THE AN NALS OF UNI VERSITY DUNĂREA DE JOS OF GALAȚI 22 [600830]

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 22
FASCICLE VIII, 2007 (XIII), ISSN 1221-459 0
TRIBO LOGY

Paper present at

Interna tiona l Conference o n
Diag nosis and Pred iction in Mecha nical
Engineerin g Systems (DIPR E’07)
26 – 27 October 2007, Galati, Ro mania

FLAMM ABILITY TES TS ON HOT SURF ACES
FOR INDUSTRIAL FLUIDS

Lorena DELEANU1, Drago ș BUZOIANU2, Minodora RÎP Ă1,
Ștefan CR ĂCIUNOIU2, Adina DRUG1

1 University Du narea de Jos of Galati, 2 ICTCM Bucha rest, Rom ania
[anonimizat]

ABSTR ACT
Industrial fluids have to be eval uated not only for their perf ormances as l oad
capacity, du rability, b ut also fo r their po tential risk in cluding ignition. Based on a
solid documenta tion the au thors po int out th e sig nificance o f testin g the flu id
flammability and prese nt some tests related to fluid ignition on hot s urfaces. T he
paper also presents prelimi nary resu lts on determi ning the fla mmab ility characte-
ristics of fluid when dropping it on a hot surface usi ng an original tester cap able of
fulfillin g the req uiremen ts impo sed by SR EN ISO 208 23:200 4 Petro leum and
related products. Determina tion of the fla mmab ility ch aracteristics o f fluids in
contact with hot surfaces.

KEYW ORDS : Flamm ability test, ho t surface, tester, hydrau lic flu id.

1. INTRODUCTION

Selecting hy draulic fluids including as
basic cri-teria of re ducing fire risk bec omes
of m ajor i nterest in part icular and gene ral
industrial applications [1 -10, 18, 28, 31].
Every i ndustrial activ ity i mplies th e
existen ce of potential ig nition sources t hat
could qui te numerous and hard to identify a
priori [21]. The specialists consi der that the
identified sources co uld and ha ve t o be
reduce d, eve n elimin ate, but the progress
adva nce a nd new technologies do not reve al
all o f them immed iately. Factory Mut ual
Engineeri ng C orporation st udied and
intabulated the ignition s ources fr om over
25000 fires. T he rel evant sources related to
explosions are prese nted in table 1 [1].
Table 1 . Ignition sources i n major fires [1 , 21].
Ignition source Occurring
percen tage
[%]
Electrical (engine c onnections ) 23
Friction (bearings, broken parts etc.) 10
Overheated parts (abnorm al temperatures) 8
Hot s urfaces (boilers, lighte ning sources etc.) 7
Flames of t he burners (incorrect de-si gn and
exploitation of burners etc.) 7
Sparks and hot subproducts from
technological proces s (metal proces sing etc.) 5
Metal cutting a nd wel ding (sparks , electrical
arches , heat etc.) 4
Mech anically g enerated sp arks (grind ing,
polishing etc.) 2
Static sparks (accum ulated electric ene rgy) 1

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 23
FASCICLE VIII, 2007 (XIII), ISSN 1221-459 0
TRIBO LOGY

2. FLAMM ABILITY
CHARACT ERISTI CS

The flammability characteristics of the subs tan-
ces are very im portant as concerning the s afety in de-
positing, tran sporting , manufacturing and using them
[1-10]. T hese characteristic s include flammability
point, flash point, autoignition temperatu re, upper and
lower li mit of flammab ility but also the substan ce res-
ponse in imposed conditions related to worst p ossible
scenari os in practice [15, 27]. All are im portant and
have to be taken into account for evaluating the risk
probability related to fire and exposure to fire or hot
environments [18, 1 9, 23, 24] . The exp erimental
values a re always necessary but the m ethods a nd the
procedures for obtaining them are sel dom similar to
real cond itions and are expensive and dangerous.
There have been developed st andards and
specifications for classifying fl uids related t o their fire
hazard [25]. T he characterist ic involve d in som e of
these classific ations i nclude the fi re point te mpera-
ture, Tf, and t he boi ling temperature, Tb (see table 2).
In these s pecifications, Tb is differently defined. F or
instance NFPA specifications define the boiling point
Tb as th e temperatu re at wh ich 20% of th e initial flu id is evaporated. DOT s pecifications de fine Tb as being
the initial bo iling point (Tib) of the fluid. Tb and Tib
were also defined in ASTM D 2887-06 Stan dard Test
Meth od for B oiling Rang e Distribu tion of Petro leum
Fract ions by Gas Chromatography .

