Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
106~Z~z
The invention relates to a method and apparatus for detec-
ting an interface of materials having different hydrogen content, present in a
metal vessel or pipe, e.g. made of steel.
Interfaces as indicated above occur frequently in the pro-
cess industry. Examples are levels of liquid hydrocarbons in settlers, in
absorption columns, in distillation columns, interfaces in pipelines for gas
transport if two-phase flow occurs, etc. Walls of columns, reactors, pipe-
lines and the like are usually of steel and it has always been of importance
to have available detection methods which do not require special constructions
on the steel walls such as sight glasses, lead-in wires for measuring equip-
ment, etc.
It is known to use the gamma-ray absorption method for de-
tecting interfaces. This method, however, has a number of disadvantages.
G = a rays are absorbed far more strongly by steel than by hydrocarbons, so
this method could only cope with large quantities of hydrocarbons behind steel
walls. Furthermore, the gamma-ray source and the detector have to be placed
in opposite positions on the container. Larga units will require large gamma
sources the handling of which is cumbersome for safety reasons. To align a
gamma source and detector in opposite positions a special construction along
the container is needed. In addition it is often difficult to find a free
optical path for the gamma rays if stirrers, baffles, trays etc. are present.
The invention provides apparatus which does not suffer from these dis-
advantages.
According to the present invention, there is provided
apparatus for detecting an interface of materials having different hydrogen
content, present in an enclosed room, provided with at least one neutron source
and at least one neutron detector, which is situated at a certain distance
from said neutron source, wherein the enclosed room is a metal walled vessel
or pipe to be used in process industry, and wherein the neutron source which
consists of californium-252 and the detector are located near or at the outer-
side of the metal wall of the vessel or pipe, said distance between source
and detector not being larger than 50 cm, and the detector(s) having a larger
sensitivity for scattered neutrons than for neutrons emitted by the source.
~ -2-
i7Z~Z
The present invention provides a method for detecting an interface of materials
having different hydrogen content, present in a steel vessel or pipe, in which
near or at the outer side of the
-2a-
1~67Z~Z
steel wall are present at least one neutron source and at least one
neutron detector, the distance between source and detector not
being larger than 50 cm, the detector having a larger sensitivity
scattered neutrons than for neutrons emitted by the source, in which
the neutron source consists of californium-252.
This method is based on the fact that neutrons are transmitted
through layers of steel and other heavy metals about as easily
as light through a glass window, but are strongly scattered by
the light hydroeen nuclei abundantly present in hydrocarbons or
in water. When a suitable source and a detector that picks up the
scattered neutrons are placed in appropriate locations outside
the vessel amounts of hydrocarbons or water in the vessel as small
as a few grams can thus be traced from behind a five centimetre
steel wall and this is sufficient to detect the pres¢nce of an
interface. Such an interface may be liquid hydrocarbons against
their vapours or air or other gases, water against its vapour or
air or other gases, the level of pulverulant solid organic material
containing hydrogen. The method does not detect an interface
between liquid hydrocarbons and water, because of the comparable
hydrogen content of these liquids.
Californium-252 is a man-made transuranium isotope with a half
value time of 2.7 year. This nuclide produces neutrons by spontaneous
fission with a yield of 2.3x10 neutrons per second per microgram.
A source containing about 0.1-1 microgram produces sufficient neutrons
for most technical applications. The dose rate from al~g-source having
an activity of about 0.5 mCi at 1 m distance in air is within the
limit of 2.5 millirem per hour, permit-ted for daily exposure of the
1067~1Z
radiological worker. This makes safe handling rather simple in
comparison with most gamma-radiation sources generally used in the
process industry. Most other neutron sources have much lower neutron
yields and therefore sources of much larger activities are needed
than Cf-252 which make the applications in process industries of
other type neutron sources like Am-241-Be less attractive.
