full size testing

8/11/2019 Full Size Testing

1/7

95 19 6

FullSize

Testing

o f

Drainage Systems

for

Burning

Liquids in Road

Tunnels

Didier

Lacroix,

Research Manager,Centre

d'Etudes

de s Tunnels,

Bron

/Dr.

Eric C asale, Scetauroute,

Annecy/

Claude Cw iklinski,

l .N.E.R.l.S.,

Verneuil

en

Halatte

/

Andre

Thiboud,So cietedes

A utoroutes

Paris-Rhin-Rhne,

Frankreich

1 .

Introduction

Most vehides transporting dangerous

substances on road are liquid

hydrocarbon tanks.

When their

passage

is permitted in

a

tunnel,

measures should be taken in order to

limit

risks in case of an accident.

Thus,

the

recent

French

road

tunnels are

equipped

with collecting

Systems fo r

burning

liquids spilled on

th e

pavement. The objective

is

to

reduce

heat and

smoke

produced

by a

fire,

to

limit

fire

propagation

to

other

vehides,a nd

to

enable rescue teams

to

approach

th e

fire.

On the occasion of the d o u b l i n g of the Chamoise T u n n e l , o n

the

A.40

motorway between Lyon and

Geneva,

ageneral s t u d y o f thesesytems

has

been

f i n a n c e d b y the French Q o v e r n m e n t and Societe

des

A u t o -

routes P a r i s - R h i n - R h n e (S.A.P.R.R.), the s e m i - p u b l i c

Operator. The

study

was d e f i n e dand f o l lo w e d u pj o i n t l y b y S C E T A U R O U T E ,contractor

to S.A.P.R.R.,and

Centre

d'Etudesdes T u n n e l s ( C E T U ) ,

w h i c h

ensured

the general guidance and conducted the

h y d r a u l i c

tests. The

Inst i tu t

N a t i o n a l de

l ' E n v i r o n n e m e n t

I n d u s t r i e l

et des R i s q u e s

( I N E R I S )

was

charged with riskstudies and f i reand explosiontests.

2 Preliminary

studies

Traditionally

water

running on

the

pavement

in

road tunnels is

collected

through open gutters along the pavement, then discharged into a

collec-

to r

through

50-100m spacedgully holes.

In

order to improve absorption

fo r

dangerous substances SCETAUROUTE use gully holes at narrower

inten/als,

about every10m.

F or

their part CETU drew elements from th e

Fig.

1: chema

of

a

drainage ystem with continuous slot gutters an d

iphons

Swiss facilities

and elaborated continuous slot

gutters which

communi-

cate

with

t he coliectorvia

gully holes

with

siphons, in order

t o extinguish

burning

liquids

(figure

1).

Very

f e w

l i ter atur edataare

available

o nthe

ef f ic iencyand possible

risks

of such

Systems,

e i t h e r in

t u n n e l

o r i n o t h e r s t r u c t u r e s ,e.g. o il industry.

T h is

just i f i ed

a

t e s t in g P r o g r a m m e .

The statistics on dangerous

substances

accidents

on

open roads an d in

tunnels have been analysed and allowed

to

consider

representative

study

cases:

c o n t i n u o u s leakage

of

30-70

l / s c o r r e s p o n d in g t oa

component

b r e a k -

age o ra p u n c t u r e of 100-150 mm in d ia m e t e r ;

sudden

release

of

a volume

corresponding to the

average recorded

spillages,

i.e.

5 m

3

excluding

t he

extreme

cases,

or 10

m

3

includingth e

fe w

most

important cases.

3 Hydraulictests

Th e

first

experimental phase was performed

in six operated

road tunnels,

in

order to make a

hydraulic

evaluation

and

comparison

of

the various

collecting Systems.

3.1.

Testingmethodology

Water

was used

f o r obvious reasons

of

safety and convenience.

The

c o n t i n u o u s leakages were o b t a i n e d b y r u n n i n g

water

o u t of th e t u n n e l

f ire

h y d r a n t s .

A prefabricated3

x

m p o o l , with one side that c o u l d be

opened

s u d d e n l y ,was used t o s im u la t ethe s u d d e n

spillages

of

5m

3

and

10 m

3

,

u n d e r g o o d

r epr oducibi l i ty

c o n d i t io n s .

Three camescopes recorded

the

tests

in

order to characterize the

we t

surface

during the spillage

operations.

