Vous êtes sur la page 1sur 33

Liquid Flowmeters

Flow Reference Section


An overview of types and capabilities, plus guidelines on selection,
installation, and maintenance
%RODUC%O
easuring the flow of liquids is a critical need in many industrial plants. In
some operations, the ability to conduct accurate flow measurements is so
important that it can make the difference between making a profit or taking
a loss. In other cases, inaccurate flow measurements or failure to take
measurements can cause serious (or even disastrous) results.
With most liquid flow measurement instruments, the flow rate is determined
inferentially by measuring the liquid's velocity or the change in kinetic
energy. Velocity depends on the pressure differential that is forcing the
liquid through a pipe or conduit. Because the pipe's cross-sectional area is
known and remains constant, the average velocity is an indication of the
flow rate. The basic relationship for determining the liquid's flow rate in such
cases is:
Q = V x A
where
Q = liquid flow through the pipe
V = average velocity of the flow
A = cross-sectional area of the pipe
Other factors that affect liquid flow rate include the liquid's viscosity and
density, and the friction of the liquid in contact with the pipe.
Direct measurements of liquid flows can be made with positive-displacement
flowmeters. These units divide the liquid into specific increments and move it
on. The total flow is an accumulation of the measured increments, which can
be counted by mechanical or electronic techniques.
Reynolds Numbers
The performance of flowmeters is also influenced by a dimensionless unit
called the Reynolds Number. It is defined as the ratio of the liquid's inertial
forces to its drag forces.

Figure 1: Laminar and turbulent flow are two types normally
encountered in liquid flow Measurement operations. Most
applications involve turbulent flow, with R values above 3000.
Viscous liquids usually exhibit laminar flow, with R values below
2000. %he transition zone between the two levels may be either
laminar or turbulent.
%he equation is:
R = 3160 x Q x Gt
D x
where:
R = Reynolds number
Q = liquid's flow rate, gpm
Gt = liquid's specific gravity
D = inside pipe diameter, in.
= liquid's viscosity, cp
The flow rate and the specific gravity are
inertia forces, and the pipe diameter and
viscosity are drag forces. The pipe diameter
and the specific gravity remain constant for
most liquid applications. At very low
velocities or high viscosities, R is low, and
the liquid flows in smooth layers with the
highest velocity at the center of the pipe and
low velocities at the pipe wall where the
viscous forces restrain it. This type of flow is
called laminar flow. R values are below
approximately 2000. A characteristic of
laminar flow is the parabolic shape of its
velocity profile, Fig. 1.
However, most applications involve
turbulent flow, with R values above 3000.
Turbulent flow occurs at high velocities or
low viscosities. The flow breaks up into
turbulent eddies that flow through the pipe
with the same average velocity. Fluid
velocity is less significant, and the velocity
profile is much more uniform in shape. A
transition zone exists between turbulent and
laminar flows. Depending on the piping
configuration and other installation
conditions, the flow may be either turbulent
or laminar in this zone.

FLOWME%ER %!ES
Numerous types of flowmeters are available for closed-piping systems. In
general, the equipment can be classified as differential pressure, positive
displacement, velocity, and mass meters. Differential pressure devices (also
known as head meters) include orifices, venturi tubes, flow tubes, flow
nozzles, pitot tubes, elbow-tap meters, target meters, and variable-area
meters, Fig. 2.
Positive displacement meters include piston, oval-gear, nutating-disk, and
rotary-vane types. Velocity meters consist of turbine, vortex shedding,
electromagnetic, and sonic designs. ass meters include Coriolis and
thermal types. The measurement of liquid flows in open channels generally
involves weirs and flumes.
Space limitations prevent a detailed discussion of all the liquid flowmeters
available today. However, summary characteristics of common devices are
shown in Table 1. (Click here to see Table1) Brief descriptions follow.
Differential !ressure Meters
The use of differential pressure as an inferred measurement of a liquid's rate
of flow is well known. Differential pressure flowmeters are, by far, the most
common units in use today. Estimates are that over 50 percent of all liquid
flow measurement applications use this type of unit.
The basic operating principle of differential pressure flowmeters is based on
the premise that the pressure drop across the meter is proportional to the
square of the flow rate. The flow rate is obtained by measuring the pressure
differential and extracting the square root.
Differential pressure flowmeters, like most flowmeters, have a primary and
secondary element. The primary element causes a change in kinetic energy,
which creates the differential pressure in the pipe. The unit must be properly
matched to the pipe size, flow conditions, and the liquid's properties. And,
the measurement accuracy of the element must be good over a reasonable
range. The secondary element measures the differential pressure and
provides the signal or read-out that is converted to the actual flow value.
Orifices are the most popular liquid flowmeters in use today. An orifice is
simply a flat piece of metal with a specific-sized hole bored in it. ost
orifices are of the concentric type, but eccentric, conical (quadrant), and
segmental designs are also available.
In practice, the orifice plate is installed in the pipe between two flanges.
Acting as the primary device, the orifice constricts the flow of liquid to
produce a differential pressure across the plate. Pressure taps on either side
of the plate are used to detect the difference. ajor advantages of orifices
are that they have no moving parts and their cost does not increase
significantly with pipe size.
Conical and quadrant orifices are relatively new. The units were developed
primarily to measure liquids with low Reynolds numbers. Essentially constant
flow coefficients can be maintained at R values below 5000. Conical orifice
plates have an upstream bevel, the depth and angle of which must be
calculated and machined for each application.
The segmental wedge is a variation of the segmental orifice. It is a
restriction orifice primarily designed to measure the flow of liquids containing
solids. The unit has the ability to measure flows at low Reynolds numbers
and still maintain the desired square-root relationship. Its design is simple,
and there is only one critical dimension the wedge gap. Pressure drop
through the unit is only about half that of conventional orifices.
Integral wedge assemblies combine the wedge element and pressure taps
into a one-piece pipe coupling bolted to a conventional pressure transmitter.
No special piping or fittings are needed to install the device in a pipeline.
etering accuracy of all orifice flowmeters depends on the installation
conditions, the orifice area ratio, and the physical properties of the liquid
being measured.
Venturi tubes have the advantage of being able to handle large flow
volumes at low pressure drops. A venturi tube is essentially a section of pipe
with a tapered entrance and a straight throat. As liquid passes through the
throat, its velocity increases, causing a pressure differential between the
inlet and outlet regions.
The flowmeters have no moving parts. They can be installed in large
diameter pipes using flanged, welded or threaded-end fittings. Four or more
pressure taps are usually installed with the unit to average the measured
pressure. Venturi tubes can be used with most liquids, including those
having a high solids content.
Flow tubes are somewhat similar to venturi tubes except that they do not
have the entrance cone. They have a tapered throat, but the exit is
elongated and smooth. The distance between the front face and the tip is
approximately one-half the pipe diameter. Pressure taps are located about
one-half pipe diameter downstream and one pipe diameter upstream.
Flow ozzles, at high velocities, can handle approximately 60 percent
greater liquid flow than orifice plates having the same pressure drop. Liquids
with suspended solids can also be metered. However, use of the units is not
recommended for highly viscous liquids or those containing large amounts of
sticky solids.
!itot tubes sense two pressures simultaneously, impact and static. The
impact unit consists of a tube with one end bent at right angles toward the
flow direction. The static tube's end is closed, but a small slot is located in
the side of the unit. The tubes can be mounted separately in a pipe or
combined in a single casing.
Pitot tubes are generally installed by welding a coupling on a pipe and
inserting the probe through the coupling. Use of most pitot tubes is limited
to single point measurements. The units are susceptible to plugging by
foreign material in the liquid. Advantages of pitot tubes are low cost,
absence of moving parts, easy installation, and minimum pressure drop.
Elbow meters operate on the principle that when liquid travels in a circular
path, centrifugal force is exerted along the outer edges. Thus, when liquid
flows through a pipe elbow, the force on the elbow's interior surface is
proportional to the density of the liquid times the square of its velocity. In
addition, the force is inversely proportional to the elbow's radius.
Any 90 deg. pipe elbow can serve as a liquid flowmeter. All that is required
is the placement of two small holes in the elbow's midpoint (45 deg. point)
for piezometer taps. Pressure-sensing lines can be attached to the taps by
using any convenient method.
%arget meters sense and measure forces caused by liquid impacting on a
target or drag-disk suspended in the liquid stream. A direct indication of the
liquid flow rate is achieved by measuring the force exerted on the target. In
its simplest form, the meter consists only of a hinged, swinging plate that
moves outward, along with the liquid stream. In such cases, the device
serves as a flow indicator.
A more sophisticated version uses a precision, low-level force transducer
sensing element. The force of the target caused by the liquid flow is sensed
by a strain gage. The output signal from the
gage is indicative of the flow rate. Target meters
are useful for measuring flows of dirty or
corrosive liquids.
Variable-area meters, often called rotameters,
consist essentially of a tapered tube and a float,
Fig. 3. Although classified as differential
pressure units, they are, in reality, constant
differential pressure devices. Flanged-end
fittings provide an easy means for installing
them in pipes. When there is no liquid flow, the
float rests freely at the bottom of the tube. As liquid enters the bottom of
the tube, the float begins to rise. The position of the float varies directly with
the flow rate. Its exact position is at the point where the differential pressure
between the upper and lower surfaces balance the weight of the float.
Because the flow rate can be read directly on a scale mounted next to the
tube, no secondary flow-reading devices are necessary. However, if desired,
automatic sensing devices can be used to sense the float's level and transmit
a flow signal. Rotameter tubes are manufactured from glass, metal, or
plastic. Tube diameters vary from 1/4 to greater than 6 in.
!ositive-Displacement Meters
Operation of these units consists of separating liquids into accurately
measured increments and moving them on. Each segment is counted by a
connecting register. Because every increment represents a discrete volume,
positive-displacement units are popular for automatic batching and
accounting applications. Positive-displacement meters are good candidates
for measuring the flows of viscous liquids or for use where a simple
mechanical meter system is needed.

