

SELECTION CRITERIA

In selecting the most promising wall materials for the protection of thehabitable compartments of helicopters against in-flight or postcrashfires, the following factors were considered:

* The wall system must provide sufficient thermalinsulation to prevent the penetration of heat

from the fire into the habitable compartment.

s The wall must maintain its structural integrityafter aluminum skin Failure.

* The materials employed should generate a minimumof toxic or irritating gases and smoke in thecabin when heated by the fire.

e The materials must be easy to apply to the wallsof the helicopter and must be easy to maintain

and repair.

e The materials used must be relatively light andinexpensive.

* The wall system should provide access to wiring

and tubing along the walls and to other servicecompartments.

e If applied to the helicopter exterior, the systemmust pass certain military specifications re-lating to weatherability and erosion.

The results of the furnace test program indicated that protective paintsand coatings applied to the exterior of a helicopter generally performedpoorly. The only promising one (Larodyne) could not be applied in place,

was relatively expensive, did not show reliable performance, and doesnot pass military specifications for exterior coatings.

By using an interior protective system, one generally expects the ex-

terior aluminum skin to be sacrificed during the fire. This is partic-ularly important it the protective materials generate toxic or irritatingproducts. The absence of the skin allows these gases to propagate into

the fire rather than into the cabin.

The test program showed that a number of composite material systems maybe used for interior wall protection. Only those offering the optimumcombination of desirable qualities were selected. Table VII identifiesthe basic composite wall systems used and their cost and weight penalty.

20

As indicated, helicopter windows are presently one of the weakest pointsin a helicopter crash fire. Current materials technology does not offera better replacement for the I/8-itch Plexiglas window presently used.One could increase the life of the hAelicopter windows significantly byincreasing their thickness. Indeed, the test program showed that a3/4-inch thick Plexiglas window appears to provide the necessary pro-tection required during a postcrash fire. But the weight penalty wouldbe high, particularly on the UH-lD where the window surface area isvery large in comparison with the total wall area. Although this studywas not intended as a developmental program for helicopter windows, testson the simple approaches to window design that were considered showedthat one should be able to design a lightweight window system from exist-ing materials that is capable of providing the necessary fire protection.

COMPATIBILITY WITH THE HELICOPTER

One of the most ilnportant considerations in choosing one system overanother is its compatibility with the particular helicopter to be pro-tected. There are marked differences between the basic structures ofthe CH-47 and the UH-ID. The walls of the CH-47 have ribs and formerswhich support the aluminum skin. Furnace tests on typical sections ofthe CH-47 wall showed that these ribs and formers retain their integrityduring the fire even though the skin melts within a few seconds. On theother hand, for the most part, the UH-lD consists of flat walls, windows,and doors, the latter consisting of a double aluminum wall. Ribs andformers are only used in the ceiling and the floor. Thus, the approachto protection was not the same for the two helicopters.

WALL SYSTEMS USED IN THE CH-47

Of the materials and composites which were evaluated, two stood out as

providing the optimum combination of properties for wall protection.Both of these systems utilized BASF fire plate as interior panelingwhile either an inorganic insulation, such as Dynaflex, or an organicfoam, such as an ICU foam, was used as additional thermal insulation

in the walls behind thR tire plate. The inorganic Insulation blanketwas used in areas containing wiring, piping, etc., where the panels mighthave to be removed for servicing of mechanical, hydraulic or electrical

components. Where no such components were present, the ICU foam wasnot only easier and faster to apply than mineral insulation, but wasalso capable of filling small cavities and cracks and was less expen--ive. Laboratory tests had shown that neither the TCU foam nor theDynaflex were likely to liberate toxic fumes into the cabin since theBASF panel provided a seal which prevented the fumes from reaching theinterior.

commercial foaming company was hired to spray the foam in place usingoff-rh9-shelf equipmen,. The only pioblem that arose during applicationof the foam was the settling of the pigmeits in the polyol portion of theMiX. Installation of an agitator on the feed tank solved this problemeasily.

21

TABLE VII. COST AND WEIGHT PENALTY OF SELECTED WALL SYSTEMS

Fire Plate/Dynaflex

Component $/ft 2

Fire Plate (1/16") .75

Dynaflex (2 layers-2" each) 4.80

Fasteners .05

Joint Strip .02

Cost 5.62

Estimated Weight, 2.27 lb/ft 2

Fire Plate/ICU Foam

Component $/ft 2

Fire Plate (1/16") .75

ICU Foam (4") .60

Joint Strip .02

Fasteners .05post 1.42

Estimated Weight, 1.31 lb/ft"

Fire Plate/Dynaflex/ICU Foam

Component $/ft2

Fire Plate (1/16") .75

ICU Foam (2") .30

Dynaflex (2") 2.40

Joint Strip .02

Fasteners .05

Cost 3.52

Estimated Weight, 1.79 lb/ft2

22

With the exception of the floor space, areas that were protectedsolely with ICU foam and fire plate were foamed up to the heightof the formers. The floor space was filled to a depth of about4 inches with foam and no fire plate was used there. Other wallareas which were to be protected with the combination of ICU foamand Dynaflex had only 1 inch or so of foam applied behind thewiring or tubing--then a layer of Dynaflex and finally fire plate.Other areas were not foamed since these were designated to beprotected solely with Dynaflex. All windows were sealed withaluminum (.03 inch) and treated as part of the wall. After foaminghad been completed, the foam was allowed to cure for at least 1hour. Excess foam was then scraped to the height of the formersbefore application of the fire plate panels.

The fire plate paneling was attached to the formers with pop rivetsMetal reinforcing strips were used to prevent the rivets from pullingthrough the paneling. Panels were overlapped at all joiuts. Theuse of rivets in the sealing strip restrained the panels in such away that when heated they would intumesce to provide a tight seal.The fire plate panels were bent to conform to the curvature of theformers at the wall-to-ceiling junction. A door was constructedbetween the front cabin and the rear compartment. The door consistedof a layer of aluminum on the cockpit side, I inch of Dynaflex anda layer of fire plate on the inside.

WALL SYSTEMS IN THE UH-ID

The systems used in the CH-47 could not be utilized in the UH-lDhelicopter. As prevJously mentioned, the UH-ID has a large windowsurface area. The doors consist of a double wall of aluminum, andmost of the UH-lD walls lack the formers and ribs which are presentin the CH-47. All the side window areas were covered with aluminum.ICU foam was then poured into the floor space to the level of thefloor formers. The floor and wall areas were then covered with alayer of fire plate. An aluminum wall was installed between thecockpit and the passenger area, and it too was covered with fireplate. Certain areas of the exterior of the helicopter were coveredwith Albi Clad 89-X mastic, while other areas were painted withFirehold 10 intumescent paint. Some were left uncoated. Thesecoatings were used to test the erosive effect of the postcrash fire.

UH-ID In-Flight Fire Protection System

ror the tests simulating in-flight fires, two compartments in therear of the UH-lD helicopter were used. These compartments had acommon wall with the habitable compartment and were approximately11 x 26 x 33 inches in dimension. One rear compartment was

23

protected with 1/8-inch Albi Clad 89-X mastic. This material wasbelieved to perform well in an enclosed area where the erosiveeffect of the fire would be small. The mastic was applied by brushto the bottom and sides of the compartment. A similar compartmenton the opposite side of the helicopter was protected by lining itswalls with BASF fire plate. The fire plate was fastened to thewalls with pop rivets. No metal reinforcing strips were used inthis case.

24

FIRE TEST PROGRAM

Four tests were conducted: two representing an in-flight firein a compartment adjacent to the habitable compartment of theUH-lD helicopter and two representing postcrash fires engulfingthe two helicopters (UH-lD and CH-47) that were provided. Ageneral description of these tests is given below. Also given aredetails of the measurements made and instrumentation used. Resultsof these tests are discussed in Results (p. 45).

