Chapter 9

Measurement of ow rate
This Chapter covers common instrumentation and techniques that areused to
measurebulk owrates, namely theamounts of uids that pass through a cer-
tain cross section of a pipe, duct, channel or other ow conduit per unit time.
Bulk ow measurement is not concerned with local velocity variations across
the cross section, neither short-time (e.g. turbulent) uctuations. Measured
ow rates can be either mass ow rates

m, or, when dealing with liquids or
low-speed gases, volume ow rates Q. Flow rate measurement is an essential
activity in a variety of industries and utility services, but it is also employed
regularly in the uid mechanics laboratory, notably in the important role of
monitoring and controlling theexperimental conditions. The operation of ow
rate measurement systems is based on diverse physical principles; with some
exceptions, such systems require calibration or empirical corrections. The fol-
lowingpresentationismainlyconcerned with bulk owmeasurement in simple
ows, which are single-phase and either steady or very slowly varying. Some
of these methods can be extended to ows of multi-phase uids, slurries and
granular materials, but the reader is advised to consult specialized manufac-
turers and references when dealing with such media. More details on general
methods for themeasurement of owrateand specicinstruments can befound
in several monographs and books [2, 4, 10, 11, 12], handbooks [7, 15, 6, 13] and
manufacturers catalogs (e.g. [8]).
9.1 Direct methods
Thesimplest ow ratemeasurement methods aredirect, which means that they
measurea typical ow velocity or theamount of discharged uid over a period
of time. Such methods aremoresuitablefor liquid than gas ows. For example,
one may obtain a rough measurement of the bulk velocity of ows in water
tunnels andopen channelsbytimingthemotionof suspended or oatingobjects.
For ows of non-volatile liquids in an open-loop conguration, the volume ow
rate can be measured by timing the lling of a container by the discharge of
245
246 CHAPTER 9. MEASUREMENT OF FLOW RATE
the apparatus; similarly, the mass ow rate can be measured by weighing the
discharged uid. I n such cases, one must take care that discharging of uid
has no appreciable eect on the operation of thesystem, as for examplewould
be the case if removal of liquid from the loop resulted in lowering the head
of a feeding tank or shifting the operating point of a pump. Direct ow rate
measurement methods, applicable to liquid and gas ows in both open- and
closed-loop congurations, includetheuseof positivedisplacement owmeters,
to be discussed in the next section.
9.2 Positive displacement ow meters
Posi tive displacement (PD) ow meters aredevices which isolatexed volumes
of the uid owing into their inlet in sealed compartments and then discharge
themto theoutlet. Neglecting leakageand other possiblede ciencies, one can
easily computethevolumeowratefromthesizeand number of compartments
in thedeviceand themeasured rateof repetition of thecycleof their operation.
I n many cases, the same instruments can be congured to measure the total
volume of uid that passes through them over a time interval, and, for this
reason, PD meters are commonly used to monitor the consumption of water,
natural gas and hydrocarbon fuels. PD meters may be operating passively,
receivingpower fromtheowing uid, or driven by an external sourceto create
the uid motion, in which case they are called meteri ng pumps. There is a
great variety of designs of PD meters, most of which can beclassied as rotary,
reciprocating or nutating. Theimportant parameters for their operation arethe
leakage and the pressure loss across them. Sealing relies on capillary action
in narrow gaps between meshing parts, thus leakage depends on the speed of
operation, the viscosity of the uid and the wear of the moving components
and thehousing. Tominimizeleakage, thecomponents aremanufactured under
small tolerances and the clearances between meshing parts are kept low; this
necessitates theuseof clean uids, and in most cases themeter is accompanied
by a lter. PD meters are suitable for uids with relatively wide ranges of
viscosity. I ncreasing viscosity improves the sealing action but also increases
pressure losses, which have to be maintained as low as possible in order to
avoid signicant loading of the uid system. Temperature variation aects the