3. TEST S FOR DET ERMINING THE
FLUID F LAMMABIL ITY ON HOT
SURF ACE S

3.1 Overview of Anterior Standard
and Non-Standard Tests

The developmen t of fire-resistan t hydrau lic
fluids required the identification of test methods that
could reproducibly differentiate the fire-resistance of
the flu ids in a m anner that wou ld relate to real
industrial con ditions [1, 12, 14, 17, 28]. Zin c elab o-
rated a classification of types of fire tests [32]
according t o four elem ents (fig. 1):
– the fire resistant p roperty that is bein g measure d as
there a re several diffe rent facets of fire resistance an d
a test can m easure one or more of them ,

Table 2 . NFPA a nd DOT specifications for flui d classifica tion taki ng int o account the fire haza rd [25].
NFP A DOT
Fluid
Classification Hazard
class Criterio n
°C Class Criterion,
°C
IA 4 bT 37.8< ; fT 22.< 8 I bT3 5≤
IB 3 bT 37.8≥ ; fT 22.< 8 3 II bT 35> ; fT2<
IC 3 f 22.8T37.8≤〈 III bT 35> ; f 60.5T 23≥≥
II 2 f 37.8T 60≤<
IIIA 2 f 60 T 93.4≤<
IIIB 1 fT 93.≥ 4
0 0 bT 815.5> for 5 minutes
NFPA – Nat ional Fire Protection Ass ociation (SU A).
DOT – Department of Transportation (SUA)

Fig. 1. Groups of facto rs influencing the fire tests [3 2].

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 24
FASCICLE VIII, 2007 (XIII), ISSN 1221-459 0
TRIBO LOGY

– the ign ition source used in the test (alm ost all fire
resistan ce test s ignite or attempt to ignite th e fluid
during the test),
– the state of the flui d during t he test,
– whet her the test si mulates an accident condition or
measures a n intrinsic property of the fluid.
Some tests are based on a sim ulated accident
(either ex plicitly o r implicitly) while others measu ring
an intrinsic property of the fluid are not based on the
conditions of any particular accid ent mode.
Both manufact urers and users ask f or tests that
certify flu id flammab ility characteristics, p referring
especially ISO or ASTM st andards. Many documents,
including EU Directiv es, give reco mmendations to
use st andard tests for estimating fl ammability
charact erisation of fluids [1 -10].
The au thors po int out that the qualitative and
quantitative quantification of fire resist ance o f a
hydraul ic fluid could not be d one by one propert y only
and t he different aspect s of f ire resi stance h ave to be
point out by different tests, including tests simulating
on sm all scale the worst scenari o that coul d happen i n
real appli cations, when usi ng hydraul ic fluids [2-5, 1 6-
19]. A part iculari ty of many of these tests is that the
test result is delivered as “p ass” or “not pass”. The fl uid
that pa ssed the tests are inclu ded in reco mmendations
or approval s, but these ones are speci fic to regi onal
reglementati ons (i n SUA: Approval Gui de or Li st of
Qual ified Fl uids, in Euro pean Union [3, 4, 6, 7, 9] or in
national normatives).
Organizations such as Occ upational Safet y and
Health Asso ciatio n (OSHA) and Natio nal Fire
Protection Association (NFPA) and some speci alists
classify flammable liquids according to their
flashpoint [13, 23, 27]. A flashpoint within the
operating temperature ra nge of the system obviously
is undesira ble as any leak would create a n immediate
fire hazard [18, 25].
The laborat ories dealing with flamm ability and
fire-resistance proper ties have t he adv antage t hat
more test s imp lies m ore work, v isibility and money
but it also requires disadvantages as: larg er invest –
ment in equipment, a greater understanding of simi-
larities and differences among related tests (inclu-
ding well-trained special ists), agreem ents for cross –
border accep tance of the results [15, 19, 20 , 26].
CETOP RP 65H 199 4 [2] „ig nition test on hot
plate for fi re-resistant fl uids” wa s consi dered
appropriate to estim ate the fire risk of this hazard
“scenari o”. During this test the fluid is dropped on a
heated tube at 704 C and the resul t has to be classified
as on e of the three fo llowing situations:
– flames or burns on the tube, but not aft er
dripping from the tube;
– the fluid does not ignite and does not burn on
the tube but it does after dripping from the tube;
– the fluid does not ignite and does not burn
neither on the tube nor after dripp ing from the tube. In 1999 the test was reco mmended by HSE
Approve d specifications for fire resist ance an d
hygiene of the hydraulic fluids for use i n machinery
and e quipment in mine [2] . But all CETOP
recom mendat ions (R ecommendations, Provisional
Recom mendations a nd Tec hnical Reports) were
officially withdrawn from 21.06.2002. [2]. One m ay
notice th e similarities b etween the CETOP test an d
the actual st anda rd ISO concerning the sam e
characte ristics of a fl uid.
Studies about fluid ignitio n on hot surfaces were
carried o ut without being the subject of a standard.
Toy [26] admitted that one m ajor s ource of
machinery fire s and acci dents may be attributa ble to
the leaka ge of flammable fluids onto hot s urfaces a nd
its su bsequent ignition.
The leak of the flammable fluids on hot surfaces
has a great risk to produce fire, affecting the work
safety, so specialists stri ctly im posed te mperature
limits for using particular fluids but accide ntally fires
still have be en occ urring even if the surface
temperatu re was lower th an these lim its. In such cases
the ignition is governe d by the conve ctive heat
trans fer towards the wetting zones a nd the process is
very complex, havi ng m any influenci ng factors a s, for
instance, the fluid rate, the d ifferen t ev aporatio n
processes, th e quality of th e gas env ironment etc.
The m ain element of st and proposed by Toy
[26] is that of a si ngle horizontal circ ular flat plate
that is const ructed from a 20 cm diameter (316 ) stain-
less steel disk (fig 2). The plate is capabl e of being
heated to surfa ce te mperatur es of 200°C…70 0°C b y a
CALROD hea ting coil embedde d wi thin the plate
assem bly. A series of 40 gauge K-type Chrome-
Alumel thermocouples are e mbedded i nto the side of
the plate to monitor th e temp erature thro ughout the
plate, wh ere it was fou nd that th e variability o f the
temperature at any radial position was less than 1%,
and that the temperature di fference between the cente r
and t he perimeter was a pproximately 5%. A si mple
gutter arrang emen t was also con structed aroun d the
circu lar plate to contain an y excess fuel spillage.

