2018 Insensitive Munitions & Energetic Materials Technology Symposium Portland, OR

Slow Heating Testing Survey and Historical Events Review

Ernest L. Baker

Munitions Safety Information Analysis Center (NATO), Brussels, Belgium

This report describes the results of an international review of the STANAG 4382 Slow Heating, Munitions Test Procedures, as well a review of heating rates and durations associated with actual fire events. The purpose of the slow heating test is to assess the reaction, if any, of munitions and weapon systems to a gradually increasing thermal environment. To perform the review, MSIAC created a questionnaire in conjunction with the custodian of this STANAG, the United States, and sent it to subject matter experts including test centers in most of the AC/326 nations. The questionnaire questions deal with the test purpose, test procedure, heating rate, actual events, oven design, oven standardization, temperature preconditioning, energetics melting, reaction temperature, test item restraints, test item orientation, instrumentation, and number of tests. This report provides an analysis of the answers received, summarizes best practice and provides some recommendations to potentially support an amendment of STANAG 4382. These recommendations are being discussed within the NATO AC/326 SG/B Slow Heating Custodial Working Group (SH CWG). The working group has already reviewed the review results and is currently drafting updates to STANAG 4382 NATO documentation, which includes the technical content of the STANAG that is being migrated into a new AOP 4382.

INTRODUCTION

This report describes the results of an international review of the STANAG 4382 Slow Heating, Munitions Test Procedures, as well a review of heating rates and durations associated with actual fire events. The purpose of the slow heating test is to assess the reaction, if any, of munitions and weapon systems to a gradually increasing thermal environment. To perform the review, MSIAC created a questionnaire in conjunction with the custodian of this STANAG, the United States, and sent it to subject matter experts including test centers in most of the AC/326 nations. Moreover, an analysis of similar standards has been done in order to achieve more consistency in the recommendations. From a NATO point of view, the requirements for the slow heating test are defined within three documents: STANAG 4439, STANAG 4382 and AOP-39. The test 7 (h) from the “UN – Manual of Tests and Criteria” specifies a slow cook-off test for the classification into hazard division 1.6. The questionnaire questions deal with the test purpose, test procedure, heating rate, actual events, oven design, oven standardization, temperature preconditioning, energetics melting, reaction temperature, test item restraints, test item orientation, instrumentation, and number of tests. This report provides an analysis of the answers received, summarizes best practice and provides some recommendations to potentially support an amendment of STANAG 4382. BACKGROUND

In 2015, MSIAC carried out a review of STANAG 4496 related to the fragment impact test This review was managed the same way as this current one, and resulted in a list of recommendations that are currently being discussed in a custodian working group to update STANAG 4496. Following the review of the bullet and fragment impact tests, MSIAC proposed

to perform a similar review for the slow heating test, on behalf of the United States who is the custodian for this STANAG. REQUIREMENTS From a NATO point of view, the requirements for the slow heating test are defined within three documents: STANAG 4439 [1], STANAG 4382 [2] and AOP-39 [3]. The test 7 (h) from the “UN – Manual of Tests and Criteria” [4] specifies a slow cook-off test for the classification into hazard division 1.6. Analysis of the requirements The table hereafter compares the STANAG 4382, AOP-39, and the UN Manual of Tests and Criteria test 7(h) regarding the slow heating test: Table 1: Differences between the STANAG 4382, AOP-39 and UN orange book test 7(h)

STANAG 4382 ed.2 AOP 39 Ed. 3 UN 7 (h)Alternative procedure Yes No

Number of tests 2 2Item configuration Bare or logistical, as

agreed by the national authority

Bare or logistical

Logistical

Test Procedure Yes Yes

Heating rate 3.3°C/hr 3.3°C/hr

Preconditioning Temperature 50°C for 8 hours or until equilibrium at 50°C

5°C below the predicted reaction temperature

Maximum Temperature 365°CReaction level acceptable Burning or no reaction Burning or no reaction

The main difference between the documents is related to the item configuration:

In logistical configuration for the UN document. This seems logical, as this document relates to the transport classification of the article;

Bare or packed, as agreed by national authority, in the STANAG, which seems logical as the national authority is able to define when a fire is more likely to impact the munitions during the life cycle.

An alternative procedure is provided in the STANAG: if no analysis has been done, a rate of 25°C per hour should be used as a default rate. With respect to temperatures specified, there are 2 main differences: the preconditioning temperature is different (higher for the UN) and the UN defines a maximum temperature. The STANAG provides more details on the test procedure and includes a basic test set-up description. Neither the STANAG nor the UN document provides a detailed example of test set-up. In addition, there are redundancies between the STANAG 4382 and the AOP-39, especially in the observations and reports part. They should be avoided to allow these 2 documents to remain independent. Indeed, the AOP-39 is linked to the STANAG 4439, and it is not automatically updated when there is a change in one of the STANAGs that defines the test procedure, like the STANAG 4382. The 3rd edition of AOP-39 includes Appendices which provided intermediate updates of all the IM full scale tests not referenced from STANAG 4382 and the contents of the Slow Heating appendix needs to be included in the review of STANAG 4382. MSIAC was requested to support AC/326 SG/B, which was agreed by the MSIAC SC, to review all these documents to remove redundancies or contradictions and to clarify where the

informatiupdated guidanceSTANAGSTANAG QUESTI The quesnation resites fromheating energeticinstrumeprovided Origin o MSIAC hthe fact twere tak(Canada(South Aand 38%group re The follo

Figure 1 The follo(duplicat

ion should documents

e on commoGs have recG 4382 need

ONNAIRE

stionnaire wepresentativm previous rate, actuacs melting

entation, andd with additio

of the Answ

has receivedthat they weken into ac

a, France, GAfrica, Swed% from privat

sponses, ra

owing chart s

: Number of

owing table ptes indicate

reside. Ass. The revieon aspects ocently been d to reflect a

was conductves for distrisurvey respl events, o

g, reaction d number oonal comme

wers

d answers fere not perfoccount for

Germany, Thden and Finte test cente

ather than in

shows the o

f answers re

presents themultiple ind

s a result, mew includes of IM testingreviewed in

a common s

ted through bution, as w

ponses. Theven design

temperatuof tests. Fo

ents.

