UNCLASSIFIED

AD NUMBER

ADBO10507

NEW LIMITATION CHANGE

TOApproved for public release, distributionunlimited

FROMDistribution authorized to U.S. Gov't.agencies only; Test and Evaluation; FEB1976. Other requests shall be referred toAir Force Aero-Propulsion Lab., AFSC,Wright-Patterson AFB, OH 45433.

AUTHORITY

AFAPL ltr, 18 Jan 1980

THIS PAGE IS UNCLASSIFIED

THIS REPORT HAS BEEN DELIMITED

AND CLEARED FOR PURL-*C RELEASE

UNDE.R DOD DIRECTIVE 5200.20,A.ND

NO RESTRICTIONS ARE IMPOSED UPON

ITS USE AND DISCLOSURE$

DISTYRIBUTION STATEMENT A

APPROVED FOR PUBLIC REtl.,%Sfj

-ION UNLIMITEDDISTRIBUI #

- -M AFL.Th.7&73.

S0 NVESTIGATON OF TITANIUM COMBUSTION

CHARACTERISTICS AND SUPPRESSION 1S• -TECHNIQUES

WC)

CC

FWIRECA REPROhTI RANCFH X~ir

PU L AN LU-C TO -DI-VISION- -- -............l . . .....

.-•

AIRNA FORCE W AYPT~WomS

Di4iY4Wi4r oi tX4W towu 6"o, tOiVflta A~tf~ n

SWhen Government drawings, speofficatinsoar ohr data are used for any purposeother than in connection with a definitely related Government proourement operation,the United States Government thereby incurs no responsibility nor any obligation

44

any way sWloed n e dd drawingso opifleatlo s, or other data, ar set fo be regpurjn

by implication or otherwise as in any manner licensing the holder or any otl-:zr peesonor corporation, or conaveying any rights or permission to manfacttre, use, or sell mypatented h•w•ft that way In any way -be relatd -thereto.. -

The distribution i% livited to U.S. Govemeut agencies only; test and eval-• 4aa-on .Joly 1975). Mt-r rcqueat for this doct~eu au be referred to theS•/•Ai Forc.e Aera,-Propulsioa 1aboratory (M]C), •;•Lght.-Patter.nou Air; rorce Base,

Ohio 45433.

This technical report has been ruvx.ewed ad appro r publcat.Lor..

V1 /UI . 1 ; app-oved

...ir e ioteect•to BraunhFuels arzd Lbzic*vt•iou Divtion/Alit Force Aero-.topula•a.eb bor4y. /ry

!i}I< ::Air Fgrce Aeko-.ifapu.soa t l•boratory, q-.

.-- Iea m thois rport should nt bo "Amwd -.in reLturn is rouirxd by Security

-. t•a Uali, or uotie oa a

UNCLASSIFIED

SECU RITY C LASSIFIC A TIO N O F TH IS PA G E (" Ohn D ata Enteo ed)_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

REPOT DCUMNTATON AGEREAD INSTRUCTIONSREPOT DCUMNTATON AGEBEFORE COMPLETING FORMt4R-PORT NUMBER 2. GoVT ACCESSION NO. 1. CIPIENT'S CATALOG NUMBER

4, TI 'in rT* Q~jE.0O COVERED

Investigation of Titanium Combustion Characteris- iJan 74

7.~ AU. PORT NUMBER(2

S. PERFORMING ORGANIZATION NAME AN0 ADI)RESS 10. PROCRAM ELEMENT. PROJECT. TASK

Air Force Aero-Propulsion Laborastory AE ORUI UBR

Air Force Wright Aeronautical LaboratoriesWright-PattersonAFB,_Ohio 45433 _______________

It. CONTROLLING OFFICE NAME AND ADDRESS 12 REPORT DAT E

Air Force Aero-Propulsion LaboratoryAir Force Wright Aero.,autical Laboratories SNM~ AE

Wright-Patterson AEB, Ohio 45433 ________

IC. MONITORING A'3ENCV NAME 4 AOORESSOI deilnts.e troaw Cwitlolllng 0911 C*) IS. SECURITY CLASS. (0 %All

UNCLASSIFIED ONRIG

SCHEDULE

16. DISTRIBUTION STATEMENT (at Ali. RopetI

Distribut"-n limited to U. S. Goveriuwnt agencies only, Tesnt dsnd Evalu~ation,(July 1975). other reque-sts for this document must be referred -to ALF Aero-Propulsion Laboratory (TBC), WPM'-B, Ohio 45433

1?. OISTUNiUYIoi. ST~ATEMENT ( W .h br,~ R t.t~ Blc U 2eS , It 4Itt*,n# fr.M 90 I

IISUPOPIMENYARY I1OVES

Titanium CoviustionCompressor FiresFire Lxtinguishment

~~ASTRC ~ *..'. al. t .0*Vo &.4 ~ih *T.I W"11 fr vab W1

Ibis test program studies the burning cbAracteristics of titanium under airflow~ conditions. The flc~t plate titaniu3 samples are ignited by molten titanit=from an electrically heated ignitor. Air flow conditioins that sup-port sustainedcombustion of a single sanple are determined. The burn rate is measured on alltest with steady state burning.

Argon gas is shou-i to be a feasible extinguxishing agent for a titaniumfire. Quick injectioni of a tufficient &%ount of argon gas to suintain a 60%1

(C~ontinuct-3 en back)LI; *..-W

UINCLASSTFII'nSECURITY CLASSIFICATION OP' THIS PAGE(Who Data*1 Enteted)

0 j. ABSTRAkCT -Continued

c~,incentration by volume of argon results in quick suppression by oxygenI d'-pletion. Carbon dioxide (C02 ), a common fire extinguishing agent, is shown

to sustain titanium burning at an accelerated rate.The ultraviolet (tJV) radiation emitted by burning titanium is shown to be

of a sufficient intensity for existing UV fire detectors to detect at reason-able distances..-.

k-N

2

-. 0 Gi ,,i

AFA.PL-TR-75-73

FOMEORD

~ I This report was prepared by Duane G. Fox of the Fire Protection Branch,

Fuels and Lubrication Division, Air Force Aero-Propulsion Laboratory (APAPL/

SFH). The work reported herein was performed under Project 3048, "Fuels,

----- ----- Lubrication, and Fire Protection," Task 304807, "Aerospace Vehicle Fire

Protection,' Work Unit 30480773, "Aircraft Fire and Explosion Prevention."

This work was performed at the request of the Components Brcnchi of the

TrieEngine Diiinin support of Work Unit 30661005, *Coim.ressor Rotor

Rub Test." Test conditions and requiremnents were supplied by Mr. Charles

W. Elrod, AFAPL/TBC.

This report covers research accomplished in-house from January 1974

to March 1975.

The author appreciates the assistance received from Mr. Jon R. ýIaneiz,

AF'APL/SPII, in designing the combustion chamber and extinguishing t~est har4-

ware and in developing the experimental tests. Special thank-.i are given to

the following individuals: Messrs. Peter Lianalak, Harvey Reeves, Glen B~oggs,

and Robert Es.zh of AFAPL fox t~heir invaluable help in the execution of the

7hisreprt as siubaiittod by thte author July 1975.

