2.

11 Mass Flowmeters, Coriolis To Receivers

FIT DIT TIT

CATHY APPLE (1995) MARTIN ANKLIN,

WOLFGANG DRAHM (2003) DE TE

FE

Flow Sheet Symbol

Measured Variables Mass flow, volume flow, density and temperature

Sizes 1/25 to 10 in. (1 to 250 mm)

Flow Range 0 to 63,000 lb/min (0 to 28,300 kg/min)

Fluids Liquids, slurries, gases (compressed, low-pressure, etc.), liquefied gases; not gas-
liquid mixtures

Output Signal Linear frequency, analog, digital (HART, Profibus, FOUNDATION fieldbus, Modbus,
scaled-pulse, display, alarm outputs, manufacturer-specified protocols)

Operating Pressure Depends on tube size and flange rating: 1400 PSIG (100 bar) typical standard rating;
5000 PSIG (345 bars) typical high-pressure rating

Pressure Drop Function of flow, viscosity, and design, varying from very low (<0.1 PSIG, 10 mbar)
to moderately high (22 PSIG approximately 1.5 bar)

Operating Temperature Depends on design: −60 to 400°F (−50 to 200°C) typical standard; 32 to 800°F
(0 to 426°C) high-temperature, special versions also used for cryogenic applications

Materials of Construction Stainless steel, Hastelloy , titanium; special materials as tantalum, zirconium and
others are available

Inaccuracy ±0.1% of rate ± (zero offset/mass flow rate) × 100%

Zero offset depends on size and design of the flowmeter; for a 1-in. (25-mm) meter
with a typical maximum flow rate of 650 lb/min (18,000 kg/h), the zero offset is
typically 0.04 lb/min (0.9 kg/h), which is below 0.01% of the maximum flow value.
Typical: 0.15% within the range of 10:1 of full-scale flow rate (FS) and 1% within
the range of 100:1 of FS

Repeatability Typical: 0.075% within the range of 10:1 of FS and 0.5% within the range of 100:1
of FS

Rangeability Up to 100:1

Cost Depends on size and design: 1/25 in. (1 mm), $5000; typical 1-in. (25-mm) meter,
$7000; 6-in. (150-mm), $27,500

Partial List of Suppliers ABB (www.abb.com)

Bopp & Reuther (www.burhm.de)
Danfoss A/S (www.danfoss.com)

225
© 2003 by Béla Lipták
226 Flow Measurement

Endress+Hauser Inc. (www.endress.com)

The Foxboro Co. (www.foxboro.com)
Krohne (www.krohne.com)
Micro Motion Inc. (www.emersonprocess.com)
Oval (www.oval.co.jp)
Rheonik (www.rheonik.de)
Schlumberger Industries (www.slb.com)
Smith Systems Inc. (www.smith-systems-inc.com)
Yokogawa (www.yokogawa.com)

In recent decades, there has been a great deal of interest in MEASURING PRINCIPLE AND THEORY
Coriolis mass flowmeters (CMFs). The market for CMFs
grew dramatically in the late 1980s and the 1990s. Today, Principle
CMFs are widely accepted in many industrial fields, and their
performance has improved steadily. One of the advantages Coriolis mass flowmeters have the proven ability to record the
of CMFs is that they measure the true mass flow directly, total mass flow to better than 0.1% for water at moderate
whereas other types measure only volumetric flow. The high velocities. Each Coriolis instrument gets its own calibration
accuracy and rangeability of CMFs is another reason for their factor that depends only on the geometrical data and material
fast growth and acceptance in industry. The commercially properties of the tube. Thus, the calibration factor is indepen-
available units show a broad variety of designs, such as dent of fluid properties. The measuring principle of CMF is
single-tube, dual-tube, bent-tube, and straight-tube. Since Coriolis force, which appears in rotating and oscillating
CMFs are available that incorporate different tube materials (vibrating) systems. Such a vibrating system is shown in
 Figure 2.11a forr a straight tube. The tube is excited by an
(e.g., stainless steel, Hastelloy , titanium, zirconium, tanta-
lum, and lined tubes), they can be used for all kinds of liquids external force FE . The excitation frequency is kept at the
or gases. CMFs are most common in the food and beverage, natural frequency of the tube, which minimizes the energy
needed forr vibration.r The
r general expression for the Coriolis
chemical and pharmaceutical, and, increasingly, oil and gas r r
industries. force is FC = 2 ⋅ m ⋅ v × ω , where q = m ⋅ v is mass flow and

FE

aE
s1 s2
E

y x yE0 yE0
a)

v FC

aC
s2
y
x E yc0
b)

− yc0

s1

FC

FIG. 2.11a
Panel a) describes the movement of a straight tube conveying a fluid, which is oscillating at the excitation frequency. The oscillation is
maintained with the excitation force FE at location E. The measuring signal is detected with the two sensors S1 and S2. When the fluid begins
to flow, the Coriolis force FC induces an oscillation as shown in panel b). The final lateral displacement is the superposition of both oscillations.