operation of PD meters in two ways: by aecting the viscosity of the uid,
with implications on leakageand pressure loss, and by aecting the density of
the uid, which is of concern when converting volume ow rate to mass ow
rate. Manufacturers would normally provide charts describing the operation
characteristics of eachmodel, such as thepressureloss for dierent owratesand
uidviscosities aswell as correction factors and uncertaintiesat dierent speeds.
A fewrepresentativedesigns of PDmetersarediscussed below[2, 13, 10, 7, 1, 3].
9.2. POSI TIVE DI SPLACEMENT FLOW METERS 247
Figure 9.1: Sketches of representative positive displacement ow meters: a)
nutating disk meter, b) reciprocating piston meter, c) rotary piston meter d)
rotary vanemeter, e) oval gear meter and f) roots meter.
P D ow meter s for l iqui ds:
Nutating disk meter s: Themain element of thesemeters is adisk which
is rotating in anutating (precessing, wobbling) fashion, whileboth its sides are
partiallyin contact with adual conical housing(Fig. 9.1a). Fluidentersthrough
the inlet port facing one side of the disk during half of the cycle, in isolation
from the outlet port; it is then swept by the precessing motion of the disk to
the outlet during the following half of the cycle, while being in isolation from
theinlet port.
R ecipr ocating piston ow meter s: These meters contain a number of
plungers or pistons, driven by a wobble plate and sweeping the volumes of
corresponding cylinders, whileat thesametime opening and closing input and
output ports or valves (Fig. 9.1b).
248 CHAPTER 9. MEASUREMENT OF FLOW RATE
R otary piston ow meters: These meters contain a cylindrical drum,
which is mounted eccentrically inside a cylindrical housing and rotates with
its outer surfacein contact with the housing, while its inner surface maintains
contact with an inner cylinder, coaxial with the housing (Fig. 9.1c).
R otary vane ow meters: Flat vanes are inserted into matching slots
around theperimeter of a rotating cylindrical drum, located eccentrically within
the housing. Centrifugal action or springs cause the vanes to slide out of the
slots until they come into contact with the housing, thus isolating a volume
of the owing uid and transporting it fromthe inlet towards the outlet (Fig.
9.1d).
R otor meter s: Thesemeters contain rotating meshingelements of dier-
ent shapes, including oval gears (Fig. 9.1e), circular gears, helical gears and
lobes. The rotary abutment meters contain both specially shaped rotors and
rotating vanes. In these devices, uid is trapped in the space between the ro-
tating elements or between an element and thehousing and is pushed towards
theoutlet in isolation from theinput.
P D ow meters for gases:
R oots-type ow meters: This is a trademark name that describes a
particular design of a lobe meter (Fig. 9.1f), developed for usewith gases.
Diaphr agm-type ow meter s: Thesemeters arecommonly used in do-
mestic gas lines. They contain bellows that ll-up with gas during part of the
cycleand dischargeit to theoutlet during a subsequent part; thegas ow from
the inlet to the outlet is controlled by sliding valves and the motion of the
bellows is linked to a mechanism that counts thecycles.
Li quid-seal ed dr um-type ow meter s: Aslo known as wet gas meters,
these devices consist of a hollow drum rotating within a cylinder partly lled
with a liquid, which provides thesealing action.
9.3 Venturi, nozzleand orice-plate ow meters
Also known as restriction or obstruction ow meters, these are devices that
force the ow through a restriction, thus increasing its velocity and decreasing
its pressure. Theowrateis estimated froma measured pressuredierenceand
an empirical correction coe cient. I n order to describe the idealized response
of these devices, consider steady, uniform, inviscid, incompressibleow, in the
absence of body forces, owing within a circular tube with a diameter D and
guided to a restriction with a diameter d. Then, one may use continuity and
9.4. OPEN CHANNEL FLOW MEASUREMENT 249
Bernoullis equation to relate the ideal volume ow rate Q
id
to the pressure
drop p between thetwo cross-sections as
Q
id
=
D
2
=4
q
1 (d=D)
4
s
2 p

(9.1)
To account for deviations from the idealized behaviour, one may introduce an
empirical dischargecoe cient C
d
to computetheactual volume ow rate as
Q =C
d
D
2
=4
q
1 (d=D)
4
s
2 p