from 11 natorming this tthe analysihe Netherlaland). 62%ers. If shoudividuals.

origin of the

eceived by n

e complete ividuals from

most of thedrafting of

g, reporting n conjunctiostructure and

the MSIACwell as a lare questions n, oven stanure, test or each qu

ions. Withintest. Therefs. Answends, UK, No

% of the answld be noted

answers by

nations

list of test cm the same

e redundanca Standardand docum

on with this d content w

web site. Arge group ofdeal with thndardizationitem restrestion, a b

n these answfore, only thrs were anorway and twers come that the ma

y nations:

centers and organizatio

cies will nos Related D

menting. OthMSIAC revith these oth

A notice waf identified lhe test purpn, temperatraints, test

breakdown o

wers, one ne 34 respon

nalyzed fromthe USA) afrom govern

ajority of res

national auon):

o longer exiDocument toher IM full scview and chher STANAG

as sent to thelaboratories

pose, test prture precont item orof the resp

nation acknonses from 10m 7 NATOnd 3 partnenmental tessponses rep

uthorities wh

ist in the o provide cale tests anges to Gs

e MSIAC s and test rocedure, ditioning, ientation,

ponses is

owledged 0 nations

O nations er nations st centers presented

ho replied

Table 2: List of facilities and nations who replied to the survey

TEST PURPOSE The survey participants were asked about the purpose of the test. They were asked if the test purpose was to provide an extreme heat rate different from the fast cook-off test, if the test purpose is to characterize the munition being tested, and/or if the test purpose is to simulate a real life accident scenario. Additionally, they were asked to comment as to the reason that the slow heating test was developed. The majority agreed with all three statements, but a larger number agreed that the test purpose was to characterize the munition being tested. Figure 2 presents circle graphs of the responses.

Organization Country StatusDOS Australia Government

DRDC Valcartier Canada GovernmentGD-OTS Canada Canada Private

AC/326 Czech Republic GovernmentTest Firing Center Finland Government

AC/326 – DGA France GovernmentNEXTER Munitions France Private

Airbus Safran Launchers France PrivateWTD91 Germany Government

MBDA Systems Germany PrivateMBDA Systems Germany Private

Centre of Excellence Weapons and Ammunition

Netherlands Government

AC/326 Norway GovernmentAC/326 South Africa Government

Bofors Test Center Sweden PrivateQinetiQ United Kingdom Private

BAE Systems United Kingdom PrivateUS Army IM Board United States of America GovernmentNSWC Dahlgren D United States of America Government

Redstone (Army) United States of America GovernmentEglin Air Force United States of America GovernmentEglin Air Force United States of America Government

AFLCMC/EBDP United States of America GovernmentNAWC China Lake United States of America GovernmentNSWC Dahlgren D United States of America GovernmentNSWC Dahlgren D United States of America GovernmentNAWC China Lake United States of America GovernmentNAWC China Lake United States of America Government

DDESB United States of America GovernmentYPG ATC United States of America Government

NAWC China Lake United States of America GovernmentNSWC Crane United States of America GovernmentNSWC Crane United States of America GovernmentNSWC Crane United States of America Government

Figure 2 For Wha A large n

•

• •

•

•

• •

•

•

•

•

: Test purpo

at Reason W

number of coThe primwhat it mthe reactSimulateIt shouldrate nowThe testspecific SCO. O“severe” testing toearly on It may nreferenceTests peTo simuhold, preAnecdotaoccurringstatemenhow thiselse. OriginallyDue to firate achiThe 6°F conditionwas baseThe test to ensuretemperat

ose respons

Was the Slo

omments wmary reasonmight do in ttion. e the slow hed be performw in use (3.3t forms a hytesting is rever time, wenvironme

o the extremin the desig

not necessaed for all cu

erformed to flate a fire in

esumably onally, it wasg aboard a nts that “one SCO cook-

y several ranancial conievable as aor C/hr he

n to reduce ed on the limalready exi

e a munitionture or with

ses

ow Heating

were received for doing athe real wor

eating of a tmed to simu°C/h). I douypothetical equired to rewe have bent. We wou

mes. SCO agn processarily be reastomers andfulfill the cusn a storagen a ship, or is to serve

ship with e US Navy -off environ

ates were testraints, this

a “bookend” eating rate wthe influencmitation of tested – it wan didn’t autohin a certa

g Test Deve

d. These coany IM test irld and then

test item dueulate a real bt that it actscenario welate to speen able to uld have neaids in down

lity but it wd munitionsstomer reque hold that sin a depot oas a worsemultiple beship burnedment can e

ested provids slowly evoto the Fuel-

was intendece of item siest equipme

as just adapto ignite undeain period o

eloped?

omments wis to charac figure out a

e to heatinglife acciden

tually does swhich is worecific scenaimprove th

ever gotten n selection o

will be a stas requiring teuirements. slowly heats

or railcar. e case testlow deck md for three d

exist (in its c

ding thermaolved into us-fire test. ed to achievize, mass, aent availableted to SCOer very extreof time. At

ere: cterize the ma way to mi

g from an adnt scenario so. rst case. Otrios that fal

he reaction to where wof design co

andard comesting.

s a nearby

t to simulamagazines. days”, but ncurrent extre

al characterising only the

ve heating and other phe in that era. Prior to thaeme conditiothat time,

munition – tonimize the i

djacent buildbut with the

ther additionl between Fof munition

we are todayoncepts or

mparator tha

or adjacent

te a seriesI’ve seen a

no other ratieme form) a

istics for a e most extre

in a near-ishysical featu. at it was useons prior to the reactio

o find out impact of

ding e heating

nal more FCO and ns to this y without materials

at can be

t storage

s of fires anecdotal ionale on anywhere

munition. eme slow

sothermal ures; and

ed to test a certain

on wasn’t

•

•

•

TEST PR The survthe SCOseparateapprovedby the Stheir natiNSWC Creasoninresponse

Figure 3 Nationa When ascommen

•

•

importanbe testedDeveloperate thanIt has bmagazinHistoricaEdition (4382).

ROCEDURE

vey participaO test as ree nationally d procedureTANAG 438ionally agreCrane divisng for develoes.