AFAPL-TV-75-73

TABLE 0F CONTENTS

SECTION PAGE

I INTRODUCTION AND SUMMARY 1

A. INTRODUCTION .1

B. SUMMARY..................... .. . . ... . . .... .. .. .. .... 1C. ADDITIONAL INVESTIGATIONS REQUIRED. .. ................... 2

II EXPERIMENTAL EQUIPMENT 4ftA. TEST FACILITY .....................

B. COMBUSTION TEST (ŽXAMBER .................................. 9C. SUPPRESSION HARDWARE....................1D.. INSTRUMENTATION.......................15

III TITANIUM COMBUSTION TESTS

A. TEST DESCRIPTION. ............................... .. ....- ... B. TEST RESULTS AND D)ISCUSSION. ......................... 24

IV BUR~N RATE ANALYSIS 4o

A. ANALYSIS DESCRIPTION. .................................... 40B. RESULTS AND DISCUSSION .. ................................ 42

V FLYA1m E SUPPRESSION TESTS 4

A. TES DSUCRIPTION. ........................................ 47B. PRLzULTS AND DISMISSION .. ................................

VI ULTRAVIOLE" (UV.) EMISSION XALALYSr S 54

A. TEST DES, IPTION ANE lDlUClUqSSN Oi RESULTS........

SIBLIOGRA.Ptil.. .................................. .. .. ....

A FAPL-TR-75-73

LIST OF ILLUSTRATIONS

FIGURE "PAGE

1 TITANIUM TEST FACILITY SCHEMATIC ..... ............. 5

2 AIR SUPPLY CONTROL INSTRUMENTATION . .................. 6

3 AIR SUPPLY PARAMETER INSTRUMENTATION ........ .......... 7

4 PHOTOGRAPH OF TITANIUM TEST HARDWARE ...... ........... 8

5 TEST CHAMBER ............. ....................... 10

6 MIRROR ARRANGEMENT FOR VISUAL OBSERVATION ..... ......... II

7 SAMPLE HOLDER FLANGE ......... ................... .... 13

8 SAMPLE AND IGNITER ARRANGEMENT ....... .............. 14

9 ARGON AND CO, INJECTION MANIFOLD ..... ............. ..

10 ARGON AND C02 SUPPLY SYSTEM SCHEMATIC ...... ........... 17

11 DATA CHA.RT RECORPER . . .......................... . .. 18

12 TEST SEQUENCE PR,.GRAKMER.. ..... ................. .. 20

13 SUSTINED, NON-SUSTAiNED BU N1NG DATA FOR SAMPLES A AND 8 36

14 SUSTAINED, NION-SUSTAINED BURNING DATA FOR SAMPLES A AND B. 37

IS SUSTAINED, NON-SUSTAIN'Eu BUKINING DATA FOR SAMLE C.......8

16 IJUW4 hATE MSSURE!4ENT TECHNIQUE. .. .. ......... . .........

17 Aiwh,-AIR CALIBRATION CURVE ... .. ........... . ...........

18 .LTRVIOET '%'&wM4 INSTRUMENTATION...........55

19 TYPICAL ULTRAVIOLET EM4ISSION .4EASU4E ..T ......... 55

AFAPL-TH-75-73

LIST OF TABLES

TABLE 0PG

1 TITANIUM COMBUSTION TEST RESULTS,SAMPLE A, 1210 C ............................................ 25

2 TITANIUM COMBUSTION TEST RESULTS,SAMPLE A, 2600 C.............................................26

3 TITAN IUM COMBUSTION TEST RESULTS,SAMPLE A, 3990 C ............................................ 2T

4 TITANIUM COMBUSTION V'ýST RESULTS,SAMPLE B, 1210C.................... . . ..... .. . .. .. .. .... 2

5 TITANIUM COMBUSTION TEST RESULTS,SAMPLE B, 260'C........................29

6 TITANIUM COMBUSTION TEST RESULTS,SAMPLE B. 399'C........................32.

7 TITANIUM COMBUSTION TEST RESULTS,

SAMPLE; C, 121C C................... . . ..... .. . .. .. .. .... 3

8 TITANIUM COMBUSTION =2ST RESULTS,SAMPLE C, 260 0C and 39 0C ................... 3

9 TITANIUM COMBDUSTION TEST RFESULTS,

10 VX~l4U AIR VEtAXITY FOR SUSTAIN'ED 01,4DUSTION .. ............

11. AVERAGE BURN RAXE DATA SUM-NARY, SZA3PLA; A. .. ..............

12 AVERAGE BURN P)ATE DATA SU?4flARY, 3A4vPLE B.........

13 AVERA4GE BURUN R~ AhSMAY APEC........

14 AVERAGE BURN RATE DATA SUKMNAIY, SAMPLE L. .. ..............

is ARGON GAS LXTINCISHIN~G LATA. .............................. 50

16 cO LXTIKGUIS&4flKG TEST !VATA..................53

17 ULT)LAWIOLET ENHISSION FROM TITANIUM A.4E .. ................ 56

..... .... ---

AFAPL-TR-75-73

SECTION I

INTRODUCTION AND SUMMARY

A. INTRODUCTION

This test program was iritiated to study the burning characteristics of

titanium under specified flow conditions and to finu a technique for extin-

guishing an on-going titanium fire in a test facility. The work was accomp-

lished prior to the operation of a full-scale, single stage compressor

test facility.

This program had two primary objectives: (1) Tests were conducted to

determine what conditions (air temperature, air pressure, and air flow) are'•t:-1

required for sustained combustion on a single compressor blade and repre-

se-tative flat plate sample. The burning rate was determined for all

cases of sustained combustion. (2) Suppiession studies were conducted to

de-terraine what concentration of an inert gas such as argon is required to

extinguish a titanium fire.

a. SUMN.RY

It was found that the burning characteristics of titni-.m samples

are not strongly dependent on air flow temperature or pressure within the

"lirmits established in this program (121 0 C < T < 399*C, 448 kPa < P e 1138 kPa).

The initiation of sustain~d burning and burning rates are more dependent

on the sample shape, thickness, and relative position to tho air flow.

Limited testing of a B alloy of titanium indicates that material coo-

position does affect the burning characteristics.

-1

AFAFL-TR-75-73

Measurement of the ultraviolet (UV) radiation emitted from the burning

titanium indicates that the UV emitted from a 2.54 x 7.62 cm sample is

at least an order of magnitude greater in intensity than from a 5-inch

diameter hydrocarbon fuel fire. Utilization of a UV fire detector for

detecting a titanium fire is thus feasible.

It was shown that, for the test conditions studied, an argon gas concen-

tration of at least 60% is required to extinguish a burning titanium sample.

The argon dilutes the oxygen concentration to t le've that will

.not support sustained combustion. Since substitution of argon for air can

be done rapidly without significantly changing the total air flow through

a test device, this extinguishment technique is applicable to turbine

i•' 1 engine compressor test facilities where titanium combustion presents a

hanard. The argon concentration must, however, be maintained until eithier

the molten material cools sufficiently to prevent re-ignition or until the

air flow is reduced so that sustained burning can not continue.