Fig. 2. A test ri g for estim ating flammab ilitz on hot
surface [12].

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 25
FASCICLE VIII, 2007 (XIII), ISSN 1221-459 0
TRIBO LOGY

Fig. 3. A sche matic diagram of the tester used in [16].

Ham ins, B orthwick and Preser [16] prese nted a
testing m ethodology on hot surface but usi ng a fl uid
jet oriented towards a m etal disk having a controlled
temperature (fi g. 3). E xperiments were first conducted
to determine the am ount of agent needed to suppress
the ignition of a JP8 spray fla me. There we re
difficu lties, howev er, in obtaining rep eatab le resu lts.
Thus, a gase ous propa ne flow re placed t he liquid
spray. U se of a gaseous fuel represents a m ost
dangerous case, when a liquid fuel has completely
vaporized. T he a uthors conside red that the
temperature was en ough uniform (±30°C) on the disk
but a fter repeated test they noticed that the dis k
temperature becom e less uniform and the plate is
damaged and has to be repl aced too often.
The ignition of hydrau lic fluids with open flame
on hot surfaces was st udied by Yuan [29], at the
Natio nal Institute for Occup ational Safety an d Health ,
SUA. A n experimental system was devel oped t o test
the ignitability o f hydrau lic flu ids using d ifferen t
spray no zzles an d heat so urces (fig . 4). A 5-liter tan k
was used t o hold the hydraulic fluid, and the fluid was
heated using an electric h eating an d control system.
The oil tank was pressu rized by nitrog en from a
cylinder. Different nozzles were c onnec ted to the
bottom of the tan k by metal tubing to generate th e
desired oil sprays or jets. An open flame or a hot
surface was used as the ex ternal heat s ources. T he
open flame was a 14-cm long, line methane diffusion
flame and was place d perpe ndicular to the centerline
of the oil spra y. The heat re lease rate for the flame
was a bout 6 kW. The ho t surface was built with
electric strip heaters connect ed to a heating control
unit. T he hot s urface was m ade of stainless steel with
a size of 50 cm × 30 cm × 0.1 cm . The hot s urface
temperature was m easured with a 0.5 m m diameter
wire K t ype, inconel shielded, grounded thermocoupl e
fastene d to the center of the surface. W ith the center
temperature at 700șC, t he edge tem peratures we re
about 50șC lower th an the cen ter tem peratu re. In this
study, the center tem perature was used a s the hot
surface tem perature. T he oil temperatures a t the top
and bottom of the oil tank were als o measured usi ng
K type the rmocouples. T he oil te mperature at the top was slig htly hig her than that at th e bottom, an d the
bottom oil tem peratu re was u sed as th e bu lk oil
temperature. No radiation co rrection was m ade for the
temperature measurem ents. Nozzles us ed for the
spray tests were impingem ent-type pressure nozzles
with di fferent orifice diameters. Thes e nozzles
gene rate a fine cone-shaped spray with a spray angle
of 90ș. Ei ght non-fire-resistant hydraulic fluids and
four fire-resist ant hydrau lic flu ids were tested in the
study. Diesel fuel wa s also tested for c omparison.

Fig. 4. Test er for fluid spray on hot surface [29].

Wright, Mowery and LePera [1, 28] prese nts an
approach of the problem s related to hydraulic fluids in
critical situ ations and po inted out that surv ivability
depe nds on fire resistance pr operties of t hose fluids.
Among other tests involvi ng ignition of the fl uid
spray t hey discuss the test related to ignition on hot
surfaces (Federal Test m ethod 791, m ethod 6053. 1).
As far as they investigated, the aut hors of this
paper con sider that this test has th e greatest similarit y
to ISO 20823:2003 as c oncerning the method, the
procedure, the tester (see Table 2 a nd fig. 5).

Fig. 5. Simulated m anifold (Federal Test method 791,
method 6053.1) Dimensions in centimetres.

3.2 The Standard SR EN ISO 20823:2004

Specialists g ive inform ation on tester and
meth odologies th at are related to flammab ility
characte ristics of fl uids on hot surface s but the
methodologies and t he procedures ha ve a h igh de gree
of differen tiatio n [11, 12, 14, 15, 22, 28, 30 ].
The st andard “SR EN ISO 20823:2004 Țiței și
produse înru dite. Determ inarea caracteri sticilor de
inflamabilitate a flu idelor în contact cu suprafe țe