© 2003 by Béla Lipták

 2.11 Mass Flowmeters, Coriolis 227

r
ω is the rotation vector. When fluid is not r flowing within a found by setting the excitation force FE(t) to zero. Inserting
vibrating tube, the Coriolis force is zero ( FC = 0). Whenr fluid the trial function for yE(t) in Equation 2.11(1), we get the
begins to flow, the Coriolis force is no longer zero ( FC ≠ 0), frequency of the first eigenmode,
and the shape of the tube is illustrated by superimposing Figure
2.11a, panel (a) and panel (b). At the inlet section, the Coriolis KE
force tends to decelerate the movement of the oscillating tube, ω E = ω E ( ρ Fluid ) = 2.11(2)
ME
whereas, for the outlet section, the Coriolis force tends to
accelerate the movement. In the middle rof the tube, the Cori-
Aside from the tube properties, ωE depends only on fluid
olis force is always zero, since
r r either ω is zero for straight density. Therefore, using Equation 2.11(2), the fluid density
tubes or q is parallel
r with ω for curved tubes, bringing the
r can directly be determined by measuring the frequency of
product q × ω to zero. As soon as the fluid begins to flow,
the eigenmode. Now, we include the excitation force FE(t) to
the Coriolis force induces a phase shift along the tube. This
determine the lateral displacement at the sensors. Solving
phase shift is proportional to the mass flow. The mass flow
Equation 2.11(1) with trial functions yE(t) and FE(t) and
can then be determined by measuring the phase shift between
Equation 2.11(2), we get
two sensor positions, S1 and S2. Since the oscillation is kept
at the natural frequency of the system, the frequency changes
with changing density of the fluid in the tube; i.e., the natural FˆE
yˆ E = 2.11(3)
frequency increases with decreasing density. Therefore, by  ω2 
knowing the actual frequency of the system, the density of the K E ⋅ 1 − 2 
 ωE 
fluid can be calculated directly. Another direct measurement,
in addition to mass flow and density, is the fluid temperature,
For commercially available instruments the amplitude for
which is measured by the CMF.
ŷE varies between 10 µm and 1 mm, and the frequency, fE =
ωE/2π, typically ranges from 80 Hz to 1100 Hz. Equation
Theory 2.11(3) also shows that the excitation force F̂E is at a mini-
mum when the driving frequency, ω, is similar to the fre-
In the literature, there are different approaches to describe quency of the eigenmode, ωE. In a real system, damping will
the dynamics of vibrating tube conveying a fluid or a gas prevent the lateral movement from becoming infinite even if
1 2
(see, for example, Païdousses and Li or Raszillier and Durst) . ω equals ωE. When the fluid begins to flow, the second mode
The general problem is very complex, and an analytical solu- is induced by the Coriolis force as shown in Figure 2.11a,
tion can only be obtained for a simple system with an ideal panel (b). For the Coriolis mode, the differential equation is
tube conveying an incompressible and nonviscous fluid. For
more complex systems, solutions can be found only through MC ⋅ ˙˙
yC + KC yC = FC 2.11(4)
approximations or using finite element methods. In this sec-
tion, we derive an analytical solution to determine mass flow where yC is the lateral Coriolis displacement of the tube at
in a simplified system. However, by solving this simple model, S1 and S2, FC is the Coriolis force, MC is the effective mass,
we gain insight into the major physical effects of CMF. and KC represents the stiffness of the tube for the Coriolis
We consider a straight tube conveying a fluid. We first mode. The trial function for the lateral displacement of the
look at the first eigenmode* of this system, which is shown Coriolis mode is yC (t ) = yˆC ⋅ cos(ω t ), and the function for the
in Figure 2.11a, panel (a). The tube is fixed at both ends, and Coriolis force is FC (t ) = FˆC ⋅ cos(ω t ). Using the same proce-
r
the velocity v of the fluid shall be zero. The movement of dure as above, we get the frequency of the Coriolis mode
the sensors S1 and S2 is described by the differential equation, ω C = KC / MC , which is typically 2.7 times higher than ωE.
The lateral displacement at the sensors becomes
M E ⋅ ˙˙
yE + K E yE = FE 2.11(1)

where FˆC
yˆC = 2.11(5)
yE = lateral excitation displacement at the sensor  ω2 
FE = excitation force KC ⋅ 1 − E2 
 ωC 
ME = effective mass
KE = the stiffness of the tube for the excitation mode
2 The Coriolis force FC is calculated by integration along the
˙ẏ = d 2y tube
dt

We are looking for solutions with yE (t ) = yˆ E sin(ωt ) and L/2

FE (t ) = FˆE⋅ sin (ω t). The eigenfrequencies of this system are FC =

∫ m˙ ⋅ y˙
0
E ⋅ aE′ ( x ) ⋅ aC ( x ) ⋅ dx
2.11(6)

* Resonance frequency or the first resonance frequency. FC = m˙ ⋅ CEC ⋅ y˙E

© 2003 by Béla Lipták

228 Flow Measurement

where CEC is a coupling factor between the excitation and the where the constant C does not depend on fluid properties.
Coriolis mode, ṁ is the mass flow, L is the length of the tube, For commercially available CMFs, this constant is deter-
aE′ = ( daE )/( dx ) is the derivative of the normalized excitation mined for each unit through calibration. Although we have
mode shape, ẏE ⋅ a′ is the local rotation velocity, and aC is the derived the formula to determine the mass flow of this system,
normalized Coriolis mode shape shown in Figure 2.11a, panel the model does not include effects such as axial pressure, in-
(b). If we define v E = y˙ E and with vˆ E = yˆ E ⋅ ω , we get ẏE = line pressure, temperature, pulsation, compressibility, and so
yˆ E ⋅ ω ⋅ cos(ωt ) = vˆ E ⋅ cos(ωt ) . Thus, Equation 2.11(6) becomes on. As mentioned before, analytical calculations including
FˆC = m˙ ⋅ CEC ⋅ vˆ E, and the lateral displacement of the sensors, such effects are very cumbersome and can be achieved only
Equation 2.11(5), becomes as approximations. The experimentally found influences of
these effects on mass flow measurements will be described
m˙ ⋅ CEC ⋅ vˆ E below.
yˆC = 2.11(7)
 ω2 
KC ⋅ 1 − E2 
 ωC 
DESIGN OF CMF
As described before, the final lateral displacement of S1 Figure 2.11b shows the tube assembly of a CMF. Generally,
and S2 is the superposition of excitation mode and Coriolis it consists of two components: the flow tube assembly and
mode. As seen in Figure 2.11a, the total lateral displacement the electronics. Typically, two electrodynamic pickups gen-
of S1 is yS1 = yE − yC, and for S2 it is yS2 = yE + yC. The time erate electrical signals containing the flow information. The
difference ∆τ between the two sensors becomes signal processing unit implemented in the electronics calcu-
lates the flow from these signals, which are very small in
∆ϕ 2 ⋅ yˆC 2 ⋅ yˆC 2 ( yS 2 − yS1 ) amplitude. The flow is split into two tubes as shown in
∆τ = ≈ = = ⋅ 2.11(8) Figure 2.11b. Sensors are mounted at the inlet and outlet
ω E ω E ⋅ yE
ˆ ˆ
vE ω E ( yS 2 + yS1 )
section of the tubes, measuring the phase difference between
these two points. The tubes are forced into oscillation by the
where ∆τ is the time lag and ∆ϕ is the phase shift between
driver, which is mounted between the two tubes. Thus, the
the two sensors. Now, we can determine the mass flow
tubes are automatically driven in counterphase, which is the
by inserting Equation 2.11(7) into 2.11(8), producing ṁ =
K C ⋅ (1 − ω E2 / ω C2 ) K ⋅ (1 − ω 2 / ω 2 ) preferred type of motion. To vibrate the flow tubes, all com-
2 ⋅ CEC
⋅∆τ , where the expression C 2 ⋅ CECE C is a mercially available CMFs use a magnet and a coil as the
constant value C. Thus, by knowing ∆τ, the mass flow of a driving mechanism. Typically, the coil is mounted on one
CMF can be determined through the simple equation tube, and the magnet is mounted on the opposite tube. To
protect the measuring system from any external disturbances,
ṁ = C ⋅ ∆τ 2.11(9) the tubes are fixed into a rigid carrier housing, which is strong