While steady-state burning data was obtainei flor the single sample.

direct extrapolation to a rotating environn.nt with co=)lcx air flow

-Atterns such as exists in a turbine engirw coore•ssor is not po•ssibýl.At best tbe steady-state buxning data obtained Xan be uted in cowter

modealing of the coimplex scgid t.-abustion phencoenon. The criticaA factor

in achieving sustained burning following ignition and localized burning

is the air flow and how it reaoves oxidized waterial from the surface.

- "This phenonenon was not txtoroughly studied in this effort.

•..-.:... C. AMDITI O"AV INV•ITI: U•

Although the testing performed provides baseline data on the burning

c-haracteristics of titanim in air flow, additional testing should be

64-':

I""

AFAPL-TR-75-73

conducted to more fully characterize and define the effects of the ignition

Source and the air flow over the sample. The effects of a stacked sample

array simulating an actual compressor also need to be adequately defined.

This information will be required when the combustion phenomenon is modeled.

Tests should also be conducted with other types of extinguishing agents.

-'p.

AFAPL-TR-- 5-73

SECTION II

•-. :EXPERIMý-NTAL EQUIPMENT

A. TEST FACILITY

The tests were conducted in a test facility located at the_ Air Force

S- Aero-Propulsion Laboratory at Wright-Patterson AFB. The facility was

- developed to test turbine eigine conmbustors. The facility provides air

at a regulated pressure, te:-.3er.ttce, and flow. A simplified schematic

indicating tne components of importance to the titanium combustion tests is

shown in Figure 1. The control instrumentation is shown in Figure 2 and

Figure 3. The overall test facility is shown in Figure 4.

The air is supplied by piston compressors and is then heated to therequited temperature by a furnace. Pressure is maintained at a prescribed

value in the test chasl~er by a feed back cntroller which opens or closes

a blood off valv. The flow is regulated by opening or closing a pluq typo

o ori fice at the end of the air flow section. Tho plug orifice was oe r%,tcd

by a rewto. switch in the control room.

by the test sequenceý used wa-s to first set the do'wirod air t.etueand

pxrissurc- and then rcqulate the pits to get the reqtttrad 1~ flw Clcinq

the pl',. for exaN,.!e, decreas the tlow throuqh the test chaacr and

increas-eis the twwit- of bleed off.The flow is eterzin-ed by teasuzixi-v the differential presrn (,i acreszt

a two-inch dianeter venturi an:! usingm the standand eqatim which relates

the tP to theass flow in -patjpds of flew par seca-n-,. For easv- of oea~n

tables were t/abulated by a coqi"ter. by etiterxn9 the tiewrature., static

AF-APL-TR-7 5-73

H 0

HH

FV

D44-0

EID4 04

00

-'-4

UU'-~z*(U

~--~ -E-4

N+

Iu

AFAPL-TR-7 5-73

C 0

IW4-

W-

r~z 43ti~j~vIg~yv 1

_ IM

-,-4

fjI

AFAPL-TR-715-73

4A2

-Al' R 4ýt44

- 13.

AF.4PL-TR-75-73

E4,

'-,4

E-4

IWOI

7..-

a4

AFAPL-TR-75-73

pressure, and differential pressure, the table gives the mass flow. The

flow is reported in both Kg/sec and lb mass/sec in this report. Air velocity

in the chamber is calculated from the pressure, flow, and temperature and is

reported in both meters/sec and feet/sec.

B. COMBUSTION TEST CHAMBER

The test chamber was designed to use readily available materials in order

to shorten fabrication time. The original chamber, prior to a few later modi-

K• fications, is shown in Figure 5. The first 24-inch long section of pipe

isolates the flow measuring venturi from the test chamber. The chamber is

-;• a standard 300 pound pipe cross. One leg of the cross contains a water cooled

jacket which houses a 7.6 cm (3 inch) diameter, 1.27 cm (1/2 inch) thick quartz

window. This window provides access for television camera coverage and also

permits measuring the ultraviolet (UV) radiation emitted from the titanium

flame. The UV radiation is of interest because it is a likely technique for

detecting a titanium fire.

The other cross leg contains a water-cooled jacket which houses a 10.2

-c (4 inch) diameter, 4.4 cm (1.75 inch) thick tempered pyrex glass window.

This window provides viewing access for high speed motion picture photography.

The window jacket is cooled to prevent the glass from weakening at the higher

temperatures. The cooling, however, causes a temerature gradient across

the glass which results in the glass fracturing at a temperature near 3990 C

(750°F). The window waa used satisfactorily for tests at 121 C (250 F) and

260"C (500'F). The glass window was replaced by a steel plate for the. tests

at 399°C (750)F). In these tests, the notion picture camera and TV caera

both view through the saller quartz window by the use of a partially reflect-

ing mirror arrangement which is showm in Figure 6. This scheme provedadequate; however, alignment is more critical through the smaller window.

AFAPL-TE-75-73

2 WU

ejj

.Y~

AF~AL-TE-75-73

MOIO

450PITR

CAMERA

H PARTIALLY REFLECTING~-WINDOW UMIRROR

FLANGETV CAMAERA

Figure 6 Mtirror Arreangeimrxt for Visual Observation

-7 -

AFAPL-TR-T 5-73

I: The sample holder and igniter are mounted on a flange which bolts into

the bottom center of the test section. This arrangement permits a fast change

of test specimens. The sample holder and igniter are shown in Figure 7. The

- sample is held in place by boron nitride blocks which are fastened toithe

*• flange by stainless steel brackets. The boron nitride blocks keep the sample

from burning past the holder. Since boron nitride is a high temperature

*: material, it works satisfactorily at the high ambient teu'perature and is not

significantly damaged by the molten titanium.

The igniter shown in Figure 7 is a 0.23 cm (0.090 inch) diameter titanium

rod which is machined to 0.15 cm (0.060 inch) diameter at the center for

0.64 cm (0.25 inch) length. This forces the igniter to bnrn first at the

center. Without the narrowed section, the igniter will usually burn first

at one end or the other because of the strain produrced at the connection

S- point. This igniter proved to be unreliable at high air flow. Analysis of

the high speed motion picture film revealed that the igniter was blowing over

the top of the sample after becoming soft prior to melting.

The igniter was modified to a 6.3 x, 1.6 mm x 7.62 cm long (0.25 inch

0 .0. G- inch x 3 inch) piece of titanium which is positioned evenly with the

top edge of the sample, as shown in the illustration in Figure 8. This

igniter is notched approximately 1.6 r.z (1/16 inch) deep on both sides at

the center. This igniter proved to be reliable and was used for the tests

described in this report. Electric current from a 200 ampere, 16 volt, 60

,fertz transformer passes through the electrical fittings to the igniter

-• holder, All conductors in the igniter circuit except the igniter are made

from copper.

~ 11 AFAPL-TR-75-73

bo

-r0

Z0

..........