In a counting tube filled with helium-3 at a pressure of about
10 bar an uncharged neutron is conver-ted by a reaction with He-3
into a tritium nucleus and a proton with a total kinetic energy of
764 keV. The charged particles produce an electron avalanche between
the wire-shaped anode (at a potential of 2000 V) and the cathode or
wall of the tube. The avalanches produce current pulses which are
counted by appropriate equipment. Such a detector is very sensitive
for scattered neutrons which owing to the collisions with hydrogen
they made have energies somewhere between the thermal energy distri-
bution and that of fast neutrons from the source. The detector
has a lower sensitivity for fast neutrons. It is therefore possible
for the detector to be in the neighbourhood of the source. The detector
receives scattered neutrons and a much larger number of fast neutrons
from the source, which cannot be shielded from radiation directly to
the detector. Owing to the properties of the detector as indicated
above this creates no problem.
The intensity of the scattered neutrons decreases by increasing
distance from the source and it is therefore of importance that
the distance between source and detector is not larger than 50
cm and preferably not larger than 25 cm. This improves the sensitivity
of the detection. For measurements on vessels with large diameters
use is made mainly of returning neutrons by positioning the source
and the detector close to each other one side of the vessel. This
avoids special constructions on opposite sides of the vessel as
-- 4 --
1~67Z12
i9 required for gamma-ray absorption with the accompanying problems
of alignment.
An attractive possibility is to locate source and detector
in such a way that the centre of gravity of both are in a plane
which is perpendicular to the centre line of the vessel or pipe.
This may be achieved by positioning source and detector along a
circle line against the outside of the steel wall of a vessel.
The two components may then be combined in one probe which can
be moved alongthatwall. Another possibility is to position the
source and the detector behind each other, the source being nearest
to the wall. Both geometries have the advantage of sharp detec-tion
of levels in a vessel or column and the combination of source and
detector in one probe promotes ease of handling.
Notwithstanding that the dose rate of a Californium-252 source
i5 already very low, it may be further decreased by means of a
shielding.
The apparatus suitable for use in the method according to the
invention therefore comprises a neutron source of Californium-252,
a detector having a larger sensitivity for scattered neutrons than
for neutrons emitted by the source? a shielding enveloping said
detector and source at least partly, which shielding consists of a
reflector~ a moderator and a layer of material between said reflector
and moderator, which prevents substantially the passage of slow
neutrons.
The method and apparatus according to the invention is very
suitable for the detection of liquid water and/or liquid hydrocarbons
present in pipelines for gas transport. The presence of liquid in a
gas transport pipeline should be avoided as much as possible because
_ . . .. . ... . ..
1(~672~2
the resulting two-phase flow decreases the capacity of the pipeline for gas
transport and it is of importance for the operator of the pipeline to be in-
formed on the occurrence of two-phase flow in order to enable him to take mea-
sures. One source and one detector may be located around the wall of a pipe-
line opposite each other for a pipeline with a diameter of less than 10 cm.
Any slug passing this location will be detected~ Still better results are
obtained with three sources and three detectors located alternatively around
the wall of a pipeline at equal distances. Apart from the detection of the
presence of liquid information is obtained on the volume fraction of liquid
present at the measuring spot. In the light of the above it is clear that
another number of sources and detectors, alternating and equally spaced, will
be optimal under certain circumstances.
The invention is very suitable for the detection of the
level of a liquid such as a liquid hydrocarbon in a vessel or a column. It is
furthermore possible to detect the foam height of such a liquid in a vessel or
a column. An interesting application is the detection of the level of a fluid
bed of particles of hydrogen containing solid material such as polymeric
material.
The detection of levels as indicated above may be used for
measurement of the height of a liquid column, a foam or a fluid bed, as well
as for alarming and control purposes.
The invention will further be elucidated with reference to
the drawings and a number of examples.
Figure l(a) shows schematically a measuring set-up for a
detector with a californium-252 source as used in the apparatus according to
the invention.
Figure l(b) and l(c) are embodiments of the detection
system as used according to the invention.
Figures 2 to 5 show several graphs, representing the
measurement-results of the apparatus of the invention.
Furthermore, in every figure the spaces, in which the
results were obtained, are represented.
-6-
1067Z12
Figure 6 shows a graph, representing another measurement-
result.
Figure 7 shows an advantageous embodimen~ according to the
invention.
Figures 8 to 11 are graphs, representing the results of
several applications of the invention.
Figure l(a) shows, schematically, a measuring set-up for a
helium-3 detector 1 with a californium-252 source 2. The detector being pro-
vided with a pre-amplifier 3, a high-voltage bias supply 4, an amplifier
analyser 5 and a digital rate meter 6.