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 26
FASCICLE VIII, 2007 (XIII), ISSN 1221-4590
TRIBOLOGY

calde. Încercarea de inflamabilitate pe metal cald” is
the national adoption by the endorsement method of
the standard EN ISO 20823:2003 Petroleum and
related products – Determination of the flammability
characteristics of fluids in contact with hot surfaces –
Manifold ignition test (ISO 20823:2003). This standard gives a testing method for determining the
relative flammability of the fluids when the fluid
contacts a hot metallic surface having at fixed
temperature. The method also allows establishing the
ignition temperature of the studied fluid by increasing
the manifold temperature.
The standard ISO 20823 was the subject of a
debate in the ISO/TC 28 [ 33]. The method is simply
but it has not to be vague, so many comments were
done on expressing the principle, the method and the
procedure. Even if the test is, in principle, simple and
the results are quantified in only three possible results,
the procedure was well established in order to obtain
repeatability, a very desired characteristic of the test
results but hard to obtain for test involving fire or
flammability characteristics. The discussions at
ISO/TC 28 from 2000 also set the test parameters
accuracy, in order to give the possibility of ranking
fluids based on a well established procedure.
For instance, there was introduced three tempe-
rature sensors, with imposed position (two on its
opposite side – near the end of the rod and one at the
center of the rod), and not one as in the first version of
the standard, in order to have a better control of the
temperature uniformity that has to be at all these three
points within 700°C ±5°C (or other temperature but
with same accuracy).
Many specifications are related to the fluid
sample: the temperature, the volume and the state
(including the fact that “any air bubbles which may
have arise on mixing shall be allowed to escape from
the fluid before testing”). It is interesting to notice that the majority of comments were done by countries that are big manufacturers and users of hydraulic
fluids (United Kingdom, Canada).
This international test method and tester has the
highest degree of resembla nce to the American Hot
Manifold Ignition Test (Federal Test Method 791,
Method 6053.1) [2, 28], the results of the test being
also quantifies in a similar way (table 3). Even the
tester dimensions are very close to those of the tester
describes in ISO standard.
The FM Standard 6930 [2] was intended to test
only hydraulic fluids. A sample of the hydraulic fluid being tested was pressurized to 6.89 MPa and heated
to a fixed temperature starting from 60°C. There were
performed four tests: test for determining the chemical heat release rate of the fluid, flame
propagation test, test for me asuring critical heat flux
for ignition, hot surface ignition test. The latest was
designed to mimic hydraulic fluid leaking under high pressure from a hose in an environment where hot
surfaces are common. Hydraulic fluid was sprayed
from a nozzle onto a steel surface heated to 700°C.
As far as the authors could have investigated,
the test included in SR EN ISO 20823:2004 has not been not performed in Romania till nowadays. This
affirmation is based on the negative answers received
from 10 laboratories having RENAR (Romanian
Accreditation Association) accreditation for analy-
sing, testing and research on fuels, oils, lubricants. These laboratories confirm that they can not perform
any one of the three tests under the requirements of
the European standards, as imposed by the Directive
92/104. These laboratories are: ICERP SA Ploie ști,
ICMET Craiova, INCERP – CERCETARE SA Ploiești, LAREX CNIEP – Centrul Na țional pentru
Încercarea si Expertizarea Produselor, ROMPETROL
QUALITY CONTROL SRL, PETROTEL LUKOIL
SA Ploie ști, RULMENTUL SA, PETROM SA
(ARPECHIM).

Table 3. Comparing two testing methods for tes ting fluid flammability on hot surfaces.
EN ISO 20823:2003 Petroleum and related products –
Determination of the flamma bility characteristics of
fluids in contact with hot surfaces – Manifold ignition test
Manifold Ignition Test (Federal Test
Method 791, Method 6053.1) [28]
Test parameters
simulated manifold is heated at 700°C, temperature
measured in three points, fluid volume rate: 10 ml in 40…60
sec; fluid temperature 20…25°C, dispenser tip at 300 mm
above the probably impact point on the manifold,
sheet metal box 300 x 300 x 450 (mm) simulated manifold is heated at 704°C
(1300°F), temperature measured at one point
in the central zone of the manifold, fluid
volume rate: 10 ml in 40…60 sec.
sheet metal box: 300 x 300 x 460 (mm)
Test evaluation
I(T), when the fluid flashes or burns on the tube but does not
continue to burn when collected in the tray below.
I(D), when the fluid flashes or burns on the tube and
continues to do so when collected in the tray below
N when the fluid does not flash or burn at any time. a. flashes and burns on the manifold but not
after dripping from the manifold
b. does not flash or burn on the manifold but
does after dripping from the manifold
c. the fluid does not flash or burn on
manifold or after dripping from the manifold

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 27
FASCICLE VIII, 2007 (XIII), ISSN 1221-4590
TRIBOLOGY

Some laboratories explai n the reasons why they
could not do these tests, especially the absence of the
necessary equipment and some of them could do one
or two of these tests but in accordance to other
standards. For instance, Total Lubrifin SA can
perform two of the above-mentioned tests but
following the requirements of ASTM standards