'7/.,rý

13

- AFAPL-TR-T5-T3

AIR FLOW

1/16 INCH

IGNITOR

~ IGNITOR HOLDER

SAMPLE

* ___________HOLDER

F~AAHGE

Figur~e 8.Sample anid Ignitor Arrwi,=,ent

I 4

AFAPL-TR-75-73

C. SUPPRESSION HARDWARE

The suppression tests primarily involved injecting argon gas into the

air flow upstream of the burning titanium sample and noting effects on the

burning sample. C02 gas was also used in a few tests. The injection mani-

* fold is illustrated in Figure 9. This manifold injects the gas as illustrated

in Figure 1. The manifr id is pressurized up to the solenoid valve by a high

capacity regulator which is manifolded to twelve, Size A argon cylinders.

The complete argon injection system schematic is shown in Figure 10. The

same hardware was also used for the CO2 studies except that only six bottles

were employed.

S..The argon temperature and pressure are measured in the injection mani-

• I fold. The injection system was calibrated by sampling the flow stream near

the sample. This procedure will be detailed later.

D. INSTRUMENTATION

The air system control instrumentation consists of a pressure gauge

reado,;t of the static wall temperature, a strip chart recorder output of the

differential pressure (AP) across the venturi, and a thexuoouple meter out-

put of the air stream temperature. These instruments were used for adjusting

-t the air flnw coditios in the test chambr.

The test chamber parameters were reacorded on a chart recorder (shown in

Figure 11) so that changes and transients could be observed. Ate conditions

during the burning tests were found to be stable and thus did not actually

require time recording. The argon suppression tests, however, do involv

rapid changes of temperatures and presserecs and the recorded data is required

for analyzing the test results.

AYAPL-TR-75-73

I4'

CC

gigi

.........

AFAPL-TR-75-73

ý-l 04

U)C

L4,-I

L

AFAPL-TR-T5-73

41L,

z--

*~al

iJLv"

AFA2L-TR-7 5-73

The air temperatui: is measured by an exposed junction Cr-Al thermo-

couple which is located about 6.4 mm (C.25 inch) from the wall in the nozzle

jast upstream of the sample. The exposed junction provides sufficient response

during the argon injection tests.

The chamber static p, essure is measured at the wall just upstream of the

sample. The pressur'e transducer is located near the chamber and provides a

sufficient frequency response to record transients in the pressure.

Two event markers are used so that both the time of ignition and the

time of argon injection can bL correlated with the pressure and temperature

traces.

The argcn temperature is mea~urad near the exit of the flow meter.

Measuranent oC the argon flow by using a turbine type flow meter proved to

be unsuccessful because the readings were difficult to interpret. The

actual argon concentraticn was determined by sampling the air stream near the

sample and analyzing the mixture for oxygen, nitrogen, and argon. Samples

were taken at the six required test conditions. The sj.acific details of this

procedure will be discussed in tho section on extinguishing tests.

The ultraviolet emismion tests were made with a spectroradiometer system

that will be described in detail in Section VI.

"4! The tests are properly sequenced by the use of a 12-step sequencing pro-

grammer which has a dwell time on each step that is adjustable from 0 to 10

seconds. This programmer (shown in Figure 12) is used to activate the motion

picture camera, TV camera, sample igniter, argon injection valve, sample

valves, and provide sync signals for the recorders and caueras. Cnce initi-

ated, the test is automatic but does have manual override on some of the

functions.

19

AFAPL--TR-7 5-73

I. It,

VIýO

A',... E

Pj

AFAPL-.TR-75-7T3

SECTION III

COMBUSTION TESTS

S-s

A. TEST'DESCRIPTAION

The test procedure used was to first install a sample in the test chamber

and then bring the air flow, pressure, and temperature to the desired values.

The test was then initiated by startingq the test sequencer. The test was

monitored on the TV system and a determination made on whether the ignition

was normal and whether sustained combustion occurred. In addition, the sam~ple

was analyzed visually after removal from the test chamber. These first hand

procedures were successful in determining if the test was satisfactory for

most tests. A more detailed analysis was later conducted by looki g at the

7 high speed motion pictures.. This analysis showed that a few tests which were

initially- thought to be good were, in fact, not valid because of i.gnition diffi-

culties. These'tests were thien excluded from the stuely.

if the TV viewing and sample anaJlysin showed that the ignition was not

normal, the test was repeated. if sustained-combustion occowa-ed, the air

flow for the next test was inergased and if non-ýsustained combustion occurred,

the 43ir flow was decreased. Eventually, tbe critiCal value of air flow was

found Uiat separated the non-suatained- combustion and the sustaina'd coambustion

regions. "Mis sequence. was repeawed at all cotb-ipnatiorns of th-i three-. press~iras

(448, 793, 1138 kPa) and the three tler*)ratures (121eC, 2600C, aiW; 3990C)'

T'vsts werp carnducted with two La-ample thicknesses (0.026 cm etn4 0-.6 cm).

Iii ~. addition, co~j essor bladeg were tested at betne of the prest~ure8 and t-emper-

atuxes.. The limited number of blades and faciiity teat time availabli' did

AFAPL-TR-75-73

not permit testing the blades over the complete range of temperature and

pressure. Enough tests were conducted, however, to allow some comparison

of results.

The combustion tests were designed to define the air flow conditions that

would support sustained combustion on a sample ignited on the edge by molten

titanium. This is both a function of the air flow conditions around the sample

and the ignition source. An insufficient amount of energy in the ignition

source will fail tn ignite a sample even though the airflow conditions are

amenable for sustained combustion. The effect of the ignition source was

not thoroughly studied in this test program, however, the high speed motion

pictures of the ignition and sample were studied to determine the character-

istics of the ignition source.

If the molten material ignited the sample along the top edge and the air

flow was correct for sustained combustion, the sample would burn completely

to the sample holder. Some burn patterns produced an even, horizontal burn

down the sample. Other burn patterns were more complex and resulted in the

flame burning down the leading or trailing edge first and then burning forward

or rearward into the sample. This was more prevelant with the actual com-

pressor blades because the blade edges are thin and burn more readily than the

center portion.

•--5. -- :A 'As the air velocity on other tests was increased, a point would be reached

such that the sample would start to burn at the point that the molten titanium

from the igniter impinged on the sample but would soon stop burniniý (usually

within a few mm, but occasionally as much as one-half cm). Since the sampleI. i would burn a short distance and then stop, it was assumed that sufficient

energy was present to ignite the sample.

.4 ;,°:

* .

S.AFAPL-TR-75-73

Several additional tests were conducted to establish the effectiveness

of the ignition source. A steady burning was established on a sample by

igniting it in air flow conditions that support sustained combustion.

The air flow was then changed to a condition that had been established as

a non-sustained burning condition. The sample stopped burning immediately

after the air flow was changed. These tests further verify that the

ignition source is sufficient to ignite the sample if the air flow is

correct for sustained combustion.

The air flow required for sustained combustion is also a function of the

angle of the sample relative to the direction of air flow. The angle used in

these tests is 40 degrees, which is typical for a compressor blade. The

effect of varying the angle between the sample and the air flow was not

evaluated in this series of tests.

