|
ULTRASONIC DOMESTIC GAS
METERS-A REVIEW
N. Bignell
National Measurement Laboratory
CSIRO Telecommunications and Industrial Physics
Lindfield, NSW, Australia
Page 1/4
Abstract: There
are many designs of ultrasonic domestic gas meters
but all the main ones use the measurement of the
transit time of an ultrasonic signal through the
flowing gas to estimate its velocity. The shape
of the duct and devices to control the waveform
of the signal passing through it are significant
parts of the design. The transducers used to produce
the signal, reciprocal operation and the main
techniques to time the transit of the signal to
several nanoseconds are discussed. The acceptance
of these meters has been restricted and their
possible future is discussed. Keywords: ultrasonic,
omestic, gas, meter, flow, transducer, transit
time, reciprocity
1 INTRODUCTION
xxxxThe purpose of
the domestic gas meter is to charge the customer
for heat used and this is done by assuming that
this is proportional to the quantity of gas used.
The best measurement of quantity is mass, but
as mass is difficult to measure for gases, the
volume has been the traditional quantity metered.
Sometimes temperature corrections are made but
not pressure, except in an average way. The usual
meter uses a bellows that fills and empties to
achieve a measurement of flow rates from a pilot
flow of 0.015 m3/h to a maximum of
6 m3/h. With a turndown ratio of 400,
operation from (at least) -10°C to 50°C, an uncertainty
over most of the range of 1.5% and a production
cost of about $20, this meter is the culmination
of a century of development. However, for at least
some systems, a large fraction of the gas consumed
is due to pilot flows that have been difficult
to measure with the bellows meter. Further development
of it seems unlikely to improve its performance.
Due to market resistance to this rather bulky
and unattractive traditional diaphragm meter,
a
competition was set up by British Gas in the 1980s
for the development of a more compact gas meter.
In the end, the successful designs were mostly
ultrasonic. During the last decade several ultrasonic
domestic gas meters have been developed to the
commercial stage and there has been much activity
in the patent literature [1, 2, 3, 4, 5, 6].
xxxxAll of the meters
seriously developed have used the transit-time
principle of operation. A pulse of ultrasound
is transmitted through the flowing gas and its
passage timed over a length L. It will have a
different time in the direction of flow Td
to that in the opposite direction Tu
and the equation
xxxxxxxxxxxx(1)
allows
the velocity v to be calculated. From this velocity
the flow may be obtained and, by integration,
the volume passed in a given time.
xxxxFor a meter that
is about the size of a common house brick, the
path length for the ultrasound is about 175 mm.
This gives a transit time of 500 ms, varying of
course with temperature and the nature of the
gas. The stream velocity cannot be too high or
the pressure drop will exceed the normal specification.
For a velocity of 22 mm/s that might be typical
of pilot flow, the difference between the upstream
and downstream times is 57 ns. Thus a resolution
of a few nanoseconds is needed for reasonable
uncertainty. In the laboratory context this is
not a difficult task using a high speed timer.
For low power battery operation that is required
for independence from the electricity mains, this
solution is not feasible as high speed oscillators
and timers use high power. Various solutions to
this problem have been found and will be discussed.
xxxxThough the times
measured allow a velocity to be measured this
is not necessarily proportional to the flow. The
correct determination of the flow from the velocity
is a problem for all ultrasonic flow meters. In
the large ultrasonic meters, based on the work
of British Gas, multiple beams are used to allow
the flow profile to be measured and hence the
flow to be estimated. This technique is not currently
feasible in a small battery operated device costing
$20 but other techniques are used and these are
described.
xxxxUltrasound propagation
in a flowing gas is not straightforward. Transducers
normally used in
ultrasonic work are made from piezo-ceramic materials
that have a high acoustic impedance. The acoustic
impedance of gases is low and this impedance mismatch
makes it difficult for ceramic transducers to
put energy into gases, or at least the process
is inefficient. The characteristics of the received
signal also depend on whether the direction of
travel of the beam is with the flow or against
it. Various meters handle these problems in different
ways.
2 TRANSMISSION
AND RECEPTION OF SIGNALS USING TRANSDUCERS
xxxxThe traditional
approach [1] uses a very low density composite
material on the surface of the
transducer, which is usually a piezo-ceramic material,
to provide a matching of the different impedances.
This layer may have stability and construction
problems but the coupling can be considerably
improved using this technique. To reduce the ringing
of the transducer, that is to reduce the Q, an
absorptive backing may be added. These transducers
usually have a frequency range of from 140 kHz
to 180 kHz.
xxxxPolyvinylidene
fluoride (PVDF) film has a naturally lower acoustic
impedance and can be made to have piezo-electric
properties by stretching and poling it. One transducer
developed using it [7] consists of a strip of
metal-coated PVDF film 25μ thick, held in a smooth
"M" shape. The curvature assists some
of the modes of vibration of the film when it
is excited by a signal applied across the thickness
of the film. The result is a transducer of low
Q and with a requency of about 115 kHz that operates
with low voltage excitation, and can be used either
as a transmitter or as a receiver, in a reciprocal
manner. A disadvantage is that the sensitivity
depends on the temperature and so the gain of
the electronics must be varied to allow for this.
xxxxA commercially
available transducer that operates at the low
end of the ultrasonic range, at about 40 kHz,
has also been used in gas meters [2]. It uses
a piezo-ceramic element and coupling to the gas
is enhanced by a small speaker cone attached to
it. This transducer has a large Q and so it rings
a lot in use. A technique mentioned later is used
to achieve the precise timing required.
xxxxOne of the operational
difficulties that the transducers must face is
that of dust. In all reticulation systems there
is some dust but in old systems there can occur
what are known as dust storms. The dust is composed
mainly of iron oxides and silica and gets picked
up by the flowing gas when there is a change in
the distribution pattern or some other disturbance.
It can be very upsetting for the operation of
the transducers for mechanical reasons, for reasons
of abrasion and for loss of linearity in performance.
xxxxLinearity in
transducers is important for the accuracy of the
meter at low flows. Non-linear behaviour can cause
the timing to be different in the two directions
even with no flow present. As a result there will
appear to be a flow when really there is not.
The need for good linearity occurs because when
transmitting the amplitude of vibration of the
transducer is orders of magnitude greater than
when it is receiving. Dust on the surface of an
otherwise linear transducer can adhere differently
at different amplitudes of vibration of the transducer
giving non-linear performance.
xxxxSome gases at
some frequencies absorb the ultrasound energy
much more than others. Gases with a simple molecular
structure such as argon and diatomic molecular
gases such as nitrogen and oxygen have low absorption.
Gases such as methane and carbon dioxide as well
as mixtures of these and water vapour with simpler
gases can cause much higher absorption. A mixture
of methane with 6% carbon dioxide is often used
as a test gas because it is especially absorptive.
The signal loss will depend on the frequency of
operation as well as the path length but can be
around 30 db.
|