GENERAL DESCRIPTION OF TESTS

In-Flight Simulation Tests

The UH-ID helicopter appeared to be the more suitable for theconduct of these tests. The aft section had several compartmentsthat could be used to simulate a fire due to a fuel or hydraulicoil line break. These were the first tests conducted. Not onlydid they provide valuable background for the conduct of the largerscale tests and for checking the instrumentation system, but theyalso allowed any damage to the UH-lD to be corrected before itsfinal total engulfment test.

To simulate a fuel line break fire, JP-4 was pumped at 10 lb/minthrough a nozzle at the end of a copper tube into one of the aftcompartments, ignited and allowed to burn for 5 minutes. The firsttest utilized the compartment protected with Albi Clad 89-X masticwith the top of the compartment completely open. The second testinvolved the fire plate lined compartment. In this case, the topof the compartment was partially covered, leaving an opening 2inches wide along its 26-inch length.

Figure 8 shows the compartment that was protected with BASFfire plate. Figure 9 shows the fire in progress in the compartmentprotected with Albi mastic.

CH-47 Postcrash Fire Test

The third test involved the CH-47. The protected instrumentedhelicopter was transported to the test site (L. G. Hanscom Field,Bedford, Mass.) and placed In an earthen dike about 50 feet indiameter. This area was assumed to represent an average fuel spillarea in a crash of a CH-47. After making the necessary connectionsfor instrumentation, gas sampling, video-observation, interiorlighting, etc., 1100 gallons of JP-4 (nearly a full load) werepoured into the dikecd area and ignited. The dike floor wassaturated with water from previous rain and was uneven. Some fuel

Figure 8. Second In-Flight Fire SimulationCompartment With Partial Cover Down.

Figure 9. First In-Flight Fire Test in Progress.

26

was trapped underneath the helicopter. Inclement weather conditionsprevailed, with rain, sleet and snow falling throughout the testperiod. Wiad of 10-20 mph was blowing along the length of thehelicopter from the cockpit toward the aft section. Figures 10-13

show the progress of the CH-47 postcrash fire.

UH-lD Postcrash Fire Test

For the final test, the protected and instrumented UH-lDhelicopter was transported to L. G. Hanscom Field and placed in a20-foot diameter earthen dike. The dike was level and saturatedwith previous rain water. Figures 14-17 show the progress of theUH-lD postcrash fire. Excellent weather conditions prevailed. Windwas approximately 4-6 mph and blowing across the helicopter in aleft to right direction.

SAMPLING AND MEASUREMENT OF TOXIC GASES AND SMOKE

General Discussion

Although substantial increase in crew safety was anticipated by theuse of the candidate fire protection systems that were evaluated inthe laboratory test program, the use of the candidate systemsintroduced a potential hazard to the crew from exposure to toxicgases which might be generated when the protective materials wereexposed to fire or when there was a partial failure of the system.

The purpose of the tests which are described in this sectionwas to provide a semi-quantitative indication of whether or not apotentially hazardous situation might arise in an actual firesituation. The accurate assessment of toxicity in situationssuch as this cannot be made by chemizal measurements alone. Infor-mation on the short-term (10-minute exposure) toxic hazards associatedwith many of the gases which might be evolved is not generallyavailable for humans. It is also impossible to account accuratelyfor the many synergistic effects on toxicity which can arise whena wide variety of toxic species are present, especially whencompounded with thermal effects. A good assessment of toxicityin cases such as this can only be achieved through well-controlled

and supervised tests involving animals. An evaluation of thatmagnitude was neither warranted nor desired in this initialfeasibility study. For those reasons, the sampling and analyticaltechniques described below were designed to provide initial insightinto the problem of potential toxicity and at the same time tominimize unnecessary complexity and excessive cost. Toxic gas andsmoke measurements were made only in the postcrash fire simulationtests. It was assumed that toxic gases and smoke generated duringin-flight fire will not affect the habitable compartment since thefire occurs in another compartment and because the doors can beopened to provide fresh air to the occupants.

27

rOWA-

Figure 10. CH-47 Before Postcrash Fire Test.

Figure 11. CH-47 Fire in Progress Soon After Ignition.

28

Figure 12. CH-47 Fire at Maximum Intensity.

Figure 13. CH-47 Cockpit Near Conclusion of Test.

29

Figure 14. 131-iD Before Postcrash Fire Test.

-".INFO

Figure 15. UH1-ID Postcrash Fire After 30 Seconds.

30

Figure 16. Fully Developed UH-ID Postcrash Fire(3 minutes).

Figure 17. UH-IlD Fire Near Conclusion (8 minutes).

31

Toxic Gases of Concern

Of primary concern were the amounts of the fire gases, carbondioxide and carbon monoxide, which tend to build up in the cabininterior. Oxygen depletion as a result of burning or slow oxida-tion within the cabin was also of interest. Concentration oftoxic gases which can arise from the decomposition of the insula-ting materials as well as from pyrolysis and combustion of fireretardants incorporated in these materials was also to be measured.

The primary toxic species which could arise from the isocyanurate(ICU) foam used in these test., were the nitrogen-containingcompounds -- HCN and NH Halogen-containing materials such as

potassium fluoroborate and halon 113 (CCI 3F) are employed in thecandidate isocyanurate foam to impart flame retardancy. Whenexposed to fire, these materials produced BF3 , free halogen, andhydrogen halides. Since these materials are of roughly the sametoxicity, sa.ples were taken and analyzed for the amount offluoride, chloride and bromide ions present.

Pyrolysis ef the isocyanurate foam also produced a variety ofhigher molecular weight, nitrogen- and oxygen-containing organiccompounds of widely varying toxicity. These materials can appearas gases or as a thick smoke (condensed liquid aerosol), as wasobserved in certain of the earlier furnace tests. These heaviermaterials are produced because much of the pyrolysis to which the

interior cabin is exposed will occur, at least initially, atrelatively low temperatures which favor formation of heavierdegradation products. Relatively higher molecular weight amines,aldehydes, ketones, acids and unsaturated hydrocarbons can beenvisioned. Because of the large number of possible compoundsand the uncertainty as to which compounds would indeed be formed,individual tests could not be conducted. However, in conjunctionwith the taking of samples for measurement of the several gasesof particular interest, samples of smoke that arose were collected.These heavier molecular weight materials were simply weighed toarrive at average particular concentration over the sampling period.

Sampling Procedures

Carbon Dioxide, Carbon Monoxide, and Oxygen

Samples of these gases were taken at two points near the center ofthe CII-47 helicopter cabin: the first at 1 foot above the floorlevel, and the second at 4 feet above the floor. Two sampleswere taken at about the 2-1/2-foot level of the UH-lD. A gasstream for sampling carbon dioxide, carbon monoxide, and oxygenwas withdrawn from each sampling point through a separate length of1/4-inch O.D. copper tubing which extended from the cabin interior

32

passed through an insulated service umbilical through the firezone to a protected area about 70 feet from the fire zone. Oneliter grab samples were taken into a glass sampling bulb from thisgas stream at 3-minute intervals during the test.

The design of the pumping and sample receiving station is shownschematically in Figure 18. A limiting orifice between thesampling bulb and vacuum pump was used to control the flow ratethrough each sampling line at approximately 3 liters per minute.The operation of the grab sampling system is described with respectto the operation of circuit 1. The operation of circuit 2 wasidentical. At the start of the test, stopcocks A and B (attachedto the sampling bulb) were opened and a gas flow of 3 liters perminute from sampling point I was allowed to sweep through flaskP. At the end of 3 minutes, the initial air that had been inflask P was completely swept out by a gas sample that was repre-sentative of the atmosphere at sampling point 1. At this time,stopcock B was closed, and, after waiting a few seconds for thepressure in flask P to equilibrate at 1 atmosphere, stopcock Awas closed and stopcocks C and D opened to begin acquiring thenext sample in flask Q. While the gas stream was sweeping flaskQ, flask P was replaced by a new flask in preparation for theacquisition of the third gas sample, and so on.

