Note: Descriptions are shown in the official language in which they were submitted.
2~ 95
, ;
D A N F O S S A t S , D K -6 34 O N O R D B O R G
Method and apparatus for monitoring a conduit system ror an
incompressible fluid for leaks
The inventlon relates to a method of monitoring a conduit
system for an incompressible fluid for leaks, wherein a testing
fluid is lntroduced into the conduit system during a testing
period when no fluid is being withdrawn from the condult
system and the conduit sy~tem is closed on the supply side by
a maln valve. The lnventlon also relates to an apparatus for
monitorlng a condult system for an lncompressible fluid,
comprisine a maln valve on the supply side of the conduit
system and control apparatus for controlling actuatlon Or the
main valve.
Fluid condult systems have to be monltored for untlght ~oints
and leaks. This applies fundamentally to all conduit systems,
regardless Or whether they are employed to convey mains water
into a house, heating liquids in heating and remote heating
systems or ga3es or fuels in distribution clrcuits.
The monltorlng Or mains water ln bulldlngs has particularly
lncreased in lmportance in recent years. The problem will be
explalned by uslng the example Or a malns water installation
in a residentlal bullding. Normally, the consumption Or
water through withdrawal by a consumer rrom a water tap amounts
to between about 50 and 1,500 l/h. In extreme cases, for
. -
2~
: ` :
:~'
-2-
~ ':
example toiletcisterns or washing machines, the consumption
could be between 30 and 2,500 l/h. Leakages accounted for by
a pipe fracture or by bursting of a supply hose for a washing
machine or dishwasher (large leak) are typically in the range
of 500 to 2,500 l/h, sometimes higher, and can therefore not
be distinguished from normal consumption. For this reason,
wlth such large volumes above a predetermined value the 3upply ~-
of water ls interrupted after a predetermined withdrawal
perlod regardles3 of whether a consumer is using the water or
there is a large leak.
In contrast, there are faults hereinafter referred to as a
"small leak". In this case, the 1099 of water i~3 in the
range Or about 1 to 25 l/h and could be caused by dripping
water taps and overflowing toilet clsterns on the one hand and
untlght pipe connections, commencement of fatigue failure in
pipes occasioned by corroslon, halrllne fractures ln pipes and
vessels or slmllar damage ln the condult system on the other
hand. Whereas the flrst set of examples may not be dangerous
but only lncrease the costs of fresh water and dralnage and
place a demand on drinklng water sources and thus on the
;; environment, ~mall leaks of the second kind could cause con-
siderable damage. More particularly, the outflowing quantity
,. ~
- . . . ~ , , , ' , . :, .
: : , : ' : .: ' ' , ' : ' ' ' . ' ' ' ' ' . . ' ,
' '
of 1 to 25 l/h may appear very small but, over a prolonged
period, walls or other parts of the building can become ir-
repairable as a result Or being saturated with moisture. The
resulting damage is often noticed too late because the dampness
starts inside a wall and does not become visible until the
whole wall is saturated. If one were to discover such a
small leak at an early stage, it could be repaired in time.
To check a central heating system for leakages, W0 ~7/04520
discloses an arrangement consisting of two flow meters Or the
vane-wheel type in the supply and return conduits Or the
system. Both vane-wheels determine the total volume flowing
through the heating system. If there ls no leak, the two
volumes must be the same. If there is a difference between
the two volumes, a leak is suspected and the circuit is shut
down by way Or a motorised valve. However, since the volumet-
ric flow meters are provided for the main flow, that i9 to say
ror large volumes, they are unable to detect small leakages
below, say, 25 l/h with the required degree Or accuracy.
W0 86/06457 disclo~es equipment for monitoring pressure con-
duits for leakage polnts, thls equipment measuring the pressure
263C8495:
--4--
in the condult system downstream of the main valve and shutting
the main valve down either when a large amount Or fluid rlows
through the main valve for a prolonged perlod or, if the main
valve is closed, the period required by the pressure to drop
from a first pressure to a second pressure is shorter than a
permitted time interval. However, since the pressure on the
supply side Or the main valve varies considerably, for example
pressure rluctuations at the waterworks that might be in the
order Or 1.2 bar or as a result Or sudden consumption in an
ad~oining conduit system where the pressure could drop by
about 0.6 bar and, on termination Or this consumption, rise
about 0.4 bar above the normal waterworks pressure and because
Or the pressure drop across the main valve as a result Or the
consumption in the conduit system being monitored, only unsat-
isfactory results can be achieved with the pressure measurement
disclased in WO 86/06457.