Reciprocating piston meters are of the single and multiple-piston types.
The specific choice depends on the range of flow rates required in the
particular application. Piston meters can be used to handle a wide variety of
liquids. A magnetically driven, oscillating piston meter is shown in Fig. 4.
Liquid never comes in contact with gears or other parts that might clog or
corrode.
Oval-gear meters have two rotating, oval-shaped gears with synchronized,
close fitting teeth. A fixed quantity of liquid passes through the meter for
each revolution. Shaft rotation can be monitored to obtain specific flow
rates.
utating-disk meters have a moveable disk mounted on a concentric
sphere located in a spherical side-walled chamber. The pressure of the liquid
passing through the measuring chamber causes the disk to rock in a
circulating path without rotating about its own axis. It is the only moving
part in the measuring chamber.
A pin extending perpendicularly from the disk is connected to a mechanical
counter that monitors the disk's rocking motions. Each cycle is proportional
to a specific quantity of flow. As is true with all positive-displacement
meters, viscosity variations below a given threshold will affect measuring
accuracies. any sizes and capacities are available. The units can be made
from a wide selection of construction materials.
Rotary-vane meters are available in several designs, but they all operate
on the same principle. The basic unit consists of an equally divided, rotating
impeller (containing two or more compartments) mounted inside the meter's
housing. The impeller is in continuous contact with the casing. A fixed
volume of liquid is swept to the meter's outlet from each compartment as
the impeller rotates. The revolutions of the impeller are counted and
registered in volumetric units.
Helix flowmeters consist of two radically pitched helical rotors geared
together, with a small clearance between the rotors and the casing. The two
rotors displace liquid axially from one end of the chamber to the other.
'elocity Meters
These instruments operate linearly with respect to the volume flow rate.
Because there is no square-root relationship (as with differential pressure
devices), their rangeability is greater. Velocity meters have minimum
sensitivity to viscosity changes when used at Reynolds numbers above
10,000. ost velocity-type meter housings are equipped with flanges or
fittings to permit them to be connected directly into pipelines.