: Test proce

lly Approve

sked for whnts received

Confirm acceptanThis is aper lates

nt, just the td at the timeed to provid

n in a fast cobeen establnes can be sally SCO wa(HD 1.6) an

E

ants were aequired by approved p

e developed82 primary t

eed test procions provideoping these

edure respon

ed Specific

hat reason were: that our tr

nce from all procedure

st requireme

ime to reace. de a scenaook-off testished than

slow heated as imposedd then intro

asked about the STANA

procedure. d. The vast test proceducedures. Aed the deslocally agre

nses.

c Procedure

was this na

rials plan mparties concreated to g

ents as well

ction. And, th

rio in which

in the casas low as 6

d by the USoduced in th

the test proAG 4382 prAdditionallymajority of rure. Germa

Additionally, ignations ofeed procedu

e

ationally ap

meet the recerned. give clear coas other Ha

he heating

h the test ite

se of fire o6°F/hr. S in UN Mahe IM policy

ocedure. Trimary test y, for what rresponders any and Frathe US Navf their localures. Figure

pproved spe

equirement

oncise stepsazard Classi

rate was th

em is heate

on the ship

anual of Tesy (MIL-STD

hey were aprocedure,

reason was conduct SC

ance providevy NAWCWlly agreed te 3 presents

ecific proced

and has o

s to conductfication Tes

e slowest th

ed at a muc

p deck, nei

sts and Crit2105 and S

sked if theyand if theya specific n

CO tests as ed the desig

WD and the test procedus circle grap

dure develo

bjective rev

t a slow coosts

hat could

ch slower

ighboring

teria 2nd STANAG

y conduct y have a nationally required

gnation of US Navy ures and

phs of the

oped, the

view and

ok off test

Locally A The com

•

•

Another former Mdefinition2105B.

Figure 4 HEATIN The survrate shouSpecificashould bwhether About haresponderespondepresents

Approved S

mments assoProceduand natreasonabSimpler dcold

interesting MIL-STD 21n. Figure 4

: MIL-STD

G RATE

vey participauld be chanally, they wbe used in pr

they had aalf of the resers were uers thought s circle graph

Specific Pr

ociated with res developtional), envble cost design and

related com05B (12 Ja

4 presents

2105B conf

ants were aged and if it

were asked rocedure 1, ny informatsponders ag

unsure or sthat item s

hs of the res

rocedure

local specifped to meevironmental

it is easier t

mment was tanuary 1994a diagram

figuration of

asked about tem size shoto commenwhy shouldion on duragreed that thstated that size should sponses.

fic proceduret STANAG

requireme

to control th

that some te4). It is desof the slow

the slow he

the test heould be a cont on what d this rate beation or ratehe heating rit depends not be a co

res were: G requireme

nts, and g

he temperatu

est areas usscribed as

w heating te

eating test.

eating rate. onsiderationheating rat

e used, shos of actual rate should on the tes

onsideration

ents, safety gather qual

ure if the ou

se the test sa very conv

est configura

They were n in definingte they useuld size be real world sbe changed

st intent. Tn in defining

requiremenlity test da

utside tempe

set-up definvenient andation from

asked if theg a slow heae, what heaa consideraslow heatingd. HoweverThe majoritg the rate.

nts (local ata at a

erature is

ned in the d efficient MIL-STD

e heating ating rate. ating rate ation, and g events. r, several ty of the Figure 5

Figure 5 What Ra The com

• •

• • • •

What Ra The com

• •

•

What Sh The comshould b Should k

• Should b

•

: Heating ra

ates Are Cu

mments receMajority Sometimsystems not be reWhateveAs requeEither 3.We use intermed

ange of Rat

mments receLess thaThis is vrate if theNearly arates areamps to

hould be th

mments recebe based on

keep the origIt should

be based onIt shouldbe based

ate response

urrently Use

eived were: response: 3

mes faster rto ensure t

eal world. er the customested by cus3°C/hr or 253.3°C/hr fo

diate rate, to

tes is Acce

eived were: an 3.3°C/hr avery depende item is not

any rate cane limited by tthe elemen

e Heating R

eived can b“real world

ginal rate: remain at 3

n “real world be realisticd on real w

es

ed?

3.3°C/hr rates – up they will wo

mer wants …stomers…ty5°C/hr or standardo allow us to

ssible with

and greater dent on the t extremely n be prograthe numberts.

Rate be in P

be divided idata” and th

3.3°C/hr.

data”: c to the scenorld data, p

to 100 °F/hork for more

…but almosypically 6°F/

tests for qo run one tes

h your Heat

than 25°C/hsize of the large. mmed. Lowr of heating e

Procedure

into three rhat the rates

nario that a spast occurre

hr – are use than the st

st always 3.3hr or 50°F/h

qualification.st per day.

ing Equipm

hr item under

w rates are eelements. A

1?

esponses: s should be

slow cook-oences, and S

sed for evatandard rate

3°C/hr hr.

. For routin

ment?

test. We ca

easily accoAt this time, a

keep the oworst and t

off test is simSME evalua

luation of me which ma

e, we often

an test at alm

mmodated, an upper lim

riginal rate, the most like

mulating. Thations of wh

mitigation y or may

n use an

most any

but high mit of 150

the rate ely.

is should hat is and

what is not a realistic heating rate in real world incidents. A suggested heating rate with some supporting data is 45-50 degrees F per hour.

• The heating rate should be changed to a rate that is consistent with reaching cook-off temperature within a reasonable time for a fire to be extinguished. If the worst case is 24 hours, then the heating rate should be determined based on the item reaching cook-off temperature in 24 hours.

Should be worst case and most likely:

• The problem with the current rate is that we don’t know if it will produce the worst reaction. No, it’s not “real world”. But, the real world rate for any munition will be highly dependent on the life cycle of that item. There should probably be at least two rates – a worst case rate and a most likely rate (one set by the specific program depending on the life cycle assessment).

Why This Rate? The received comments are:

• Maintain compatibility and comparability with previous test data. • We should be assessing for IM compliance over a range of slow rates, rather than

just at a single point. Having a single point may enable developers to focus on passing just that single requirement, whereas passing over a range of rates might make them more focused on a better IM solution.