FIG. 2.11b
Tube assembly of a typical Coriolis flowmeter.

© 2003 by Béla Lipták

 2.11 Mass Flowmeters, Coriolis 229

enough to isolate the system from the environment. This value but in counterphase directions. Ideally, this results in
carrier housing is not shown in Figure 2.11b. zero force acting on the flanges. The perfect symmetry of the
The tubes are vibrated at their natural frequency. As two tubes is unaffected by changes in fluid density, temper-
shown before, this frequency requires the least amount of ature, pressure, viscosity, and so on.
energy to excite the system. Even large meters can be vibrated The sensors shown in Figure 2.11b can be mounted
with only a few milliamps of excitation current. The natural between the two tubes and do not have to be supported by
frequency depends mainly on the mass of the system and the the housing. This results in maximum common-mode rejec-
elastic properties of the measuring tubes. The total mass of tion and maximum suppression of externally induced vibra-
the system includes the mass of the tube itself, the mass of tions. The mounting of the driver and sensors must be done
the fluid within the tube, and the mass of any attached items in such a way that the overall mass balance of the tubes is
such as driver and sensors. Therefore, since the material maintained.
properties remain constant, a change in natural frequency If the flow is not split completely symmetrically into the
directly indicates a change in the density of the fluid. As two measuring tubes, no additional error will occur, because
described before, this change in frequency can be used to the flow signal, which is due the Coriolis forces, is composed
determine the density of the fluid. of the displacements of each tube separately and therefore is
independent of the exact flow distribution. Thus, a well-
Balancing Systems for CMF defined flow profile is not a requirement for the design of a
CMF. This also indicates that no special precautions are
CMF are among the most accurate flowmeters on the market. needed for installations near devices that may generate flow
This accuracy is achieved over a wide measuring range, turbulences.
which is required because, for example, liquids with high The majority of the commercially available CMFs use a
viscosities do not reach high velocities and have low total double-tube design, because this offers the best performance
mass flow. A high turndown from maximum flow is also with regard to accuracy and insensitivity to external distur-
needed for gas flow measurements because, even at high bances. However, the dual-tube design requires flow splitters,
pressure and at high velocity, the total mass flow rate for gas which are not recommended for applications with fluids that
is small in comparison to mass flow rate of fluids. The accu- are prone to plugging. Such fluids are often used in the food
racy for lower flow rates is limited by the zero-point errors. processing industry, where single-tube meters are required.
An error of 0.005% of full scale due to zero-point instability
is typical. Single-Tube Meters
In the previous section on “Theory,” it is shown that mass
flow induces very small displacements along the measuring Generally, there are two different designs of single-tube flow-
tube. These displacements have to be measured accurately, meters. In the first design, the tubes are bent to form a double
even though the instruments are often mounted in a harsh loop. This design behaves similarly to the dual-tube flowme-
process environment. A key parameter to achieve a precise ter with the difference that the tubes are in series rather than
and stable CMF reading is the decoupling of the internal parallel. Such single-tube flowmeters offer the same advan-
measuring system from any environmental and external dis- tages as dual-tube meters, and they do not have the disad-
turbances. If CMFs are not decoupled to near perfection, the vantage of employing flow splitters. However, with this
oscillations from the measuring tube will be transmitted to design, the tube length increases dramatically, which results
the connected process piping, which in turn begins to vibrate in increased pressure loss. Furthermore, easy drainage of the
as well. Vibrating process piping can then cause the CMF to instrument is impossible with this design. The second single-
be excited by undefined vibrations. Depending on the mag- tube flowmeter design contains a straight, or fairly straight,
nitude and the strength of such external excitations, this can single tube. From the customer’s point of view, these designs
lead to a disturbed reading of the CMF. Therefore, it is an are preferred, since they offer the best cleanability and the
important requirement of a CMF to be a balanced system, in most prudent fluid handling. A challenge is to find a balancing
which oscillations of the measuring tube are well defined mechanism for such flowmeters that allows accurate mea-
within the meter and are not transmitted to flanges and pro- surements for various process conditions and changing fluid
cess piping. This requirement is also a general rule to ensure densities. Nevertheless, straight (or fairly straight) single-
a good zero-point stability. tube CMFs are available that offer comparable performance
to that of dual-tube flowmeters.
Dual-Tube Meters
Tube Geometries
Designs with dual tubes offer the best performance for the
decoupling of the measuring system from the process envi- A variety of tube designs are currently available, a small
ronment. Similar to a tuning fork, the two tubes vibrate in selection of which is shown in Figure 2.11c. Most designs
counterphase. While the oscillation is maintained, the forces aim to magnify the effect of the Coriolis force by the geo-
at the fixation points of the two tubes are identical in absolute metrical form of the tubes. The larger the Coriolis effect

© 2003 by Béla Lipták

230 Flow Measurement

a certain value, which depends on the material of the tube.