Reactive Gases

The other gases of interest, NH , HCN, and the halogens and halidesare more reactive and would hav a strong tendency to be lost onthe surfaces of long sampling lines if one attempted to withdrawa sample from the cabin to the sampling station as planned forCO , CO, and 02. For this reason, gases were trapped in 25-mlmiiget impingers in which the reactive toxic gases of interestwere absorbed into appropriate reagent solutions. The impingers,each containing 15 ml of reagent, were located in the cabin inclose proximity to the sampling point. They were installed priorto the test, actuated just before the test began, and retrievedfrom the cabin after the test had been completed.

A set of two impingers was required for each of the two samplingpoints. An impinger pair along with its associated filters andvacuum service, is shown schematically in Figure 19. Forcollectin, NH3, and other amines and basic species, one of the twoimpingers was filled with 0.lM sulfuric acid. The second impingerwas filled with O.lM sodium hydroxide for trapping the acidicspecies: HCN, halogens, and halides. To prevent any particulatesfrom entering the impinger and to obtain a sample of the particu-lates for subsequent analysis, each impinger was equipped with adual filter assembly composed of a drying tube filled with glasswool followed by a conventional paper filter. Both filters were

33

ILat0

0a.

C-))

> 0-

v-4

C 0-)

C-a)

34

preweighed so that a second weighing after the test permittedan estimate of the particulate concentration in the cabin to bemade. Individually calibrated orifices were installed after eachimpinger to control the gas flow through each impinger at a rateof 3 liters/min.

Gas Analysis Techniques

Carbon Dioxide, Carbon Monoxide, and Oxygen

These three gases were measured by temperature pVogrammed gas/solidchromatography using columns of molecular sieve on a gaschromatograph. A helium carrier gas flow rate of about 80 ml/mmnwas used; detection was via a thermal conductivity unit maintainedat 175%C. The molecular sieve column was 2.5 feet x 0.25 inchdiameter packed with 100/110 mesh 5A* molecular sieve. This columnwas held at 60*C for the first 6 minutes of the analysis, thenheated ballistically to 170 0C at a rate of about 15*/minute andfinally held at 170*C until the CO2 peak emerged.

A 3-ml sample was withdrawn from the flask containing the gassample and injected into the chromatograph. This sample size gavea detectability for each of the three gases of about 0.05-0.1%.

Calibrations were accomplished using appropriate gas mixtures con-taining CO, CO, and 02 in He which were procured from a commercialvendor. Tte composition of the calibration gases was certifiedby the vendor to be accurate to better than + 5%.

Hydrogen Cyanide

The cyanide content of an aliquot of the sodium hydroxide impingersolution was measured by using the colorimetric method for cyanideas described in the Standard Method for Examination of Water(American Public Health Association, 1960). This method involvesthe reaction of cyanide with Chloraminc-T and then with methyl-phenylpyrazolone to yield a blue color which is measured byspectrophotometry. For a 30-liter gas sample (3 liters per minutefor about 10 minutes), this analytical method for a cyanide had asensitivity of better than 0.1 ppm HCN in the test atmosphere.

This method was standardized using accurately prepared solutions-I reagent grade sodium cyanide.

Ammonia

w.u wi_,onia analysis was performed on an aliquot of the sulfuricid iiipinger solution. The measurement utilized Chloramine-T

k_ ''s methylphenylpyizolone color reagent similar to that used

35

F

L

ORIFICE R

-- 0.1 M H2 SO4 SAMPLINGPOINT

TOVACUUM

SERVICE

Figure 19. Interior Sampling System for Reactive Gases.

36

for the cyanide analysis. The pcocedure is described in"Colorimetric Determination of NLn-Metals," D. F. Boltz, Editor,Interscience (1958). For a gas sample volume of 30 liters, thismethod was again sensitive to less than 0.1 ppm of ammonia in thegas sample. Standardization involved measurements of calibratedsolutions of analytical reagent grade (NH4)2So4 in distilled water.

Fluoride Ion

The analysis of fluoride was made with a solid-state fluoride ionelectrode. A 2-ml aliquot of the caustic implnger sample waspipetted into 10 ml of 15% sodium acetate solution to providethe proper pH and ionic strength. The fluoride elect-de has a

selectivity for fluoride over a chloride, bromide and carbonate ofat least 1000. Calibrations were carried out and checked afterevery 5-6 samples using sodium fluoride standard solutions dilutedin the same way as the samples. This technique was capable ofdetecting fluoride ion at less than 1 ppm and has a relativeprecision of approximately + 5%.

Chloride and Bromide Ions

A silver/silver bromide electrode (prepared by electrolyzing silverwire in dilute Hbr) was used for the direct measurement of bromideion in the NaOH impinger liquid. The selectivity of this methodfor bromide was 300-400 over chloride and greater than 1000 overfluoride. Only Lodide and sulfide ions could interfere if present,but these were not expected. Calibration was carried out andchecked after every 5-6 samples using fresh analytical 5tandardsolutions of aqueous HBr. The measurement had a relative precisionof approximately + 5%.

CnJ¢cide ion content was measured with a silvei/silver chlorideelectrode prepared in a manner similar to the bromide electrodeused above. However, prior to the chloride measurement, ifbromide was present it was removed by mixing 10 ml of sample with10 ml of 3N HNO , evaporating (without boiling) on a hot plate to10 mi and-then diluting to 25 ml with deionized water for subse-quent measurement. -Pis procedure had been tested on a solutioncontaining 1.1 x 10 M HCl and 1.1 x 10-2 M HBr. After carrying;t the bromide removal procedure, the concentrations of HX leftI-orrected back to the initial 10-ml sample volumes) were 1.1 x 10 - 2Y HC1 and <5 x 10-6 M HBr. This treatment should also removeiodide and sulfide which would be serious interferences if present.'ibration was performed by using fresh analytical standard

iu irns of aqueous nC..

37

Measurement of Smoke

Smoke buildup in the cabin was monitored by two simple photometerswhich responded to the increasing optical density of the cabinatmosphere as smoke developed. Each photometer consisted of a

light source and a photocell positioned at opposite ends of thecabins. A collimated light beam from the source was directed downthe length of the center of the cabin and impinged upon a photo-tube detector in the receiver. The decrease in light intensityreaching the detector as a result of smoke buildup was read out as adecrease in percent transmission on a calibrated millivolt scanner

recorder. For the CH-47 helicopter, two photometer systems wereutilized: one positioned 1 foot above the cabin floor and thesecond 4 feet above the floor. The photocell was 20 feet away fromthe light source. Only one photometer was used in the UH-IDpostcrash fire test. This was placed at the 2-foot level from thefloor with the photocell 4-1/2 feet away from the light source.

The photometer systems were similar to the one used in the small-scale furnace measurements except for modifications in the trans-mission optics to minimize divergence of the light beam over thelonger path lengths involved in these tests.

Each photometer was calibrated immediately before the test by

generating a calibration curve of percent trar&inission versus

recorder reading in millivolts. In addition to the obvious pointsof 0% and 100% transmission, neutral dcnsity filters were used togenerate calibration points at five intermediate transmission values.

THERMAL INSTRUMENTATION

Temperature

Since the air temperature of the habitable compartment of ahelicopter engulfed by flames or threatened by a fire in an adjacentcompartment is of particular concern in determiaing its habitability,

a number of chromel-alumel thermocouples were placed at carefullyselected points inside each helicopter and at strategic locationson the interior walls. Other thermocouples were used to measureflame temperatures outside the helicopter or in the burning com-partment. For the CH-47 p,3tcrash fire test, thermocouples werealso placed in the video camera compartment.

The distribution of the thermocouples used in the UH-ID in-flightfire tests is shown in Figure 20 for one test. In the second test,the thermocouples were placed in symmetrical positions, in theother half of the helicopter. The distributions of thermocouplesin the helicopters during the postcrash simulation fires areshown in Figures 21 and 22.