(9.2)
The value of C
d
depends mainly on the geometry of the apparatus and the
Reynolds number, with 0 < C
d
< 1. For Reynolds numbers su ciently largefor
theow to bein thefully turbulent regime, Cd becomes insensitiveto Reynolds
number and depends only on the shape of the device. I t is desirable to keep
its valueas largeas possible, in order to reducethepermanent pressureloss in
theowmeter. On theother hand, largevalues of thedischargecoe cient may
only beachieved with carefully shaped and relatively long devices, which tend
tobebulkier and moreexpensivethan devices with lower C
d
: Common low-loss
restriction ow meters are the Venturi tubes (Fig. 9.2a,b) and the Dall tubes
(Fig. 9.2c), while relatively high-loss meters include ow nozzles (Fig. 9.2d)
and orice plates (Fig. 9.2e). The designs of such devices are regulated by
various standards (e.g. ISO and ASME standards), so that they may be used
interchangeably and without the need for individual calibration. A limitation
of obstruction ow meters is their relatively narrow dynamic range, which is a
result of the non-linearity in the Q p relationship. Thus, thesensitivity of
these ow meters decreases rapidly as the ow rate drops below about 25% of
thefull-scalevalue.
9.4 Open channel ow measurement
Thevolumeor mass ow rateof liquids in open channels or partially lled pipes
and ducts is often measured with the use of direct methods (see section 9.1).
Another common approach is the use of ow restrictions, including weirs and
Venturi umes [10, 16, 9, 5, 15].
Weirs consist of obstructions positioned across the channel and over which
theliquid is forced to ow. Thereareseveral types of weirs, both sharp-crested
and broad-crested, with rectangular, V-shaped or trapezoidal openings (Fig.
9.2f ). Themost common typeused in uid mechanics and hydraulics laborato-
ries is thesharp-crested, V-notch weir. An approximate expression for the ow
rate over such weirs is [5]
Q 2:5tan

2
H
5=2
(9.3)
250 CHAPTER 9. MEASUREMENT OF FLOW RATE
Figure 9.2: Sketches of obstruction ow meters: a) classical Venturi tube, b)
ASME Venturi tube, c) Dall tube, d) ow nozzle, e) oriceplate and f) weirs.
where is the full angle of the notch (usually equal to 90

) and H is the
easily measurablehead of theliquid abovetheweir, namely thevertical distance
between thefreesurface and thelowest point in theweir opening.
Venturi umes are converging-diverging channel constrictions, analogous to
Venturi tubes used in pipes ows. They generally have low pressure losses
and areavailablein several dierent designs, includingumes with rectangular,
trapezoidal and U-shape cross sections; commonly used designs for sewageand
irrigation ows arethe Parshall and Palmer-Bowlus umes. Their advantages
over weirs is that they do not cause a water back-up and are less likely to be
aected by deposited solids that may betransported by the water.
9.5 Averaging P itot tubes
These instruments consist of a tube spanning thecross-section of the pipeand
having multiplefrontal openings such that it measures a total pressure roughly
averaged over thecross section (Fig. 9.3a); they also havea second tube, facing
backwards and monitoring the local static pressure. The volume ow rate is
estimated fromthepressuredierence p as
Q =C
d
D
2
4
s
2 p