By choosing a material with a low expansion coefficient, the
axial forces can be kept below the critical value, even for
straight tube CMF. Unfortunately, the rather high expansion
coefficient of stainless steel, which is the most common mate-
rial for measuring tubes, allows only a very restricted tem-
perature range for a straight-tube design. Therefore, stainless
steel tubes need to have a curved shape to reduce the maxi-
mum stress, since the tube can expand into the curve. All
commercially available CMFs with straight tubes use tita-
nium or zirconium for the measuring tubes, since these mate-
rials offer a small temperature expansion coefficient. With
these materials, even great temperature differences between
the measuring tube and the housing result in only small
additional axial stress. Moreover, titanium offers higher stress
limits than stainless steel. CMFs with single straight tubes
are available for use up to 150°C.
With regard to corrosion, erosion, and pressure rating,
the wall thickness of the measuring tubes should be as thick
FIG. 2.11c as possible. However, the sensitivity of the instrument to
Selection of geometries of various Coriolis flowmeters. flow-induced Coriolis forces decreases with increasing wall
thickness. Therefore, tube dimensions have to be optimized
for several considerations, including the overall pressure loss.
becomes, the larger the time or phase difference between the
For a 1.5-in. (DN 40) dual-tube design, a typical size of the
flow sensors becomes, and the easier it is to determine the
measuring tube is 1 in. (25 mm) inside diameter with a wall
mass flow. Such magnifying geometrical forms often result
thickness of 1/16 in. (1.5 mm).
in large tube loops that take up much space and have no
Flowmeters are commercially available with stainless
advantage in zero-point stability, because external distur- 
steel, Hastelloy , titanium, zirconium, and tantalum as tube
bances are also magnified. Thus, the signal-to-noise ratio 
material. Exotic materials such as glass or Tefzel -lined tubes
remains the same. As electronics have become more and more
are also available for special purposes.
efficient, the need for such geometrical magnification of the
Coriolis effect has disappeared. Therefore, the large loops
Sensors
can be replaced by compact tube designs that require little
space. An example of such a compact design is shown in As shown in Figure 2.11b, two motion sensors are needed to
Figure 2.11b. In addition, the compact design shortens the measure the displacement of the tube at the inlet and outlet
tube length, which results in higher oscillation frequencies sections. The phase difference or time lag between the two
of about 300 to 1100 Hz. Higher oscillation frequencies have sensor signals is a measure of the mass flow. The sensor could
the advantage of a better decoupling performance from pipe- be of any type that can represent the motion of the flow tubes,
line vibrations and external disturbances, which are predom- measuring position, velocity, or acceleration. At present, the
inantly in the range of about 50 to 180 Hz. most commonly used device is the electrodynamic sensor, in
For the dual-tube design, symmetry is the key factor, so which a coil is mounted on one tube and a magnet on the
a pair of tubes are chosen that are nearly identical in terms other tube. The relative motion between the tubes induces a
of mechanics. The two tubes have to be assembled in such a voltage in the coil, representing the differential velocity of
way that tube symmetry is not altered. Therefore, the produc- the tubes. Electrodynamic sensors have the advantages of
tion of these tube assemblies needs to be done very accurately, offering very good phase accuracy and high reliability.
with a good understanding of the production process itself.
The main reason for using bent tubes is the thermal Temperature Sensors
expansion of the measuring tube. While the fluid temperature
may change by several hundred degrees Celsius, the temper- As described previously, mechanical properties change with
ature of the supporting structure changes much less, due to temperature. This leads to axial stress and also changes the
thermal transport, convection, and radiation. This can lead to Young’s modulus. An increase in temperature decreases the
large temperature differences between measuring tube and stiffness of the tube by lowering the Young’s modulus. To
housing, which increase the axial forces of the tube. For a compensate for the influence of thermal effects on CMF read-
straight-tube CMF, the axial forces are largest and mainly ings, each flowmeter needs to be equipped with at least one
depend on the expansion coefficient of the tube material. To sensor to measure fluid temperature. Furthermore, because a
prevent the tube from damage, the axial force must stay below temperature difference between the measuring tube and the