%urbine meters have found widespread use for accurate liquid
measurement applications. The unit consists of a multiple-bladed rotor
mounted with a pipe, perpendicular to the liquid flow. The rotor spins as the
liquid passes through the blades. The rotational speed is a direct function of
flow rate and can be sensed by magnetic pick-up, photoelectric cell, or
gears. Electrical pulses can be counted and totalized, Fig. 5.
The number of electrical pulses counted for a given period of time is directly
proportional to flow volume. A tachometer can be added to measure the
turbine's rotational speed and to determine the liquid flow rate. Turbine
meters, when properly specified and installed, have good accuracy,
particularly with low-viscosity liquids.
A major concern with turbine meters is bearing wear. A "bearingless" design
has been developed to avoid this problem. Liquid entering the meter travels
through the spiraling vanes of a stator that imparts rotation to the liquid
stream. The stream acts on a sphere, causing it to orbit in the space
between the first stator and a similarly spiraled second stator. The orbiting
movement of the sphere is detected electronically. The frequency of the
resulting pulse output is proportional to flow
rate.
Vortex meters make use of a natural
phenomenon that occurs when a liquid flows
around a bluff object. Eddies or vortices are
shed alternately downstream of the object.
The frequency of the vortex shedding is
directly proportional to the velocity of the liquid flowing through the meter,
Fig. 6.
The three major components of the flowmeter are a bluff body strut-
mounted across the flowmeter bore, a sensor to detect the presence of the
vortex and to generate an electrical impulse, and a signal amplification and
conditioning transmitter whose output is proportional to the flow rate, Fig. 7.
The meter is equally suitable for flow rate or flow totalization measurements.
Use for slurries or high viscosity liquids is not
recommended.
Electromagnetic meters can handle most liquids and
slurries, providing that the material being metered is
electrically conductive. ajor components are the flow
tube (primary element), Fig. 8. The flow tube mounts
directly in the pipe. Pressure drop across the meter is
the same as it is through an equivalent length of pipe
because there are no moving parts or obstructions to
the flow. The voltmeter can be attached directly to the
flow tube or can be mounted remotely and connected
to it by a shielded cable.
Electromagnetic flowmeters operate on Faraday's law
of electromagnetic induction that states that a voltage
will be induced when a conductor moves through a
magnetic field. The liquid serves as the conductor; the
magnetic field is created by energized coils outside the
flow tube, Fig. 9. The amount of voltage produced is
directly proportional to the flow rate. Two electrodes
mounted in the pipe wall detect the voltage, which is measured by the
secondary element.
Electromagnetic flowmeters have major advantages: They can measure
difficult and corrosive liquids and slurries; and they can measure forward as
well as reverse flow with equal accuracy. Disadvantages of earlier designs
were high power consumption, and the need to obtain a full pipe and no flow
to initially set the meter to zero. Recent improvements
have eliminated these problems. Pulse-type excitation
techniques have reduced power consumption, because
excitation occurs only half the time in the unit. Zero
settings are no longer required.
Ultrasonic flowmeters can be divided into Doppler
meters and time-of-travel (or transit) meters. Doppler
meters measure the frequency shifts caused by liquid flow. Two transducers
are mounted in a case attached to one side of the pipe. A signal of known
frequency is sent into the liquid to be measured. Solids, bubbles, or any
discontinuity in the liquid, cause the pulse to be reflected to the receiver
element, Fig. 10. Because the liquid causing the reflection is moving, the
frequency of the returned pulse is shifted. The frequency shift is proportional
to the liquid's velocity.
A portable Doppler meter capable of being operated on AC power or from a
rechargeable power pack has recently been developed. The sensing heads
are simply clamped to the outside of the pipe, and the instrument is ready to
be used. Total weight, including the case, is 22 lb. A set of 4 to 20
millampere output terminals permits the unit to be connected to a strip chart
recorder or other remote device.
Time-of-travel meters have transducers mounted on each side of the pipe.
The configuration is such that the sound waves traveling between the
devices are at a 45 deg. angle to the direction of liquid flow. The speed of
the signal traveling between the transducers increases or decreases with the
direction of transmission and the velocity of the liquid being measured. A
time-differential relationship proportional to the flow can be obtained by
transmitting the signal alternately in both directions. A limitation of time-of-
travel meters is that the liquids being measured must be relatively free of
entrained gas or solids to minimize signal scattering and absorption.

Mass Flowmeters The continuing need for more accurate flow
measurements in mass-related processes (chemical reactions, heat transfer,
etc.) has resulted in the development of mass flowmeters. Various designs
are available, but the one most commonly used for liquid flow applications is
the Coriolis meter. Its operation is based on the natural phenomenon called
the Coriolis force, hence the name.

Coriolis meters are true mass meters that measure the mass rate of flow
directly as opposed to volumetric flow. Because mass does not change, the
meter is linear without having to be adjusted for variations in liquid
properties. It also eliminates the need to compensate for changing
temperature and pressure conditions. The meter is especially useful for
measuring liquids whose viscosity varies with velocity at given temperatures
and pressures.
Coriolis meters are also available in various designs. A popular unit consists
of a U-shaped flow tube enclosed in a sensor housing connected to an
electronics unit. The sensing unit can be installed directly into any process.
The electronics unit can be located up to 500 feet from the sensor.
Inside the sensor housing, the U-shaped flow tube is vibrated at its natural
frequency by a magnetic device located at the bend of the tube. The
vibration is similar to that of a tuning fork, covering less than 0.1 in. and
completing a full cycle about 80 times/sec. As the liquid flows through the
tube, it is forced to take on the vertical movement of the tube, Fig. 11.
When the tube is moving upward during half of its cycle, the liquid flowing
into the meter resists being forced up by pushing down on the tube.