(9.4)
where C
d
is an empirical correction coe cient accounting for deviations from
theideal response. Simplicityof operation andlowcost arethemain advantages
9.6. LAMI NAR FLOW ELEMENTS 251
Figure9.3: Sketchesof common owmeters: a) averaging Pitot tube, b) laminar
ow elements, c) rotameter, d) vortex shedding ow meter, e) drag ow meter
f ) turbine ow meter and g) paddlewheel ow meter.
of averaging Pitot tube. Their limitations include theneed for clean uid and
a narrow dynamic range, extending to only about 30% of full-scale.
9.6 Laminar ow elements
Laminar ow elements are pipe sections or devices of a larger diameter which
contain tube bundles or relatively long honeycombs (Fig. 9.3b). The uid is
subdivided to pass through these elements, which are su ciently narrow for
the Reynolds number in each element to be lower than the transitional value
of about 2300. Thus, the ow in the elements is laminar and the pressure
drop across the elements is related to thevolumeow ratethrough theHagen-
Poiseuilleexpression, which is linear, in contrast with the quadratic expression
for turbulent pipe ow. For fully-developed laminar ow in a circular tube of
length l and diameter D, this expression becomes
p =
128l
D
Q (9.5)
Due to their linear response, these ow meters are suitable for very low ow
rates. Their disadvantages are large frictional pressure losses and bulkiness.
They also tend to beclogged by impurities in theow.
252 CHAPTER 9. MEASUREMENT OF FLOW RATE
9.7 Rotameters
Rotameters, or, moregenerally, variablearea owmeters, aresimpleand versa-
tiledevices that can be used with a widevariety of liquids and gases over wide
ranges of ow rates. They consist of a vertical tube, tapered such that its cross
section increases linearly upwards and a oat, which is pushed upwards by
the owing uid and stops at a position at which the drag, the boyancy and
the weight are in balance (Fig. 9.3c) [2, 15, 10]. The height of the oat is
proportional totheow rate, which is displayed in appropriateunits on a scale
engraved on the tube. Variable area ow meters arenot very sensitiveto uid
viscosity and can becorrected for density variations. They arepopular because
they requireno external power, can bepositioned near pipe bends and present
relatively low pressure losses. They are fairly accurate, except in the lower
end of their scale, typically below 10% of full-scale reading. Dierent tubeand
oat materials, sizes and shapes are available for dierent applications. Tubes
arecommonly madeof glass or transparent plastic, but stainless steel variations
with magneticsensingof theoat position arealsoavailablefor corrosiveliquids
or high temperatures and pressures.
9.8 Vortex shedding ow meters
The main component of the vortex shedding ow meters [2, 13, 10] is a blu
object immersed in theowing uid and spanning the pipe cross section (Fig.
9.3d). Their operation is based on the periodic shedding of vortices (Krmn
vortex street) from the edges of the object; this occurs at a frequency f (in
cycles per second), which is related to the frontal width h of the object and
theow velocity V . I n dimensionless form, theshedding frequency is called the
Strouhal number
S =
hf
V
(9.6)
For Reynolds numbers greater than a certain value(typically, about 5000), the
Strouhal number maintains an essentially constant value in the range 0.14 to
0.21, dependingon theshapeof theobject and independent of V . Theshedding
frequency is detected by a variety of means, including piezoelectric pressure
transducers, straingauges, self-heated resistanceelementsandultrasonicbeams.
9.9 Drag ow meters
Also referred to as target ow meters, drag ow meters [7, 8] (Fig. 9.3e) are
based ontherelationshipbetween thedragforceF
D
onan immersedblu obj ect
and theow velocity. I n general,
F
D
=
1
2
C
D
AV
2
(9.7)
9.10. TURBI NE FLOW METERS 253
whereC
D
is thedrag coe cient and A is thefrontal area of theobject, namely
the area of its projection on a plane normal to the ow velocity. C
D
is essen-
tially constant for an object with sharp corners immersed in turbulent ow at
su ciently large Reynolds numbers, typically greater than about 1000. Thus,
thevolumeow ratethrough a pipe would begiven as
Q =k
p
F
D
(9.8)
where k is a constant. I n practice, the target, which is a disk-like object, is
inserted in thepipeand mounted on asupport instrumented with strain gauges
or LVDTs (linear variable dierential transformers), which measure the drag
force through deection. Such instruments arevery sensitiveand bidirectional
and can be used at high pressures and with a variety of uids. As the target is
usually positioned in the centre of the pipe, they do not get easily clogged by
suspended impurities.
9.10 Turbine ow meters