© 2003 by Béla Lipták

 2.11 Mass Flowmeters, Coriolis 231

housing results in an axial force, a second temperature sensor multiplying the time difference or the phase shift with the
is needed to adjust the reading of the flowmeter for this effect. calibration constant of the flowmeter. Furthermore, thermal
Instead of a second temperature sensor, the axial stress can also effects on the mass flow and density reading have to be
be detected by a strain gauge attached to the measuring tube. included as well. This is commonly done with a micropro-
Temperature sensors have uses beyond merely accounting cessor. However, analog circuitry can also be used. Today,
for thermal effects. Because they measure the temperature of much analog circuitry is being replaced by digital signal
the fluid, temperature information is used as the third direct processors, which offer powerful mathematical functions to
process signal of a CMF, in addition to mass flow and density. allow, for example, filtering of the flow signals. With digital
processing, the response time of a CMF becomes faster, and
Security the reproducibility of the flow reading improves. Thus, with
digital signal processors, CMFs become capable of control-
The oscillation amplitude of a CMF is very small (typically, ling formidable applications such as rapid batching, where
100 µm). Stress in the measuring tubes is limited to ensure fast response and high accuracy are critical.
reliable operation of the meter for many years and to protect
the meter from damage due to tube oscillation.
Communication/Output
The whole vibration system, including driver and sen-
sors, is fixed in a solid housing, typically constructed of The primary output from a CMF is mass flow. However, most
stainless steel. This housing can act as a secondary contain- electronic designs are also capable of providing temperature,
ment. The more compact the CMF, the smaller the housing density, and volumetric flow data. Furthermore, totalizers
can be and, possibly, the higher the pressure rating of the provide mass or volume totals.
secondary containment. Housings with pressure ratings up to Most electronics are equipped with configurable alarm
1500 psi (100 bar) are available. outputs. Sophisticated relay functions are available whereby
Because they employ a small excitation current, intrinsi- the CMF directly controls a valve in a batching process.
cally safe CMF versions are available for use in hazardous Many digital output protocols are supported (e.g., Profi-
areas. The electronics must be tested for electromagnetic bus, FOUNDATION fieldbus, HART, Modbus, scaled pulse,
compatibility (EMC), fulfilling general EMC requirements and others), allowing a choice of communication solutions.
according to applicable guidelines. However, current (4- to 20-mA) and frequency outputs for
mass flow are still the preferred and most common output
Electronics signal formats.
The drive circuit initiated the tube oscillation and maintains
the oscillation at a certain amplitude. This circuit needs to be
built to provide a fast response to changing fluid properties. TECHNICAL DATA
Air bubbles, for example, cause a sudden increase in excita-
tion power. This information has to be supplied to the driver Measuring Accuracy/Range
quickly so as to keep the amplitude of the oscillation constant.
The driver circuit also controls the excitation frequency. Figure 2.11d shows the excellent measuring accuracy and the
The sensor signals are very small sinusoidal signals, large rangeability of CMF. During the 12-h test run, the zero
which have to be amplified to make them processible in the point and the calibration factor remain stable and are well
succeeding signal processing stages of the electronics. These within the specification of the instrument. Note that the read-
amplifiers need to have a very broad bandwidth to prevent ing remains accurate even at low flow rates, even below 1/100
the mass flow signal from containing additional zero-point of the maximum flow rate specified for the CMF.
errors.
The electronics can be mounted on the flowmeter directly, Pressure Drop
forming one compact flowmeter unit, or the flowmeter can
be interfaced to the electronic via a cable. This permits the The pressure drop depends on tube design and mainly
electronics to be located remotely from the sensor. The remote depends on the length of the tube and its inner diameter. For
assembly may be necessary for high-temperature meters, or the pressure drop of CMF with dual tubes, the design of the
it may be convenient if the sensor is installed in a place that flow splitter is also important. The lowest pressure drop
is not easily accessible. occurs with single straight-tube flowmeters, where the inner
diameter of the measuring tube is identical to that of the
Signal Processing connected process pipe. Typical pressure drops at the maxi-
mum flow speeds specified by manufacturers are 7 to 20 PSIG
The sinusoidal signals from the two sensors are compared to (0.5 to 1.5 bar) referred to water. For the measurements
determine either the time difference or phase shift between shown in Figure 2.11d, the pressure drop at 80 in./sec (2 m/s)
the two signals. The mass flow rate is calculated directly by is only 0.4 PSIG (30 mbar).

© 2003 by Béla Lipták

232 Flow Measurement

Error [%] of Reading

1.5

Typical Error Limit

0.5

-0.5

-1

-1.5
0 10 20 30 40 50 60 70 80 90 100
Flow [%], 100 % Flow Corresponds to v=80 inch/s

FIG. 2.11d
This figure shows the measuring uncertainty for a 1” (DN25) Coriolis flowmeter. The maximum flow speed is 80 in./s (2 m/s), which is
20% of the maximum specified flow speed of the flowmeter. The curves show the specified error limits.

Influences on the CMF Reading Mounting Pipe stress is introduced not only by in-line pres-
sure and temperature, as described before, but also by different
While improving the accuracy of CMFs during the past mounting conditions. These conditions may cause compres-
decade, many effects, mostly secondary, can be identified that sion, tension, or shear forces to be applied to the flowmeter,
influence the performance of a CMF. These effects can be which may affect the zero offset of the CMF. The influence
roughly separated into two groups. of these effects has been greatly reduced during the last
decade so that, today, a zero-point calibration is needed only
1. Effects such as changing fluid temperature, for which for special applications as described below.
CMF can directly account
2. Effects like external vibration, for which CMF cannot Vibration In most applications CMFs are exposed to some
directly account external vibrations. Such vibrations can occur as a result of
the pumping system or nearby vibrating devices, or they may
The latter effects are minimized by either the design layout be flow induced as observed in pipeline systems. External
or, if that is not possible, by special installation or correction vibrations typically occur at 50 to 180 Hz. As mentioned
instructions. In this section, we will briefly describe different previously, CMFs are designed such that the effect of external
effects. influences is minimized. Therefore, external vibration plays
a minor role and generally has no effect on the accuracy of
Temperature As mentioned previously, changing fluid and the CMF reading. However, if the external vibration is close
housing temperatures will affect the elastic properties of the to the working frequency of the CMF, measurement errors
CMF and thus influence the mass flow and density readings. will occur. It has been shown that pulsation is critical not
We can account for this effect directly by measuring the fluid only at the working frequency ( fE) of the CMF but also at
3
and housing temperatures separately. On the other hand, tem- frequencies f = fC − fE, where fC is the Coriolis frequency.
perature changes can also influence the zero offset and the Therefore, CMFs with high working frequencies are much
performance of the electronic components to some degree. less sensitive to pulsation and external vibrations than others.
The drift in electronic components will usually lead to This is because both fE and the difference fC − fE are high;
changes in the zero offset of the flowmeter. Both influences i.e., above roughly 200 Hz. For severely vibrating applica-
can be minimized by using a special design that does not tions, where the low working frequency of the CMF might
require any further corrections or installation instructions. become critical, the influence of the external vibration can
be greatly reduced by using flexible piping and vibration-
In-Line Pressure With changing in-line pressure, the tube isolating pipe supports.
becomes slightly deformed, which influences the stiffness of
the layout and thus can affect the reading of the CMF. With Humidity Because CMFs are typically enclosed in sealed
special designs, this effect can be minimized. cases that are completely isolated from atmospheric conditions,