Having been forced upward, the liquid flowing out of the meter resists
having its vertical motion decreased by pushing up on the tube. This action
causes the tube to twist. When the tube is moving downward during the
second half of its vibration cycle, it twists in the opposite direction.
Having been forced upward, the liquid flowing out of the meter resists
having its vertical motion decreased by pushing up on the tube. This action
causes the tube to twist. When the tube is moving downward during the
second half of its vibration cycle, it twists in the opposite direction. The
ammount of twist is directly proportional to the mass flow rate of the liquid
flowing through the tube. agnetic sensors located on each side of the flow
tube measure the tube velocities, which change as the tube twists. The
sensors feed this information to the electronics unit, where it is processed
and converted to a voltage proportional to mass flow rate. The meter has a
wide range of applications from adhesives and coatings to liquid nitrogen.
%hermal-type mass flowmeters have traditionally been used for gas
measurements, but designs for liquid flow measurements are available.
These mass meters also operate independent of density, pressure, and
viscosity. Thermal meters use a heated sensing element isolated from the
fluid flow path. The flow stream conducts heat from the sensing element.
The conducted heat is directly proportional to the mass flow rate. The sensor
never comes into direct contact with the liquid, Fig.
12. The electronics package includes the flow
analyzer, temperature compensator, and a signal
conditioner that provides a linear output directly
proportional to mass flow.
Open Channel Meters
The "open channel" refers to any conduit in which
liquid flows with a free surface. Included are tunnels,
nonpressurized sewers, partially filled pipes, canals,
streams, and rivers. Of the many techniques available for monitoring open-
channel flows, depth-related methods are the most common. These
techniques presume that the instantaneous flow rate may be determined
from a measurement of the water depth, or head. Weirs and flumes are the
oldest and most widely used primary devices for measuring open-channel
flows.
Weirs operate on the principle that an obstruction in a channel will cause
water to back up, creating a high level (head) behind the barrier. The head
is a function of flow velocity, and, therefore, the flow rate through the
device. Weirs consist of vertical plates with sharp crests. The top of the plate
can be straight or notched. Weirs are classified in accordance with the shape
of the notch. The basic types are V-notch, rectangular, and trapezoidal.
Flumes are generally used when head loss must be kept to a minimum, or if
the flowing liquid contains large amounts of suspended solids. Flumes are to
open channels what venturi tubes are to closed pipes. Popular flumes are the
Parshall and Palmer-Bowlus designs.
The Parshall flume consists of a converging upstream section, a throat, and
a diverging downstream section. Flume walls are vertical and the floor of the
throat is inclined downward. Head loss through Parshall flumes is lower than
for other types of open-channel flow measuring devices. High flow velocities
help make the flume self-cleaning. Flow can be measured accurately under a
wide range of conditions.
Palmer-Bowlus flumes have a trapezoidal throat of uniform cross section and
a length about equal to the diameter of the pipe in which it is installed. It is
comparable to a Parshall flume in accuracy and in ability to pass debris
without cleaning. A principal advantage is the comparative ease with which it
can be installed in existing circular conduits, because a rectangular approach
section is not required.
Discharge through weirs and flumes is a function of level, so level
measurement techniques must be used with the equipment to determine
flow rates. Staff gages and float-operated units are the simplest devices
used for this purpose. Various electronic sensing, totalizing, and recording
systems are also available.
A more recent development consists of using ultrasonic pulses to measure
liquid levels. easurements are made by sending sound pulses from a
sensor to the surface of the liquid, and timing the echo return. Linearizing
circuitry converts the height of the liquid into flow rate. A strip chart
recorder logs the flow rate, and a digital totalizer registers the total gallons.
Another recently introduced microprocessor-based system uses either
ultrasonic or float sensors. A key-pad with an interactive liquid crystal
display simplifies programming, control, and calibration tasks.
SELEC%G A FLOWME%ER
Experts claim that over 75 percent of the flowmeters installed in industry are
not performing satisfactorily. And improper selection accounts for 90 percent
of these problems. Obviously, flowmeter selection is no job for amateurs.
The major steps involved in the selection process are shown in Fig. 13.
The most important requirement is knowing exactly what the instrument is
supposed to do. Here are some questions to consider. Is the measurement
for process control (where repeatability is the major concern), or for
accounting or custody transfer (where high accuracy is important)? Is local
indication or a remote signal required? If a remote output is required, is it to
be a proportional signal, or a contact closure to start or stop another device?
Is the liquid viscous, clean, or a slurry? Is it electrically conductive? What is
its specific gravity or density? What flow rates are involved in the
application? What are the processes' operating temperatures and pressures?
Accuracy (see glossary), range, linearity, repeatability, and piping
requirements must also be considered.
It is just as important to know what a flowmeter cannot do as well as what it
can do before a final selection is made. Each instrument has advantages and
disadvantages, and the degree of performance satisfaction is directly related
to how well an instrument's capabilities and shortcomings are matched to
the application's requirements. Often, users have expectations of a
flowmeter's performance that are not consistent with what the supplier has
provided. ost suppliers are anxious to help customers pick the right
flowmeter for a particular job. any provide questionnaires, checklists, and
specification sheets designed to obtain the critical information necessary to
match the correct flowmeter to the job.
Technological improvements of flowmeters must be considered also. For
example, a common mistake is to select a design that was most popular for
a given application some years ago and to assume that it is still the best
instrument for the job. any changes and innovations may have occurred in
recent years in the development of flowmeters for that particular application,
making the choice much broader.
A recent development is the availability of computer programs to perform
the tedious calculations often necessary for selecting flowmeters.
Calculations that used to take an hour can be performed in a matter of
seconds (see accompanying section, "Selected Reference aterial").
Click Here For Questions To Ask When Selecting A Flowmeter
Cost Considerations
There are a wide range of prices for flowmeters. Rotameters are usually the
least expensive, with some small-sized units available for less than $100.
ass flowmeters cost the most. Prices start at about $3500. However, total
system costs must always be considered when selecting flowmeters. For
example, an orifice plate may cost only about $50. But the transmitter may
add an additional $500 or $600, and sensing line fabrication and installation
may cost even more.
Installation, operation, and maintenance costs are important economic
factors too. Servicing can be expensive on some of the more complicated
designs.
As with many other products, a plant engineer generally gets what he pays
for when he purchases a flowmeter. But the satisfaction that he receives
with the product will depend on the care that he uses in selecting and
installing the instrument. And that gets back to knowing the process, the
products, and the flow-metering requirements. "Overbuying" is not
uncommon. Plant engineers should not buy a flowmeter more capable or
complicated than they need.
WORKG W% FLOWME%ERS
Although suppliers are always ready to provide flowmeter installation
service, estimates are that approximately 75 percent of the users install
their own equipment. But installation mistakes are made. One of the most
common is not allowing sufficient upstream and downstream straight-run
piping for the flowmeter.
Every design has a certain amount of tolerance to nonstable velocity
conditions in the pipe, but all units require proper piping configurations to
operate efficiently. Proper piping provides a normal flow pattern for the
device. Without it, accuracy and performance are adversely affected.
Flowmeters are also installed backwards on occasion (especially true with
orifice plates). Pressure-sensing lines may be reversed too.
With electrical components, intrinsic safety is an important consideration in
hazardous areas. ost flowmeter suppliers offer intrinsically safe designs for
such uses.
Stray magnetic fields exist in most industrial plants. Power lines, relays,
solenoids, transformers, motors, and generators all contribute their share of
interference. Users must ensure themselves that the flowmeter they have
selected is immune to such interference. Problems occur primarily with the
electronic components in secondary elements, which must be protected.
Strict adherence to the manufacturer's recommended installation practices
will usually prevent such problems.
Calibration
All flowmeters require an initial calibration. ost of the time, the instrument
is calibrated by the manufacturer for the specified service conditions.
However, if qualified personnel are available in the plant, the user can
perform his own calibrations.
The need to recalibrate depends to a great extent on how well the meter fits
the application. Some liquids passing through flowmeters tend to be
abrasive, erosive, or corrosive. In time, portions of the device will
deteriorate sufficiently to affect performance. Some designs are more
susceptible to damage than others. For example, wear of individual turbine
blades will cause performance changes. If the application is critical,
flowmeter accuracy should be checked at frequent intervals. In other cases,
recalibration may not be necessary for years because the application is
noncritical, or nothing will change the meter's performance. Some
flowmeters require special equipment for calibration. ost manufacturers will
provide such service in their plant or in the user's facility, where they will
bring the equipment for on-site calibration.
Maintenance
A number of factors influence maintenance requirements and the life
expectancy of flowmeters. The major factor, of course, is matching the right
instrument to the particular application. Poorly selected devices invariably
will cause problems at an early date. Flowmeters with no moving parts
usually will require less attention than units with moving parts. But all
flowmeters eventually require some kind of maintenance.
Primary elements in differential pressure flowmeters require extensive
piping, valves, and fittings when they are connected to their secondary
elements, so maintenance may be a recurring effort in such installations.
Impulse lines can plug or corrode and have to be cleaned or replaced. And,
improper location of the secondary element can result in measurement
errors. Relocating the element can be expensive.
Flowmeters with moving parts require periodic internal inspection, especially
if the liquid being metered is dirty or viscous. Installing filters ahead of such
units will help minimize fouling and wear. Obstructionless instruments, such
as ultrasonic or electromagnetic meters, may develop problems with their
secondary element's electronic components. Pressure sensors associated
with secondary elements should be periodically removed and inspected.
Applications where coatings may occur are also potential problems for
obstructionless instruments such as magnetic or ultrasonic units. If the
coating is insulating, the operation of magnetic flowmeters will ultimately be
impaired if the electrodes are insulated from the liquid. This condition will be
prevented by periodic cleaning. With ultrasonic flowmeters, refraction angles
may change and the sonic energy absorbed by the coating will cause the
meter to become inoperative.







Velocity-Profile Deviations Influence Flow meter Performance
Selecting the right flow meter for the liquid

Close your eyes and imagine a world in which all Iluids Ilowing down pipes are
perIectly homogeneous, having no disturbances or eddies in the Ilow stream. A world
in which pipes always contain Iully developed turbulent Ilow and pipe bends and
obstructions are non-existent.
Sounds silly right? But these are the hypothetical conditions upon which Ilowmeter
accuracies are based. II deviations Irom these conditions exist, accuracy and/or
repeatability will degrade. This article will discuss the Iirst oI Iour major inIluence
quantities that aIIect Ilowmeter accuracy and perIormance: velocity-proIile
deviations, non-homogenous Ilow, pulsating Ilow, and cavitation. With this
background, we will then look at how these inIluence quantities Iactor into Ilowmeter
perIormance and what can be done to minimize their eIIect.
Imagine a horizontal pipe a Iew Ieet in Iront oI you at about eye level. II a vertical cut
were made through the center oI the pipe to remove the Iront halI oI the pipe, it would
be easy to see how "nicely behaved" Iluid Ilow develops.