• A suggested heating rate with some supporting data is 45-50 degrees F per hour. • With modern day fire fighting equipment aboard ships, fires should be completely

extinguished in less than xx-hours. I do not know what that reasonable timeframe is but we should be able to determine it from the experts. The example I cited above was for a 24-hour fire. Using a cook-off temperature of 180-Deg C, starting at 50-Deg C, 130 divided by 24 hours equals 5.5-Deg C per hour.

• Various energetic materials will have their worst case reaction at different rates. There will not be one rate value that will invoke the worst case reaction in all or most articles or even components. Retaining rates of 3, 4, 5° C etc. will only be valuable for scientific research. A real fire will be extinguished well before any reaction occurs (and won’t last for two days).

Should Size be a Consideration? The received comments are:

• The size of the item to be tested is not foreseeable. It is as big as it is. • The rate should be based on what can be expected in a real world application. • It might be necessary/prudent to tailor the heating rate(s) to item size in order to

assess the effectiveness of reaction mitigation features across a range of credible stressing conditions (see previous comment).

• We already have an artificial rate…don’t make it worse by adjusting the rate based on size. How does that relate to anything real?

• We should remain standardized for all, rather than variable. • Worst case should be used, whatever that rate is.

OVEN DESIGN The survey participants were asked about test oven design information, including oven construction material and thickness, the oven heating system, the oven airflow and oven photographs. They were also asked about oven design issues that affect the testing and potential test outcome, including the item spacing to the oven wall, the observed temperature homogeneity while testing and about protection for energetic material exuded out of the item.

General Figures 6The ovenon the sc

• •

•

•

Figure 6

Figure 7

Oven Desi

6-17 presenn designs vcale and typ

Our ovenDesign fbeing tesIt's an expipe beininsulatedWe have

: NSWC Ch

: Qinetiq, U

ign Informa

nt provided pary significa

pe of test iten designs arfor oven: Ssted. xpandable dng covered d box. e various ov

hina Lake, U

UK.

ation

photographsantly and it im. Below are tailored fo

Square oven

design conswith an ele

vens.

US Navy.

s and diagras clear that

are the commor individualn or cylinde

stituted of a ectric blank

ams of variotest areas u

ments receil items.

er oven dep

metallic pipket. This set

ous testing fuse differenved on gene

pendent on

e surroundit is enclose

facility oven t designs deeral oven de

dimension

ng the muned itself in a

designs. epending esign:

of asset

ition, this a thermal

Figure 8

Figure 9

: Bundeswe

: Yuma Prov

ehr Technica

ving Ground

al Center for

ds, US Army

r Weapons

y.

and Ammunnitions, Germmany.

Figure 10

Figure 1

Figure 12

0: Explosiv

1: NEXTER

2: GD-OTS

es Centre, F

R Munitions,

S, Canada.

Finnish Defe

, France.

ence Forcess, Finland

Figure 13

Figure 14

Figure 15

3: Airbus Sa

4: BAE, UK

5: BAE, UK

afran Launc

.

.

chers, Franc

ce.

Figure 16

Figure 17 Oven Ma A very lalisting of

• • •

•

• •

•

6: MOD NLD

7: Redstone

aterial and

arge variety the differen

16 and 14 inch thUsually insulationMade frowith alumRegular The oveform of sacrificiaDouble w1” Cerampresent t

D / KCW&M

e Test Cente

Thickness

of materialsnt responses

8 gage steeick insulatiowe use larn we use mom 4 ft x 8 fminum tape.steel, arounns used areinsulation.

al. wall design: mic Thermalto interfere

M, Netherlan

er ATEC, US

s and thickns: el dependingon material.rge air condineral wool ft x 3’’ Hom. The oven ind 1mm thice thin doubThese oven

1” thick, fol blanket ratwith potent

nds.

SA.

nesses are

g on the siz

ditioning pip(thickness 1

me Insulations mounted o

ck. le walled alns are cust

il-faced, fibeed to 2400Ftial fragmen

used for the

e of the ove

pe/tube or 10-20cm), an sheeting pon a steel fr

luminium detom built fo

erglass ductF. (“Soft” ovet flight) for t

e oven cons

en for structu

“chicken waluminium fopolyisocyanurame/table.

esigned withor size and

t board lineden design wthe inner bo

struction. B

ural integrity

wire” and fraoil and tape. urate R3 as

h fibreglassare all dee

d with singlewith no hard ox, which is

elow is a

y.

ame. For

ssembled

s or other emed as

e layer of surfaces

s covered

with an outer box made from the same duct board allowing 3 inches airspace between the inner and outer to allow for circulation.

• Standard Stone Wool inside steel grid. • 1/16” aluminium walls with a 1” aluminium angle skeleton to produce a box which is

then covered with high temperature insulation as needed and then covered by a clear tarp to protect the insulation from dew or rain.

• Double-wall construction with inner wall of 1-in thick fiberglass duct-board and outer wall of 2-in thick rigid polystyrene foam.

• Usually steel sheet, appx 1/16” (China Lake) or appr 1/8” (Eglin) • Bespoke oven created for each individual test scenario. • Mild steel with insulating material sandwich between inner and outer layer. • Double-wall design (i.e. inner chamber/outer chamber). Inner chamber constructed

from 1-inch thick duct board. Outer chamber constructed from 1 ½ inch thick foam insulation.

• We use reinforcement mats for the framework of the oven and 200 mm thick mineral wool for insulation. The oven is protected from wind and rain with polyethylene foil.

• Heat resistant wool. Thickness is very much dependent on the outer conditions, i.e. thicker insulation in the winter time than in the summer time.

• Oven material and thickness 1mm steel plate on light frames, rockwool insulation • Thin steel sheet metal, see NAWCWD 473000D for detail info • Ceramic, approx. 3 inches.