38

Total Heat Flux

Another important factor in determining the habitability of ahelicopter compartment during a fire is the total heat flux that

may impinge on the human skin. A calibrated heat flux meter was

located at the 6-foot level in the CH-47, facing the roof of the

helicopter. Additional simple calorimeters were utilized to

measure heat flux at other points. These calorimeters consistedof small blackened brass spheres of known mass. A thermocouple

was imbedded in the sphere and the heat flux to the sphere at any

moment was calculated from the temperature record during the fire

and the equation

ATQ = MCp A- qA

where

Q = total rate of heat transfer (Btu/hr)

M = mass of the brass sphere (ib)

Cp = specific heat of brass (Btu/lb*F)

ATAT slope of the temperature time curve ata particular instant in time (*F/hr)

q = heat flux (Btu/hr sq ft)

A = area of sphere

Figures 20, 21, and 22 show the locaticns of these heat flux meters.

PHOTOGRAPHIC AND VIDEO iQVERAGE

Exterior photographic coverage of all four tests and interior videocoverage of the CH-47 postcrash fire :est were carried out. Color

16-mm movie film and 35-mm slides were taken of the helicopterexteriors before, during and after each fire. The interior wasalso photographed before and after each fire.

A wide-angle video camera housed in a well protected enclosure wasinstalled in the aft section of the C -47 helicopter. A 1-inchvideo recorder and monitor was used to tape and monitor the sequenceof events in the interior of the CH-47 during the postcrash simula-

tion fire.

39,

-LU 40

I co X U)co 0 34~

(Au

X4

HH

fX)

000

00

00

CCo

40

04

44

03

I F

ow-1

12 -

Go I

C30

~'0

cldi

004 J

V) r

L a

dli -

410

-Ir

w0

w 41

m 4-i

44 1

A)J

j4

0.

00 N 4

00 OD44

N

42

DATA ACQUISITION

The output of the thermocouples, heat flux meters and smoke detec-tors was channeled to a terminal strip within the helicopter.Lead (extension) cables as well as gas sampling tubes and powerlines for the vacuum pump and light sources were run through a3-3/4-inch I.D. iron tube which extended some 50 feet from thehelicopter interior to a protected instrumentation area. The tubewas insulated to protect its contents from the fire.

The electrical extension cables were two conductor-shielded cableswhich minimized noise from extraneous electrical signals. Thermo-couple and heat flux leads were connected to a VIDAR scanner which

had been calibrated (NBS traceable).

The VIDAR recorded the data on magnetic tape which was read outlater and the data printed out. It also recorded the time at thebeginning of each 3can. In addition to the VIDAR, the output ofselected thermocouples was recorded on a strip chart recorder asa safeguard against the loss of the VIDAR data. Output from the

smoke detectors was recorded on a separate strip chart recorder.

POWER SUPPLY

Power was needed to run three vacuum pumps, four 100-watt lamps

in the interior of the CH-47, the light sources fo, the smokedetectors, the VIDAR scanner, a cooling water circulation pumpfor the heat flux meter, and the video system. A gasoline-drivenpower generator was used after it had been tested for constancyof voltage and frequency and shown to operate satisfactorily withall the recording equipment.

43

RESULTS

IN-FLIGHT FIRE SIMULATION TESTS

Figures 23 and 24 show temperatures of the various locations inthe UH-ID helicopter throughout the two separate in-flight simu-lation tests. These locations have been identified in Figure 20.Figures 25 and 26 show the heat flux variation with time for eachtest as measured by the brass spheres and heat flux meter. Asindicated earlier, both tests were conducted with the doors closedto allow better thermal measurements in the absence of wind.Smoke and toxic gases were not measured in these tests because itwas assumed chat in real-life the occupants would have opened thedoors of the helicopter in case of an in-flignt fire in preparationfor egress after landing or to prevent smoke accumulation.

It should be noted that the heat flux data from the brass sphereis lower than the data from the total heat flux meter. This isbecause the meter was facing the hot wall of the fire compartmentthroughout the test, whereas the sphere could see the rest of thecompartment which remained cold. Thus, the radiant component ofheat flux was greater for the meter than for the sphere while theconvective component was about the same.

Comparison of the data from the two tests shows that there aremarked differences in the behavior of the two fires and in thecorresponding temperatures and heat fluxes recorded inside thehelicopter. It should be recalled that the in-flight fire com-partment used In the first test tied no cover, thus allowing morecomplete combustion of the fuel and higher temperatures to berecorded in both the fire and habitable compartments. The secondcompartment was partially covered, Lhus restricting the flow ofair into the fire compartment. The flame was sootier than in theprevious test. It was also noted that some unburnt fuel hadaccumulated in the bottom of the compartment at the conclusion ofthe second test. Although such a fuel-rich fire was expected toproduce lower temperatures in the fire compartment, it is believedthat the liquid JP-4 spray impinged on thermocouple 1 during thistest, providing an unrealistic low temperature for the fire com-partment.

The first in-flight test which utilized a coating of Albi masticon the walls of the fire compartment resulted in the accumulationof large amounts of white smoke within the cabin. This is believedto have resulted from heating the ICU foam which had already beensprayed in the floor space in preparation for the large-scale testsand from the mastic itself. Also, there were direct penetrations

44

i nu In B~~min - n n I~!JU N~ll i 'e . . .=n=uMInnE

TC1 ATC 2 0

1200 - TC.3 0 . .TC 4 V -j§TC 5, 6 0 A"

TC 7,-8 A A

A1000 - -

tlo

I ODO

BOO

00

0

.... . o.. .. --- o -- - - .. .

A

0C 00

0 0 0

1 A V 4A0- - --- - V - -

0 100 20300Tune (conds)

Figure 23. Temperatures Recorded During theUH-iD First In-Flight Fire Test.

45

TC 1

TC 2 0

700 -- -- --- . . . . . .TC3 TC 4 V

TC 5,6, 7, 8

600 ---

A

5001 .... ....................----- - -

400----

E

300 .. .... . . .. .. .

200---- A- -- A

A A 0 0

100--_-

0 0 0 0 0 V v

0 100 200 300Time (seconds)

Figure 24. Temperatures Recorded During the

UH-1D Second In-Flight Fire Test.

46

800

Heat Flux Meter 1

Brass Sphere A

700

0 00

600 . .............. ..

5004--_ _ AA §S ~ j

4826400O0

o0

- 0

0

U 0 0

-100

0 40 80 120 160 200 240 280

Time (seconds)

Figure 25. Heat Flux Data Within the UH-ID CabinDuring the First In-Flight Fire Test.

47

60

Heat Flu', Meter 0Brass Sphere A

050

40 -,n -©..

000 0 0

-o 0 oZ( 0 ) o

0 0 0 0 0 00

-20

0 00

AA

A A

- 10 1 1.......1) 40 80 120 160 200 240 280 320

Time (seconds

Figure 26. Heat Flux Recorded in the UH-ID Duringthe Second In-Flight Simulation Test.

48

between the left compartment, the floor space and the habitablecabin. No such penetrations were present in the second case whenBASF fire plate was used.

In both tests, air temperatures (and temperature rises) and heatfluxes recorded within the habitable cabin were far below humantolerable levels. One of the authors remained within the heli-copter with the doors closed for the 5-minute duration of thesecond test without any discomfort.

It is concluded that in-flight fires can be protected against bylining the walls of potential fire compartments with intumescentmastic coatings (e.g., Albi Clad 89-X) or with an intumescent in-organic panel (e.g., BASF fire plate).

CH-47 PCSTCRASH FIRE

Thermal Measurements

Temperatures taken at the various locations identified in Figure 21and as recorded on the Vidar are plotted in Figures 27, 28 and 29.Heat flux measurements made with the brass spheres and the heatflux meter are shown in Figure 30. It should be noted that theheat fluxes recorded by the heat flux meter in this case were belowthose of the brass spheres at the same position. This is becausethe meter was pointing at a cooler spot on the wall while thesphere was receiving radiation from all hot walls.

It is noted fr m these plots that the duration of the data is only160 seconds*, when the fire actually proceeded for more than 20minutes. The reason for this is that power was lost due to ashort, in the line leading to a water circulating pump placed withinthe helicopter for cooling the heat flux meter.