The following five samples were used in the combustion tests:

1. Sample A - Size: 2.54 cm x 7.62 cz, x 0.06 cm

(1" x 3" x 0.025")

Material: Titanium Alloy

6% Aluminum, 4% Vviadium

2. Sample 13- Size: 2.54 cm x 7.62 cm x 0.16 cm

(W2 x 3" x .064")

Material: Titanium Alloy

6% Aluminum, 4% Vanadium

3. Sample C - Compressor Blade

6% Al=minum, 4% Vanadium (thickness less than Sample D)

4. Sample D - Compressor Blade

S446% Aluminum, 4% Vanadium (thickness greattr than Sample C)

13

AFAPL-TR-75-73

5. Sample E - Size: 2.54 cm x 7.62 cm x 0.11 cm

Material: Titanium A.loy

B structure

B. TEST RESULTS AND DISCUSSION

The test results are tabulated in Table 1 through Table 9. The critical

value of the air flow for supporting sustained combustion is taiulated in

Table 10. The critical value is determined by looking at all tests of one

sample at a fixed set of air flow conditions and estimating the break point

* ' between the sustained and non-sustained burning regions. This is not necessarily

the midpoint between the data points, but is the result of analyzing the

individual tests and considering factors such as ignition. It should be

apparent that this value could vary somewhat due to the interpretive factors

involved in making the determination. This value is, however, the best that

can be obtained from this series of tests. The results are sufficient to

establish trends in the burning characteristics over the temperature and pressure

range of interest in this study.

. Sample A generally burned at a higher air velocity than Sample B, which is

-: expected because of the difference in thickness. The effect of an increase

in air tempar.,ure is an increase in the air velocity at which sustained

"combustion can occur, as illustrated in Figure 13. The effect of pressure

7- ,varies, as illustrated in Figure 14.

The data on Sample C is presented in Figure 15. It should be noted

that two data points are not actual break points between non-sustained and[.-i."; sustained burning, but only measured data points in the sustained burning

region. There were net aufficient tests conducted to determine the actual

break point. The blade shows a definite effect of temperature and

4~ 7i

7A 7. -----

AFAPL-TH-T5-73

-4

0) w

o Zn -4

4) -4 .-14 VO 0 r

0.44 Q.4 0)V 03

4 02.,j ,4

ý0 0 0 04 0043 V i4011 1"01. 0; t 2 w

430 40

0 N 0 U,4

to c )- w 4 ) 443 rtp o :3: I 3 4 odo ý I

U N

%Dd~¶ trI-u cn

;~41 -4.4-m44-44% v -4 m ON 104.) 4 -4r"w

44

N N 4t0~0 av '0% nC,' 0VOn4'

%9 106' %00 0 4 4 4 4 .a 4w % %

414 e o q o qo(- '.1

AFAPIT-75-7

IQI0 d)~ CJ) )

4' w

4.) 0 0 0 00 00 0 0 0w 0 00 W--'.1 0410 4JO0 09

43

41 0 V 4t 4-) 4J V V 4) V 4 1~J)

0

IA M11 n D V 0 V W N m

04J

U9 44

1 n0 n 0 tA 0 t-4-4H Ln tA 0 N N N

-4 4 N 0 kV &fl V' tfLA LA

P4 .4 4 .4 A4H 4.4 .-4 r4

0~~ 0, 0 V VnFc

AFAPL-TR-75-73

14 Itle,

$4 p $wow wo I'llB -4 Qý 01

4)4 410 0 C,0 41

000 00 00 104 0 r. 0

.0 Or 0 ~ 0 0E4 V 1 000 04

ou 0 ' U~ 8 §

, 8~j V 40 0'g ~. 00

14 .-

00

0 oin0 o01% 00 t-,r 1 t- 0a 40c

V fn (n M

0 n tm 41oC4N4Dk1N Dinr o C

41 0 414 414 04( 4.40 CIC r-.C40 N(I

Hm .4.N r4CNH. NC

41C4O0..4.0 C14 .4M 0 O~C'Af MC 4 r4

.- I H .-..4 HH4 'DH

4 -w

AFAPL-TR-75-73

0

4) .40 .04)

a) 0 4)

-dW V4~44 41 0

21- 44.)4r.-d 0 '~-

>'- :4)44 ). 4)