In a known apparatus ror monitoring installatlons ror tightness
(DE-OS 21 58 901), when no fluid is being wlthdrawn, a leak is
detected by withdrawing compressible fluid from the source,
for example the supply mains, upstream Or the main valve,
compressing lt wlth a compressor, and feedlng lt lnto the
conduit system downstream of the main valve. Arter reaching
2~G~ 95
a test pressure, the compressor is shut Orr. One now contols
whether the pressure loss does not exceed a predetermined
value within a predetermined time. In a difrerent embodiment,
one checks whether the compressor can build up the required
testing pressure during a predetermined runnlng time. Since
the volume conveyed by the compressor alway~ depends on the
pressure difference between the inlet and outlet Or the com-
pre3sor, it is practically impossible to come to any conclusion
about the conveyed quantity if both of the pressures are not
also monitored. The known equipment is therefore only suit-
able ror determining whether there is a leak. One cannot say
how large this leak might be. In addition, a separate drive
i8 required for the compressor and this might lead to undesir-
able noise. After detecting a leak, the main valve is locked
in the closed position but fluid can neverthele!ss penetrate
into the conduit system by way Or the compressor and continue
to escape through the leak.
.
; It ls the problem Or the lnvention to provide a method and
apparatu~ wlth whlch even small leaks can be reliably detected.
; This problem is solved in a method Or the aforementioned kind
ln that, without the entry Or testlng fluld from the supply
slde Or the maln valve, a predetermlned test volume Or the
testing fluid is introduced into the conduit system under
2~1G8~95
pressure and the time required by the test volume to flow into
the conduit system is measured.
During testing, exactly the same amount of fluid leaves through
the leakage point as during normal operation. Since this
escaping amount Or fluid is immediately replenished, the
volume Or leakage flow can be measured exactly lf one assumes
that it does not fluctuate appreciably over a period of tlme.
It is simply detected by dividing the known test volume, i.e.
the replenished quantity, by the measured time. In addition,
a constant pressure is maintained through replenishment Or the
testing fluid in the conduit system to be monitored, whereby a
sufricient amount of fluid is immediately available upon
commencement of consumption.
In one embodiment of the method, the pressure with which the
test fluid is introduced into the conduit system i9 Or the
same order as the fluld pressure on the supply slde Or the
main valve. In a known apparatus, an increase in pressure is
required which might increase the leak caused by a weakness in
the material. To detect a leak, no excess pressure i9 there-
fore applied but testing rluid is introduced into the conduit
system under normal pressure. Since the normal pressure,
;~ . . ~ . ~. I . . . . . . . ... ..
.~. .,, - - ~ ~, . . - . :. .
~ ! . , . . . . . . ..
~' ' ' ' . . " ' , ' ~ ' , ' ', ' ~, . ', ' ': .
2~3~8~95
-7-
i.e. the supply pressure Or the mains, for example the water
mains of a clty, i9 applied to the leakage points when the
main valve is open, replenishment of the testing volume under
the same pressure can practically reproduce the escape of
leakage fluid under normal conditions. Since excessive
pressure is avoided, the conduit system to be monitored is
not stressed any more intensively than during normal opera-
tion.
Preferably, the testing fluid is withdrawn from the conduit
system berore the testing period. For testing purposes,
therefore, exactly the same fluid is employed as that which
is normally distributed by the conduit system to be monitored.
No special testing fluid has to be provided and this makes
the method considerably cheaper. Accordingly, all of the
testing fluid that is employed has also already passed through
the main valve and any upstream metering clock- 90 that no
difficultles are encountered for example when accounting for
the mains water consumptlon to the waterworks. Also, no
additional rilters or llke equipment have to be provided for
the testlng fluid. Since the withdrawal of the testing
fluid takes place immediately prior to the testing period, it
is also practically impossible ror any error to occur because
Or the time difference between withdrawal of the testing
2~R~9s
, :
rluid and the testlng period.