The

pavement was

squared

by

retroreflective, self-adhesive

strips

and numbered

studs

which

also

allo-

wed t

measure

the water depth. When allowed by the site, the

water

height

w as

permanently mea sured in

the gutters.

3 2 Test results

The observation

with time of th e

liquidsheet let

appear

three successive,

interacting

f lowing

zones (figure

2):

A

first

zone

can be q u a l i f i e d

s

iner t iazone,because it is

a

f u n c t i o n

o f

the

l iqu idrelease

c o n d i t io n s ,

p r inc ip al ly

its e ject ionspeed;

The

second zone

is

d e t e r m i n e d

b ythe

g r a d i e n t ,the transverse

slope

and

thesurface

characteristics o f

the

p a v e m e n t ;the liquid

f l o w s inthe

g e n e r a l

d i r e c ti o n of

the

pr inc ipal l ine u n d e r the ef fect of

gravity;

I n

a t h ir d

zone,

also

a

g ravitat ion one, the l iq uid n o tabsorbed into

the

gutter f lows

along

th esidewalk

edgebefore

b e i n g

removed t h r o u g h th e

d r a i n a g eSystem.

We

called

them zone

B the f irst two

surfaces together: it corresponds to

th e

minimal

sheet

extension

that

would be obtained ifall

water reaching

CongresMondial

des

T u n n e l set

Jo ur nees

S T U V A ,Stuttgart,

6-11

mai1995,

p.

230-236

p

Author manuscript, published in "World Tunnel Congress et STUVA (Tagung'95), Stuttgart : Germany (1995)"
http://hal.archives-ouvertes.fr/


2/7

Fig.

: D iagram

of

f lowing zones

the

edge

of the

pavement was immediately absorbed. The

third

surface,

called

zone

A, expresses

th e

factthat

the device does

not

absorb the

liquid

perfectiy:

it

therefore allows to compare th e

efficiency of

th e

various Systems.

Table

1 presents the

characteristics

of the tested tunnels, and gives s

an

example the main

resultsobtained

with the

sudden

release

of

10

m

3

.

Figure

3 shows the extent of the wet zones

in this case.

The continuous

leakages an d

sudden

releases

of

5

m

3

led

to

comparable

situations,

but

obviouslywith

smaller

w et

surfaces

and lengths.

3.3. Main conclusions

Inall cases

it appears

that the conventional

System

moreover badly

maintained

of

the

Monts Tunnel

is

particularly

little efficient. The wet

surface exceeds 1000

m

3

,

evenfo racontinuous

leakage

o f35 l/s

The comparison

of

zones A shows

that

the slot gutters prove to be the

most efficient facilities. Among

th e

two tested types of slots (figure

4)

those with th e

opening

in the verticalplane

benave

best: the zone A is

non-existent

in theCornil and Siaix Tunnels,

even for

a

sudden

release

of 10 m

3

. This arrangement indeed allows to utilize directiy the flow

dynamics

to removeth eliquid. In

the

Grand MreTunnel,where

the

slot

opening is

in

the

horizontal

plane,theliquid

is

collected

under

the action

of gravity onlyand

thereforen ot s o well intercepted.

Inthis

last

tunnel, the slot gutter is also hindered

by

its inner diameter

of 200 mm, which proves insufficient and forces the liquid

back

in case

of

heavy release.

The diameter of 400 mm

installed

in the two other tunnels is satisfying.

Siphons there also showed

a

proper

hydraulic

behaviour: in the

Siaix

Tunnel, the volume of 10

m

3

was

drained

through the

gully hole within

50

s,

which corresponds to an average flow

of

20 0

l/s.

The

lowerperformance

of Systems

with

gully

holes

a tsmall

interval

in the

Chti llon and Saint-Germain-de-Joux Tunnels ca n be explained by the

very

principle of

the gully holes and their design,

which

only allow an

insufficient

absorption.In addition

th epresence

o f

small open

gutters,

3 0

cm

in width and 3

cm

in depth, along each sidewalk (including at the

transverseslope top) induces

th e

formation of very long liquid "tongues"

on

both sides

of

the pavement.

3.4.Other S tatements

Except the case of the MontsTunnel, water needs less than 2 minutes

beforebeing removed from the pavement

after

t he en d

o f

the continuous

or suddenspillage.In the Grand-MareTunnel,however,where the pave-

ment

is coated with pervious

bituminous

concrete, someliquidco ntinues

to flow below the wearing course for over

15minutes.