By the time the shorted line was disconnected (after a lapse ofabout 5 minutes) the fire was of such intensity that the aluminumfoil/glass wool mineral insulation on the iron pipe protecting thelead cables and thermocouples had melted and the PVC insulation onmost thermocouple extension wires within the pipe had melted, thusshorting these leads.

The cause of the power failure could not be identified until thevideo tape recording of the interior was played back in slow motion.The tape showed that early in the fire an explosion had occurred

* Since the Vidar was set to record data every 10 seconds, themaximum duration of the data may have been 180 seconds (3 minutes).

49

200 __l_ _

TIC 1 ATC 20

TC 3 0

180 TC4 V

TC 5 0TC 6 0

160

140

120

EI-

100

60--

40 20 40 680 100 120 140 160

60 8r 160

Time (seconds)

Figure 27. Temperatires Recorded in the CH-47Cabin During the Postcrash Fire Test.

50

180

TC7

TC8 A

Tc 9 0160 TC10 3

TCll V

TC12 *TC13 0

140

120 t

100 __ _ -Cp-

0

1780 0_ I

60__

4

0 20 40 60 80 100 120 140 160

Time (seconds)

Figure 28. Temperatures Recorded in the CII-47 Cabin Ouringthe Postcrash Fire Test.

51

10O0

900

800 - +-

700- .

So -- -- - -

c -

E

I-I400- --- A--A

300 - ,

200 -

1 - - - ,

1t

TC 14 *'TC15 A

U 40 80 120 160Time (seconds)

Figure 29. Temperatures Recorded by ThermocouplesExposed to the CH-47 Postcrash Fire.

52

2000

Brass Sphere 3B 0Brass Sphere 8B 0' (2527)

180 - Brass Sphere 10B A --

Heat Flux Meter 0

(Near Brass Sphere 3B)

o1600 . ..... . ... ........ ....

140 0 . . . . . .. .... . .. .... ... ...

1200 ---- ---------

I A A

0 A

~1000- V

80 0--0

-----00----_

LA 0 C>

0 - ___ -

00

o II o0-- -

40 -- t - ---- -_ _ _

2W->

0 4k

0 _ 0 0-20C: .. . . .. .

0

0 20 4 12 140 16(

Figure 30. Total Heat Flux Recorded During theCH-47 Postcrash Fire Test.

53

within the helicopter. By back-timing the videotape from themoment power was lost, it was found that the explosion must havetaken place within 15.7 seconds from ignition. An aluminum ribbedpanel had been used to cover the floor space where the cargo doorhad been sealed and foamed to a height of about 4 inches, leavingan air space of about 7 inches. Apparently, fuel vapor hadpenetrated into this air space between the foam and the panel andeventually ignited. Figures 31 to 33 show the position of thisspace, the panel, and selected scenes from the video sequence.This floor space continued to burn after the explosion and was theonly position within the helicopter that required extinguishment

by the standby fire fighters. The electrical wiring leading tothe pump motor was immediately above this space and is believed tohave burned because of the fire there.

Examination of the temperature data during the first 160 secondsshows that the cabin air temperature (thermocouples 5, 6, 7) wasnot excessive (<200*F) and by itself would not have been fatal.The heat flux cat- as recorded by the brass sphere near the ceilingwas above tolerable limit for humans (60O-800 Btu/hr/sq ft). Themaximum heat flux level recorded (2527 Btu/hr/sq ft) was far belowthe level ('10,000 Btu/hr/sq ft) necessary to ignite cellulosicmaterials whir may be present in a helicopter. Although the totalheat flux level would have caused serious burns to the exposedskin surfaces of occupants, simple shielding of the hand. and headwould have prevented such burns. It is difficult to predict thermalconditions beyond the first 160 seconds because of the lack ofdata. It can be assumed, however, that temperatures and heatfluxes increased after that time and may have reached levels whichwere harmful to occupants.

Smoke and Toxic Gases

Figure 34 shows percent light transmission across a 20-foot pathversus time. It can be seen that visibility dropped sharply afterthe first 20 seconds. Light transmission dropped to 50% within50 seconds at the 4-foot level and 63 seconds at the l-foc level.The video recording showed that the smoke developed immediatelyafter the explosion.

Table VIII summarizes the concentrations of oxygen, C02, CO, andother reactive species generated in the CH-47. These results showthat while oxygen concentration remained adequate, the concentra-tion of the reactive species Fl-,Br-, CN-, and CO were above thethreshold limit values (TLV) for 8-hour exposure as adoptedby the American Conference of Governmental Industrial Hygienists(ACGIH). These concentrations are not necessarily lethal for theshort duration involved. The major offender in this test was the high

concentration (average of 350 mg/m3 ) of particulates in the cabin air.

54

Figure 31. Cargo Door Where Explosion Occurred.

Figuie 32. Cargo Floor Space With Aluminum Ribbed PanelUsed as Cover(Damage to Walls From Firefighters'Water Streams).

55

samplingprobes

a

smokedetector

b

Figure 33. Sequence of Explosions in CH-47 Posterash Fire.

56

C

d

p anel

Figure 33. (Continued).

57

100_" ____ ,

90 i ... ...

Bottom Smoke De:ector 0

Top Detector A

80 . . . .. . . .. . .. - - - - - -I

0

70 . .... . ........

-

O

~60 -

40 0

AA

I0

20 - -,I - -

0

0 20 40 60 80 100 120 140 160 80

Tpme (seconds)

Figure 34. Light Transmission at the 4-Ft and 1-Ft Levels

During the CH-47 Postcrash Fire Test.

58

Even though nothing is known about the synergistic effects of thespecies identified, especially when combined with the temperaturelevels recorded, it is felt that an occupant of the CH-47 heli-copter could have survived this fire for the first three minutespossibly with 1st or 2nd degree facial and hand burns. Theunavailability of data beyond that time does not allow furtherpredictions.

CH-47 Integrity After Fire

The helicopter walls were inspected carefully after the test.Figures 35-41 show various scenes of the CH-47 interior and exterior.Figure 35 shows the condition of the interior fire plate panelingafter the fire (the hole was made afterwards). Of particularinterest are Figures 36 and 37 which are of the same locationduring application of the protective walls and after the fire.Dynaflex was used in the lower half, whereas the upper and adja-cent sections were foamed with ICU foam. Both sections werepaneled with BASF fire plate on the interior. During the fire,the Dynaflex fell to the outside as soon as the aluminum skinmelted. The BASrt panel remained in place until fice fighters usedthat opening to reach the floor space fire in the interior nearthe end of the test. The overall photographs of both sides of thehelicopter (Figures 38 and 39) also show that it remained fairlyintact except where Dynaflex was used or where the foam had norib support. The cracks in the charred ICU foam (Figures 40 and41) were not deep enough to reach the interior. There is no doubtthat if ICU foam could have been used on all sections of the heli-copter walls, the conditions within the helicopter would have been

m@uable for a longer period of time.

59

TABLE VIII. TOXIC GASES AND PARTICULATES GENERATED IN THE CH-47 FIRE

Concentration

Reactive Species* 4-ft level 1-ft level TIV**

Fluoride (Fl-) 10 ppm 1.6 ppm 3 ppm

Chloride (CI-) 5 ppm <2.D ppm 5 ppin

Bromide (Br-) 8 ppm 2.5 ppm 3 ppm

Cyanide (CN) 13 ppm 3 ppm 10 ppm

Ammonia (NH3) 40 ppm 20 ppm 50 ppm

Particulates* 440;170 mg/m3 600;200 mg/m3

Permanent Gasest

Oxygen (02) 20 % 17 %

CarbonMonoxide (CO) .05% .05% .005%

CarbonDioxide (CO2 ) 0.3 % 0.2 % .5 %

* Total time for sample was 10 minutes. Sample collected for 3minutes after ignition until power was lost. After return ofpower, sample collection continued for another 7 minutes.Total volume of gas sampled = 30 liters.

t Sample taken 3 minutes after ignition.