Heating Systems and Airflow There appears to be much less variation in the heating systems, with two main approaches: internal heat source convection oven, or external heat source convection oven both using electric heating elements. Almost all responses indicated that they used forced airflow in order to try and achieve temperature homogeneity. So the primary difference between the two heating systems approaches is the location of the heat source: internal for convection ovens and external for heat source convention ovens which employ a pipe to transfer the heat.…are that for the internal heat source convection ovens, the electric heating elements are within the oven, and for the external heat source convection oven, the electric heating elements are in a separate unit from the oven and heated air is piped into the oven. Below are some associated comments from the survey:

• Heating system is like large “fan-oven”. • Convection oven heated by four 120 Volt, 500W strip heaters protected by thin

aluminium witness plates. The heaters may me reusable from test to test. • Electrical resistances. • One or more heating elements off the ground and away from the item under test

within the oven system and fan assisted for air circulation. • ( 4) 500 Watt heating elements. • Hielkema Air Heaters. • Typically 3 each tubular heaters controlled by a Watlow control device that regulates

the time that the heaters receive voltage, thereby, providing heat to the box. Wall Distance All survey responders indicated that they maintained the distance between the test item and the oven wall to be >200mm per the STANAG requirement. Temperature homogeneity The STANAG lists a requirement for temperature homogeneity to be within 5°C. Most respondents indicated that this requirement was not difficult to meet, except for large test items.

Associat•

•

•

• •

• Protecti The survprematuthey did Associat

•

•

•

•

•

•

•

Figure 18

ted commenWhen tevalid testTemperaoccasionTypicallywith relaUsually 4Usually can be aWhy sho

on for Exud

vey asked rely ignited use some p

ted commenA “drip tthe heatiWe havebe insulaDependsto have cWe put prompt igOur ovensurface. I undersbad. As that you right undWe cons

8: Protectio

nts are: esting new it. ature homogns, this tolery we can mtively large, 4-7°C. If insless than 5

at least 30°Could there be

ded Energe

if any proon contacti

protection, a

nts were: ray” is locating elemente a tray addated from ths on the testcontact withthermally ingnition whenns are built We try also

stand why thwe don’t hawon’t have

der the weapsider the pyr

on from exud

tems, dry ru

geneity hasrance has beeet the reqelongated iulation is no°C but with

C hotter. e uniform te

etic Materia

tection wasng the oven

as seen in Fi

ted below tht. ed to the ine hotter ovet item, in so

h oven wallsnsulated recn energetic of heat resi

o to insulate he concern ave a valid rhot spots o

pon? Then trolysis gase

ded energet

un with iner

s almost alween exceeduired homoitems (e.g., ot done well

strip heate

emperature?

als

s used to n wall. . Abigure 18.

he test item

side of the oen wall. me occasio

ceptacle belmaterial fallistant wool wthe fixtures exists abou

rationale foron the floor?the floor will

es as having

tic material.

rt dummy is

ways been ed.

ogeneity. Holarge surfac, it can be 1

ers the oven

? Based on w

prevent exbout 40% of

m to avoid e

oven so fill

ons, “trays” a

low tested ml on the ovewhich is insin the oven

ut exuded mr the test, ho? What if the be the hott

g more influe

s always pe

kept below

owever, we ce-launched0°C or more

n wall, espe

what real wo

uded energf the respon

nergetic ma

that exudes

are added to

munition. Inen wall. ulating and

ns which holmaterial: it ow do you me fire is in thest part.

ence on pre

rformed to

w 5°C. On v

have had pd rocket mote.

ecially on th

orld rational

getic materndents indic

aterial dropp

s out of the c

o prevent m

n the aim to

do not act d the test itemakes the

make the asshe portion of

mature ignit

ensure a

very rare

problems tors).

e bottom

le?

ial being cated that

ping onto

case can

elted EM

o avoid a

like a hot em test look sumption f the ship

tions.

OVEN DESIGN ISSUES The survey asked that any issues with the oven designs be raised. The primary topics discussed by the respondents were that the oven should be: 1) very well Thermally insulated and 2) designed so as not to significantly confine the reaction, fragmentation or blast. Thermal issues Comments associated with thermal insulation are:

• Forced air flow ovens have resulted in varying responses from the same munition. It is believed the air inlet is creating localized hotspots.

• The convective oven wall is much hotter than the oven air so when fill contacts it ignites. This not only causes a slightly earlier reaction but ignites the remaining fill in the case in a location that may not be realistic.

• We use heat resistant wool as construction material and place most metal parts (heating elements, fans etc.) beneath the test item.

• The ultimate SCO oven material would be a heat resistant, light colored (better “in oven camera” coverage) affordable, light weight / density and environmental friendly plate.

• Air flow in our test set up does not have high velocity. On my opinion that does not have significant influence to test item and test result.

• Several different cameras that look through a window in the oven and the most difficult problem is determining how to keep the camera cool to keep the internal view camera alive. I have had successes and failures, but I haven’t found the perfect answer for that yet.

Confinement and Fragment Flight Issues Comments associated with confinement and fragment flight are:

• Heavy wall construction can influence the flight of fragments. • I feel that the greatest issue is ensuring that the design of the oven truly provides the

minimum confinement that can be achieved practically, in order to minimize suppressive effects on the ejection of debris and attenuation/focusing of blast

• Oven walls should be constructed of foam/fiber panels with minimal structural integrity to lessen their effect of slowing fragment projections

• Even if the confinement exacerbates oven throw distance we do not believe it throws test item parts farther.

• Confinement will always be an issue, but items can be compared using the same test setup and a “calibration” shot can be used to measure/determine the full detonation properties of the test item.

• Oven design should have minimal effect on the projection of fragments from the oven.

OVEN STANDARDIZATION The survey asked whether the oven design should be standardized. The results split fairly evenly as seen in Figure 19. The problem with standardizing the design is that different munition types require different considerations. Differing sizes of munitions required different sized ovens. Rocket motors will need to be restrained to prevent flight in case of a strong propulsive reaction. Also, munitions tested inside a shipping container or canister can affect the oven design. One thing to consider is the use of the oven itself as a surrogate shipping container or canister. This could avoid duplicate confinement. Another point to discuss will be the consideration of the effect of heating bands on the reaction if material is extruded out of the test item. Heating bands create localised hot surfaces which are eliminated if forced air is used. So, guidance could be provided to use forced air if energetic extrusion is anticipated. External conditions affect the design as well: if it’s very cold outside additional insulation will be needed. Recommendations could be

developethe poten TEMPER The survthat prioprecondiwhichevemelt casbe chang Precond As seen the respmaterialsAbout 1/responde

Figure 19 Should a Commen

•

•

•

•

•

Should t

ed for oven ntial of seco

RATURE PR

vey participor to comitioned at 5er occurs fi

st energetic ged.

ditioning Re

in Figure 19ondents thos, whereas /3 of respoents though

9: Tempera

a Melt Cast

nts associatIt is usefas we adown beI don’t seto 120⁰F If the soaround 8case nowIf the tesabove thIf it is inneeded.

the Require

design to londary fragm

RECONDIT

ants were amencement0°C for 8 hrst. The sushould be p

esponses

9, almost allought that mabout ¼ ofndents thou

ht that it shou

ature precon

t Energetic

ted with pre-ful to have re performi

efore the tesee any reas(~50⁰ C) be

oaking temp80°C for the w, the soakist is to be uhe melting pontended to

ement to P

limit maskinmentation.