Questions may be

raised about the risksresulting fromthis lasting presence of dangerous

substances inside th etunnel.

Lastly

it

must be mentioned

the

f lowing

problems noted in certain

drai-

nage

Systems, which

induced liquid to be

forced

back onto the

pavement, thus

raisingapotentialrisk.

Such

phenomena resulted

either

from anarrowerhydraulic section at

th e

gully hole

entrance, or

from the

presence of

debris due

to

insufficient

maintenance. Obviously the

strict

avoidance o f

such

situations

mustb ereached.

lilillil^

|||||[|Mi(roel|am||

|||||loca

M o n t s

(Chambery)

Chti l lon

(A

4 0)

SaintGe nnain

de

Jo u x (A

40)

Grand Mre

(Rouen)

Corni l

(RN

8 9

-Co rreze)

Siaix

(RN

90

Svoie)

IIBII

liliientll

0.7

0.6

1,8

3.5

3.2

1.5

an

IIIBBII

1

1.8

3

2.5

2.5

2.5

iilllillllllllranage|||||||||||^

||||||j|||||| |system|||||||||

Conv.grid

gully

hole

75x30

cm

2

every

50

m

Grid

gully

hole

30x30

cm

2

every

l im

Grid

gully

hole

13x50 cn^every

l im

Horizontal

slot

gutters

and siphons

Verticalslot gu tters

an d

siphons

Vertical

slot

gutters

and siphons

||||||||||Se

ili

Wsrtacs^

1480

38 5

405

535

300

335

illiilii

^ i i t e n g t I i S ^ ^

llllmlllil

> 2 0 0

105

120

90

35

40

IIZonellKl

llsurll'cell

IlMl

1080

140

155

120

0

0

||Zoinej|i|

lienjgltK

lilnli

> 2 0 0

80

95

40

0

0

- 2-

p


3/7

100m

50m

Om

Les Monts

(gully holes every 50m)

100m 50m

Om

C h a t i l l o n (g u l ly h o l e s every

l i m )

100m 50m

100m 50 m O m

Grand

Mre

(horizontal continuous slot)

50m

Om

00m K

m

E 3.2%

l

K

"

(s

P

S A

r

Cornil (2 vertical

continuous

siots)

100m 50m Om

St Germain

(gully

holes every 11m)

ES ZoneA

S

Zone

Pool

usedfo r spillage

Siaix

(vertical c o n t i n u o u s s lot)

Fig. 3 : W et

zones

A

a nd B

fo r

a

sudden

spillage o f 10m

3

4

Fire and explosion test

As the Systems including

slot

gutters proved

to be properiy operating

from

the hydraulic

point of

view,

it appeared

necessary to test then

regardingfireand explosionrisks.

The basic

purpose

of

these tests

w as no t only

to check

th e

efficiency of

the

facility fo r extinguishing and removing

a

burning

liquid,

but also to

characterize the mechanisms

implied

in dysfunction situations.

This

explains

that

numeroustests

have

been devoted

to

such

cases,

although

they

a re exceptional.

4.1. Methodology

4.1 .1 .

T he

experimental

facility

F o r

safety

reasons, tests could not be made U n d e r g ro u n d The

selected

site

was the

I N E R I S

M o n t l a v i ll e open

site i n Verneuil-en-Halatte.

As no similarities

allowing a reduced scale study were

available,

a

section typical of a drainage System w as

reproduced in fll size. It

included tw o Siphons separated

by

a 48

m long

slot gutter communi-

cating with a collector which crosses

th e

facility

from on e end to the

other

(figure

5).

Povi

H o r i z o n t o l l y

opened

s l o t

Verticolly opened

slot

Fig.

4 :

Po ssibles c h e m e s fo r continuousslotgutters

Fig.

5 : Diagram o f the

experimental

System

- 3

p


4/7

Th e

gully

holes and their siphons were th esame s thoseinstalledin

th e

Siaix

Tunnel.

Th e

internal

volumes were numbered from 1 to 4 in

the

liquid

flowingdirection.

A

second facility,

consisting

only of tw o

coupled gutter

sections

(6

m)

allowed to characterize more stationary phenomena.

Th e

System

is

sea-

led

at both ends, so

that the combustion corresponding

to various filling

up

levels can

be

tested.

4.1 .2 .M easuring p rinciples

Th e

facility

is

provided

with

a

complex

measuring

equipment

including

thermocouples, flame detectors, pressure sensors, explosimeters,

etc.