0 >2- M 0

* ~~~~~4)10 14 4.931 4 4140> to- k 4~ 0'~) 4) ~

.0~' 4) '0 V 4J 4 41.0" A A .9 0' A)0 O4) 4)-d m44"4 w444' N44 0A, U04 LT. 4)04 .-4 r.)-44) 1( d4)14 "-1 0) )

0 10 H 0H14

00400 +1 -

~~-1 040 14)00O0 000> 0 040 4

~10 %8U Aow1 10 A) 4) H (44)0q4.) 4j1 .)j H-H10o 04 00: 4

(0 -d41 H 4 04 I44-~ B4 0-V) (00'-1 0 0 0 '.- 4 r. H0

V No 40C' 4J'Ir'M4 j 41) 0

E' 'A~ 0

0H~~ (0r0 I- 0)

V4 to0 04 x'

q m ) H(

0H ,1 f.OD 00(I CO N.-4 0 0 0%%CD0) 0)4) 00v0iuC4)

E-4

LA00 0 n 0N~~ 0H M %DN.4 0 00I nn C n%

C44

m lr- t- en m 1H-4M m ) '4' v %Dm 4 M -4HMU 9( f.. m )(okNN Nn 000): qr %Q MM r.4 (1 r O 4O%-tM N NNC

%D% D% -00 0 .40 0 004 . 4 0O O 0 0

44.4 ~ ~ ~ ~ ~ ý 4~ru4f (0 ý.O')f A)1'14 .4 -)4 4fl) I) IU

10 4 H 0000 I f fn1 CD W MMW ) Cp ( 0 c) v occ m coc

N M H4N rm 104m r- 0 HN4144 H N0 0 0 0 00 0 ~ 00 a '.44 -4H

AFAPL-TR-75-73

A0 -P A ~ V A r4Jr00

0 D 0 0)0 -H H H H 4H H *4,HP4 V 0.'d 04 P, Fu0 U P

tp tp w H '4) 1W0) p 0 P-H a)0 t" a),~41P '0 $4 $4 P4,Wt (D w w

> . 0.. 0 o.0 0~4) 4A)

000 000a 0000 0 00 0 a000 00 0 00 0 000 9C041t 4J 9 T 9C 00i 41 0

4) 0 00 q m a40 w 0 0 0 0 0 W04.) V 0 . 4) A 4J 6H.r4 V4. V 443

0 u -H -A

V 4J H l - -§H 4J .4 4) HH0 AH H41 41 tp 44,-.1

OQH H -40' M - t. '4W ~ ~ ~ ~ ' 11 L( ( nHMt A N A CAt 1 at

H4

u~ 6 0

N Aj W

E- 4 to

0" A22 0)404) fl 4 r

4-~ mrOO or- rALU oQ '0 vLLA D8DQ o noN -c 3iN w D C 0 q 0 , NLn ei . 4 . V A i A n 0 fn

0 d w% nO 4% coLWL fnA0L 0O' L 1OA InM A M %D MmOLA

N4 0 N -4 0 0 0~L 4)4 0H H0 02 N 2 0 0H0 0

6O Q ON' % 4 4-4 %D 100000 ý 0at0%00n-4 10HHOV H

-. 4 ~ ~ .LALAAL 414 'ý4 LAALLA OOýe' .HN

14HH L Z 4O P mp C 0 C, L W 0 ~ '0- r- H In UvU iri i, r -rlr 4 .4N 20 o 0 A i7%00II~ 0~E4 000Do 00. a4

.. y -,c

v*. 4 -0 "l 0 4

0 a

0000 0 0 0

E4b 0,0 0 0 -,1 4 0

0 'o00

20 t0 0003 444

r-4 1000 r 0 00 0

0

fI ~ 43C4J4C

.0 v

.~~~~.4 .~ . ....

114 %D% 1

cc~~~ w.MQ ,0 awm w

0 0

AFAPL-TR-75-7341 3

Y

4

Id 0 ' fd.0 60 Qt v tv4

41I 4)

ua8i i 8]8 vq~ 43 4300A

q-4 4~ 0 ~ ~ ¶f

4~~1 V 0 fu'~ e

3)3~~~4 t- 4 ' 1of -4

j~~ 4j 1 8 t. lz ~r.-Ž 00 r- 0. '0 ' Qt - A( 0 w'' .

C'D

:n4 ; 4 4

YQQ.IteIV

It ~ ~ ~ c ItC - L r.

It2%

tlt41 A.)

piV V V V

. 0 .0 0 .0 .0

44 A.)4 .4 .

ar a a t2 Xvi v3 43 4 4).

........-.. '-

0' iS 0. 0 0. V 0~In

-1-a4 (D

'1 -4 -'q-4

.0 10 0n',

.0 .

z 0 0 0 04 0t4 . 4-J 4J V V 41

00 0 '0 0 0

z 2 0 4 '0 a a2 0, ro0 (v 8

884 14.

0H r 0 G %0 a 0

H .4 ý 9 4 H 4

V44 0 40 0

E41 a, 0 0n 0T 0n 0n 0 0

>, 04 I-4 " -4 -,- .4 -. N ('

'4- '4 4~ ~ ' 4) a ~ ' j -

.~ . .~ . .~ 4-'

8'4 4 0 0 .- 400 --

a 0

4.4 co

-AWr ~0 - a,,1-~ 2 ~ ~ *33

-P8

A.PLTR757

I44411 CD4 e Q 4

0 43co 4 +)

>2 .14 14'

041- 0 3, H2.4-0 3.0 )

8: v

4)001.

-.;4 - 3-O 0".103 03 to

uV 03 U ''W03d

031-, 08 00 0 C0

0i 0 0 003 03v0 0 03 0 0 03

0 3 003.

HI4jli 4 A-)0 000 V -0

w4 0''0l Tha' C!0en

c- A n -w

&~ -4". v~a

~ aj~ nQQ

-A . Q

221

~'.4 Q, 0d 0 ~ -.

1* 4

.::~< 4 ~ W 0't6.3

AFAPL-TR-75-73

-')r.0 U) w ,

M J '- '- ~ ~A4J 4iJ

HO 0) U) In T I

H- 0' Cl) m Lf H N r- LO 0f L) %0C) 0' NCO In D WA vA V) 11 %

0 ) ( -1 H (1 1- V1 LA NA ~ X

0 4N 0 0) en 0 v' 0 ko

E-4 to V In 1 NP H~ vA %D) 0LA C

04Q) 0 r- LAq Hn 0 L COA~ HoMrI D V04-LW)

41

E4 00(14 V) H 14 (n H- LA Z

-27-

In L ILALn % ull In A 0

CO fn -e OD C13 co 0

Ch mA co a%) en LA MLA

~L) H '0 '9 - ' ~ )

) AFAPL-TR-75-73

+ 0.06 Gm(SAMPLE A)

00.16 cm(SAMPLE B)

-448 kPa

80 -- 793 kPaI -- -1138 kPa

I ~(NON-SUSTAIN.ED BURNING)

I 60

..~- .LA.A.... .... -J 40

20 (SUSTAI NED BURNING)

100 200 30040

TEMPERATURE (OC)

/Figure 13. Sustained, Non-Susta-ined Burning ]D~taf~or Samples A and B

AFAPL-TR-75-73

80 (NON-SUSTAINED BURNING)

+-

0

20

(SUSTAINED BURNING)

4oo 600 800 1000 1200

pfwl~SUHE (ko.)

+ SAPLE (0.611)121 0 CSWý~~ Li A 00-' 2600C

O SAMPLF 3 (0.16"1) -- -39C

Figure 14. Sustajined, Non-6 -tined Burnin~gData f'or S~mples A and B

37

APAPL-TR-T5-73

8011

(NON-SUSTAINED BURINING)

40.

20-

.................-448 kPa

(SUSTAINED BURNIN~G) 9 P

100 200 30400

TEMPERATURE (0 C)

F'igure 15. Susta~ined, Non-Sustained BurningData for Sample C

38

AFAPL-TR-75-73

also pressure. The pressure effect may be causad by a change in the air

flow across the sample at higher pressures.

There are not sufficient data points with Sample D to establish an5

trends in the burning characteristics. The data on Loth types of blades

indicate that the blades will bura generally in the same air flow conditions

as the flat surface samples.

A few tests were conducted with a ( structure titanium alloy (Sample E).

The thickness of this sample was between that of Sample A and Sample B. It

was not possible to achieve sustained burning with this sample at an air

velocity that supported sustained combustion with Samples A and B. In addi-

tion, the burn rate of this samnle at 2600 C, 1138 kPa, and 20 m/sec air

velocity is 10 cm/min, which Iq somewhat less than the burn rate measured for

Samples A and B at the same air flow conditions. The burn rate of the other

samples is discussed in the next section.

These limited test results indicate that the alloy and structure of

titanium have a definite effect on the burning characteristics.

4 -3

I'

AFAPL-TR-75-73

SECTION IV

BURN RATE ANALYSIS

A. ANALYSIS DESCRIPTION

The burning rate for samples which exhibited sustained burning is

determined by analyzing the high speed motion pictures. Each test was

photographed at 60 frames per second. Burning rate can be expressed in

several ways, such as weight change or volume change. The speed of the

burning edge as it burned down the sample is used for this analysis.

Difficulties occur, however, with this approach because the sample often

does not burn at a constant rate over the whole sample. What is described

here is a determination of an average burn rate on the portion of the

sample which d4.d experience a fairly steady and even burn rate. The top

edge near the igniter and the bottom edge near the sample holder were excluded.

As illustrated in Figure 16, the burn rate is determined by measuring the

v time it takes the sample to burn between two points on the sample. The end

points were varied somewhat from test to test to exclude variations such as

an unusually slow initial burn rate or a burn pattern which developed an odd

shape. If less than 2 cm length of sample burning could not be approximated

by a straight line burn pattern, the burn rate was not determined for that

specific test.

Analysis of the motion picture film reveals some general characteristics

of the burning. If the sample ignites across the entire top edge, it

generally burns from top to bottom. Sometimes the sample will burn faster

down the frot edge and then burn back toward the trailing edge. Other

~4O

AFAPL-TR-75-73

START POItNT

I AL

END POINT

t to t tl t t2

* ~BURN RATE A=

t2rtFigure 16. Burn Rate Measurement Technique

AFAPL-TR-75-73

tests result in the sample burning faster down the trailing edge and then