© 2003 by Béla Lipták

 2.11 Mass Flowmeters, Coriolis 233

external humidity has only a minor influence. Also, the flow-

meter electronics are commonly enclosed in a housing that
provides protection against external humidity. However, in
CMFs with inadequate case seals or damaged housings,
extremely humid environments can create condensation on the
flow detector coils, which may lead to corrosion and compo-
nent failure.

Fluid Velocity It is well known that the velocity of the fluid

1
can slightly influence the accuracy of the CMF reading. This
is a minor effect, which is below the specified accuracy of
most CMFs and does not necessarily require any correction.
Nevertheless, given that the velocity of the fluid is known, a
CMF can directly account for this effect.

Gas Measurements Only in recent years has it been shown

that the compressibility of gas can affect the accuracy of the
4 FIG. 2.11e
CMF reading. Although this effect can be neglected for most Not recommended mounting location of a CMF.
fluids, it becomes relevant for gases in which the speed of
sound is diminished. Knowledge about this effect allows us
to correct the reading of CMF.

Two-Component Flow A CMF may be suitable for homo-

geneous two-phase (solid/liquid) flows and for heterogeneous
flows. Such applications include many food processes, sand
in water, pulverized coal in nitrogen, water in oil, and many
others. To measure two-phase fluids, single-tube meters may
be preferable.

Corrosion, Erosion Corrosion and erosion diminish the

FIG. 2.11f
wall thickness and therefore change the stiffness of the tube,
Orientation of CMF with curved tubes; the orientation shown in
which can lead to faulty CMF readings. Since CMFs are the left panel is not suitable for fluids with solids content; the
available with different tube materials, corrosion can signif- orientation shown in the right panel is not suitable for outgassing
icantly be reduced by choosing the appropriate material for fluids.
each application. To reduce erosion caused by highly abrasive
media, it is necessary to keep the flow velocity low. Erosion
also depends on the design of CMF and is smallest in straight,
single tubes. tubes to be completely drained and protects them from solid
build up.
Reynolds Number Although the accuracy of a CMF gen- When measuring liquids, the CMF should not be installed
erally does not depend on the flow profile, the sensitivity at the highest point of the system, because gas may accumu-
changes slightly from laminar flow to eddy flow. Knowledge late in the flowmeter as shown in Figure 2.11e. Installation
of the Reynolds number allows us to determine the state of in a vertical pipeline directly upstream of a free pipe outlet
the flow regime and thus to account for it directly. should also be avoided.
With curved tubes, the CMF orientation should be
Installation adapted to type of fluid used. Figure 2.11f illustrates prob-
lems with outgassing fluids and with fluids containing solid
Some general recommendations for installations are applica- particles.
ble to all CMFs. The measuring tubes should remain full of
the process fluid. Mixtures of gas and liquid should be Mechanical Installation Modern CMFs offer good balance
avoided. For gas measurements, the tubes should be filled in the vibration system and therefore have no specific instal-
with gas only, with no fluid droplets present. lation requirements. The CMF can be installed easily in a
The preferred installation orientation is vertical, with an pipeline. When heavy CMFs are used, mechanical support
upward flow direction. With this orientation, entrained sol- of the pipeline has to be considered. Pipeline supports should
ids can sink downward, and gases can escape upward, when not be attached directly to the sensor, and the CMF should
the medium is not flowing. This also allows the measuring not be used to support process piping directly.

© 2003 by Béla Lipták

234 Flow Measurement

TABLE 2.11g
Examples of Common Coriolis Flowmeter Applications
Food and Beverage Chemical and Petrochemical Petroleum Products
Beer, soda Adhesives Hydrogen peroxide Asphalt
Chocolate Alcohol Latex Bunker C
Fruit juice Ammonia Nitric acid Crude oil
Honey Catalysts Phosgene Diesel fuel
Ice cream Caustic Phosphoric acid Fuel oil
Margarine Cyclohexane Polyol Gasoline
Milk Ethylene Propylene Hydraulic oil
Molasses Formaldehyde Resins Jet fuel

Peanut butter Freon Solvents Kerosene
Pet food Glycerine Styrene Lube oil blending
Tomato paste Glycol Sulfuric acid Oil/water emulsion
Animal, vegetable fat Hydrochloric acid Toluene Tar
Pharmaceutical Pulp and Paper Other
Alcohols Antifoaming agents Compressed gases: nitrogen, helium, carbon
IV bag filling Black liquor dioxide, CNG
Palm oil Cellulose slurry Dyes
Perfume Paper pulp Ink
Pill coatings Red liquor Liquefied gases: carbon dioxide, LPG, LNG
Soap Titanium dioxide Magnetic tape coating
Sodium methylate Paint
Talcum powder Photographic emulsion
Vitamins Wax
Filling airbags (automobile industry)

Zero-Point Adjustment (Static/Dynamic) After factory cal-

ibration of a CMF, the calibration factor and the zero point
are stored in the electronics. CMFs that have good balance,
and thus are decoupled from connected piping, are not affected
by the installation into the process piping. As a result, the zero
point will not change, and no special zero-point adjustment
is necessary. Practical experience has shown that a zero-point
calibration is required only in special cases; for example, to
achieve the highest measuring accuracy possible in the pres-
ence of very slow flow rates or in the case of extreme process
conditions such as very high fluid temperatures.
Zero-point calibration is carried out using completely
filled measuring tubes with no mass flow. During the zero-
point adjustment, care has to be taken that no gas or solids
are present in the measuring tube. Keeping the in-line pres-
sure high during the zero-point calibration reduces the risk
of gas formation in the CMF and thus increases the accuracy FIG. 2.11h
of the zero-point calibration. This picture illustrates a CMF installed into a compact space. The
shown CMF is a single-tube Promass I. (Courtesy of Endress+Hauser
Flowtec AG.)