Turbine ow meters measurethevolumeow rateof uids in pipes as propor-
tional to the angular velocity of an immersed vaned rotor [10, 15, 8]. A very
common typeutilizes an axial turbinewith its axis aligned with thepipecentre-
line(Fig. 9.3f). Thepassageof eachrotatingbladeis sensed electromagnetically
by an externally mounted sensor and theow rateis given by
Q =kn (9.9)
wheren is thenumber of pulses per unit timeprovided by thesensor and k is a
constant depending on the impeller design and size, the pipe diameter and the
number of blades. Turbineow meters aresubject tosignicant pressurelosses
and are prone to cavitation, when used with high-speed, low-pressure liquids.
A low-cost version, called the paddlewheel ow meter (Fig. 9.3g), utilizes a
partially immersed rotor, with its axis normal to the ow direction. Besides
ow rate, turbine ow meters may also provide the total uid volume that
passed through over a timeinterval. Thecommon domestic water meters areof
theturbinetype.
9.11 Ultrasonic ow meters
Ultrasonic ow meters [2, 10, 13] utilizehigh-frequency (typically of theorder
of 10 MHz) pressurewaves to compute thevolumeow rateof liquids in pipes.
Thereare two distinct types of such meters: the Doppler ow meters and the
time-of-ight owmeters.
A representative Doppler ow meter consists of two piezoelectric crystals,
a transmitter T, which transmits an ultrasonic wave through the pipe, and
a receiver R, which receives the ultrasound reected by solid particles or gas
bubbles transported by the owing uid (Fig. 9.4a). The frequency f
r
of the
254 CHAPTER 9. MEASUREMENT OF FLOW RATE
Figure 9.4: Sketches of representativeow meters: a) Doppler untrasonic ow
meter, b) time-of-ight ultrasonic ow meter, electromagnetic dlow meter and
d) Coriolis ow meter.
reected sound is shifted fromthefrequency f
t
of thetransmitted sound by an
amount f, called theDoppler shift, which is proportional to thevelocity V of
thereector, as
f =f
t
f
r
=
2f
t
cos
c
V (9.10)
wherec is thespeed of sound. Such devices arecalibrated toprovidean output
that is equal to the average velocity of the uid in the pipe, assuming that the
owis fullydeveloped. Theyarenon-invasiveand can behand-held or strapped
to theoutsideof a pipe.
A representativetime-of-ight owmeter consistsof twoexternally mounted
pairs of piezoelectric transducers. Each transmitter emits sound waves towards
thecorrespondingreceiver, oneof whichislocateddownstreamof itsmate, while
the other is upstream of it (Fig. 9.4b). Each transmitter emits a sound pulse
each time the corresponding receiver receives theprevious one. Becausesound
waves aretransported by theowinguid, sound propagates faster downstream
than upstream and the frequencies of pulsation of the two pairs dier by an
amount
f =
2cos
l
V (9.11)
9.12. ELECTROMAGNETI C FLOW METERS 255
wherel is thedistancebetween thetransducers of each pair. This conguration
makes theowmeasurement independent of thespeed of sound and, thus, ow
temperature.
9.12 Electromagnetic ow meters
Theseinstrumentsprovidethevolumeowrateof electricallyconductingliquids
in pipes. Their operation isbased on Faradays lawof electromagnetic induction,
which states that, when a conductor with length l moves with speed V in a
direction normal tothedirection of a magnetic eld with magnetic ux density
B, an electric potential E is generated across it as
E =BlV (9.12)
Practical electromagnetic ow meters [8, 10, 2] consist of an insulated pipesec-
tion of thesamediameter D asthepipeof interest, surrounded byanalternating
or pulsed magneticeld andhavingtwosurfaceelectrodes embeddedon thewall
across adiameter normal tothemagneticeld direction (Fig. 9.4c). Thevoltage
dierencebetween theseelectrodes is related to the volumeow rateas
E =
4kB
D
Q (9.13)
wherek isanumerical coe cient. Electromagnetic owmeters haveanaccuracy
that exceeds 0.5% and are not overly sensitive to the velocity prole. On the
other hand, they arebulky, heavy and relatively expensive.
9.13 Coriolis ow meters
Coriolis ow meters [10, 2, 14, 15] were developed relatively recently, but have
become increasingly popular in a variety of industries, due to their versatility
and their capacity to measure true mass ow rate, essentially independent of
uid properties and ow conditions. There are several dierent geometrical
designs, all based on the Coriolis forceprinciple. Consider a uid element with
mass m owing with velocity V in a tube that rotates with angular velocity
! about an axis normal to its own axis; assume that the uid element is at a
radial distance r from the axis of rotation and, during time t = r=V , moves
away fromit toa radial distancer+r. Then, theangular momentum(m) !r
2
of this uid element would increase to (m) ! (r +r)
2
(m) !