Advantageously, the test volume is less than 0.5 1. Even
the largest test volume of 500 ccm is still relatlvely small;
it only fills a cylinder of about 8 cm diameter and 10 cm
high. This reduces the construction costs and very conslder-
ably reduces the space taken up by the testing apparatus.
The smaller the test volume, the higher wlll be the time
resolution.
Prererably, arter the complete introduction of the testing
fluid, further testing fluid with the test volume is withdrawn
rorm the conduit syste~ and held ready for renewed Lntroduc-
tion into the conduit system. This makes continuous measure-
ment Or the leakage volume possible. The time taken ror the
leakage flow can thus be more efriciently monitored.
Preferably, the entire volume Or the lntroduced testine rluid
is determined by adding the lndivldual volumes that were
introduced. Thls not only gives an indication Or the indl-
vidual volume Or leakage rlOw but also determines the quantity
the total discharged fluld as an additlonal criterlon for
consideratlon.
It ls also advantageous ror the leakage flow to be determined
., .. -.-, . . -
.: ; - ,
X~8~95
continuously. This enables rapid recognitlon Or a change in
the leakage behaviour of the conduit system being monitored
and hence suitable protective or countermeasures can be taken
in good time.
Preferably, an alarm signal is produced when the entire
volume exceeds a predetermined fir3t value and/or when the
flow of leakage volume lies between a rirst and a second
value.
The alarm may be an optical or acoustic signal. To release
the alarm, two criterla are therefore available, namely the
entire volume that has escaped at the leakage point on the
one hand and the actual leakage flow on the other hancl. Ir
the actual leakage flow lies below a certain limlt, for
example 1 l/h, the system is declared to be leakproof. With
a flow Or leakage volume between, for example, 1 l/h and 3
l/h, a small leak i9 assumed whlch, although lt has to be
monltored, will not cause much damage. With a flow Or
leakage volume between for example 3 l/h and 20 l/h, one
assumes a large small leak which could glve rise to serlous
dama~e. A small small leak can, for example, be indicated
immediately. However, lt could al90 be indicated only when
the amount discharged through the small leak has exceeded a
.. . . . ~
;~ 495
-10-
predetermined first value.
Preferably, the supply of fluid to the conduit system i9
completely interrupted when the total volume exceeds a prede-
termined second value. Irrespective Or the size Or the
leak, an escaped amount Or water can present a grave danger
to the bullding and therefore lt ls better to close the maln
valve completely in order to avoid further damage. Natur-
ally, shuttlng down can also be made dependent on the actual
flow Or leakage volume.
In a preferred embodiment, the total volume is returned to
zero when the flow of leakage volume has been reduced by a
predetermined extent. For example, it can happen that the
leak is caused by a dripping water tap which the user failed
to close completely. When the user notices his mistake and
closes the water tap, the leak will also disappear. In this
case, lt is senslble to correct the total volume that was
assumed to escape lnto the wall from a faulty polnt in the
condult system. Thls wlll then enable one to work wlth
realistlc parameters during the next testlng period.
.
In an apparatus Or the aforementioned kind, the problem is
solved ln that the conduit system communicates with a chamber
: .. ~ . . - ,.
~........ ' .-- , , .: . :
-, ~ ,: :
,, . :
X~(~89~95
--11--
for introducing a testlng fluid with a predetermined test
volume, the volume of the chamber being variable between a
predetermined first larger value and a predetermined second
smaller value. The chamber communicates only with the
conduit system and time measuring means are provided for
measuring the time during which the chamber reduces in volume
from the first value to the second value.
Thus, the chamber serves as a store for testing fluid with-
drawn from the conduit system. Since the chamber can a3sume
two termlnal conditions, namely one with a larger volume and
one with a smaller volume, during the time between these two
conditlons it must be the exact dlrference ln volume that has
flowed out of the chamber lnto the conduit system or out of
the condult system lnto the chamber. Slnce the chamber
communicates only wlth the conduit system and not wlth the
supply side of the main valve, everything flowlng out of or
into the chamber must also flow through the conduit system.