I|gure 1 2ero L|qu|d V|scos|ty
In the hypothet|ca| s|tuat|on
shown here the ||qu|d v|scos|ty
|s zero and the ve|oc|ty prof||e |s
a stra|ght ||ne 1he vert|ca| sheet
of f|u|d moves forward at
ve|oc|ty V
The cross-sectional view shown in Figure 1
illustrates a hypothetical situation in which the Iluid
has zero viscosity. In this example, the velocity
gradient is constant and equal across the entire
cross-section. That is, the Iluid velocity at the pipe
walls is equal to the Iluid velocity at the pipe center
and at all points in-between. However, every liquid
has some measure oI viscosity; thereIore, the
velocity proIile becomes distorted Irom this "ideal
case scenario."



-ewton|an Ve|oc|ty rof||es
xamples oI Newtonian Iluids include water, milk, sugar solutions and mineral oils.
Distortions can occur Ior homogeneous Newtonian Iluids Ior a number oI reasons.
The viscosity oI a Newtonian Iluid is dependent only on temperature, but not on shear
rate and time. The Ilow oI a Newtonian Iluid will exhibit a "sticking" eIIect on the
walls oI the pipe. In Iact, the boundary conditions placed on the stationary pipe wall
demand that the Iluid in direct contact with the pipe walls have zero velocity (or at
least that is what is assumed in the calculations within this article). It is this boundary
condition that distorts a Newtonian Iluid's velocity proIile.
II the velocity vector at the pipe wall is zero, the maximum velocity occurring in the
pipe can be worked out mathematically.

I|gure 2 Samp|e -ewton|an
I|u|d Ve|oc|ty rof||e
Ior steady f|ow |n a p|pe of
rad|us k conta|n|ng a sma||er
cy||ndr|ca| e|ement of ||qu|d
rad|us r a shear force
deve|ops equa| and oppos|te to
the cy||ndr|ca| water e|ement
pressure
Consider a pipe oI radius "R" that contains a smaller
cylindrical element oI liquid oI radius "r." See
Figure 2.
Assuming the cylindrical element oI Iluid is moving
uniIormly through the pipe, at constant velocity, the
shearing Iorce on the pipe wall is equal to the
pressure Iorce oI the liquid moving through. The
shear Iorce, which is directly proportional to the
velocity gradient, is given by:
(dv/dy)
where:
is the shear Iorce or shear pressure.
dv/dy is the velocity gradient (where y is the
distance Irom the pipe wall).
is the constant oI proportionality, also reIerred to
as the dynamic viscosity.
Because the Iluid pressure can be simply given as
the pressure diIIerential through the pipe times the
cross-sectional Iluid area, and because the Iluid
pressure Iorce is equal to the oppositely directed
shear Iorce on the pipe wall, the equation can be written:
Shear Wall Force (2nr)L 2nrL(dv/dy) APA APnr
2
Fluid Pressure
where:
r is the radius oI the cylindrical Iluid element.
L is the length oI the cylindrical Iluid element.
AP is the pressure diIIerential through the pipe.
A is the cross-sectional area oI the Iluid element
Solving Ior t, we get
APr/2L
It is more convenient in this example to give the shear Iorce as a Iunction oI the pipe
center instead oI the pipe wall. This is easily accomplished easily by rewriting the
shear Iorce as:
-(dv/dr) (dv/dy)
quating these two expressions Ior the shear stress t gives us the Iollowing
relationship:
APr/2L -(dv/dr)
Solving Ior the velocity gradient yields:
dv/dr -APr/2L
By integrating this expression across the pipe to radius R, the velocity (V) as a
Iunction oI the radius r can be determined.
V AP/4L(R
2
- r
2
)
Now it can be seen mathematically that the maximum velocity in the pipe occurs at
the center oI the pipe, or when r 0. Actually, these two Iactsthat the velocity at the
pipe wall is zero and the maximum velocity occurs at the pipe centerare the
boundary conditions used to determine the constant oI proportionality Ior the above
integration.
This equation is in the Iorm oI a parabola. Graphing it gives the velocity proIile
shown in Figure 3.

I|gure 3 Lam|nar I|ow rof||e
for -ewton|an I|u|d
1he ve|oc|ty |s zero at the p|pe
wa|| and |ncreases parabo||ca||y
w|th f|ow reach|ng |ts max|mum
at the p|pes center
This proIile develops at certain velocities Ior
Newtonian Iluids and is called laminar Ilow.
Laminar Ilow is characterized by Iluid movement in
layers or streamlines with very little mixing
occurring between those layers. The Iluid layers are
restrained Irom mixing by the viscous Iorces within
the liquid.
Laminar Ilow is very well behaved and predictable.
However, as the velocity oI the Iluid increases
within the pipe, inertial Iorces start to overcome the
viscous Iorces, and small eddies and Iluid
oscillations will Iorm, causing the Iluid layers to
start mixing.
For a while, the Iluid might shiIt to being somewhat laminar, but the eddies keep
working to mix the Iluid layers. This partial mixing oI the streamlined layers is known
as transitional Ilow.
Transitional Ilow is not really laminar Ilow but it is not Iully turbulent Ilow either. In
Iact, the Iluid may exhibit laminar Ilow at some pipe radii and turbulent Ilow at other
radii. II a dye stream were to be injected into the pipe, it would waver and mix
slightly as it moves down the pipe. The velocity proIile Ior transitional Ilow is
diIIicult to predict because it changes with time, being laminar-like in some instances
and turbulent-like in others.
As the velocity oI the Iluid within the pipe increases Iurther, Iully turbulent Ilow
develops. Turbulent Ilow is characterized by the laminar Ilow streams completely
breaking up and mixing together as the Ilow moves down the pipe. Although the
average Ilow still moves down the pipe, there will be small Ilow velocities in the
radial r direction as well. As eddies Iorm in the Iluid stream, they will break down
randomly into smaller eddies, swirls and vortices, leading to increased shear within
the liquid.
Is there any way to predict when each oI these regimeslaminar, transitional and
turbulent Ilowwill dominate the Iluid Ilow? The answer, oI course, is yes. The
Reynolds number helps determine what type oI Ilow can be expected. Named aIter the
Iluid dynamics research oI Osborne Reynolds, the Reynolds number is simply the
ratio oI momentum to viscous Iorces. It takes into account the Iluid density, viscosity,
and pipeline velocity. II the viscous Iorces are large (low Reynolds number), the
viscous eIIects oI the Iluid will dampen out any tendency oI streamline mixing. For
large inertial Iorces, there will not be enough viscous damping Ior laminar Ilow to
continue, and the Ilow will become transitional or turbulent. The equation Ior the
Reynolds number is:
R vD/
where:
v is the average velocity.
D is the pipe diameter.
is the density.
is the dynamic viscosity.
Using SI units, it becomes apparent that the Reynolds number is dimensionless. The
units oI vDp are (meters/sec) x (meters) x (kg/meters
3
), and the units oI are
(kg/meters x sec). The rule oI thumb is that the Ilow will be laminar below an R oI
2,000. Between an R oI 2,000 and 4,000, the Ilow is transitional; above 4,000, the
Ilow is turbulent. In actuality, there are no distinct breaks between laminar,
transitional, and turbulent Ilows. For example, depending on the Iluid, Iully developed
transitional Ilow might not occur until R equals roughly 7,000.
So, what is the velocity proIile Ior turbulent Ilow? This is a diIIicult question as the
proIile is not a Iixed geometry. It changes with the wall roughness and Reynolds
number. The velocity equation Ior homogeneous Newtonian Ilow in the turbulent
regime (see Figure 4) is:
V
r
V
av
1 + 1.44\f + 2.15\f log
10
(1-r/R)]
where:
V
r
is the velocity at radius r.
V
av
is the average Ilow velocity.
R is the radius oI the pipe.
f is the Iriction Iactor.
The Iriction Iactor is diIIicult to calculate because turbulent Ilow is unpredictable, but
a lot oI experimental data exist on this subject. The Iriction Iactor comes into play
when the pipe wall is not smooth. For a rough pipe interior, the Iluid closer to the
walls is held back because oI the additional shear. This, in turn, causes the velocity
proIile to become Ilatter and blunter in Iront.