r
2
+2rr

.
This increase of angular momentumis attributed to a torque rF
c
, whereF
c
=
2(m)V ! is called the Coriolis force. The direction of the Coriolis force is
circumferential and opposite in sense to the direction of rotation, for outward
motion. I n vectorial notation, theCoriolisforceis writtenas
!
F
c
=2(m)
!
V
!
! .
I t is also written as
!
F
c
=(m)
!
a
c
, where
!
a
c
is called theCoriolis acceleration.
The owing uid receives this force from the tube walls; by reaction, the uid
256 CHAPTER 9. MEASUREMENT OF FLOW RATE
applies a force upon the containing tube, which is equal in magnitude and
direction to the Coriolis force, thus aecting thetube motion.
Practical ow meters do not rotate the tube but set it in vibration at its
natural frequency by subjecting it to an alternating magnetic eld. As repre-
sentative of this class of instruments, Fig. 9.4d shows a sketch of the U-tube
Coriolis ow meter. The uid is passed through a bent tube, whose ends are
clamped, while its tip is set to vibration. The instantaneous angular velocity,
and therefore theCoriolis force, increase towards the tip. The two legs of the
tubereceiveforces in opposite directions and, thus, the tube is twisted in one
sense during half of the cycle and in the opposite senseduring the other half.
Thetwist angleis measured by magnetic or optical position sensors sensing the
time delay t between the passage of the two legs through a transverse plane.
This time delay is related to themass ow rateas
t =
8r
2
t
K
s

m (9.14)
where r
t
is the radius of the tube and K
s
is a constant that, ideally, depends
only on the tube material. Small deviations from this relationship may be
caused by multi-phase eects and other variations in uid properties. Even so,
Coriolis ow meters are suitable for conventional as well as contaminated and
non-Newtonian uids.
9.14 Thermal mass ow meters
Thermal mass ow meters are used to measure the mass ow rate of gases.
They are not used for liquid ows due to the much higher power required to
heat aliquid than agas. For relatively lowmass owrates, theentiregasstream
is passed through the meter, while, at higher ow rates, only part of the gas
is heated by passing it through a bypass tube. There are two types of such
instruments, theheated tube ow meters and thei mmersion probe ow meters
[8].
I n the heated tube ow meters, the owing gas is passed through a piece
of tube that is heated electrically and is instrumented with two temperature
sensors, commonly thermocouples or resistance temperature detectors (RTD).
The rst sensor is located upstreamof the heated section and the other one is
downstreamof it. Therate of heat transfer

H to theuid is

H =

mC
P
T (9.15)
where

m is themass ow rateof the gas, C
P
is its specic heat under constant
pressureand T is thetemperaturedierenceacross theheated section. Thus,
themass owratefor agiven gas can bemeasuredfrommeasurements of