These

appliances are

distributed

in

th e

sensitive

zones

according to the

expectedphenomenologies.

Al ldata

are

sent back to a collecting Station.This

has

a completely self-

sustaining

Operation, since

th e personnel has

to stay beyond

a

safety

limitduring mosttests.

The visualobservations

are

strengthened b y fourvideo cameras focused

onto

th e

sensitive points

o f

t hefacility.

4.1 .3.

Spillage principles

The

product

selected for these

tests

was

unleaded gasoline, because

this

liquidcountsamongth e

mostdangerous

ones which are representa-

tive of

the risk connected

with

th e transportation of inf lammableliquidsin

tunnels.

For

safety reasons,

and in order to ge t

a

better characterization

of

th e

phenomena,

th e

Spillage

implying burning

gasoline

didn ot

use quantities

so

high

s

fo r

the

hydraulic tests.

Al l tests

were conducted

with f lows

of

about3

l/s.

4.1 .4 .

Survey

o f the

experimental testing series

Tw o

experimental

testing

series

were conducted at

an

interval of one

year.

Inafirst

approach

( in

1993) onetried

to

characterize

the behaviour

pecu-

liar

to

each component

o f

the drainage

System.

It

appeared

then

thatt he

exchanges occuring at

the

limits

of

each element were determining

fo r

the general phenomenology. Th e second series (in 1994) therefore fo -

cused on more

global

behaviours.

In total

twenty-five tests

were conducted.

This

figure does not include

a

few preliminary tests made

without

ignition.

4 2 Demonstrated basic phenomena

4.2.1.

Fire

The fire developing

ina drainage System issubmitted to

constraints in

relation

to t he confined environment.

This

latter limitst he oxygenquanti-

ties

available for the

combustionmechanisms a nd hinders the

removal

o f

burntgases.

These conditions

m ay lead

to self-extinction.

Maintaining

the

combustion

requires

a

sufficient movement

of

gases.

Th e

mainspring of this

System

consists

in

the Archimedean

Stresses

developed

mainly at t he

fire level.

This process of

fresha ir

supply

will

be

calledbelow

"respiration".

4.2.2.Ex p l os ion

The risk analyses

show

that gasoline ca n

enter

the drainage

System

without

burning.

Vapours

therefore

can develop

there. The

confined

en-

vironment generally succeeds in

maintaining

them at

a

high concentra-

tion.

The explosion risk is

a

function

of gasoline vapour concentration

at th e

instant

of

ignition. For th e unleaded gasoline used for the tests, the

atmosphere

is

made

deflagrable by hydrocarbon contents

between

1.4

and 7.6 %.

Fig. 6:Co m bustio n o r

self extinctionm echanisms

in the slot

gutter

4 3 Behaviour

of the facility

in case

of fire

It

wa s demonstrated

that the

phenomena which occur

at

various

pointsof

th e

facility

cannot be

dissociated

in

a

globalphenomenology. It is pos-

sible, however, to describe

th e

response peculiar to

each

system

element

to

a f ire.

Al l mechanisms

described

here

have

been brought out by the observa-

tiona nd

analysis

of recordings made

during

th e

experiments.

4.3.1.

Gutter

In anormalO peration

Situation

th e gutter is the only element in the drai-

nage System to have a

volume

in direct contact with the

tunnel

at -

mosphere.

Al l

tests

show that

a

gasoline

fire

in the slot

gutter

is little

violent and does

not

produce

much smoke.

In

a

tunnel

the generated

opacity

will

therefore be lower than

that could

be expected in such an

event.

This

is

therefore

a

rather favourableSituation.

Various

observations show

thattw o

respiration

modes are in rivalry:

a transverse

mode

in which

fresh

air

flow

and burnt

gases flow must

cross

in the slot

thickness

(6

cm);

a longitudinal

mode which supposes that no combustion exists

in

a

zonenearthe flame (figure

6 ).

A number of tests

with blocked

longitudinal respiration movement

(small

gutter

length, or f lame

on

the whole slot length) lead

to

self-extinction.

They

tend

to

demonstrate

that

th e transverse

respiration

mode is

very

fragile.

These tests generally

correspond

to very particular experimental proto-

coles. The other tests show that the maintainedcomb ustion remains the

most probable

hypothesis.