I|gure 4 1urbu|ent I|ow rof||e
for a -ewton|an I|u|d
1he ve|oc|ty |s zero at the p|pe
wa|| but the face ve|oc|ty |s
stra|ghter and squared up
One widely used method oI determining the Iriction
Iactor is the Moody diagram. Moody-diagram-
related reIerences are included in the bibliography.
The velocity equation shows that the maximum Ilow
will occur at r 0, just like it did Ior laminar Ilow.
At r 0, the equation becomes:
V
max
V
av
(1 + 1.43\f)
Figure 4 shows that the Iace velocity is more
uniIorm and squared up. As the velocity oI the Iluid
continues to increases, the Iace velocity will
continue to straighten up until all particles are
moving at the same velocity (except at the pipe walls, where the Ilow will remain at
zero). This hypothetical situation, where the Iace velocity is completely Ilat, is known
as inIinite Reynolds number proIile, or plug Ilow.


-on-ewton|an I|u|ds
How does the velocity proIile change when the ratio oI shear stress to shear strain is
no longer constant? BeIore attempting to answer this question, it is helpIul to outline
the diIIerent types oI non-Newtonian Iluids.
lass I: time-independent non-Newtonian fluids. These Iluids have a viscosity at a
given shear stress that does not vary with time. Class 1 nonNewtonian Iluids include
pseudoplastic, dilatant and plastic Iluids.
Pseudoplastic Iluids, also known as shear-thinning Iluids, decrease in viscosity as the
shear rate increases. xamples include paints, shampoos and water suspensions oI
clay.
Dilatant Iluids, also known as shear-thickening Iluids, increase in viscosity as the
shear rate increases. xamples include corn starch in water, titanium dioxide, and wet
sand.
Plastic Iluids, also called Bingham or plug-Ilow Iluids, behave as a solid until a
critical shear ratecalled the yield valueis reached. At the yield value, the Iluid
will start to Ilow. As the shear rate continues to increase, the Iluid then might exhibit
Newtonian, dilatent or pseudoplastic characteristics. xamples include tomato paste,
toothpaste, hand cream, chocolate, mayonnaise, and grease.
lass II: time-dependent non-Newtonian fluids. These Iluids have a viscosity at a
given shear stress that will vary with time. They include thixotropic and rheopectic
Iluids.
Thixotropic Iluids, also known as time-thinning Iluids, experience a decrease in
viscosity over time while the rate oI shear is constant. xamples include yogurt and
paint.
Rheopectic Iluids, also called time-thickening Iluids, exhibit an increase in viscosity
over time as the shear remains constant. Rheopectic Iluids are pretty rare. xamples
include gypsum paste and printers ink.
What makes the Ilow oI nonNewtonian Iluids so interesting is the required boundary
condition oI zero Ilow at the pipe walls. This boundary condition imparts a varying
shear rate into the Iluid as the Iluid velocity changes. And because a nonNewtonian
Iluid exhibits changing viscosity with changing shear rates, the Iluid can react in a
wide variety oI ways. Like Newtonian Iluids, non-Newtonian Iluids have laminar,
transitional, and turbulent Ilow patterns.
For time-independent nonNewtonian Iluids, the Iluid viscosity does not change with
time, assuming a nonchanging shear rate. As the Iluid moves down the pipe, the
velocity distribution between the Iluid layers varies and changes with the varying
shear that results Irom the zero velocity boundary condition at the walls oI the pipe.
Another important condition is the apparent viscosity oI the liquid. Because the Iluid
viscosity changes with the shear rate, the apparent viscositythe viscosity at a
particular rate oI shearmust be used to calculate the Ilow proIile instead oI the
dynamic or absolute viscosity discussed Ior Newtonian Iluids. This makes the velocity
proIile impossible to predict completely and correctly.

I|gure S 1|meIndependent
-on-ewton|an Lam|nar I|ow
rof||es
Shown are prof||es of
pseudop|ast|c d||atant and
p|ast|c ||qu|ds over the prof||e of
a -ewton|an ||qu|d
However, the velocity proIile still has some inherent
symmetry. For the laminar Ilow region, the velocity
proIile is calculated as a deviation or departure Irom
the laminar Newtonian proIile. Figure 5 lays the
typical velocity proIiles oI pseudoplastic, dilatant,
and plastic liquids over the proIile oI a Newtonian
liquid, revealing the distortions that can occur Ior
time-independent nonNewtonian laminar Ilows.
The Iace proIile Ior a plastic Iluid is very Ilat. The
entire Ilat Iace is known as the plug diameter and
can vary greatly, depending on the type oI plastic
Iluid moving down the pipe, as well as on the
upstream pressure. Remember, plastic Iluids will
behave as a solid until shear stress reaches the
critical yield value. They then will start Ilowing.
NonNewtonian Iluids can be characterized by a Reynolds number whereby they will
move Irom laminar to transitional to turbulent Ilows. This Reynolds number,
however, is calculated diIIerently Ior pseudoplastic, dilatant, and plastic Iluids. For
Iully developed turbulent Ilow, Ilow proIiles are similar enough to Newtonian Iluids
that the Newtonian Reynolds number equation can be used iI apparent viscosity is
substituted Ior dynamic viscosity.
Thixotropic Iluids can experience some interesting eIIects, depending on whether the
shear increases, decreases, or stays constant. A thixotropic Iluid that increases in shear
rate up to a constant value, dwelling at that value Ior a time beIore dropping back
down to zero shear, can experience a decrease in apparent viscosity during the dwell
time. When the shear rate is brought back down to zero, a yield value can be reached
that will mimic the eIIects oI a plastic Iluid. The Ilow proIile will go Irom parabolic
shaped to plug-Ilow shaped. OI course, this type oI pipeline scenario is extremely
rare.
Because rheopectic Iluids increase in apparent viscosity with time at constant shear,
they are sometimes known as negative thixotropic Iluids. Rheopectic Iluids will
exhibit the same type hysteresis curve as thixotropic Iluids, but generally return to the
same initial viscosity once the rate oI shear drops to zero. Like thixotropic Iluids,
rheopectic Iluids moving through a pipe are normally in the laminar regime and will
typically exhibit the normal-to-slightly distorted parabolic curves already discussed,
assuming that the Ilow has had enough time to Iully develop through a straight section
oI pipe.
An excellent and much deeper discussion oI Ilow prIiles and how they develop in pipe
Ilow can be Iound in the Flow Measurement and ngineering Handbook, written by
Richard Miller (Cole-Parmer #00545-08).