APPLICATIONS precision filling of containers, custody transfer, and other

applications. An overview of some of them is presented in
CMFs are currently used in many areas, including chemical, Table 2.11g. CMFs are suitable for many applications, because
petroleum, petrochemical, pharmaceutical, food and bever- they can be very compact and do not have any upstream or
age, and pulp and paper industries. Because of their versa- downstream piping restrictions. An example of a compact
tility, CMFs are used for process control, batching, inventory, application is shown in Figure 2.11h. The photo shows a

© 2003 by Béla Lipták

 2.11 Mass Flowmeters, Coriolis 235

single-tube CMF Promass I from Endress+Hauser Flowtec 4. CMFs are not available for large pipelines; the largest
AG. Note that inlet and outlet parts are bent at a 90° angle CMF has a maximum flow rate of 63,000 lb/min
and that the available room is very limited. (28,300 kg/min) using flanges with 10-in. (25-cm)
diameters. To measure higher flow rates, two or more
CMFs must be mounted in parallel.
5. CMFs are not suitable for gas applications with low
ADVANTAGES OF CMFs
in-line pressure, since low-pressure gases have low
densities. To generate enough mass flow to provide a
1. One of most important advantages of CMFs is that
sufficient Coriolis signal, the velocity of the gas must
mass flow is measured directly. This can be per-
be quite high. This may lead to a large pressure drop
formed with high accuracy, typically with 0.1% error.
across the meter.
High accuracy is also maintained over wide ranges
of temperatures (typically from −50 to +200°C) and
in-line pressures. Furthermore, CMFs are extremely
References
linear over their entire flow range.
2. CMF rangeability is extremely high. Measurements
1. Païdoussis, M. P. and Li, G. X., Pipe conveying fluid: a model dynamic
can still be performed at low flow rates, 100 times problem, J. Fluids Struct., 7, 137–204, 1993.
lower than the maximum flow rate specified. 2. Raszillier, H. and Durst, F., Coriolis effect in mass flow metering, Arch.
3. In addition to direct measurement of mass flow, tem- Appl. Mech., 61, 192–214, 1991.
perature and density are measured directly. Knowl- 3. Koudal, O. et al., High frequency Coriolis meter performance under
pulsating flow, in Proc. FLOMEKO 1998, 9th International Conference
edge about density allows us to convert mass flow
on Flow Measurement, Lund, Sweden, 239–242, 1998.
data into volume flow data. 4. Anklin, M. et al., Effect of finite medium speed of sound on Coriolis
4. The measuring principle is independent of the flow mass flowmeters, in Proc. FLOMEKO 2000 10th International Con-
profile of the fluid or gas. Therefore, no flow condi- ference on Flow Measurement, Salvador, Brazil, 2000.
tioner or special upstream or downstream pieces are
required. A CMF can also be used with a pulsating flow.
5. The accuracy of a CMF is independent of fluid prop- Bibliography
erties such as viscosity or density. Therefore, a CMF
can measure all kinds of fluids, including Newtonian Adiletta, G. et al., Twin rigid straight pipe Coriolis mass flowmeter, Mea-
and non-Newtonian fluids, slurries, and gases. surement, 11, 289–308, 1993.
Babb, M., New Coriolis meter cuts pressure drop in half, Control Eng.,
6. CMFs do not have any moving parts that wear out October 1991.
and require replacement. This reduces the need for Baker, R. C., Flow Measurement Handbook, Cambridge University Press,
and the cost of maintenance. UK, 2000.
7. Single-tube CMFs do not have internal obstructions Baker, R. C., Coriolis flowmeters: industrial practice and published infor-
that could be damaged or plugged. mation, Flow Meas. Instrum., 5(4), 229–246, 1994.
Birker, B., Theory, design and performance of the straight tube mass flow-
8. CMFs are designed to measure forward and reverse meter, in Mass Flow Measurements Direct and Indirect, IBC Technical
flows with high accuracy. Services Ltd., London, 1989.
9. Because CMFs are available based on different con- Blickley, G. J., Mass flow measurement aided by coriolis methods, Control
struction materials, they can be used for many dif- Eng., April 1991.
ferent applications, including corrosive fluids. Blumenthal, I., Improving productivity through mass flow measurement and
control, enhancing productivity, in Proc. Pacific Cascade Instrumen-
10. The design of CMFs allow them to operate with low
tation 1984 Exhibition and Symposium, 163–168, 1984.
power requirements. Bugher, G., Coriolis flowmeters, Meas. Control, September 1990.
Cascetta, F. et al., Experimental intercomparison of Coriolis mass flowme-
ters, Trans. Inst. Meas. Control, 14, 99–107, 1992.
Corser, G. A. and Hammond, G. C., A combined effects meter, InTech, April
LIMITATIONS OF CMFs 1993.
Cheesewright, R., Clark, C. and Bisset, D., The identification of external
1. CMF prices are rather high as compared to other mea- factors which influence the calibration of Coriolis mass flowmeters,
Flow Meas. Instrum., 11, 1–10, 2000.
suring device types. However, to measure mass flow
Cheesewright, R., Clark, C. and Bisset, D., The effect of flow pulsations on
with a volumetric meter, it is often necessary to install Coriolis mass flow meters, J. Fluids Struct., 1025–1039, 1998.
an in-line densitometer, which brings the cost up to Cox, B. M. and Gonzales, F. A., Coriolis mass flow rate, U.S. Patent,
roughly the equivalent of a CMF alone. 4127028, 1978.
2. There are no CMFs available for medium temperatures Drahm, W., New single straight tube Coriolis mass flowmeter without instal-
above 800°F (426°C). lation restrictions, in Proc. FLOMEKO1998 9th International Confer-
ence on Flow Measurement, Lund, Sweden, 243–248, 1998.
3. CMFs cannot be used for liquids with any significant Durst, F. and Raszillier, H., Flow in a rotation straight pipe, with view on
gas content. This effect can be reduced by increasing Coriolis mass flow meters, J. Fluids Eng., Trans. ASME, 112, 149–154,
the in-line pressure. 1990.