In other words the fire self-extinction in the

gutter,

although

possible,

does not

appear

s aprobable phenomenon

in

a realistic scenario.

4.3.2. Siphon

a)

Firebreak

role of the

Siphon

In presence of

a

hydraulic Isolation

even gasolineth e siphon en-

sures its

firebreak role.

Tests show

that

the

f lame

never succeeds

in

reaching

th e

downstream compartment.

b)R espiration mode innormalOperation

Th e tests whichimply asiphon fire show that the combustion

process

is

being

maintained there

provided

that

the longitudinal

respiration

mode

h as already

settled

in th e gutter (figure7) .

T he

secondcondition

supposes

that

th e fresh air f low and the burnt ga s

flow cross

in th e

entrance hole.This movement ca n be easily disturbed and several

cases of

self-extinction

have been

observed. In

the

general

case,

where

fire

lasts,the

combustion

process

is low.

- 4 -

p


5/7

Fig. 7:

Diagram o f f l o w s between

the

gutterand th e

siphon

c)

Dysfunction

situations

The hydraulic Isolation divides th e

siphon

an d

more

generally th e

drainage System

into

tw o

distinct

volumes with

respect

to th e

gaseous

phase. As soon s

this

separating role is

altered

t he

System

enters

a

disfunction regime. Several

plausible

situations have been

tested:

If th e Isolation is composed

of

gasoline

its

leveldecreases when com-

bustion

is being

maintained in

th e upstream 1-2

compartment.

In

a

test

where

this

isolation

has

become

insufficient

to

preserve its separating

role,it has been observed that th eexplosible atmosphere developedin

the downstream3-4

compartment

was sucked

then burnt

by th e f lame,

withouta ny explosionphenomenon.

This

testw as conductedinsiphon

B

where

th ehead loss

inth ecollector is lo w (figure 5 ).

For

comparable tests conducted

in

siphon

A,

th e

head

loss in

th e

collector wa shigher, and it

has

been

demonstrated

that th e f lame had

been able to

extend

in the

counter-direction from

th e

upstream

1-2

compartment to

the downstream

3-4

compartment, following the free

space under

the

blade. Inthemost disadvantageous casean explosion

wa s

initialized

in volume

4.

In the other cases

this

compartment was

the

seat

o f

a well-marked

flash.

If

placed in

Situation

of double

dysfunction (no

hydraulic isolation

and

cover open on th e downstream3-4 side) th e siphon does not play

any

more its role of f irebreak protection towards aburning leakage. In

return

if

th e

hydraulic

Isolation is

absent,

but

th e downstream com-

partment

still airtight,

the absence

of standing f lame in th e down-

stream

part

of th e

facility

is

guaranteed,

although some ignition ha s

been

established during

the very f irst instants

during

th e tests. The

contest between the

speed

of

self-extinction and th e raise

of

th e

liquid

which

arrives into

th e

retention

pool

canexplainthat, before the

overflowing, th e f lame has not found th e conditions

requirecTfor exi-

sting

and has

extinguished.

In th e inverse

case

it

is not

irrealistic to

suppose

a

collector

fire.

The firebreak role

of

th e

siphon

is then limi-

ted to maintaining

conditions

unproper to combustion in the down-

stream compartment.

d) Layering o f

gasoline vapours

After spillage,

layering of

vapours can induce

a

rapid Variation in

explosibilityco nditions in

the

same volume and completelychange

the

assessment

of

risk

fo r

th e

same type

of

experience. Severa tests

show

that

this gravitationalmovement is relatively quick

in

th e

siphon,

although

it appears

necessaryt o

relate it

to

volatility

of

gasoline

which

Islikely

t o

vary

accordingto the product manufacture (this depends on

the periodin

the year).

e)

Effects of

explosion

In the mostviolent cases

of

explosionth e concreteplates coveringthe

Siphons

are slightlydisplaced

under

the effect of overpressures. The

cast-iron coversare thrown up and

a

short f lame appears.

In

th eother

situations

only the metalcovers

a re

ejected. When the phenomenon

is

limrted to a flash,

this

generally goes

with

a small Jump of covers

which

fallagain

in

their

seats

(dang).The airtightness of

th e siphon is

then

maintained.

F i g 8 : Combustion

process

in

the collector

4 .3 .3 .Co l lector

a

Fire

Th e experimental configuration selected

fo r

th e collector may seem

particular, since due to

its

small length

this

duct cannot

be considered

s representative

of

an actual facility (especially the aeraulic head

los-

ses

are

underevaluated).