-6a-
_
1067Z~2
Fig. 1b and c show the detection system, like reference
numerals denote like parts, which may be placed in a shielding 7,
consisting of a graphite reflector o, a paraffin wax moderator 9
snd a layer 10 of cadmium between the graphite reflector 8 and the
paraffin moderator 9. The graphite 8 reflects the neutrons from the
source 2 towards the medium (not shown) to be measured. The neutrons
which pass through the graphite ô are absorbed by the paraffin wax 9.
To prevent slow neutrons returning from the paraffin wax 9 to the
detector 1 a layer 10 of cadmium is provided. This shielding in-
creases the background signal at the detector 1, but this increase
is more than compensated for by an increase in the signal from the
measuring medium. The inclusion of the shielding 7 is thus equivalent
to an increase in the source strength.
This shielding thus has two advantages:
1. Reduction of the dose rate by at least a factor of 3.
2. A reduction in background signal variation due to variations in
the immediate surroundings. For example, if the shielding is omitted,
a person approaching the detector causes a slight increase in the
background signal.
This set-up is preferably used in the examples I to V which will
now be discussed.
EXAMPLE I
A settler made of 2 cm thick steel, with an outlet at 2.55 m
height and with a diameter of 2.00 m was used to separate an alkylate/
propylene mixture from water. The neutron scattering method was used
to detect the alkylate-propylene/gas interface for which purpose a
probe was used containing 2.5 ~g californium-252 as a neutron source
106721Z
and a helium-3 detector mounted on an aluminium pipe which could be moved
along the outer surface of the steel wall. The neutron-scattering intensity
is given in counts per second along the horizontal axis as a function of
height in meters above the bottom of the settler along the vertical axis in
fieure 2. The plotted numbers are net counts beine the difference between
observed number of counts and the number of counts (~700 c.p. 10 sec) without
any scattering medium present. A strong increase in signal is observed when
the probe is moved outside the settler in downward direction. The level found
at 2.55 + 0.05 m corresponds exactly with the centre of the alkylate/poly-
propylene outlet.
The settler is indicated in figure 2 as 11, with an inlet 12, a bottom
outlet 13 for water and an outlet 14 for alkylate/propylene. The height
scale in the graph corresponds with the actual height at the settler,
EXAMPLE II
Similar measurements as described in example I have been carried out on a
flasher of a propane deasphaltine unit in order to detect foam and liquid levels.
In figure 3 the flasher is indicated by 14 and consists of a column of
10 m height and 3 m diameter made of steel. Two trays 15 and 16 are present.
The feed inlet with a tray is indicated by 17. A wire mat 18 is present to pre-
vent liquid droplets leaving the column. An outlet 19 for gas is present at the
~Je~iC~
~ 20 top and an outlet for liquid 20 near the bottom. The ~i7-r~-~ axis of the
ho~i20~Jt~L/
graph corresponds with the actual height at the flasher in meters, the vcrtical
axis corresponds with the counts per 10 seconds.
It is shown in the graph that the liquid level in the column is
present at ~l,6 m, The smaller peaks in the counts indicate liquid present
on the trays and the feed inlet tray, There is virtually no foam above the
feed inlet as is clear from the low counting rate above 7 m height.
,, . , ,, _ ~ _
~0~7Z~Z
EXAMPLE III
Similar measurements as described in examples I and II have been
carried out on an asphalt flasher unit. In figure 4 the flasher is indi-
cated by 21 and consists of a column of 7 m height and 2.3 m diameter
made of steel. The feed inlet is indicated by 22, A wire mat 23 is
present to prevent liquid droplets leaving the column. An outlet 24
for gas is present at the top and an outlet for liquid 25 near the
bottom. There are no trays in this flasher. The vertical axis of the
graph corresponds with the actual height at the flasher in meters and
its horizontal axis with the counts per 10 seconds.
The liquid level is present at 1.1 m as is clear from the sharp
increase in counting rate. ~o foam is present above the liquid.