I|owmeter erformance and I|ow rof||e D|stort|ons
In each case discussed previously, the Ilow is through a long, straight section oI
piping, and the velocity proIile had time to develop and stabilize as it moved down the
pipe. This stabilization and development oI the velocity proIile is known as Iully
developed Ilow. Although it is great Ior doing Iluid dynamics calculations, in the real
world, pipes undergo bends, turns, reductions, and enlargements as they twist their
way through a typical plant. What happens to these velocity proIiles under some oI
these conditions, and how do the velocity proIiles aIIect Ilowmeter placement and
Ilowmeter perIormance?
In general, because the Ilow proIile is smooth and Ilat over most oI the pipe diameter,
turbulent Ilow is preIerred Ior a Ilowmeter over laminar and transitional Ilows. It's
much easier to calibrate and check the repeatability oI a Ilowmeter using a Iully
developed turbulent Ilow proIile than it is using the parabolic Ilow proIile oI laminar
Ilow or the unstable Ilow proIile oI transitional Ilow.
This is not mean laminar Ilow is bad Ior Ilowmeters. In Iact, stable laminar Ilow is
Iairly easy to measure accurately, providing the appropriate technology is used.
The Iirst step in choosing a particular Ilow technology is to Iind out its acceptable
Reynolds number range and then to calculate the Reynolds extremes oI the
application. A minimum Reynolds number is calculated by using the minimum
expected Iluid Ilow and density and the maximum expected viscosity. A maximum
Reynolds number is calculated using the maximum expected Iluid Ilow and density
and the minimum expected viscosity.
Another way to assess the conditions oI a Ilow application is through use oI Doppler
or transit-time Ilow technologies.
Paddle-Wheel Flowmeters: These use a small rotor usually having Iour blades
embedded with magnets. The paddle is inserted into the pipe at a Iixed depth; the rate
at which it turns is proportional to the Ilowrate.
One oI the advantages oI the paddle-wheel meter is it will work with a large range oI
pipe sizes. Typically, the paddle-wheel sensor can be installed in plastic or metal
pipes Irom 1/2 inch (in.) up to 36 in. or larger. The paddlewheel itselI is only an inch
or so in diameter, so iI it is installed in a large pipe, it would have problems
measuring laminar Ilow (see Figure 6).


I|gure 6 add|eWhee|
I|owmeter Insta||at|on Lxamp|e
A|though they can be used |n
stab|e |am|nar f|ow app||cat|ons
to read max|mum or average
f|ow padd|ewhee| f|owmeters
are more accurate when used
w|th fu||y deve|oped turbu|ent
f|ow
Adjustable insertion depth sensors are available to
allow exact positioning oI the paddle-wheel within
the pipe, but they have limited value with laminar
Ilow and assume that the ratio between the
measured velocity and the average Iluid velocity is
stable, which is not always the case. II an insertion
sensor must be used in a laminar Ilow application,
plant operators should try to Iind the point where the
Iluid velocity is related to the average velocity on a
one-to-one basis. This point will be roughly one-
eighth oI the pipe diameter Irom the inside pipe
wall, and should be Iairly independent oI the
Reynolds number.
For smaller pipe sizes oI a Iew inches or so, this is a
lesser problem, and center-line insertion might be
the best option. For a 20-inch pipe however, the
Ilow proIile gradient can be too sharp to allow
accurate or consistent measurement by correlating a
speciIic velocity within the pipe to an average pipe
velocity. Similarly, Ior transitional Ilow, because the Ilow proIile is unstable, the
paddle-wheel design does not oIIer accurate or repeatable results. Because turbulent
Ilow proIile is blunted, Ilat and stableassuming it has had time to Iully stabilize
while Ilowing down the pipeit is the best proIile when using a paddle-wheel sensor.
In Iact, manuIacturers will calibrate their paddle sensors to Iully developed turbulent
Ilow within their Ilow calibration loop; thereIore, paddle-wheel manuIacturers have
very speciIic upstream and downstream straight piping requirements.
Venturi-one Differential Pressure Flowmeters: These Ilowmeters oIIer quite a
design contrast to paddle-wheel Ilowmeters. Developed in the early 1980's, the
Venturi-cone, or V-cone Ior short, uses a cone positioned in the pipe's center. The
cone geometry and placement Iorce the Iluid to move around the cone, generating a
pressure drop across the length oI the cone. See Figure 7. This pressure drop is
proportional to the Ilow velocity within the pipe.






I|gure 7 Vent
ur|Cone I|owmeter Insta||at|on
Lxamp|e
In th|s |nsta||at|on the
suspended cone reshapes the
approach|ng f|ow prof||e
upstream
The cone is always suspended at the pipe center, so
Ior laminar Ilow, the maximum Iluid Ilow is driven
into the "point" oI the cone. This eIIectively Iorces
the high-velocity portion oI the laminar Ilow stream
to mix with the lower velocity layers, causing the
upstream Ilow proIile to mix and Ilatten into a more
well-developed turbulent pattern.
The V-cone's ability to Ilatten irregular Ilow proIiles
is quite strong and even allows the V-cone meter to
be installed Iairly close to pipe elbows and bends,
something usually avoided with other Ilow
technologies. The meter acts as its own Ilow
conditioner, shaping irregular or laminar Ilow
proIiles into more stable turbulent-like Ilow proIiles. This allows the technology to be
used over a wide range oI Reynolds number.
%urbine Flowmeters: These Ilowmeters use a rotating turbine placed directly in the
Iluid path. Although designs vary, a pulsing signal usually is generated as the turbine
blades pass a magnetic pickup coil. The Irequency oI the pulses is proportional to the
Iluid velocity in the pipe. For laminar and transitional Ilow patterns (i.e., having low
Reynolds numbers), the number oI pulses per unit volume, the K-Iactor, can vary,
causing a loss oI accuracy and repeatability. For this reason, turbine meters are best
used in turbulent Ilow conditions. Many turbine manuIacturers that build Ilow
straighteners directly into the body oI the Ilowmeter to minimize any upstream swirls
generated by pipe bends, which also can degrade accuracy and repeatability. Insertion
turbine meters, like the paddle-wheel sensors, will measure the velocity at only a
particular point within the pipe. ThereIore, they are better suited to Iully developed
turbulent Ilow.
Magnetic Flowmeters: These Ilowmeters measure the average Ilow over the pipe
diameter and show very little perIormance change in going Irom laminar to
transitional to Iully turbulent Ilow. Insertion mag meters will measure the Ilow within
a localized point around the sensor.
In both designs, however, swirling eddies or vortices that occur in pipe bends and
turns can pose problems. Also, because velocity proIile distortions can occur in
several Iluid planes simultaneously, velocity averaging Ior Iull-bore magnetic
Ilowmeters has its limitations. For this reason, manuIacturers usually recommend Iive
pipe diameters oI straight run pipe upstream and two pipe diameters downstream to
stabilize any variations in the Ilow proIile.
Gear Flowmeters: These Ilowmeters work by Iorcing speciIic volumes oI Iluid
between two coupled gears. The Iluid normally is trapped between the gear and the
inner Ilowmeter housing. AIter the number oI Iluid pockets passing through in a
particular time period is determined, the Ilow can be calculated. See Figure 8.