** Threshold limit values for 8-hour exposure.

60

Figure 35. CH-47 Interior Aftor Fire LookingToward Cockpitl Door (Hfole in WallMade by Firefighters).

61

Figure 36. CH-47 Interior Typical WallConstruction Before Fire.

'rVI

Figure 37, CH-47 Interior Typical W4allConstruction After Fire.

62

Fiur 38 Rih Sid of CH4 tr ie

Figure 38. RLgt Side of CH-47 After Fire.

63

Figure 40. Unribbed Section Where ICU Foam Could Not SupportItself (Uncharred Foam Shows on Top Where Char WasRemoved After Fire).

;W I,

Figure 41. Typical Cellular Structure of Charred ICU Foam.

64

UH-lD POSTCRASH FIRE

Therial Measurpments

Because the UH-lD is smaller in size than the CH-47, fewermeasurements were taken during the postcrash fire test. Locations

of the various measurements are shown in Figure 22. Figures 42, 43,

and 44 are plots of temperature versus time for the 8-minute dura-

tion of the test. Unlike the CH-47 test, all data were received

and recorded since the umbilical iron pipe containing all leads

and extensions was better protected. At the height of the fire, a

cooling water stream was directed at the pipe for further protection.

The temperature data show that the maximum cabin air temperature(335*F) was that recorded by thermocouple 2 at 5 minutes after

the beginning of the test. This thermocouple was close to thewall which received the maximum intensity of the fire because it

was on the lee side of the wind where hot vortices were observed.Most other thermocouples recorded their maximum temperatures atabout the same time with the exception of thermicouple 9. Thisthermocouple was placed on the floor above the ICU foam layer, andthus read very low temperatures throughout the fire with its

maximum reading of 120*F occurring at 9 minutes 50 seconds. Thermo-couple 7, which was placed in the center of the habitable compart-

ment, gave a maximum cabin air temperature of 273*F.

The heat flux meter was not used in this test to avoid the needfor a circulating pump for cooling water in the interior (and arecurrencu of the power failure encountered in the CH-47 test)and because this test was expected to cause extensive damage tothe helicopter and its contents. Only one brass sphere was usedand its data (Figure 45) gave a maximum reading of 478 Btu!hr sq ft.The sphere was located near the upwind side of the helicopter (seeFigure 22) where wall temperatures were not as high as those on

the other side. It is expected that a sphere closer to the leewall would have received a higher level of thermal radiation and

convection and would have registered a higher total heat flux.

In summary, thermal measurements indicate that air temperatures

and the heat flux within the UH-ID cabin during the first 3 minuteswere lower than those encountered in the CH-47 fire during thesame period. However, after 5 minutes the cabin air temperaturewas too high for human tolerance and may have caused respiratoryinjury to the occupants.

Smoke and Toxic Products

The variation in light transmission with time during the UH-ID

test is shown in Figure 46. Table TX shows the concentrations of

65

E4

0

4

00

0 0 0 4i

0 0 0 0 r

40 0 C>

40 00

00 40 0

43 0 10 t

0

3D0

I cIc

0 0 V> 0

I H4

66

rn0

E-4

-'.4

P-4

72

C> 04

r,. 00C>

-4

t> 0

jj ) imejdwai-4

670

1800_ _

TC13 ITc 14 * j

1600 - -0r

1400 __ ___ ____ ___ _

60

800 _ _

I

0

200 - 0

2000

0 j

0 40 80 120 180 200 240 280 320

Figure 44. Temperatures Recorded by ThermocouplesOutside the UH-lD Cabin.

68

5000

300 .0

0

0~00

oo 0

0 0 0 00

o Il

0

-300 L0 40 80 120 160 200 240 280

Time (seconds)

Figure 45. Total Heat Flux Measured by Brass SphereCalorimeter in the UH-ID Postcrash Fire Test.

69

toxic gases and particulates generated in the UH-I-D cabin duringthis fire. It can be seen that large amounts of smoke and partic-ulates were generated very early in this test also. The averaleparticulate concentration over a 7-minute period was 1600 mg/m(see Table IX). Figure 46 shows that 50% light transmissionoccurred at 30 seconds from ignition. The high levels of smokecan be attributed to the absence of a tight seal between the doorsand the floor and the interior walls. Furthermore, unlike theCH-47, the UH-ID bottom was raised above the ground by about I foot,which is the height of tha landing skids. This allowed JP-4 toburn underneath the helicopter (until the skids collapsed), thuspyrolyzing the ICU foam in the floor. Because of the high tempera-tures to which the BASF fire -late was exposed in this test,larger quantities of steam u generated within the cabin thanin the CH-47 where fire pla.. was not exposed directly to thefire. The steam generated in the U1-1D may have contributedsignificantly to the reduction in light transmission.

Because a smaller quantity of ICU foam was used in the UH-lD thanin the CH-47, smaller concentrations of the reactive species wereobserved. However, because the IJH-iD was not as air tight as theCH-47, higher concentrations of carbon monoxide and carbon dioxidewere recorded, especially during the first three minutes. Never-theless, the concentrations of the reactive species, CO, and CO2were not considered to be high enough by themselves to pose athreat of injury or fatality to occupants for the short durationof the fire. However, combined with the high cabin air tempera-tures and the presence of irritant particulates, permanent respi-ratory system damage is very likely to have resulted.

UH-ID Integrity After Fire

Figures 47 and 48 show the general condition of the UH-lD afterthe fire. The skids collapsed after about three minutes fromignition. Figure 47 shows the upwind side of the helicopter. Thealuminum sheets covering the windows were protected with Albimastic, while the section just below that was painted with Fireholdintumescent paint. Because of the wind direction, it is difficult to

assess the contribution of these wall treatments toward the survivalof the aluminum wxlls on this side, when the aluminum on the lee(hotter) side was completely melted (cor-are Figure 47 and 48). TheBASF fire plate on the lee side sagged iuecause of the absence ofany formers or ribs. However, penetrations were small and apparentlydid not contribute much towards increasing overall air temperaturesor toxic product concentrations in the cabin.

Of the two compartments that were used to conduct the in-flightfire simulation tests, only the one that was paneled with fire plateiurvived. The unprotected front and aft sections of the helicopter

wt rv ( omp lute ly dest roved in the fire.

70

TABLE IX. TOXIC GASES AND PARTICULATES GENERATED IN THE UH-ID FIRE

Concentrationt

Reactive Species* Probe #1 Probe #2 TLV**

Fluoride (Fl-) 0.9 ppm 1.6 ppm 3 ppm

Chloride (CL-) <2.4 ppm <2.4 ppm 5 ppm

Bromide (Br-) 2.3 ppm 3.7 ppm 3 ppm

Cyanide (C"-) 8 ppm 10 ppm 10 ppm

Ammonia (NH3) 40 ppm 20 ppm 50 ppm

Particulates* 2000, 1700, 800, mg/m3

Permanent Gases Carbon Carbon

Sample Oxygen Monoxide Dioxide

Time Point (02) (CO) (CO2)

3 minutes Prube #J 21% 0.13% 1.6%

after ignition Probe #2 20% 0.16% 1.6%

6 minu e Probe #1 20% <0.03% 1.1%

Probe #2 20% <0.03% 1.8%

9 minuces Probe #1 21% <0.03% 0.3%

Probe #2 21% <0.03% 0.2%

TLV** -- .005% .5%

*Total time for sample was 7 minutes. Totai volume of gas

sampled = 21 liters.

Threshold Limi: Values for 8-hour exposure.

t Sampling probes were close together and near the middle of the

hel. copter.

71

100

90

00

80

070

p60-

.2

.~50- __ __

E0.

-40 - _

030

200_

0 0

100

0l0 10 20 30 40 50 60 70

Time (seconds)

Figure 46. Light Transmission in the UH-ID Cabin

Durin, the Postcrash Fire Test.

72

Figure 47, Upwind Side of UH-lD After

Postcrash Fire Test.

Figure 48. .ee Side of 13W-iD After Test ShowingAbsence of Aluminum Walls and SaggingFire Plate Walls.