3.3 Comments of ISO/TC 28 for ISO 20283

Before approval of the standard ISO 20283,
comments done by the ISO members have had
imposed the modification of the initial text, the voting
being achieved at 21.10.2002. Among the 15
members of TC 28 (Technical Committee 28) that
agreed to the final text were states known as having
representative mining and heavy industries as USA, United Kingdom, France, Poland, Romania [33].
The discussions on the initial text were focused
on the following issues:
– maintaining the fluid drop rate at 40…60 sec
between drops; a solution could be to repeat the dropping before testing the fluid on the hot manifold
in order to obtain the desired value; other option
would be the use of the controlled dispenser as in the
electronic assembling processes;
– the high between the dispenser tip and the
manifold; the UK member outlined that he did not
know papers that study the influence of this parameter
but recommends the distance of 300 mm as will be
required in the approved standard;
– the box design for mounting the tube: this
enclosure has to facilitate the taking out of the tube for cleaning.
– the representing specialist of India suggested
to measure the temperature w ith three sensors, two of
them at the ends of the active zone of the heated
manifold and one at its cent re, the accuracy being kept
at de ±5°C for a tested te mperature of 700°C at all
three points of temp erature measuring;
– if there are done more than one test without
cleaning, the standard establish the position of first
test and the height of the dispenser;
– to the initial text was added the following from
the representing specialist of UK in order to extend the utilisation “It may be used with other types of more
flammable fluids at lower surface temperatures, but this
could significantly increase the hazard of this
procedure”;
– restrictive conditions for protecting the heating
element: the cleaning could be done without
extracting the heating element but with caution;
– the tested fluid can not contain air bubbles,
thus the dispenser has to allow the elimination of air bubbles that could occur during the sampling of the
fluid.
Analysing the standard and the comments [33]
done by TC 28, one may notice that the principle is quite simple but the test is complex due to the
requirements imposed to the tester and to the
procedure:
– the imposed requirements for temperature
measuring, its accuracy recommend that the sensors could be withdrawn without manual intervention as
this could be dangerous due to high temperatures
involved;
– the fluid drop rate could be achieved only by using
an automatic dispenser taking into account even that
the fluid viscosity could vary a lot;
– it is necessary a positioning system allowing to
obtain the imposed accuracy for the dispenser tip position and also allowing the movement parallel to
the manifold for having a new testing position;
– according one of the three qualifications may be
supported by photos or films taken during the test.

4. THE TESTER

The design meets the requirements imposed by
SR EN ISO 20823:2004. The tester comprises (fig. 2)
a robotic system 2 that ensures automatically the
positioning of the fluid dispenser 4 above the
simulated manifold, at a position desired by the operator (high and position along the manifold),
within a high temperature enclosure 5. The manifold
is made of corrosion resistant steel and it could be
heated by an electric system up to 700°C. The
temperature control is done by the help of three digital
thermocouples mounted on a transversal guiding
system that allow them to be in contact with the
manifold and then withdrawn during the fluid dropping. All the above mentioned subsystems are
enclosed in the ventilated enclosure 3. The tester may
be controlled by the help of its own specialized soft
that could run on a PC (6). The soft also allows the
operator to control the tester.
The tester may function in two different ways
1- manually controlled cycle
The operator may command the robot arm and the
arm supporting the thermocouples’ system.
2– automatically conducted cycle
1. The doors of the ventilated enclosure are closed 3.
If during the functioning the doors are accidentally
open then the robotic arm and the thermocouples’
platform are withdrawn at their initial position.
2. Once the ventilated enclosure has the doors
closed, the ventilator is started.
3. The heating system is started. This will function
only if the doors of the ventilated enclosure are
closed.
4. The heating system is ma intained till reaching the
temperature previously introduces in the computer.
5. Reaching the imposed temper ature is the start of
the engine acting the thermocouples’ platform.
The motion will continue till the thermocouples

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 28
FASCICLE VIII, 2007 (XIII), ISSN 1221-459 0
TRIBO LOGY

reach t he heated m anifold. T he data re gistered by
temperatu re sen sors will be display o n the PC
screen.
6. Whe n the temperature sen sors give the desi red
values, th e sen sors’ p latform is au tomatically
withdrawn in the initial po sition.
7. By the hel p of the s pecialised s oft, the e ngine is
commanded t o move t he robotic arm and t he
dispenser in the position alread y estab lished at th e
beginning of the testing procedure.
8. The dispenser is com manded to drop the fluid as
estab lished at the start o f the au tomatised
procedure (including rate flow).
9. After dropping the fluid the soft gives the
comman d for withdrawing the robotic arm at the
initial p osition.

5. PRELIM INARY RE SULT S

Table 4 prese nts three tests done at elevated
temperature a nd in the last colum n there are prese nted
observations that co uld give supplementary
characte ristics of the tested fl uids.
For instance, a chara cteristic that is not
mentioned in the ISO st anda rd, but was recorded
during these preliminary test s, was t he time in terval
till the flu id starts to burn, time measured fro m the
starting of the fluid dispensing process. Fr om table 3,
on e m ay notice that the gra de oil ACEA-2002 A2/B 2
– API SL/CF had the larg est ti me in terval between
drop starting an d burn starting on the manifold. This
larger tim e may b e cru cial in activ ating th e fire
suppression systems. Another im portant aspect
noticed during these tests was th e quality an d quantity
of the residues left on t he manifold, in the tray an d on
the metal box walls.
Figure 7 prese nts images du ring the test done for
the fluid AC EA-2002 A2/B2 – AP I SL/CF.
Vary ing the manifold temperature, a nd repeating
the test methodology it is possible to estab lish a
temperature (or a very na rrow range of te mperature)
to ignite th e fluid, un der the test co nditions. Th is
temperature c ould be com pared to the ignition point
or flash poi nt, this comparison bei ng a criterion fo r
establishing t he degree of hazardous probability for
the fluid.
A complex stud y [29] base d on sam ple principle
as IS O 20823, gave a range of hot surface ignition
temperatures for differe nt hydraulic fluids. The t est
results are very diffe rent a ppreciated: the lower value
is the m inimum hot surface temperature at which t he
oil sp ray was n ot ignited at least o nce, while th e
higher one is the m aximum hot surface temperat ure at
which the sam e fluid spray was not ignited at least
once . The conclusions of this study reveal :
– the viscosity o nly seem s to affect th e oil ato mi-
zatio n, not the combustion, – the flash point has little effect o n the minimum
hot surface ignition tem perature.
– ho t surface ig nition is affected both b y
chem ical reacti vity and the volatili ty of the fluid.
Babrauskas [in 29] assum ed that the hot s urface
ignition tem peratu re is ap proximately 2 00°C abov e
the autoignition temperatu re, and future tests will try
to relate th is characteristic to the tem peratu re range
that passes o ne fluid from N category (when the fluid
does not flas h or burn at an y time) to I(T ) or I(D)
categories.