In the absence

o f

inflammable vapours already

present in

the collector

th e fire can last there only under the form

of

a front flame which fol-

lows

th e

progression

ofa gasoline runnel.

Behind this

front

extinction

is

imposed

by the

concentration

o f

combustion

products (figure8 ).

The speed

of

this

displacement

depends

on the

speed

of the

gasoline

runnel at the collector bottom. The existence of the f lame therefore

depends on the

competition

between this speed, the collector cross

section and the production

of

burnt gases. Unbalance in

these para-

meterscan lead

to

self-extinction.

The inside

surtace of

th e collector

a PV C pipe

in the experimental

facility

warped on

th e

occasion

of

a temp orarily

stationary

fire.This

incident

resultedin the complete sealing of th e

duct

and the outage o f

th e whole drainage System, which can

hardly

be repaired. Obviously

sucha

materialmust

be avoided,

even

s alostformwork.

b)Explosion

This phenomenon

is

expressedbya deflagration

along

th e collector. It

supposes that

explosible

conditions first grow up,

s those generated

by the presence before the

ignition

o f

a gasoline

runnela t the collector

bottom. The demonstrated propagation speeds ar e higher than int he

previouscase

a nd

reach severalmetres p er second(figure 8).

Although the explosion regime never exceeded

that

of

deflagration,

the collector is an element in

which

the f lame front speed is

likely

to

speed

up,

not at t he scale

o f

a collector

section b ut rather

at the

scale

of a

complete facility.

This

risk

remains

at th e

state of

assumptions,

but

it shows that it

is

wishable to break

th e

continuity of

this

duct on

the total length

o f

th e

System.

p


6/7

5

Analysis

o f

the risks raised

by the

collecting System

Th e

previous

experimental results

allowed a comprehensive

analysis of

risks that may

raise

in

the

investigated

facility,

fo r

all

assumable scena-

rios.

It

appears

from

this

assessment

that a

very

high safety level is

brought in caseo f accidentalspillageo f

an

inflammableliquid:

The vertical slot gutter allows the very quick removal of the hydrocar-

bon sheet and therefore reduces

th e

potential

fire

to

a

gutter fire, often

little

violentand whichin

addition

does not

produce

much smoke.

Very quickly,

maximum

after

several

tens

metres, the hydrocarbon

flow,

burning

or not, reaches

t he

siphon.

In presence

ofa hydraulic Isolation

thesiphon proves to be an ess ential safety factor by stoppingth e pos-

sible

f lame, also by isolating the f low from the tunnel environment, s

the hydrocarbon

pursues

its way in the collector. In some cases

a little

violent

fire

characterized by

a

low combustion speed may settle down

in

th e

upstream part of the siphon. Its effects remain small during

several hoursand extinctionwith foarn

is easy.

Th e study

of

risks also considered the dysfunction situations of the

System,

such s

partially

damaged

siphon

covers,

or

th e

absence

of

hydraulic Isolation.

Most

of time

th e

effects

are

limited

in

their

gravity.

The most severe scenario occurs.in

case

of hydrocarbon

and

inflamm-

able vapours associated inth e

collector.

T he

starting

and propagation

of

an explosion in

this

part of

th e

facility can lead to

projected

covers and

flames comingout. These effects will then regardthe immediate environ-

ment

of

the Siphons,and their gravitywilldependo n

the

collector

length

affected by th e explosion development.

Such

dysfunctions

must

be

prevented

by adapted supervision and

main-

tenance, in

order

not to affect the safety

potential

provided by the

System.

6 Conclusion

Th e

experimental results summarized above ar e closely

related

to the

geometrypeculiar to the tested System, and

extrapolating

them to other

layouts requires

a

lot

o f

precautions.

Reflections

are

still necessary for

drawing

maximal advantage from the

experimentation

and optimizing

th e

drainage

Systems. Among

t he items

to be examined, on e

canmention

th e

design of

siphons

in

order to re-

duce their volume (nearly2

m

3

at present) without

altering

their

quali-

ties, their Installation

so

that they ca n

"break"

the way

of

a possible

explosion in

the

collector, th e

minimal inside diameter fo r the

slot

gut-

ters,

the design

of vertically

opened slots consistent with low

sidewalk

edges,

etc.