EXAMPLE IV
Similar measurements as described above have been carried out
on a flasher for deasphalted oil. In figure 5 the flasher is indicated
by 31. It is provided with a feed inlet and tray 32, a gas outlet 33
and an outlet for liquid 34. A wire mat 35 is also present and the
graph again shows the relationship between the actual height at the
flasher and the counts per 10 seconds like in fig. 4.
This flasher below the feed tray is filled with foam as is clear
from the counting rates below a height of 8m. In this flasher no
liquid level could be detected because of the presence of a concrete
support at the location where the level was expected. The measuring
points show a distinct scatter which is probably caused by density
fluctuations of the foam.
EXAMPLE V
Measurements have been carried out on a stripper which
is used to dry polypropylene powder with nitrogen in order to detect
the surface of the fluid bed. The stripper is a steel vessel with
~067Z~Z
a height of 3.9 m and a diameter of 1.6 m. Fluidi~ation of the contents
was carried out with 750 kg N2 per hour. ~esults of measurements along
a vertical line outside the wall are shown in figure 6, which represents
along the vertical axis the height in meters and along its horizontal
axis the counts per second, it is clearly shown that a strong signal
increase, due to the presence of polypropylene is found if the probe
is moved down the reactor wall.
From the examples I to V can be concluded that the gauge,
consisting of a californium-252 source and a helium-3 detector, is an
effective tool for external detection of hydrocarbon levels in vessels.
Further experiments for testing the neutron gauge as an external level
detector were carried out with an artificial hydrocarbon level, which
was created behind a steel wall by piling paraffin wax bricks behind
various steel plates with thicknesses from 2.5 to 14 mm. The gauge was
located on the other side of the steel plate at a fixed position. By
removing rows of paraffin wax bricks the level was effectively moved
with respect to the gauge. It appeared to be possible to detect levels
behind steel walls of up to 14 cm thickness.
Experiments also showed that depending upon the response time of
the gauge the source strength may be reduced to only 0.1~ gram for
several applications.
It will further be shown that the present method, as well as
apparatuses are extremely useful in acquiring a better understanding
of two-phase flow phenomena.
In the production and processing of natural gas, the heavier
components of the gas tend to form a liquid phase called condensate.
This liquid may be either water or hydrocarbons with five or more carbon
atoms in the molecule. The condensate is, in general, transported along
with the natural gas until it is removed by gas condensate separators.
-- 10 --
lQ67ZlZ
Whilst in the pipeline, however, the condensate affects the transport
of the natural gas. A better understanding of two-phase flow phenomena
is thus especially important in the design of natural gas transport
systèms.
In a bench-scale set-up the detection of hydrogen containing
materials with one Cf_252 source and a He-3 detector was studied. F'or
this purpose use was made of a piece of 30 cm 3" I.D. gas transport
pipe with a steel wall thickness of 15 mm. The ends were closed by two
metal prices welded on the pipe. This container was connected to a supp]y
of water so as to be able to fill it gradually, while it had an open
connection with the atmosphere. The source and detector were located
with respect to the pipe diametrically opposite each other. Gamma-ray
atte~uation measurements were made on the same pipe. The signals at
zero hold-up, i.e. only air in the pipe were normalized to 100%; an
increasing hold-up gave an increaseinsignal for the neutron scattering,
while an exponential decrease was observed, as expected, for the gamma
signals. The gamma-ray absorption only yields information about the
liquid hold-up in the optical path of the gamma-ray beam while the
neutron scattering gives a signal related to the total amount of liquid
present in this pipe,
It appeared that the gauge has its lowest sensitivity for small
hold-ups. For a hold-up increase from zero to 0.1 a signal increase of
50% is found while from zero to Q.2 liquid fraction, the signaL
increases by 300%.This initial low sensitivity is probably due to the
fact that several neutron-proton collisions are needed to sufficiently
slow down a neutron and to achieve a substantial increase in detection
probability, If only a small amount of hydrocarbon is present the
relative probability of more than one collision gets very small.
10~7ZlZ
This low initial sensitivity can be a disadvantage if small
hold-ups have to be measured accurately when no other scattering ma-terial
such as a high pressure gas-phase is present. To solve this problem
polyethylene covers were constructed. The amount of atomic hydrogen (13 g)
and the geometry needed to achieve optimal conditions have been determined
experimentally. It shall be clear that any material, other than poly-
ethylene, which contains hydrogen will be suitable for this purpose.