Since flowing out of the chamber for testing purposes only
takes place with the main valve closed and without any pres3-
ure increase in the conduit system, the chamber introduces
exactly aq much fluid into the conduit system as escapes from
the conduit system through a leakage polnt. The tlme measur-
ing means measure the time required by the testlng fluld to
2~ 95
-12-
rlow into the condult system. In other words, they measure
the time required by a certain volume to flow out Or the
conduit system through the leakage point. This enables one
to obtain an indication about the actual rlow Or leakage
volume ir one assumes that this rlOw i9 not sub~ected to
marked varlations in time.
Preferably, the chamber is closed on one side by a movable
wall. The chamber is sealed at all sides with the exception
Or the aperture to the conduit system, the volume being
changeable by the wall. The volume therefore changes linear-
ly with the displacement Or the wall which makes the evalua-
tion simple.
Preferably, the wall i9 movable towards the smaller value Or
volume against the rorce Or a spring. If the same pressures
obtain on both sides Or the wall, the sprlng wlll move the
wall so that the chamber wlll assume lts largest volume.
The sprlng thererore asslsts resetting.
Prererably, on the slde remote rrom the chamber, the movable
wall is subJected to a force which ls constant throughout the
path Or movement. Regardless Or the dlsplaced dlstance,
thererore, always the qame pressure will act on the wall and
thus on the chamber ir one disregards the counter-force Or
the spring which becomes more intensively compressed as the
~ :
: :
2~3~8'~95
-
di3placement increases. However, since the spring is rela-
tively weak ln relatlon to the force actlng on the slde of
the wall remote from the chamber, the change ln the counter-
force Or the sprlng may be disregarded.
With particular advantage, the side Or the wall remote from
the chamber communicates with the supply side Or the maln
valve. Thus, the supply pressure Or the source, for example
Or the mains water from the waterworks, acts on the chamber
wlthout establlshing communlcation between the source and the
conduit system that ls belng monitored, l.e. wlthout enabling
fluid to enter the conduit system by bypassing the main
valve. In addltion, no auxiliary energy is required.
Instead, an available pressure 1~ utilised. The pres~ure in
the conduit system being monitored can become no higher than
the supply pressure from the source and this avoids excessive
stressing Or the monitored conduit system during monitoring.
In a prererred embodiment, a sensor ls provlded which trans-
mits a slgnai to the control apparatus when the volume Or the
; chamber has reached the smaller value. Thls termlnal po31-
tion ls, for example, requlred for the tlme measurement.
It is also preferred that the control apparatus will open the
main valve in respon~e to this signal. When the chamber
2 ~ ~ 8 ~ 9 5
-14-
volume has reached its smaller value, a pressure drop mu~t
have occured ln the conduit sy3tem. This pressure drop may
be caused by consumption or by a leak. For consumption, the
maln valve must open so that the user can withdraw fluid rrom
the conduit system. In the case of a leak, monitoring has
to occur.
~or monitoring a large leak, the time element preferably
produces a command to close the main valve a predetermined
time after the main valve was opened. Since the leakage
monitoring system is unable to differentiate between consump-
tion and a large leak, this measure ensures that only a
maximum amount of fluid can leave the conduit system. A
consumer who wlshes to withdraw more fluid can glve prlor
lndlcation of this to the control apparatu~ or he wlll inter-
rupt the consumption momentarily to make it known to the
control apparatus that there is no large leak.
In a further preferred embodiment, the control apparatus
compriseq an lntegrator whlch lntegrates the volumes Or fluld
red into the conduit system from the chamber. Thls enables
a value to be avallable at all tlmes that indlcates the
amount Or leakage flow that has escaped up to that time.
. ' . ~ ':
. . - ~ ,
8~95
-15-
It is also advantageous if the control apparatu~ locks the
main valve in the closed position when the integrator has
determined a total volume which lies above a predetermined
value and/or the rlow of volume exceeds a predetermined
value. When the flow volume exceeds a predetermined value
there will, as prevlously explained, be no fear Or extensive
damage even ln the case Or a small leak. Another criterion,
which could also be combined with the fir3t criterion, i3 the
fact that a certain amount Or leakage fluid has escaped
altogether. This amount can be adapted to sult the condi-
tions. Upon exceeding thi3 predetermined leakage rlOw~
however, the main valve should be closed to avold extensive
damage.
.