73

CONCLUSIONS AND RECOMMENDATIONS

The results of the in-flight and postcrasb fire tests conducted onthe UH-iD and CH-47 lead us to the following conclusions:

1. Fires occurring during :light in compartments otherthan the habitable compartment can be protectedagainst by lining potential fire compartments withan intumescent mastic coating or a sodium silicatehydrate panel.

2. Presently used Plexiglas windows are the weakest

points in a helicopter during a postcrash fire.

3. If windows can be protected appropriately, thesurvivability of occupants in postcrash fires canbe increased significantly by insulating theinterior walls with isocyanurate foams and sodiumsilicate hydrate panels. However, the unreliabilityin performance of the protective wall materials(especially in areas where wiring and tubing haveto be left accessible or where the walls have beendamaged by the crash), the unknown effects onhumans cf smoke and particulates, and the weightand cos. penalties render such an approach subjectto sexious questions.

Our recommendations are as follows:

1. The protection against in-flight fire should bepursued further by the U. S. Army. Recommendedmaterials should be further tested against firesin various types of compartments and the weather-ability of the materials evaluated.

2. Protection o' the interior walls of the habitablecompartment against postcrash fires should not bepursued any further.

3. Approaches toward enhancing the survivability of

occupants should involve the developMent of personalprotective clothing and masks, the improvement of thecrashworthiness of the fuel tanks, and the utiliza-tion of active fire suppression devices on board.

74

APPENDIXTEST DATA

f N 00 C'4 N NC14 +4J4~ W ' 0 0 0 N'

co 44 -1 4 H v4 H-

4.J

a)

p.4

-x 4c-

C.Ca

0)-

rnu 00. C)NN CJ '