The location of the liquid phase in a gas transport pipe, with
respect to the source and detector will affect the measured sienal. The
signal will also be affected by the position of the source with respect
to the detector. These dependences are referred to as geometry effects.
In actual gas transport pipelines there is no a priori knowledge
of the two-phase flow geometry. This will give a large inaccuracy in
the hold-up determination if only one source and one detector are
used. In the following a measuring set-up using several sources and
detectors is discussed which almost completely eliminates geometry
effects.
Fig. 7 shows, schematically, a measuring set-up employing three
sources 42 , 42 and 42 and three detectors 43 , 43 and 43 placed
alternatingly and symetrically around a pipe 41, which is partly
filled with water 44.
The signals of each separate detector will be different, although
ll ll l
to a lesser extent for the detectors 43 and 43 , however the sum
of the signals will be a measure for the quantity of water present.
Experiments showed that for pipe diameters up to 5 an arrangement
with three detectors and three sources is within 1% standard deviation
independent of geometry effects. Said experiments were carried out by
displacing the sources and detectors with respect to the water, as well
as by means of three concentric rings of glass tubes, which were placed
inside the pipe tnot shown) and could separately be filled with water.
_ .
~06721Z
It was found that there is a well-defined correlation between
the sum signal of the three detectors and the l;quid hold-up in a
gas transport tube. By calibration a signal can be converted into a
hold-up value. It was further possible, by comparing the three
separate detector signals, to determine -the location of the liquid
present in the pipe.
No scattering due to the gas phase occurred in the experiments
which were hereinbefore discussed with respect to pipes. However,
in actual gas transport lines, where pressures may be as hieh as 250
bar, the density of the gas phase is su~ficiently high to contribute
significantly to the neutron scattering.
To simulate an actual eas transport pipe on a laboratory scale
a 15 l. gas cylinder with an I.D. of 14 cm, was pressurized with
methane.
Fig. 8 shows the signal increase as a function of gas pressure,
in bar along the horizontal axis, indicating that the neutron
scattering gauge can also be used as a completely external pressure
gauge. It should be noted, however, that the neutron scattering
signal is only indirectly related to the gas pressure, the gauge
in fact determines the density of hydrogen atoms inside the cylinder.
To simulate the two-phases in natural gas transport lines as
closely as possible, the condensate was replaced by water which has
about equal hydrogen content, on a volume basis, is easier to h~ndle
and has a negligible solubility for methane. The gas phase was methane
at 100 bar. In this experiment the detector-covers were not used since
the gas phase has sufficient initial hydrogen at zero hold-up.
Fig. 9 shows the sum signal of the three detectors along the
horizontal axis versus liquid hold-up in the cylinder, The signal
offset of about 5~ is due to neutron scattering in the gas phase, This
- 13 -
10672i2
result indicates that hold-up measurements for two-phase flow can
be made on actual gas transport lines. If the gas pipeline is con-
siderably larger than 14 cm, I.D. then more than three detectors
and sources may be applied.
Time-dependent two-phase flow measurements were made with a
pipeline for experimental purposes of 100 m length and 5 cm diameter.
Air was pumped through this pipe together with water in various
amounts. Three sources and three detectors were placed around the
pipe in an arbitrar,v location, Scattering measurements were carried
out as a function of time, It depends very much on the air velocity
and the amount of water in which way the water passes through the
pipe, The air blows over the surface of the water and causes waves,
the magnitude of which varies from ripples to surges, The measure-
ments of the three detectors have been added and used to calculate
the hold-up of water with the aid of a calibration carried out pre-
viously. ~esults are shown in flgure 10 and figure 11, both giving
the relationship between volume fraction along the vertical axis
and time ln seconds along the horizontal axis, In figure 10 the
arrival of water at the measuring location occurred at 1,5x103 sec,
The volume fraction of water was virtually constant after 2,5x103 sec.
at a level of 0,33, Small thin waves were visible at more or less
regular intervals,
In figure 11 the same amount of water was used and a much higher
air velocity, Large peaks were detected? almost reachlng through
the entire cross-section of the pipe,
_ 14 -