An example Or the invention will now be described in conJunc-
tion with the drawing which is a diagram~atic representation
Or an apparatus ror monitoring a conduit system ror an incom-
pressible fluid ror leaks.
A conduit system 8, ror example for mains water in a residen-
tial building, i9 red by way Or a main valve 1 from a source
7, for example the mains water from a waterwork3. The main
valve 1 18 remote controlled by an actuating apparatus 5
whlch is operated by a control apparatus 6. When the main ~ ~.
.. ' . ~ ' . '
.. . . . . .
~8~95
.
-16-
valve l i9 closed, no water can reach the conduit sy3tem 8
from the source 7. Parallel to the maln valve 1 there is
the actual leakage monitoring equipment. This consists Or a
cylinder 2 which ls divided by a movable wall 4 lnto a pre3s-
ure chamber 11 and a chamber 12. The pressure chamber 11
communicates with the 3upply side of the main valve l. The
chamber 12 communicates with the discharge side Or the main
valve 1, l.e. with the conduit system 8. The wall 4 seals
the chamber 12 from the pressure chamber 11.
The wall 4 is movable in the cylinder 2 90 that the volume Or : ;
the chamber l2 is variable between a larger value at which
the wall 4 abuts the left-hand end of the cylinder Z and a
smaller value at whlch the wall 4 abuts the right-hand end of
the cylinder 2. The wall 4 is pressed towards the left-hand
end of the cylinder 2 by the force Or a spring 10.
It will now be assumed that the main valve 1 13 open wlthout
water being wlthdrawn from the condult system through a tap9.
No water therefore flows through the maln valve 1 and there
wlll be no pressure drop. The pres~e Pl on the supply side of the
maln valve and equal to the pressure of the source 7 ls
therefore equal to the pressure P2 on the discharge side of
the main valve, i.e. equal to the pressure in the conduit
~, .
.
::
-
.
8~L95
:~
system 8. The pressure P1 also obtains ln the pressurechamber 11 whilst the pressure P2 obtains in the chamber 12.
Accordingly, the same pressures act on both sides of the wall
4. However, since the wall 4 is additionally Rub~ected to
the force of the spring 10 on the side facing the chamber 12,
the wall 4 will be displaced towards the left-hand end Or the
cylinder 2. At the rlght-hand end of the cylinder 2, there
ls a sensor 14 whlch i9 actlvated by a generator 13 ln the
wall 4 when the wall 4 is at its right-hand terminal position,
l.e. when the chamber 12 has assumed its smallest volume.
When the wall 4 ls pushed towards the left under the force Or
the sprlng 10, thls 19 detected by the sensor and notlfied to
the control apparatus 6. A time element now runctlons in
ths control apparatus 6 and, after a predetermined time,
signals the valve actuator 5 to close the valve. If no
fluld 19 being withdrawn from the conduit system 8, the
pressure will remaln constant there, i.e. the wall 4 remalns
in lts left-hand termlnal posltlon.
.. ..
However, lr a small leak occurs, rluid wlll trickle rrom the
condult system 8 to the exterior thereby gradually reducing
the pressure P2 in the conduit system. Slnce the wall 4 i9
subJected to the pressure P1 of the source obtaining in the
2~8'~9~
-18-
pressure chamber 11, lt will wander to the right, whereby the
test volume Or fluid located in the chamber 12 i8 replenlshed
to the condult system. After a certain tlme, which is
measured by the time element 16, the wall 4 will reach its
right-hand terminal position, which is detected by the sensor
14 which may be in the form of a reed relay. Since the test
volume is known, the test volume and the time required by the
test volume to rlOw into the condult system 8 will enable one
to calculate the flow of volume, i.e. the volume per unit
time, that has escaped the conduit system ô through the
leakage point. Since the te3t volume enters the conduit
system 8 without elevated pressure, no higher pressure loads
will occur ln the condult system 8 than lf the maln valve 1
were to open and allow the pressure from the source 7 to pass
dlrectly into the conduit system 8.
When the sensor 14 has recorded the ract that the wall 4 ls
in lts rlght-hand terminal position, the control apparatus 6
will give a signal to the valve actuator 5 ror the main valve
1 to open agaln. The wall 4 will now be displaced to lts
left-hand terminal position agaln ln the manner described
above and the testlng cycle wlll start afresh.