a) The first fl uid drops reach the m anifold.

b) The first flash es, after few secon d after startin g the
fluid dropping.

c) the fluid burns on the manifol d, during its drip and
it burns in the tray below.
Fig. 7. Images during t he test.

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 29
FASCICLE VIII, 2007 (XIII), ISSN 1221-459 0
TRIBO LOGY

1 2 3

Fig. 6. The tester for flamm ability characteri stic of fluid on hot manifold. Original solution from ICTCM and
Galati Un iversity.

Table 3 Preli minary re sults in determ ining the flammabilit y characte ristics of fl uids in contacts with hot surfaces
(methodol ogy as required by EN IS O 20 823:2003)

Tested
oil
Num-
ber of
test H
(mm)
Ave-
rage
height Man i-
fold
tempe-
rature
oC Oil
vo-
lume
(ml) Drop-
ping
time
(sec)
Perm issible
deviation of
parameters
Comments
1 300 620 10 45
2 300 620 10 45 Oil
grade
H.G.
46
3 300 620 10 45 ±5 oC;
±0.5 ml;
± 10 s;
300 mm
± 10 mm – the fluid ign ites after 5-7 s and- after
dropping 1…1,5 ml of f luid;
– test result: I(T ), when the f luid flashes
or burns on the tube but doe s not
continue to burn when collected in the
tray belo w.
1 300 700 10 50…60
2 300 700 10 50…60 Oil
grade
T 90
API
GL-2 3 300 700 10 50…60 ±5 oC;
±0.5 ml;
± 10 s – the fluid ign ites on the manifol d after
5…10 sec o f start dropping and it burns in
the tra y below .
– test result: I(D ), when the f luid flashes
or burns on the tube and c ontinue s to do
so when collected in the tray below
1 300 700 10 50…60
2 300 700 10 50…60 Oil
grade
ACEA-
2002
A2/B2
– API
SL/CF 3 300 700 10 50…60 ±5 oC;
±0.5 ml;
± 10 s – the flu id ign ites and burns after 22 sec
from drop startin g< it burns on the
manifold, du ring its dr ipping but it does not
burn in the tray,
– a lot of r esidue was noticed after burning
on the tub e but also in the high-temperature
metal enclosure,
– test result: I(D ), when the f luid flashes
or burns on the tube and c ontinue s to do
so when collected in the tray below

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 30
FASCICLE VIII, 2007 (XIII), ISSN 1221-459 0
TRIBO LOGY

Resu lts fro m laboratory tests h ave the potential
to solve, often on ly partially the fire hazard in real
industrial activ ities b ut specialists (b oth laboratory
and risk asse ssment expert s) do not offer clear
methods of how to use the laboratory results [18].
The eval uator has to do a de tailed examination
of both technical system and i ts environment and to
perform a “full” inventor y (many accidents are the
resul t of a f orgotten aspe ct) of p otential ignition
sources. Th is will in clude the lo catio n and
temperature of an y hot surface, th e presence of lagg ed
pipes or t he existence of unseale d electrical
equipment etc. In normal funct ioning, if the surface
temperatures the fluid m ight cont acts, were below the
hot manifold ignition temperatu re (that means for ISO
20823 the fluid is in N cate gory), the n the source m ay
be discoun ted in all bu t fault conditions.
Tribo logy could help specialists to evaluate
temperatures of diffe rent surfaces in relative m otion,
in critical s cenari os (lea k of l ubricant, plastic
deformations etc., to tal or partial re moving of frictio n
coatings etc.). The e valuator m ay develop the
following scheme fo r ignition on hot surfaces:
– if norm al functioning tem perature s exceed the
test hot manifold temperature Tc, th e ig nition
probability is 1,
– for functioning temperatures below Tc, the
evaluator h as to designate an ignition probability on a
scale that could be ze ro for som e fraction from Tc
(som e specialists give 0.5T c or 0.75Tc [18]),
– add itional rising of ignition probability when
fault conditions are possible to happen.
In some cases the evaluat or coul d apply risk
index methods, cal culating a “ri sk val ue” or ri sk
index, base d on the relationship:
n
risk jj
j1I ar
==⋅∑ (1)
where is the attrib ute j related to risk evaluation
(for insta nce, ignition te mperat ure on hot surface,
smoke production, electrical ignition sou rces etc),
j=1,…,n, and is a v alue associated to likelihood of
occu rrence a nd consequence s. Both and has to
be introduce d in a norm alised scales. may have t he
following values: 0 – the occurrence is not cred ible, 1
– unlikely, 2 – medium probab ility, 3 – highly lik ely.
For the attribut e of ignition on hot surfaces, this could
be related to the ignition tem peratu re of the flu id
involved, but in an ind irect pro portio nality to this one.
For insta nce, if the e ngineer had to select an industrial
fluid among several with differen t hot surface ig nition
temperatures , , after tested un der the pro –
cedu re of ISO 20823, the fluid j has the normalised
attrib ute ja
jr
jajr
jr
12 n TT …T〈〈jnaT /Tj= (2)
It is o bviously th at a lo wer value of this attri bute
is desi red for safet y functioning and for a l ow
probability of hazardous ev ents. The probl em to be
solved is th e compromise between initial costs and
performances of t he fluid to be sel ected. Several
decade s ago the ratio betwee n high sec urity fluids a nd
hazardous flui ds was as great as 5…3 to 1. For
instance, a fluid-power syst em will b e more ex pensive
when using water-based fl uids due to the materials
involved in designing (especially co rrosion resistan t
steels, sealings etc.) as com pared to a sy stem with
similar performances but using m ineral oils. Recently,
specialists g ive d esign solutions that overpass on ly
with 30…50% th e classical o nes, fun ctioning with
more hazardous flui ds [13, 31].
The fi re resi stance of s ome hydraulic fluids m ay
change with time o r with operatio nal serv ice. Man y
fire-resistan t fluids rely on their water con tent or their
chemical co mposition and physical pro perties to
provide fire resistance. Circum stances that coul d
result in th e red uction of th e water con tent below its
original value or chem ical or physical cha nges i n the
fluid could produce hard-to-estim ate fire resistance.
Such situ ations cou ld arise through persisten t high
temperatu res, flu id spillage where evaporation or
separat ion could occur or breakdown of fluid
chem ical properties duri ng use. No specifi c test has
been designated to cater for th ese situ ation s, which
should be addressed through regular fluid monitor-
ring and goo d housekeeping procedures [15, 18 , 23].