14. rj e

*0

H

0

79

- -~ - - ~ '~~0 - -

*<*1.

~~~~~-~0~~

Ii *0

U 0

2

.0

0 0 C.H U~ 0.0' .0 .0 H. U0 '4

I, '4 H.. g .0 "4 .0 '4C '4 PP P4 0 0 0 0

H.. CI H.. P4 H.. A0 = P4 '4 '4 0 .0H., '4 .0 '4

* U .0 '4 '4 . . . 'P'0 H.P U 0 CI CI PH

'4 .~ Vo oa U ~0 -

p. a - o0 CI V CI 4 U - a - 0

0 H.P PP PP 'P t 'P 'P 1.. H.. I.'4 H, .4 P4P4.0

~ - - N P4 N -~ .0 4% 0. 4% 0' N* . P * . . . . .

'~-0 0 0 0' OH 0' OP OH OH OH 0' 0. 0'400 -

0=I 2 lIP 0 PP

-' H.. N eq t. Co N '0 OC N 4 P. N

CI~ P~ I I .4' 0~ ~'P P4 .4 'P .0 -P eq -

O N N HO Co P~4 N N N eq NOH 0' 0 CI 0' '0 - N N N P"

2.0

.JP40 eq N ." 0 0 .~ N .0 '0

- N N N N 4%

I'P - - - - N - H.' pH' eq eq N N

.0 0 0 OH 0 0 .0 N - 4%.0 - . . - .1 - o~ .0 N.0 N 4% 0' .0 - 0 .4 .0

4 ~ . H.. '4'4 - ~..

4P4PH 0 '4 0 0 IPH..4- - '4 - H.H - 0. 0.~H

~P4 - - .0* 0~ H" - .0 .0 0'4.41 H.. 0 '4 eq P. 0 44 0 CC C. - -

~-CI 0. '0 0. .0 0. '.0 0. .0 P.4% H..'0 ~0Q ~0

'P .0 ~' " - - -0'0 S H.P S'4, SX*. S~<. 50.' P~N 0.0 600 '4o.p C. oCI C.N.0 CCN.0 ON-P .0N.0 P * P * '444 N N C..0.00-p ~0.4 .00-p .0*000 0'

0. . ' *.0N .. 40 *4 .4~.4 **.4 pp pc .CH#%8 '.4-4 .4-4 *o-'.0 .0 - .0H.. .00 .06 -. 44

PH-fl 0 0,00 'P.4-P 'P.4. 'P. 0-..' 0. - OPHUC - - 0,..- - eq

CO N 4% '.0 N 00 '0 'P Co N .0

L;Li ~.__________ ___i

+,~

• !

oo 0

- '•

A 6.

'UI , -m 0 +,

ml

-- - - A.o.~o. ~ - 42044:4~4*4~ 4'?* - ~44,0~

.4 44

. '4,

4-

I; 00 -~

,. ~ >4o 0 44

>4

* 0' 0,- 444

A 0. .4 .t. '.0

44 .4~4 44 4'= .4 ~ ~i0

~ 0 0 '44'0 - -

44 . '8 ~0 .4 .4 .4 .4 0

A A A 4' 4 0 00 A 84.4 44 .4 .4 .40 U> - .4 1 -4' .4 *4 0'0 '0 44444.4 a44 0 .4 .00 .4 444 44 440

8 >444 4008 ~' 'Of

4 A 8 ~', ~

ta 04.04' '~ 8 ~Oh 8 81 0 4

44 '0 040 - - - :2 >4 140 4~ - h.44 ~to A '4 0- .~ '~' U. U A £8

11

p.44 .4

0 '*~~'U 0 44.4 .

0o4

0 4 4- 44 40 N 0~ 40 44 - N ~'. 44 44 -40 0.~0- 4

40 40 0'. 04 04 04 40* a ~ ~' ~ ~ 04 '4' -. -

40 4.4'.,

4~

0' 40 04 44 4 04 44 .4 .4 "4 0'43 to,., N 4 - 0 04 ~0 N - 44 44 04 N0' 47 0' - - 4 40 '0 '0 -

to

.8 ~ N N ~.

>4 44 . 0 40O 44'..4 -

-4 404444to

4 4 4'.

0.4.4 00 -4 0 44 44 40 0 .4 44 04 .4 '0 40 40 0 044'.. '4 0) .4 - 0 0' '0 8 44 - '4 4' ' 0 04

.4 44 '4 40 44 .4 ~0 40 40 04 04 04 40 '1 0 ~4

0 N 04 40 0 0' .0 @4 40 N 04 '0 N 0 0~ .44 .0 N N 0 04 - .4 '4 0' ~ 0' '4~ '0 .4

N 04 04 0) '4 40 0 '.0 44

- . 11

C '4 4 .4 0

.4 11 0.4 .4 '.' ~U ~ A 0.>4 .4 044 0' .4.'0 0 42.0 .00 '.4 '.4 0 .4 0 '40 4' 1 40 .444 4' '040 4 400 044 4' 044.4 444 - 44 0 44 .444 '>4 - 44 0%' 04.4 0 0 .40 40.4 .44 .4 44 ~ 44' .4 00 440. 04 * A 0.0 "0>'. 0.~ 440.4.0 >"0.4o4 '44 .0040 04 4.0 '.404 044 .1I 4' '.4 ~ ~~'.'.t~ 44o 0 '0 440 '.4 .0 '444 4444

'.4 0' .4 4.0'.4 .0 044 A .4 .4 .44- U C 44 40'0~ 44'. 'OVO 44 0.A 4444~ '44 00 04444444044 44 44 0440 .444.4 0" (440 4'. -4 *444 4444,444.4 044~4 *04 '4 444 .4.4

4. 0 0 0. 42-4 44 44 .40'040' X ." .00 ~ 0 044.44' .4 .4 44 .4 ~ 0.444 00 4) .4 0 40 4144-4.4.4>4 44 '~4444~ 44400 U .0444 44 44 4 4.4 - .4.4'... 0 *44'4 A A4' 0 "04 444444'4 44 44 40 ".44 40 0 44004 04454 .0 .4 *A 444. k 44 0..4 11>4.40. -4.4 4044.4 44444 -4.4 A .4 .4.44 0'44.440 0. 4'4 04.440.4'..4-440'.4 .4.0.40 40 44) .4440 44 44 40.0 .440'"*4* 4 ~o4~4~4 '...-44. 444 44 '4.4 0-4 0 44 .44 .400

>4- .'4 .4 04 0' 0' 40 0 40 '0 40 N 04 0 404 4. 04 40 0' 04 40 40 40 .0 40 40 4.4 40 40 0'. N N 4

S '4

78---- 4 ----.-..--.--.-.---.---.----.------.-------.---- ~.--..--------...--.-----------.....-----4-.-., -

.4

4 44 .4 .4 .4

a . a a Ka a a. N- a a a

.4 ., .4 - a.4 .4 *4 - - .4a a. - .. - .4..4 .S ~ a o a a - a- a .4.

~ : : A a a ~ ~a a a> S S A -

o ~ ,~ ~

* a a a

a a a go

a a a a a a

4. 4. 4. .4 .4

.4 ~ ~0 ~ ~4 ~ a ... 4~

'4 A A A A A A A A

a a a a a

~ ~ ' ' ~

~ a. K K 0 o a a~a

K K .04

4.44a NI, N N 44 44 - 44 #4 44 #4 04 N 444 4~....................45* 44 44 0 44 4.~'a.- - - 2 2 44 44 0 04444

A

0 4 0 N 044444.44a. .'4 . 44 44 4 N 4 0 -i 0 a. 04.4

$ 40 N.~* ~. 44 ~ $ ~ a. 44~04.4.4 #4 4 4 4

~

2.0 #4 #4 0 a. A 0. 44 .4 444 44

- #40 4'2 a. .0.4-a - -

.4 .4 .4,.4 g

~K~4.4 400.40 ~

0 ..g ~44 1

0 4.- 4.* 4..*~ ~ . ~0 aa 00 .444 4K 4. -- ~ K -0.4.. K~ - 0 #4 44. KO #40 #4 ('hA4 0.z .~ .g ~

- ~ ~ ~ ~2A1 ~ ~a. ~0a -I ~.ggg gg-~'~<4~4. <4~0 ~ 0.%0X.~'444 ~4.4 <0

79

00

> 0 0q 44,

ui'4 '0i H 0r 40 r4 1 -4 w C)4

4 o0a ) 0 4-34 $ 4 $4 ,

41 4J.I 41, (a' 4-0a 4wi) 40-4 0N 00 O

U)j (a0 0 04 C-. 0.n~ $40 4 4.) -) 4to V1 (- ~

o o 0-' 10 .,0 z. 00 0

0JU 0~ 0 ~ (A 00'i 0t

Z~ZH

4j~ 00U

x 430 0 0 0

P- 41 0 0 C 0 0 ) 0C0

C)-3 0 0 0 .. t 0 0

C).*,- 0q r

E-44

0 z u

~i 0 0 0.~ 0

U) 0

H- - 4 J X 4 0) 0~ 44 4JO $ 4

P4 13 M0 ~0 ~ ~ rI ' H 44- 0

-4 'r H. H'r IH' ~ ~ H~

03g-o 0. 10 $oJ

H I- 41 i- 0 0 H04 00 , .P 4 o aiL

80 1 r q )Q

j-

W cu0' r. C-) u

a)-. H0 . - )I

41 >

04 -r4 10 fq r;r:34 H 00 H 4 H H.4HCO C

0 0 Z, 0 oH 0 0

:j N -6 - U

4 H H- - H H~d

00zz

:3 0 o ~)-z C) -11 0'T 0' V) 0' 0'

004~~~ r__ H_ _

IA___ _ __ _ __ _I_

rjrr- -41 h o A 1) ) 81 U

b- p

-H

0o)) Q ) 41 )

Ni P, CCO&Jr

to HW H HH - 4.wo b4) :3 -0-t~l

0- .*I 3 -i a 0c4.) 04 a) p 9

U) to 0 H ~.0~~ ~ 414 .r, 4 *H.

0 0 :3

0- 0)-

4J

>00 0 ) 00*1H, ;pJ C', 0 00 C'4 0: H

#r4 0") 4-4 H C

ca Ti

fl).U 0 0) (1) 0 0 00 AN )

r, ;i. LA 0 rq HH Ent -j a) H

0 ) .. W ) V >* ..1

T4 0 ICr) -% HQ H p r4

r-i 0 0 Qr-i 0 0 4 r. > 4 o 0U) M

0 4M 0 '0 -r 0. m

0> I- col 00'-H4 ca U) 0 x~

U) 0 O r-i 01 0 0

Q) r4)Ua *~4- 4 .0

P4 o t4 C) 00 4) X -

.0 X ~,4i rI~ 0 ~H82'r~

STCUiity classification,DOCUMENT CONTROL'DATA. R& D

S (eSC tltf" flcalIon 61 title, body oftabstract and indexing anno tioni must be entered When the overali report Is cliteeliled)I. ORIGINATING ACTI'ITY (Cotpotato author) .20,REPORT SECURITY'CLASSIFICATION

Arthur D. Little, Inc. li ,ASSTSIFIED'20 Acorn Park 2b. GROUPCambridge, .assadhusetts

3. REPORT TITLE

INVESTIGATION AND EVALUATION OF- NONFLAMMABLE,, FIRE-RETARDW'T MATERIALS

4. DESCRIPTIVE NOTES (Typeof teport and Incluslve dates)

'Final ReportS. AUTHOR(S) (First name, middle initial, fast name),

Sami AtallahHenry L. Buccigross

0. REPORT DATE 7a. TOTALNO.Op PA0ES; 7N. OF REFS

November 1972 93[0CONTRACT OR GRANT NO., ORIGINATOR-S REPORT NUM3IESIS)

DAAJ02-71-C-0042b. PROJECT NO.

IF162203A529 USAAMRDL Technical Report 72-52C. Sb. OTHER REPORT NOW|| (Any other numbera tOl may be aaldlned

thi report)

d. ADL-73588-I0. DISTRI"UTION STATEMECNT

Distribution limited to U.S. Government agencies only; test and evaluation; November1972. Other requests for this document. must be referred to the Eustis Directorate,U.S. Army Air Mbility Research and Development laboratory, Fort Eustis, VA 23604.

1.. SUPPLEMENTARY NOTES SP2. SONSORING MILITARY ACTIVITY

Eustis, DirectorateI US Army 'Air Mobility R&D Laboratory

• j Fort Eustis, VA 23604 4113, A STRACT

A comprehensive survey was made of present materials technology. A number ofmaterials and composites were selected and tested in a specially designed furnacecapable of providinga thermal flux equivaleit to that encountered in JP-4 fires.Final candidate wall systems were compared for protection effectiveness, cost, andweight penalty. Various combinations of isocyanurate foams,, sodium silicatehydrate panels, a mineral insulation and intumescent mastic paints were thenapplied to the walls of two crash-damaged helicopters (UH-ID and CH-47) andexposed to full-scale fires simulating in-flight and postcrash fires. The helicopterwere fully instrumented tomeasure temperature, heat flux, smoke density and toxicgases.

The results of the in-flight simulation tests indicated that it should be possibleto protect the habitable compartment against a fire occuding in an adjacentcompartment resulting from a fuel or hydraulic oil line leak. Sodium silicatehydrate panels placed on the fire side appeared to give the best pertormance.

The large-scale tests indicated that total wail protection of existing helicoptersagaInst postcrash fires is not feasible and should not be pursued any furtherbecause of cost, unreliability, and lack of assurance that the walls will maintaintheir integrity in a 'crash.

D 1473' S O FooRM ,A,. I JAN 0. WHIC..IS,I ..ov 4 7 OLlA. o , 1O7A31V U". UNCLASSIFIED

Scurity clsltication

UNCLASSIFID,,Security, Classification _

14. ,LINK 'A.L. LiN CKEY WPROS ' I 'W

MOLE WT R'OLE WT' 'ROIE W

Postcrash fires

Fire-resistaft niaterials-In-flight firesHelicopter fires,Intumescent materialsIsocyanurate foams'

L

ii

I I

UNCLASSIFIEDSecurity Classification 11252-72

FrD 27%31