TIONING

asked about of the hhours or unurvey askedpre-soaked

l respondedmelt cast enf respondenught that thuld remain t

nditioning.

be Pre-soa

-soaking methe same sng the testst starts.

son to. Durinefore firing.perature is TNT whichng should bused as chaoint until all be used a

recondition

ng of primar

ut temperatuheating contil the test

d if they did differently a

d that they pnergetic sho

nts thought the requiremthe same.

aked Differe

elt cast enerstarting temps in autumn

ng firing test

much lowe is the lowe

be the samearacterizatiomaterial hareal-life su

n be Chang

ry blast and

ure precondndition (ratitem reachepreconditio

and should t

recondition ould be prethat they sh

ment should

ently?

rgetic are: perature forn and winte

ts we would

er than thest of the me

e. on test, the as melted. urrogate, the

ged?

fragment e

ditioning. Tte), the teses thermal on per the Sthe precond

per the STAe-soaked idehould be pre

be change

r all f munitier times and

also condit

e typical meelting explos

material sh

en all pre-s

effects whils

The STANAst item shequilibrium STANAG, wditioning req

ANAG. Aboentically to e-soaked d

ed and abo

on types. Ed the muniti

tion these e

elt temperasive used) a

hould be soa

soaks are n

st limiting

AG states hould be

at 50°C, whether a quirement

out half of non-melt ifferently. ut 1/3 of

Especially ion cools

nergetics

ature (i.e. as it is the

aked just

no longer

Commen

•

•

• •

• •

•

MELTIN The survtesting renergetic

Figure 20 PROPUL As seenprevent believe t

nts associatSome guitems mathink thesize (diaThe onlyheating iConsideWe do nensure tambient howeverenergeticI’m not aI would deg. Whmunitionmap to thI think thbetter fit

G

vey participarequirementcs during a

0: Melting e

LSION

n Figure 21propulsion. that they po

ted with poteuidance shoay reach the STANAG meter, weigy reason I cs a driver ofr making it enot see thethat initial tair and ov

r be much c materials o

aware of the consider eshat is gaines life cycle his requiremhat one shoone of the m

ants were ast. As seen test should

energetics.

, almost haMore than

ose some s

entially chanould be proermal equilshould prov

ght, amount can see forf the reactioeven higher e need to ctemperatureven walls telower than or other ma history of wstablishing

ed by doing or threat p

ment? ould take a most commo

sked whethin Figure 2not affect th

alf of the rehalf said th

safety risk.

nging precoovided on thibrium in 8 vide a guidof container

r it is if it afon severity.r if that can bchange thises are idenemperature.the temperterials.

why 50°C waand documthis? Why

profile? How

look at the on accident

er melting o20, about 7he requirem

espondents hat they do Many of the

nditioning ahe minimumhours but l

de for pre-cor/item insulaffects the o

be shown to. Precondit

ntical from . The precorature at w

as chosen amenting the

50 degreesw does a po

entire heatt scenarios.

of energetics75% of resent.

stated thatrestrain pote described

re: m time to plarger itemsonditioning ation). nset of self

o not changeioning shoutest to testonditioning hich degrad

as the soak need for p? What doeotential ther

ting rate pro

s during a tpondents th

t they do ntentially prod restraint m

recondition.s definitely wtime based

f-heating an

e results. uld be perfot, regardlestemperatur

dation begin

temperaturereconditionies it represermal threat

ofile so that

est should ahink that m

not restrain pulsive item

methods allo

. Smaller will not. I

d on item

nd if self-

ormed to ss of the e should ns in the

e. ng at 50 ent in the scenario

t it would

affect the melting of

items to ms if they ow some

item movitem. Some of

• •

• •

•

•

•

Figure 2 ITEM OR When asusually rthis is thby end cthe ejectstraight uproblems INSTRU The survresponsenormallypressuremeasurefacility lis

vement, pro

f the receiveTest itemThe motflying awthe test sMostly bSystem thrust ceWe onlyrange-sarocket mIn spite othat takeWe norm

1: Item restr

RIENTATIO

sked about representative horizonta

caps, plugs, tion of theseup, which cas evaluating

MENTATIO

vey requestes were ver

y used: digite gauges anement and psted bikini pr

oviding the t

ed commentm is normallyor is placed

way during resite. y using vertdependent,

ells, are usedy restrain iteafety hazard

motors. of a nearestes ordnancemally always

raint from pr

ON

item orientave of the orl orientationetc. that m

e pieces of dan result in g the results

ON

ted informary similar frotal data acqnd thermocopiezoelectricressure gau

testing facili

ts are: y strapped t

d in relativelyeaction due

tical orientat range safed as neededems in instad. Typically

t neighbor a off the stati

s allow prop

ropulsion.

ation duringrientation du

n, which canight be ejecdebris by dia misleadin if video cov

ation on insom all of thequisition recouples. Typc pressure guges for blas

ty has some

to the test sty open but sto the onse

tion when teety depended ances wherthis is uniq

t 20 miles, wion boundarulsion in IM

g testing, theuring most n facilitate thcted from threcting dow

ngly short apverage is po

strumentatioe respondenorders, videically type Kgauges for bst measurem

e informatio

tand with stestrong steel et of thrust –

esting rocketent. Tetheri

re propulsioquely assoc

we must be ries. tests. We h

e majority sor all of thehe assessme test item.

wnward into pparent ejecoor; or the re

on used for nts in terms eo coverageK thermocoblast overprment.