forward into the sample.

4 AB. RESULTS AND DISCUSSION

The burn rate data are tabulated in Tables 11 through 14. Although

the burn rate varies considerably even at the same awýr flow conditions, it is

possible to establish trends and the lack of trends from these data.

The temperature effect on the burn rate is too small to be determined. Not

much effect is expected because the ambient temperature change is small com-

pared to the temperature at the burning surface.

A definite trend due to pressure is apparent with Sample B and somewhat

apparent with Sample A. The samples burn faster at higher pressure. The air

is more dense at the higher pressure and thus more oxygen is available for

4 Icombustion.

In general, Sample A burned faster than Sample B under the same test

conditions.

S..The burn rate of the compressor blades (Samples C and D) does not show a

trend due to pressure or temperature. Since the blades were not tested over

the full temperature and pressure range, it is difficult to determine a trend.

The blades appear to generally burn at a rate between the rates of Samples A

and B. The blade thickness varies from the thin edges, which are approximated

• .by Sample A, to the thick center, which is approximated by Sample B. The

imasured burn rate for the blades is an average rate over the complete blade,

and thus would be expected to be within this range. In a few tests, a blade

04--,. edge burned several times faster than the center portion.

S! 4

AFAPL-TR-T5-T3

\,O. coC\CC)0 cfl I-- Ht- O-~-T- n H\0 mw _r

0 t- C'JC 01 k -t 10 Lr\ _z\O t'- ztr 0 \10 _:t!t- 0\o1) oW'\0 k t(U\Lr\ 0 - M 0 (Y) \DLr\ \0~Q~ f~f, ~ 'Wl HHH Hqr - - rIr -H 04 HHH co GHr HC\ OH

Cl) 4-

C.)

M~ --I\O h--r SID U-U4 ) -4 ( _r-T \ C\I\C t J\\U IA

K E-4E-A-

O\s m0 (h 0a _00 00h 00 000 0

t^L\U\ L 0 000 000 LA~u utuA UNV\ký4' cu ('OC'i N N40 L^ UN.Um 'UN~' UNV - -- t- t- t' t--

HH e-H 000 000 ggh ON

AFAPL-TR-75-73

I-4

0 - tco uN w COC Oe\ Q( m(V

4ý t r--

-~--CY V *

F-.UN

iC i

1 -1C4C

II

4V

0- *4C\J 0 r-

u 1Ur r'1\r4 A 4 A -

ci

p' E-4\ -U\\~ .

at u,4 H '~ H H '

~ H H1-i.~ Cx V.~~J~t- ..

g 0 00 0 c

vN UNjHH u- 00 0 C)HN .-4 .4J Ml VN x xi(3

hl 0

AFAFL-TH-75-73

.d - 4 Lr M.IJ-

U H UH \ HH:

14c

m El 0 0 \

v\ar

C-I

C..al

Ch IO C- ®r

.. ...

I AFAPL-TR-75-73

SECTION V

FLAME SUPPRESSION TESTS

A. TEST DESCRIPTION

The titanium flame extinguishing tests were conducted to determine what

concentration of argon gas is required to extinguish a titanium fire. Argon

is believed to be inactive in the presence of the high temperature titanium

flame and can thus be used to decrease the oxygen concentration to a level

below that required to support combustion. The tests were conducted at

air flow conditions of 260 0 C (5000 F) and 1138 kPa (165 psia). The air velocity

past the blade prior to argon injection was 24 m/sec (80 ft/sec). The tem-

perature and density of the mixture changed the velocity in the test chamber

after injection of the argon gas, but these changes were not sufficient to

cause the flame to extinguish. The mixture velocity after argon injection

dropped to 19 m/sec (61 ft/sec) for the worst case (100% argon).

Attempts to calibrate the concentration of argon by using a rotary vane

I ,flow meter did not prove successful because the meter output did not

change sufficiently over the range of use in these tests and thus czused

inaccurate data. With a sufficiently higher pressure in the argon ganifold

than in the test chambor, the flow is held constant by a critical orifice

located after the solenoid operatc~d itnjection valve. After a short transient

upon opening the valve, tle fl.Iw is controlled by the argon manifold pressure

;> < •.which also experiences a short transient (less than 0.1 sec) and then reaches

a• steady state.

J

*'~-~'~~ -~47

i~N

SYAPL-TR-75-73

The following procedure was used to calibrate the argon injection system.

The test chamber was operated at the air flow conditions required for the

- L4 flame extinguishment tests. Argon was injected at six different values of

manifold pressure, and the argon-air mixture was sampled near the test sample

at these six values. The values of pressure required were estimated from

calculations of critical orifice flow. The range of argon concentration of

interest was between 50% and 100%. The six gas mixture samples were analyzed

for composition. The percent argon mixture can thus be determined from the

argon manifoid pressure. The calibration data are illustrated in Figure 17.

The amount of argon injected is assumed to be linear with] respect to the

manifold pressure. The highest pressure data point does not fall in line

with the other points. Most likely, an excess of the 100% argon is actually

injected at this pressure and the excess is bled off through the pressure

regulating bleed valve.

The argon-air mixture .teaches a quasi steady-state temperature within a

few seconds foll-wing injection of the argon. The argon cools after expansion

through the critical orifice. The pipe between the injection point and the

test chamber is, however, at the initial air flow temperature and contains a

larg'i mass which heatA the argor Air mixture. The temperature of the mixtur"

did not change significantly after reaching the steady-state value.

B. RESULTS AND DISCUSSION

"j} The test IVstilts are tabulated in Table 15. Four tests were conducted

with Samp)le B, and the flame was extinguished in each test. Seven tests were

conducted with the thinner sample (Sample A) with the argon concentration

I• ranging from 45% to 100%.

CJ.

:" •-

.. - ..>-.

AFAPL-TR-T5-T3

100+

90 +

I' +

8o

+

70-

+

¼ 60 +0

50

4o

30

20

1300 kPa 00 ~ 2760 kPzi

A RGOIN MANIFOLDI) ESUF

~i~t &Y.Arn injection Calibration Curve

* 41i-

*4

KRIM a~

AFAPL-TR-75-73

TABIZ 15

ARGON GAS EXTINGUISHING DATA

TEST SAMPLE %ARGON TIME TO EXTINGUISHMENT

8AZO1 B 90 z-0.1 see

6AZ02 B 98 .0.1 sec

8Azo4 B 100 '01see

8AZ0)4 B 100 < 0.1 sec

8AV08 B 48 1.0 sec

BAV1O B 62 0.5 sec

8AV11 B 50 0.5 sec

~f I8A705 A 100 < 0.1 sec

8Az06 A 90 <. 0.1 sec

-I8AZ07 A 73 0.15 sec

8AZ08 A 66 0.15 sec

8AZO9 A 55 7 sec (See text)

8 AZIO A 45 Continiued to buirn

8AZI 2 A 100 <- 0.1 sec

Pressure 113"t kFa (165 psia)

Temprerature 2600C (5000F)

Air Velocity 24m!s~e (80 ±'t/sec) prioi to argon injecti.on

45

AFAPL-TR-75-73

It is somewhat difficult to determine exactly when the flame is extin-

"A qguished since the material remains hot for a period of time. The argon

I arrival time at the sample was calculated by assuming that the argon and

air were completely mixed at the injection point and that steady state con-

1 ditions were reached immediately after the valve opened. The flame was

considered extinguished when all molten material stopped leaving the sample

surface and activity on the surface slowed considerably. On the tests with

high argon concentration, this effect took place within 0.1 second after the