6. CONCL USIONS

The original so lution for th e tester allo ws safe
and repeatab ility o f the procedure and also the
variation of some param eters as fl uid, manifold
temperatu re, heig ht of th e dispenser tip and position
along the manifold.
There is no test th at en sures a h igh level o f
safety for fire resistance b ut a particular se t of tests,
selected after an actual risk assessm ent could give a
better solution for a sy stem using hydraulic fluids in
high risk environments.
Determination of fl uid flammability on hot
surfaces im poses particula r solutions for im proving
the secu rity of the designed syste m.
The l ist of hy draulic fluids p ossible to be
selected and t he tests that these fl uids ha ve to pass ,
will h ave to be known and set even in the d esign stage
of the eq uipment in orde r to introduce necessary
solutio n for reducing fire ris k. It is also im portant to
analyse similar accide nts related t o the real
application s in o rder to notice p ossible improvements
in equipment, process and environment cont rol and
for workers’ training .

THE AN NALS OF UNI VERSITY “DUNĂREA DE JOS“ OF GALAȚI 31
FASCICLE VIII, 2007 (XIII), ISSN 1221-4590
TRIBOLOGY

AKNOWLEDGEMENT

This research was supported by National Authority
for Scientific Research (ANCS), Minister of Education
and Research Romania, under the grant CEEX-M4-452
“Adoption and Implementation of Test Methods for
Lubricant Conformity Assessment.”