The control apparatus 6 comprlses an lntegrator 15 whlch
summates the number Or cycles Or the wall 4 and, since the
' ' . ' ~
B~95
-19-
test volume is known, thereby enables an indlcation to be
obtained about how much rluid has escaped through the leak
altogether.
Slnce the test volume is always again introduced into the
conduit system it is possible to obtain a continuous indica-
tion about the actual flow of leakage volume. In addition,
there i8 an indication of the leakage flow that has already
escaped so that, with the aid Or these two leakage loss
criteria, an indicator can be reliably actuated and/or the
main valve 1 can be closed. For example, an indicator is
actuated when the leakage volume exceeds a first predetermined
value, e.g. 1 l/h. When the leakage volume exceeds the
predetermined first value, e.g. 60 1, and the escaped leakage
volume exceeds a predetermined first value, an indication can
likewise be given and the integrator 15 is returned to zero
again. Naturally, the number of times for which-the integra-
tor is reset to zero can be llmlted so as to prevent an
excessively large amount Or leakage fluld to escape through
the leak. For example, one can ensure that on the third
occa~ion the integrator 15 is not reset to zero but the main
valve 1 19 locked in the closed condition.
If the leakage volume is larger than a predetermined second
.. ,
-20-
value, e.g. 3 l/h, the integrator is not reset to zero when
reachlng the predetermined first leakage value but only an
indicator i9 actuated. Integration, i.e. summation of the
lndividual test volumes, is continued. If the lntegrator 5
finds that an amount of leakage fluid has escaped that is
larger than a predetermined second test volume, e.g. 180 1,
the control apparatus 6 likewise locks the main valve l in
the closed position. In addition, the main valve can like-
wise be locked in the closed position when the leakage volume
exceeds a predetermined second value. Preferably, however,
the closing crlterion will also be made dependent on the
previously escaped leakage volume, i.e. the amount of leakage
rluid that has escaped.
The individual values for the flows of leakage volume can, as
mentioned, for example be 1 l/h for the first value and 3 l/h
for the second value. Below a value Or l l/h; the conduit
system is oonsldered to be leakproor. Above 3 l/h, one
derines a large small leak with which only a certaln amount
Or fluid can pass berore the main valve l is closed.
If a consumer wishes to withdraw water at the point 9 he may,
for example, turn a water tap 9, whereby the pressure P2 will
suddenly drop in the conduit system. The wall 4 is ~ery
~'~ '', ~ - '
.
.
,
::
X~38~5
-21-
rapidly pushed to the right-hand terminal wall Or the cylinder
1 under the pressure P1, whereupon the main valve 1 opens.
The water can now flow into the conduit system 8 from the
source 7. The same will take place if there is a large
leak, for example if a pipe breaks or there i9 a burst in the
supply hose for a washing machine or dishwasher. To prevent
too much water from escaping in ~uch a case, the time element
16 will close the main valve 1 agaln a predetermined time
after the pressure drop. This time is, for example, suffi-
cient for filllng a bath or having a generous shower, e.g. 15
minutes. or course there are also cases in which the consu-
mer will want to withdraw water throughout a longer period,
e.g. to wash his car or water the garden. In this case, he
can signal this to the control apparatus 6, for example by
actuating a switch whereby the apparatus will fix the maximum
withdrawal time ror the next consumption to, say, two hours.
For all subsequent consumer activities, however, the original
time Or, say, l5 minute~ will apply. Another posslbillty 19
ror the tlme element 16 to transmit an acoustical optical
slgnal Just berore expiry Or the predetermined period, where-
upon the consumer can close the tapping point 9 momentarily.
The pressure P2 will thereupon rise to move the wall 4 to the
lert again. At the instant when the control apparatus 6
detects that the wall has left its rlght-hand terminal posl-
: . ~. . : : ' ~ . : .
263~8~95
~ . .
tion, i.e. the volume Or the chamber 12 has increased again,
the maximum tapping time can start afresh. Such a pressure
ri~e would be most unlikely in the case Or a large leak.
One therefore ensures that damage caused by a large leak will
likewise be reliably kept relatively small.
~. ~. ~ ' '.',. '' :
.. . .: . , : ,, . .~ .: ,.
:: :,