argon mixture reached the sample; however, the time to extinguishment listed

in Table 15 can only be considered accurate to within 0.1 second due to t!he

interpretive nature of the data analysis.

Concentrations of argon above 65% effectively extinguished the flame,

.- . although slightly longer times are required for the tests at 66% and 73%

-" •than for the tests with 90% and above. The test conducted at 55% concen-

1 tration showed a different characteristic. A definite slowing of the com-

bustion took place within 0.1 second after arrival of the argonj however, the

sample continued to burn for another 8 seconds before surface activity stopped

and the sample cooled. The test conducted with Sample A at 45% argon con-

centration showed the same initial decrease in burning activity; but in this

. . test, the burning continued for considerable time and the sampla almost

-~ completely burned.

I Additional tests were conductod to determine w~hat would happen if the

argon concentration wum decreased after initially extinguishing a burning

sample but prior to Uhe sample cooling. In these tests, the chamber air flow

conditions were the same as sho-n in Table 15. The test sample (Sap.lle A)

was ignited and allowed to achieve a steady burn. The flame was then

................................... :-

AFAPL-TR-7 5-73

extinguished with 75% argon air mixture, which was maintained in the chamber

4 for 3 seconds. The argon was then turned off and the chamber returned to

the initial conditions within 1 second. At this point, the sample which

was still hot re-ignited immediately and started to burn. When the argon

concentration was maintained for a sufficient time to allow the sample to

cool to a dull red glow, the sample did not reignite after removal of the

argon. The time required for the sample to cool was as great as ten seconds

for these tests.

These tests point out a definite design requirement on such a technique

utilized for protection of a test facility. In a large scale test, a large

amount of molten titanium and other materials would be present following a

fire. Sufficient argon would need to be supplied until the ignition sources

cooled (possibly a long time) or until the air flow changed to a condition

that will not support combustion. Most likely a combination of both these

techniques would be used to effectively protect a facility.

Several tests wre conducted with CO2 gas, which is a co=n fire extin-

guishing agent for hydrocarbon-type fires. The C02 gas was injected and

calibrated with the same hardware that was described for the argon extinguish-

ing tests. In these tests , t:e sample was ignited and achieved steady burning

. the. burn rate iocreased consid~erably after the C02 was injected. The- tests

with 23% C.O_ show about a 50% increase in burn rate, while the tests with4

nearly l0O0 C02 show an increase in burn rate of about 300%.

"N"J... 5

AFAPL-TR-T5-'T3

N C4

N N

to U

44

-v4

t9 N

tn ~ ~ ' UntA.4 o4

a;o o 021

~~53

44 53

AFAPL-TR-75-73

SECTION VI

ULTRAVIOLET (UV) EMISSION ANALYSIS

A. TEST DESCRIPTION AND DISCUSSION OF RESULTS

A good method for detecting the occurrence of a titanium fire is to sense

the ultraviolet (UV) radiation emitted from the burning titanium. Spectral

emission data from metal fires are available and indicate that energy is

• • i emitted from 2000 • and 3000 • (Reference 1). This spectral region is of

interest because solar blind ultraviolet detection systems that operate in

'Liii I this wavelength region are available (Reference 2). These systems are presently

used for fire detection of hydrocarbon fuel-type fires.

The tests conducted were designed to determine if these developed detection

systems are applicable to titanium fire detection. Actual spectral lines are

not measured with the spectroradiometer equipment used in the tests. The

!- test equipment set up is shown in Figure 18. The spectroradiometer was set

at one of five wavelengths and the incident UV po%er on the detector measured

as a function of time while a sample burned. The UV emission from a typical

"test is shown in Figure 19. The initial peaks are caused by the igniter burn-

ing. As illustrated, the output varies considerably with time due to the

fluctuations in the burning. The measured spectroradiometer detector output

current is averaged after the initial ignition and prior to the trailing off

as the sample is consumed. This average value is then used to calculate the

UV power at taspcfcwvlnhas shown in Table 17 .~ vial

Although spectral peaks are not maasured, the average power available

in the solar blind UV spectral region can be used to aproximate the

1_4:

AFAPL-TR-75-73

SANPLE

DETECTOR

L-~~ 600

Figure 18. Ultraviolet Measurement Instrument tion

102

INI TION iTIMiE (SECONDS)

Figuire 19. Typical Ultraviolet Emission b~casuremen

L 0

-44

P-1 H

N IV rH m-

0 4)

-k-

814

4J 00 co 0~ m~ 0 MA

00 0 0 o LA IU-4 r4 rH 14 H -''

1.14

t~o

03 03 03 3

r_ 4 0 0 0 0 C H Q-

03 00 0 0 kA

ý44

co co O

AFAPL-TR-75-73

sensitivity of a UV detector to a titanium flame. This calculation requires

knowing the spectral response of the detector, the spectral output of the

source, and the distance between the source and detector. The UV emission

from the 2.5 x 7.6 cm titanium sample is compared to the UV emission from a

5 inch diameter pan fire of JP-4 fuel in Table 17. A typical hydrocarbon

flame UV detector can detect the pan fire at a distance of 10 feet. This com-

parison shows that existing UV detectors can be adapted to detect titanium

fires, since they have adequate sensitivity. The detector is, however,

limited to a line of sight operation. The UV detector cannot distinguish

between sparking, which might occur as a result of rubbing, and an actual flame

because both generally have the same spectral output, however the signal level

might be used to discriminate between sparking and a titanium fire.

An engine test facility can be protected by the installation of detectors

at key locations where a titanium fire might occur.

Further analysis with equipment capable of measuring emission over the

UV spectrum and capable of resolving the narrow, spectral lines would be

;:•:?.•..required to dotermine if selective wavelength-detecters can be used to

disinuih aV~anumflame from a hydrocarbon type- fla e. Texingih-~

"-inj technique for the two types of flam-s wy be difernt a.d thus dis-

2- crimination betwe.en the two flames may be required.

5 7

AFAPL-TR-T5-T3 BBIGAU

ft 1. Leen, Albert, "An Ultraviolet Sensing Flame Detector for Use On

High Performance Military Aircraft," Technical Report AFAPL-TR-

69-107, Air Force Aero-Propulsion Laboratory, Wright-Patterson AFB,

Ohio; February 1970.

2. Runyon, C. C.; Moulder, J. C.; and Clark, A. F., "Time-Resolved

Spectra of Bulk Titanium Combustion," Combustion and Flame,

~ j Volume 23, Number 1; August 1974.