H and
T . Manufacturerssupplyinstruments with an output that has been calibrated
in air, nitrogen or someother gas. When used with dierent gases, this output
9.15. SELECTI ON OF FLOW METER 257
has to be corrected by multiplying it by the ratio of specic heats of the two
gases.
I mmersion probe ow meters consist of a probe with two RTDs connected
in a Wheatstone bridge conguration. One RTD is used to measure the gas
temperature, whilethe other is provided with a current so that it is heated to
a temperature higher than the gas temperature by a xed amount T . The
electricpower required toheat thesecond sensor isrelated tothemass-weighted
velocity V of thegasby anon-linear relationship, called Kings law. Electronic
circuitry is employed to linearizetheoutput so that it is proportional to V for
a given gas. To obtain the mass ow rate, one has to multiply this output by
the pipe cross-sectional area. Corrections for use with dierent gases are also
available.
9.15 Selection of ow meter
Consideringthediversity of designs and properties of ow meters and thewide
ranges of ow conditions encountered in a uid mechanics laboratories, it is
advisable to compare carefully the dierent options that are available before
purchasing a ow meter. Although it is possible that several devices may
be equally suitable for a given application, it is also quite certain that sev-
eral others would be totally unsuitable. As an aid towards the selection of
the optimal ow meter, we provide the following table, which was based on
information supplied by dierent manufacturers and contained in the website
http:/ / www.geocities.com/ ull_ km1980/ owmeterselectionguide.html
258 CHAPTER 9. MEASUREMENT OF FLOW RATE
ow me-
ter type
dirty
uid
1
dyn.
range
press.
loss
2
uncer-
tainty
3
upstr.
pipe
4
visc.
eect
2
cost
2
PD N 10:1 H 0:25 r none H M
Venturi Y 4:1 L 1 fs 5-20 H M
nozzle Y 4:1 M 1-2 fs 10-30 H M
orice
plate
Y 4:1 M 2-4 fs 10-30 H L
weir
(V-notch)
Y 100:1 VL 2-5 fs none VL M
Parshall
ume
Y 50:1 VL 2-5 fs none VL M
Pitot N 3:1 VL 3-5 fs 20-30 L L
rotameter Y 10:1 M 0:5 r none M L
vortex Y 10:1 M 1 r 10-20 M H
drag Y 10:1 M 1-10 fs 10-30 M M
turbine N 20:1 H 1 r 5-10 H H,M
Doppler Y 10:1 none 5 fs 5-30 none H
time-of-
ight
N 20:1 L 1-5 fs 5-30 none H
electro-
magnetic
Y 40:1 none 0:5 r 5 none H
Coriolis Y 10:1 L 0.4 fs none none H
thermal Y 10:1 L 1 fs none none H
1
Y: yes, N: no
2
H: high, M: medium, L: low, VL: very l ow
3
percent of full scal e(fs) or reading (r)
4
in diameters
References
[1] Anonymous. 1969 guide to process instruments elements. Chem. Eng., 76
(12):137164, 1069.
[2] R.C. Baker. Flow Measurement Handbook. Cambridge University Press,
Cambridge, UK, 2000.
[3] R.C. Baker and M.V. Morris. Positive-displacement meters for liquids.
Trans. I nst. M.C., 7:209220, 1986.
[4] R.P. Benedict. Fundamentals of Temperature, Pressureand FlowMeasure-
ments, Second Edition. Wiley Interscience, New York, 1977.
[5] E.F. Brater and H.W. King. Handbook of Hydraulics (Si xth Edi tion). Mc-
Graw-Hill, New York, 1976.
[6] J . Saleh (Editor). Fluid Flow Handbook. McGraw-Hill, New York, USA,
2002.
[7] N.P. Cheremisino (Editor). Encyclopedia of Fluid Mechanics, Volume.
Gulf Publishing Company, Houston, USA, 1986.
[8] Omega Engineering. The Flow and Level Handbook. Omega Engineering
I nc., Stamford, Connecticut, 2001.
[9] P. Ackers et Al. Weirs and Flumes for Flow Measurement. Wiley, New
York, 1978.
[10] R.A. Furness. Fluid Flow Measurement. Longman, London, UK, 1989.
[11] R.J . Goldstein. Fluid Mechanics Measurements (2nd Edition). Taylor and
Francis, Washington, 1996.
[12] A.T.J . Hayward. Flowmeters. McMillan Press, London, 1979.
[13] S.P. Parker (Editor in Chief). Fluid Mechanics SourceBook. McGraw-Hill,
New York, 1987.
[14] K.O. Plache. Coriolis/ gyroscopicowmeters. Mechanical Eng., 101(3):36
41, 1979.
261
262 REFERENCES
[15] J .G. Webster. The Measurement, I nstrumentati on and Sensors Handbook.
CRC Press, Boca Raton, Florida, 1999.
[16] F.M. White. Fluid Mechanics (Fourth Edition). McGraw-Hill, New York,
1999.