on on propu

eel bandingcage to pre

– this is a ge

t motor or ping, catch b

on of the caciated with

prepared fo

handle that w

stated that the item logist

ment of the dVertical oriethe oven floction range.eaction occu

the slow hof the instr

e, microphouples are uressure mea

lsive potent

. event the meneral requir

propulsive iteboxes, barr

ase might pitems that h

or a propuls

with our risk

he test orientics life-cycldebris hazaentation canoor or direct The latter c

urs at night-t

heating testumentation

ones, witnesused for temasurement.

tial of the

otor from rement of

ems. ier walls,

present a have live

ive event

k areas.

ntation is le. Often, rd posed n impede ting them can pose time.

ting. The that was

ss plates, mperature

One test

Number of Thermocouples The STANAG is somewhat inconsistent in that it states “A minimum of four thermocouples should be used to be sure that the oven is uniformly heated and to monitor the surface temperature of the test item.”, but goes on to state “In general, there should be at least two thermocouples mounted on opposite surfaces of the test item, one each in the air space near the air inlet and exit, and one each in the air space on opposite sides of the round (see Figure 1).”, implying at least 6 thermocouples should be used. The number of thermocouples used by test facilities appears to vary greatly from 4 to 100. Below are some of the responses:

• 4 as per the STANAG. • 6 (minimum) installed in accordance with the STANAG to assess compliance with the

heating conditions and provide an indication of reaction of the test item. • Between 8 – 16: near the oven wall, near the item, and when possible inside the item

(charging tubes). • Typically fifteen. This includes the typical air temp at various points near the item,

oven wall temp, skin temp in several locations including just outside the oven wall, and one or two internal to the item to detect self heating.

• Ten to thirty thermocouples are typical. Some dictating factors include, size of the oven (ensure temperature homogeneity), specific test information about a location, efficiency of heat transfer, STANAG requirement, engineering considerations, etc. We are equipped to use as many as 100 thermocouples.

VISUALIZATION ISSUES A number of respondents described visualization issues:

• Occasionally an internal camera will fail or be obscured by fill exudate prior to initiation. This will compromise diagnostics.

• A minimum of four cameras are used. Two are fielded to view the test store inside the oven and two are deployed to view the outside of the oven.

• We use two cameras outside and one camera inside the oven to record the reaction of the munition.

• Cameras are used external to the event. Disposable internal cameras are often used, as many as four in a single test.

• We have a camera outside oven, shooting through window to inside oven, to the test item. It has been very useful and has given information during test and just before reaction.

• The window allows the ability to confirm reaction of the item but our current set-up does not allow for visualization of the test item reaction, since we prefer using a general surveillance camera. A system using bigger window, mirrors, and high speed camera is possible. Trigger is an issue with some instruments, although we used bridge-wire to acquire data (ex.: pressure) at a higher speed rate during reaction.

SUMMARY OF RECOMMENDATIONS This is a summary of the recommendations, the explanations have been provided in the core of this document:

• Develop a group consensus as to the intent of the test and document it. • Query all of the MSIAC nations to provide information on actual event durations and

rates. • Based on consensus test intent and supporting data, develop a consensus as to

changing rate or leaving the rate unchanged. • Clarify the minimum number of required thermocouples and thermocouple

positioning.

• •

• • •

ACTUAL There wactual sloHeating review ofrom thethreats. Areport sereport “Amany exresult, firminimumhave occfires thatday firestemperatsummaryassessmas 25°C/thermal which insimilar.

Figure 22

CONCLU This sturecomme

ObservaDevelop scales ofCharacteProvide aRemovebetweenprocess)

L EVENTS

were 32 respow heating Custodial W

of actual evee MSIAC saA search of earch resulteAssessing Txamples of re modelling

m heating racurred over t have occu

s appear to ture rise raty of this a

ment conclud/hr. Lower rmodelling. cluded sup

2: Thermal t

USIONS

dy and theendations to

ations of eveand provid

f test items. erize the heaa best pract redundanc

n the STANA).

HEATING R

ponses thatincidents. AWorking Grent heating afety databa

the MSIACed in a largehermal Threfire duratio

g results of tes. The vamultiple day

urred sequebe completetes remainsassessmentde that acturates appeaDr. David Hporting fire

threat categ

e associatedo further im

ents inside thde a best p

ating equipmtices exampcies or contrAG 4382 an

RATES

t the individAs a result, roup (SH Crates and d

ase regardinC MAD-X acce number ofeats” [33] thns and veractual even

ast majority ys, but it ne

entially. Thee in much ss the same t is presenal credible a

ar very difficuHubble frommodelling [

gorization.

d historical mprove the

he oven. ractice oven

ment and peple test confradictions and the AOP

duals had nduring the N

CWG) meeturations anng real-life scident databf referenceshat was pubry few actuants are requof fire event

eeds to be ree individual

shorter timesas that ass

nted in figuadjacent comult to justify

m NSWCDD[34]. The r

events assSTANAG 4

n design ex

erform calibiguration.

and clarify wP-39 (AC/32

no informatiNATO AC/3ting, MSIACd share anyslow heatinbase provids [5-32], inclulished in 20al temperatuired in ordets are compecognized thfire events

s. The sumsessed in thure 22. Tmpartment fbased on a

D, USA receresults and

sessment w4382. Thes

xamples for

ration testin

where the in6 SG/B has

ion on the 326 SG10-1C was requy available hng events aed no applicuding the pr

003. The revure rate meer to estimaplete within ahat these ev

s associatedmmary inferrehe 2003 MS

This review,fires could p

actual fire evently compleconclusions

was an efficse recomme

r different ty

ng.