This

work

is

in process andwill

result

in guidelines

to

be applied to

Frenchtunnels. It already becomes obvious that these guidelines leadto

generalize the us e

of

slot gutters

and

Siphons

fo rall tunnels

of

a

certain

length

permitted

to

the transport

of dangerous substances. A special

care

must be brought to the maintenance of these Systems,which have

an essential role

in case

o f

severe

accident.

Summary

Tests

in

six road

tunnels have permitted to compare the hydraulic

effi-

ciency of several

Systems planned

to drain possible burning liquids

spilled

on

th e

pavement aftera n

accident.

Moreover an experimental, fll

size

facility

of

50 m in

length including

slot gutters,

gully

holes

with

siphons and

a

collector has

been

sub jected to twenty-five f ire

a nd

explo-

sion

tests. Lessons are drawn

on

the interest and safety of such

a

System.

Kurzfassungen

/A bstracts/

Resumes

Versuche

im

Mastab

1 : 1 mitA bleitsystemen

fr brennende

Flssigkeiten in

Straentunneln

Eine

zunehmende

Zahl

von franzsischen Strassentunneln genehmigen

de nTransit

v on gefhrlichen

Gtern. Diese

be treffen

hauptschlich

f ls-

sige Kohlenwasserstoffe in Tanklastwagen. Im Fall eines Unfalles mit

Leckage muss unbedingt da s

Abwassersystem

die rasche Ableitung

dieser

Flssigkeiten

erlauben, um die Entwicklung eines mglichen

Feuers

z u

vermindern

u nd

seine Ausbreitung

z u

vermeiden.

Verschiedene

Ableitungssysteme wurdenseit mehreren Jahren

in Frank-

reich in neueren Strassentunneln

installiert

und untersucht: Sie

sind

entweder klassische Wasserrinnen den Gehweg

entlang

mit Sinkksten

in

geringem Abstand, oder

kontinuierliche Schlitzrinnen,

die vo n den

Sammlern

durch

Siphons getrennt

sind, um

die

bertragung

vo n

Brnden

zu verhindern. Bevor solche Systeme zu verallgemeinern

sind,

schien

es

notwendig, die

Wirksamkeit

jeder

Lsung experimentell

zu

vergleichen und

die

Feuer-und

Explosionsrisiken,

die von der

Einrich-

tung ausgehen konnten,z u analysieren.

Eine erste Versuchsreihe betraf

die

hydraulische Wirksamkeit

der

Sammlungssysteme. In sechs betriebenen Strassentunneln wurden

Ver-

suche mit einer

kontinuierlichen

Strmung von reduzierter Leistung

(20

bis 30 l/s) und raschen Freigaben vo n5

und

10

m

3

durchgefhrt.

Wasser

wurde verwendet, um d en gefhrlichen

Stoff zu simulieren.

Ein zweiter Versuchstypbetraf eine e xperimentelle,

mehr

als 50 m lange

Einrichtung in natrlicher Grosse,

die

mit

einer

kontinuierlichen Schlitz-

rinne, Sinkkasten und einem Sammler versehen wurde.

Bleifreier

Superkraftstoff

wurde verwendet, um die verschiedenen

Feuer-

und

Explosions-Ausbreitungsphnomene

in

der

Einrichtung zu

analysieren.

ZahlreicheVersuche

wurden durchgefhrt,

u.a.

auch unter

Bedingungen,

fr welche das

Ableitungssystem

fehlerhaft war. Die

Ergebnisse

wurden

verwendet, um d ie

Risikend erEinrichtung

zu

analysieren,und daher

d ie

Betrachtung aller mglichen

Betriebsflle

de s Ableitungssystems zu

erlauben.

Der Vortrag stellt die Resultate dieser Untersuchungen vor

und die

Schlufolgerungen, die

fr

Tunnelprojekte ber die

Rinnengestaltung,

die Knienutzbarkeit

und

-gestaltung,die Sammler, Einrichtungsunterhal-

tung, usw. gezogen werden

knnen.

- 6-

p


7/7

Fll

Size Testing o f Drainage

Systems for B urning Liquids

in

RoadTunnels

Th etransit of dangerous goods is permitted in an increasing

number

of

French roadtunnels.

Most

o f

these

substances a re liquid hydrocarbons in

bulk.In case

o fspillage,the

drainage System mustdischarge

these liquids

very

quickly,

in

orderto reduce the

severity

of

a

possible

fire and

prevent

its

propagation.