REFERENCES

1. ***Approval Standard for Flammability Classification of
Industrial Fluids (Class 6930), Factory Mutual Global, January
2002.
2. *** CETOP RP 65H 1994, Manifo ld Ignition test for Fire-
resistant Fluids.
3. *** Council Directive 92/104/EEC of 3 December 1992 on the
minimum requirements for improving the safety and health
protection of workers in su rface and underground mineral-
extracting industries (twelfth individual Directive within the meaning of Article 16 (1) of Directive 89/391/EEC),
OJ L 404,
31/12/1992 P. 0010 – 0025.
4. *** Directive 94/9/EC of the European Parliament and the
Council of 23 March 1994, on the approximation of the laws of the Member States concerning equipm ent and protective systems inten-
ded for use in potentially explosive atmospheres, (OJ L100,
19.4.1994, 1), amended by Regul ation (EC) No 1882/2003 of the
European Parliament and of the Council of 29 September 2003, OJ
L284 1 31.10.2003, Corrected by Corrigendum, OJ L21, 26.1.2000,
42 (94/9/EC), Corrigendum, OJ L304, 5.12.2000, 19 (94/9/EC).
5. *** Fire Testing Services Offered by Southwest Research
Institute in Support of the Inte rnational Maritime Organization
(IMO) and Fire Test Procedures (FTP) Code, October (2004).
6. *** HSE Approved specifications for fire resistance and
hygiene of hydraulic fluids for use in machinery and equipment in mines, (M) File L11.6/3, (1999).
7. *** Directive 1999/92/EC of the European Parliament and of
the Council of 16 December 1999 on minimum requirements for
improving the safety and health prot ection of workers potentially at
risk from explosive atmospheres (15th individual Directive within
the meaning of Article 16(1) of Directive 89/391/EEC), OJ L 23,
28.1.2000
8. *** Regulations (the CHIP ACOP on Test Methods), on-line
(07.05.2007), at:
http://www.ueil.org/documents/h se/PrepDirectivePosition.doc
9. *** The 7th Luxembourg Report, March 3, 1994, Doc. No.
4746/10/91, Requirements and Test Applicable to Fire-Resistant
Hydraulic Fluids used for Power Transmission and Control; L-
2920m, Luxembourg, Commission of the European Economic
Communities DG\E\4.
10. *** US Department of Labor Mine Safety and Health
Administration, 30 CFR 35.20; Autogenous- ignition temperature
test, 30 CFR 35.21 Temperature – pressure spray-ignition tests, 30
CFR 35.22; Test to determine effect of evaporation on
flammability, 30 Code of Fede ral Regulations, SUA, on-line
(17.12.2006), at: h ttp://www.msha.gov.
11. Brandao A.V., 1998, New hydraulic fl uid flammability test by
Factory Mutual, Iron and Steel Engineer , December, pp. 27.
12. Deleanu, L., Rîp ă, M., Drug, A., 2007, Flammability Testing
in Risk assessment, Proc. of European Conference on Tribology,
ECOTRIB, Ljubljana, Slovenia, pp. 449-462. 13. Gere, R., Hazelton, T., 1993, Rules for choosing a fire-
resistant hydraulic fluid, Hydraulics & pneumatics , vol. 46, no. 4,
pp. 37-42.
14. Goode, M.J., Phillips, D.W., Winkeljohn, R.D., 2000, Fire-
Resistant hydraulic Fluids and the Potential Impact of New Standards for general Industrial Applications, Proc. of the 48
th
National Conf. on Fluid Power, paper 100-1.12.
15. Grosshandler, W. (editor), 2001, Work on Fire Testing
Measurement Needs, Proceedings.
16. Hamins, A., Borthwick, P., Presser, C., 1990, Supression of
ignition over a heated metal surface, Int. Conf. on Fire research and
Engineering, Sept. 10-15, Orlando, SUA.
17. High, M.S., Danner, R.P., 1987, Prediction of upper
flammability limit by a group co ntribution method, Industrial and
Engineering Chemical Research, 26, pp. 1395–1399.
18. Jagger, S.F. Nicol, A., Sawyer, J., Thyer, A.M., 2004, The
incorporation of fire test data in a risk-based assessment of hydraulic fluid fire resistance, INTERFLAME, pp.
569-576.
19. Khan, M., 1991, Spray Flammability of Hydraulic Fluids and
Development of a Test Method, Technical Report FMRC J.1 0T0W3.RC, Factory Mutual Corporation, Norwood, MA.
20. Korb, T., 2006, The effect of metal-oxide scale son hot surface
ignition of hydrogen-air mixtures, PhD Thesis.
21. Krishna, K. et al., 2003, The use of aerosol formation,
flammability, and explosion info rmation for heat-transfer fluid
selection, Journal of Hazardous Materials, 104, pp. 215–226.
22. Phillips, W.D., 2004, The Development of a Spray Ignition
Test for European Hydraulic Fluid Specifications, Symposium of Fire-Resistant Test Methods, ASTM D02 N. 06 Meeting, Tampa,
Florida.
23. Satyanarayana, K., Kakati, M.C., 1991, Correlation of flash
points, Fire and Materials, 15, pp. 97–100.
24. Shermann, J.V., 2005, Current Standards for Certification of
Fire-Resistant Hydraulic Fluids Used in General Industrial
Applications, Fluid Power Journal , pp. 2-9.
25. Tewarson, A., 2005, Thermophysical and fire properties of
automobile plastic parts and engi ne compartment fluids, Technical
Report, FM Global, Project ID 0003018009-3.
26. Toy, N. Nenmeni, V.l., Bai, X., Disimile, P.J., 2004, Surface
ignition on a heated horizontal flat plat, 4th Intern. Aircraft & Cabin
Safety Research Conference.
27. Wierzbicki V., 2006, Flash Point testing: A User Guide,
Petrotest, on-line (15.04.2007):
www.petrotest.com/media/en/FlashpointGuide.pdf
28. Wright, B.R., Mowery, R.B., LePera, M.E., 2000,
Survivability? It Could Be a Matte r of Hydraulic Fluid!, Proc. 11th
Annual US Army Ground Vehicle Survivability Symposium, 27-30
March (2000).
29. Yuan, L., 2006, Ignition of hydraulic fluid sprays by open
flame and hot surfaces, Journal of Loss Prevention in Process
Industry , vol.19 (4), pp. 353-361 .
30. Yule, A.J., Moodie, K., 1992, A method for testing the
flammability of sprays of hydraulic fluid, Fire Safety Journal , vol.
18, pp. 273-302.
31. Zink, M.D ., 2006, New Fluid Technologies for Water-Free,
Fire-resistant Hydraulic Fluids, Iron & Steel Technology, vol. 3
(4), pp. 98-101.
32. Zinc M.D ., 2000, Fire Resistant Hydraulic Fluids: Shifting
Definitions and Standards, Proc. of the 48th National Conf. on Fluid
Power, paper 105-8.3.
33. *** ISO/TC 28 Petroleum products and lubricants, ISO/TC
28N 2139, Results of voting ISO/CD 20823 Petroleum and related
products – Determination of the flammability characteristics of
fluids in contact with hot surf aces / Manifold ignition test,
24.10.2001.