nformation ss already be

duration or1 April 2017

uested to cohistorical infnd potentiacable informrevious 200view results easurement

ate credible a day. Somvents are ald with theseed actual firSIAC report, and the produce ratevents and aseted a simils appear to

cient way toendations a

ypes and

should sit egun this

r rates of 7 /B Slow onduct a formation l thermal

mation. A 3 MSIAC provided

ts. As a actual or

me events l multiple

e multiple re events t [33]. A previous

es as low ssociated lar study, o be very

o identify are being

discussed with AC/326 SG/B who has already chartered a working group to review and update the STANAG. The working group has already reviewed the survey results. According to the new requirements in the NATO documentation, the technical content of the STANAG will be migrated into an AOP. ACKNOWLEDGEMENT MSIAC would like to acknowledge the custodian of the STANAG, USA, and particularly Dr. Stephen Struck for his review and amendment of the survey. MSIAC also acknowledge the individuals from all the nations who have contributed to the survey by providing an answer to the questionnaire or information related to this test. REFERENCES 1. STANAG 4439, edition 3: Policy for introduction and assessment of Insensitive Munitions. 2. STANAG 4382, edition 2: Slow heating, munitions test procedures. 3. AOP 39, edition 3: Guidance on the assessment and development of Insensitive Munitions. 4. UN, Recommendations on the transport of dangerous goods, Manual of tests and criteria,

sixth revised edition, 2015. 5. Back, G., R. Darwin, J. Scheffey, “Propellant fires in a simulated shipboard compartment:

Project HULVUL Phase III, Naval Research Laboratory NRL/MIU6180 --99-8394, 1999. 6. Boggs, T., K.P. Ford, J. Covino; “Realistic Safe-Separation Distance Determination for

Mass Fire Hazards”, NAWCWD TM 8668, March 2013. 7. Booz-Allen & Hamilton, Inc.; Threat Hazard Analysis prepared for Program Executive

Office for Theater Air Defense, 26 November, 1996. 8. Budnick, E., D. Evans, H. Nelson, “Simplified fire growth calculations”, SFPE Handbook of

Fire Protection Engineering, Section 11, Chapter 10, 1997. 9. Bukowski, R., “Fire Hazard Assessment”, SFPE Handbook of Fire Protection Engineering,

Section 11, Chapter 7, 1997. 10. Doolan, C.;”Ordnance Heating Rates in Shipboard Magazines”, DSTO-TR-1188, July

2001. 11. Fontenot, J. and M. Jacobson; "Analysis of Heating Rates for the Insensitive Munitions

Slow Cookoff Test", NWC TM 6278 (1988). 12. Fournier, A., C. Lallemand, “Phénomènes thermiques induits par un accident de munition

à bord d’un navire”, AGARD-CP-511, paper 22, 1992. 13. Frey, R.; “Appropriate Slow Cookoff Heating Rates” Army Research Laboratory, 18 April

2000. 14. Guinn, W.R., “A study of heat transfer in liquid pool fires from steady burning through

boilover”, Paper for Fire Protection Engineering -520. 15. Kennett, S., G. Gamble, J. De Li, “Modelling of the HMAS Westralia fire”, DSTO -TR-0698,

1998. 16. Heimdahl, O., L. Bowman, “Standoff distance effect on cookoff of ordnance stowed in

MDCS ship cargo holds”, Twenty-ninth DoD Explosives Safety Seminar, 2000. 17. Leblanc, D., “Fire environments typical of Navy ships”, Thesis, submitted to the Faculty of

the Worcester Polytechnic Institute in partial fulfillment of the requirements for the Degree of Master of Science in Fire Protection Engineering,1998.

18. King, B., “Unplanned Explosions at Munitions Sites”, Small Arms Survey Organization (http://www.smallarmssurvey.org), 2016.

19. Lundstrom, E., “Slow cookoff analysis”, Proceedings of the NIMIC 1993 Workshop on Cookoff, paper TP-7, 1993.

20. Mansfield, J., "Preliminary Analysis of the Heating of Ordnance in Ship Magazines Due to a Fire in an Adjacent Compartment", NAWCWPNS TP 8186 (1996).

21. Madrzykowski, D., R. Vettori, “Simulation of the dynamics of the fire at 3146 Cherry Road NE, Washington D.C., May 30, 1999”, Buildings and Fire Research Laboratory, National Institute of Standards and Technology, NISTIR 6510 (2000).

22. Null, G., “Computer simulation of the thermal effects on a concentric canister missile launcher with a fire in an adjacent compartment”, Naval Postgraduate School Thesis, ADA337018, 1997.

23. Pitts, W., “Improved real-scale fire measurements having meaningful uncertainty limits”, W. Pitts, Workshop on Fire: Testing measurement needs: Proceedings, W. Grosshandler, Building and Fire Research Laboratory, 2001.

24. Stokes B.B., Fitzgerald-Smith J. DeFourneaux M., Kernen P., “Summary of NIMIC 1993 Workshop on Cookoff”, NIMIC-BS-307-94, 1994.

25. Taylor, T.N., “Ammunition Accident at the Evangelos Florakis Naval Base, Zygi, Cyprus 11 July 2011”, MSIAC Report O-150 Rev. 1, September 2014.

26. Victor, A., “Threat Hazard Assessment for Insensitive Munitions”, Presented at NIMIC Hazards Analysis Workshop, 1996.

27. Victor, A., “Exploring cookoff mysteries”, JANNAF Propulsion Systems Hazards Subcommittee meeting, 1994.

28. Wallace, I., “Cookoff – a UK naval perspective”, Proceedings of the NIMIC 1993 Workshop on Cookoff, paper TP-5, 1993.

29. Wharton, R., J. Harding, A. Barratt, R. Merrifield, “Measurement of the size, duration and thermal output of fireballs produced by a range of pyrotechnics”, 21rst international pyrotechnics seminar, pp.916-931, 1995.

30. Wilkinson, A., “Ammunition Depot Explosions”, SAS Conventional Ammunition in Surplus Book 15 Chapter 13.

31. “Recent Explosive Events in Ammunition Storage Areas,” South Eastern and Eastern Europe Clearinghouse for the Control of Small Arms and Light Weapons (http://www.seesac.org), 2007.

32. “Dangerous Depots: The Growing Humanitarian Problem Posed by Aging and Poorly Maintained Munitions Storage Sites Factsheets”, Fact Sheet produced by the U.S. Department of State’s Bureau of Political-Military Affairs, 2010.

33. Peugeot, F., “Assessing Thermal Threats” MSIAC Technical Report L-097 published in 2003.

34. Hubble, D., “An Investigation into a Proper Heating Rate for Slow Cook-off Testing”, Insensitive Munitions & Energetic Materials Technology Symposium, Portland, OR, 2018.