This ist he

reason

wh y various drainage Systems

have been studied

fo r

several

years

an dapplied in recently

built

tunnels in France:

they

ar e

either

co nventional water gutters along t he sidewalk,provided with gully

holes at short inten/als,

or continuous

slot gullies separated from th e

collectors by Siphons aimed

at extinguishing

th eburning liquids. Before

generalizing suchSystems

it

appearednecessary to

make

an experimen-

tal comparison of the

efficiency of

the various solutions and to

examine

the

fire

an d

explosion risks thatmay be generated

bythe

facility.

A first testing

Programme

concerned the hydraulic efficiency of

the

drainageSystems. In six roadtunnels underOperation, tests have been

performed

with a continuouslimitedflow (from 20

to

30 l/s)a nd

asudden

release

of

5 and

10 m

3

. Water was used

to

simulate the dangerous

goods.

A second series

of tests was conducted

on

an

experimental fll size

facility exceeding 50

m

in

length,

provided with

a

continuous

slot

gully,

gully

holes

with Siphons and collector. Unleaded super

fuelwa s

used

t o

study the

various

phenomena of firepropagation and

explosion

risk in

th e

facility. Numerous

situationswere tested, including

m alfunctioning

of

th e

drainage

System. The gathered data have been used to analyze

th e

riskso f

the

facility,allowingth e consideration o fall

possible scenarios

for

operating of the drainage System.

The

paper

presents

th e

results

ofall these studies and draws conclusi-

on s

fo r tunnel projects, regarding

th e

design of

gullies,

the

usefulness

and design of

Siphons, the

collectors, the maintenance

of the

facility,

etc.

Essais

en vraie

grandeurde

systemes de

recueildes liquides

enflamm es

en tunnel

routier

Un nombre croissant

de

tunnels

routiers

francais est autorise au pas-

sage des

matieres

dangereuses. Celles-ci sont

majoritairement

constituees d'hydrocarbures

liquides

en citerne. En cas de deversement

accidentel,il

es t

indispensable

qu e

la Systeme

d'assainissement puisse

evacuer tres rapidement

ces

liquides,

afin

de reduire l'importanced'un

eventuelincendiee t d'eviters propagation.

C'est

pourquoi

depuis plusieurs

annees

differents

systemes d'evacua-

tion

ont

ete etudiees et mis en

oeuvre dans

les tunnels recents en

France:

l

s'agit

soit

de fils d'eau classiques en bordure de trottoir,com-

portant des regards avaloirs intervalles rapproches, soit de caniveaux

fente continue, separes des collecteurs par

des Siphons

destines

eteindre les matieres enflammees.

Avant

de generaliser

de tels sy-

stemes,

il

a

paru

necessaire de

comparer

experimentalement

l'efficacit

de chaque

solution

et d'examiner

les

risques d'incendie

et

d'explosion

qu e

l'installation

pouvait

engendrer.

Une premiere campagned'essaisa concerne l'efficacite

h ydraulique

des

systemes

de

recueil. Dans

six

tunnels

routiers en

exploitation,

des

essais ont ete

realises

avec un ecoulementcontinude debit limite (de 20

30

l/s),

et avec des relchements brutaux de

5

et

10

m

3

.

De

l'eau a

ete

utiliseepoursimuler la

matiere dangereuse.

Un

deuxieme

typed'essais

a

ete realise

sur une Installation experimen-

tale en vraie grandeur de plus de 50

m

de longueur, comportant

caniveau

fentecontinue, regards

munis

de Siphons e t

collecteur. C'est

du supercarburant

sans plomb qui

a

ete utilise pour

examiner

les

differents

phenomenes de propagation d'incendie et de risque d'ex-

plosion

dans

l'installation. De nombreuses situations

ont

ete testees,

y

compris des conditions

o

le Systeme d'evacuation

presentait

des

defauts.

Les

donnees

ainsi

obtenues

o nt

e te

utilisees

pour

faireune ana-

lyse des

risques

de

l'installation,permettant

d'envisager tous

les

scena-

rios

possibles

de

fonctionnementdu

Systeme d'evacuation.

La

communicationpresente

les resultats

de l'ensemble de ces

etudes

et

les enseignementsqu'on peut en tirer pour les

projets

de tunnels quant

la

conception des caniveaux, l'utilite et

la

conception des Siphons, les

collecteurs,

l'entretien

de

l'installation

etc.

. 7-

p

full size testing

Documents