© 2003 by Béla Lipták

236 Flow Measurement

Eide, J. M. and Gwaspari, S. C., Comparison Test and Calibration of Coriolis Raszillier, H., Allenborn, N. and Durst, F., Effect of a concentrated mass on
Meters, North Sea Flow Measurement Workshop, Peebles, Scotland, Coriolis flowmetering, Arch. Appl. Mech., 64(6), 373–382, 1994.
1996. Rezende, V. A. and Apple, C., Coriolis Meters for LGP Custody Transfer
Ginesi, D. and Annarummo, C., Application and installation guidelines for at Petrobras, North Sea Flow Measurement Workshop, Kristiansand,
volumetric and mass flowmeters, ISA Trans., 33(1), 61–72, 1994. Norway, Paper 30, 1997.
Grini, P. G., Maehlum, H. S. and Brendeng, E., In situ calibration of Coriolis Rieder, A. and Drahm, W., A new type of single straight tube Coriolis mass
flowmeters for high-pressure gas flow calibration, J. Meas. Instrum., 5, flowmeter, in Proc. FLOMEKO 1996 8th International Conference on
285–288, 1994. Flow Measurement, Beijing, China, 250–254, 1996.
Hemp, J., A theoretical investigation into the feasibility of Coriolis mass flow- Robinson, C., Obstructionless flowmeters: smooth sailing for some, rough
meters for low density fluids, in Proc. FLOMEKO 1996 8th International passage for others, InTech, 33(12), 33–36, 1986.
Conference on Flow Measurement, Beijing, China, 265–270, 1996. Schietinger, M., Mass flow vs. volumetric flow, Meas. Control, September
Hemp, J., The weight vector theory of Coriolis mass flowmeters, Flow Meas. 1990.
Instrum., 5(4), 247–253, 1994. Sipin, A. J., Mass Flow Metering Means, U. S. Patent 3329019, July 4, 1967.
Henry, M., On-line compensation in a digital Coriolis mass flow meter, Flow Smith, J. E., Method and Structure of Flow Measurement, U. S. Patent
Meas. Instrum., 12, 147–161, 2001. 4187721, Feb. 12, 1980.
Kalotay, P., On-line viscosity measurement using Coriolis mass flowmeters, Spitzer, D. W., Flow Measurement, Practical Guides for Measurement and
Flow Meas. Instrum., 5, 303–308, 1994. Control, ISA, Research Triangle Park, NC, 1991.
Keita, N. M., Behaviour of straight type Coriolis mass flowmeters in the Sultan, G., Single straight tube Coriolis mass flowmeter, Flow Meas.
metering of gas: theoretical predictions with experimental verifications, Instrum., 3, 241–246, 1992.
Flow Meas. Instrum., 5, 289–294, 1994. Tokyokeiki, K. K., Straight tube type mass flowmeter, Patent Par. Applu.
Kolahi, K., Gast, Th. and Rock, H., Coriolis mass flow measurements of Disclosure 137 818/82, Tokyo, Japan, Aug. 25, 1982.
gas under normal conditions, Flow Meas. Instrum., 5, 275–283, 1994. Vetter, G. and Notzon, S., Effect of pulsating flow on Coriolis mass flow-
Medlock, R. and Furness, R. A., Mass flow measurement—a state of the art meters, Flow Meas. Instrum., 5, 263–273, 1994.
review, Meas. Control, 23, 100–112, 1990. Watt, R. M., Modelling of Coriolis mass flowmeters using ANSYS, in Proc.
Menke, D., Use of Coriolis mass flowmeters in custody transfer, in Proc. ANSYS User Conference, Pittsburgh, 1990.
FLOMEKO 1996 8th International Conference on Flow Measurement, Withers, V. R., Strang, W., and Allnutt, G., Practical Application of Coriolis
Beijing, China, 232–237, 1996. Meters for Offshore Tanker Loading from the Harding Field, North
Nicholson, S., Coriolis mass flow measurement, in Proc. FLOMEKO 1994 Sea Flow Measurement Workshop, East Kilbride, NEL, Scotland,
Conference on Flow Measurement in the Mid 90s, NEL, Scotland, 1994. 1996.
Pawlas, G. and Patten, T., Gas measurement using Coriolis mass flowmeters, Yamashita, Y., Development of Coriolis mass flowmeter with a single straight
ISA Advances in Instrumentation and Control: International Confer- tube as flow tube, in Proc. FLOMEKO 1996 8th International Confer-
ence and Exhibition, 50(3), 781–790, 1995. ence on Flow Measurement, Beijing, China, 265–270, 1996.
Plache, K., Coriolis/gyroscopic flow meter, Mech. Eng., 36, 1979. Young, A., Coriolis Flowmeters for Accurate Measurement of Liquid Prop-
Raszillier, H. and Durst, F., Coriolis effect in mass flow metering, Arch. erties, ISA Advances in Instrumentation and Control: International
Appl. Mech., 61, 192–214, 1991. Conference and Exhibition, 45(4), 1891–1898, 1990.

© 2003 by Béla Lipták