I|gure 8 Gear I|owmeter
W|th each gear rotat|on a pocket of prec|se|y
measured f|u|d |s formed he|d t|ght|y aga|nst
the chamber wa|| then re|eased
This positive-displacement Ilowmeter
has an accuracy that is independent oI
the Reynolds number. It typically works
best when there is a little back-pressure
downstream. This can be achieved by
the use oI a control valve downstream,
throttled just enough to allow a slight
back-pressure. The back-pressure, in
turn, helps ensure that the Iluid pockets
within the Ilow cavity are Iull
throughout meter operation.
Because the Ilowmeter is grabbing speciIic volumes oI Iluid and moving them Irom
the inlet to the outlet, the meter is not aIIected by the Reynolds number. The Ilow can
be laminar, transitional, or turbulent and still work well with gear design. One
caution, however: Because thin, water-like Iluids have the ability to slip past the
gears, accuracy will not be as high as it is when the Iluid is thicker. A general rule oI
thumb is not to use gear meters Ior liquids under 5 centipoise.
Co Lo 1op


Assess|ng |pe I|ow Cond|t|ons
lbows, reducers, chemical injection ports, Iilters, screens and valves also can cause
radial, tangential and axial swirling eIIects within the pipe. In combination, these
changes can rapidly distort the velocity proIile, degrading the Ilowmeter's accuracy
and repeatability. This is one reason why manuIacturers give straight-length upstream
and downstream piping requirements Ior installation.
It is important to gain an understanding oI the conditions oI a pipe's Ilow to select a
Ilowmeter that will be acceptable Ior worst-case conditions. II the Iluid conditions
within a pipe are unkown or the actual Ilow proIile in certain piping sections needs to
be studied, a special type oI non-contact ultrasonic Ilowmeter can be oI great beneIit.
Known as pulsed-Doppler Ilowmeters, the devices work by allowing true velocity
measurements to be made at multiple points along the pipe diameter. This can help
assess the Ilow proIile in larger pipes, but because the pulses are typically a Iew
centimeters long, they are oI little help in smaller pipes. Also, the Iluid within the pipe
must contain either particulate matter or bubbles Ior the Doppler signal to reIlect
properly.
An insertion meter also can be used to help determine the Ilow proIile in a pipe. It
measures the point velocity at several transverse locations within the pipe. Although
the meter does a pretty good job in determining the velocity proIile, any change in the
Ilow conditions can alter the Ilow proIile itselI. Once some inIormation on the Ilow
proIile has been gleaned, meter selection and installation will be improved.
II proIiling the Ilow within the pipe is not practical or not desired, the Ilow must be
stable enough to get the maximum accuracy possible with the Ilow technology being
used. This can be accomplished through strict adherence to the manuIacturer's
recommended upstream and downstream piping requirements. II these straight-length
requirements are not practical in the plant, Ilow conditioners and straighteners could
be used.
Co Lo 1op


I|ow Cond|t|oners
A stable Ilow proIile can be compromised quickly iI any pipe bends, reducers, valves,
strainers, or a variety oI other velocity-proIile killers exist. See Figure 9.
In addition, it is not always possible to adhere strictly to the manuIacturer's
recommended straight-piping requirements because many plants have minimized
production spaces to eIIect cost-eIIiciency.

I|gure 9 Common I|owrof||e
D|stort|on Cu|pr|ts
|pe e|bows reducers and a
var|ety of other ve|oc|typrof||e
k|||ers qu|ck|y can comprom|se a
stab|e f|ow prof||e
Normal upstream piping requirements aIter a 90
degree elbow can range Irom 15 to 25 pipe
diameters. For two 90 degree elbows in diIIerent
planes, the straight piping run beIore the Ilowmeter
can rise to 50 or more pipe diameters. Through Ilow
straightener use, these straight-piping runs can be it
can be decreased by a Iactor oI Iour or more.
Downstream piping runs are normally Iar less than
upstream requirements and are in the range oI 5 to
10 pipe diameters.
A properly installed Ilow conditioner will isolate the
liquid Ilow disturbances Irom the Ilowmeter while
minimizing the pressure drop across the conditioner.
A good Ilow conditioner will remove swirls Irom
the Iluid stream and allow the proIile velocity to
come up to Iully turbulent Ilowor at least a stable
acceptable geometryquickly. The popular
honeycomb/tube Ilow conditioners are great at
eliminating swirl, but their small size makes them prone to becoming clogged. In
general, as the eIIiciency oI a Ilow conditioner correcting Ilow proIile distortions goes
up, so does the corresponding pressure drop.
Remember, although Ilow conditioners can reduce the amount oI straight piping
required signiIicantly, they should not be used to cut corners. The manuIacturer's
requirements Ior straight piping beIore and aIter the Ilow conditioner should be
Iollowed.
Co Lo 1op


Conc|us|on and kecommendat|ons
In many cases, the Iluid Ilow through a pipe will be Iast enough to be in the turbulent
regime. Allowing the required straight piping runs will eliminate swirls and vortices
and will allow the velocity proIile to be stable and well blunted. This is the perIect
condition Ior Ilow measurement. But, as discussed previously, many Iactors can alter
this ideal condition. Here are some items to keep in mind as you assess the conditions
within your pipe:
Pipe Assessment: How old is the pipe in which the meter will be installed? II the pipe
is old, what is the inside surIace condition? II the pipe has a rough interior or scale or
coating buildup, the Ilow proIile will be distorted and the accuracy oI the meter could
be degraded.
II the velocity proIile is in question, you can use a pulse-Doppler meter or an
adjustable insertion meter to measure the velocity at diIIerent pipe depths. This will
allow you to see iI there is a stable correlation between the measured point velocity
and the average velocity Ilowing through the pipe.
The expected minimum and maximum Reynolds numbers should be calculated. The
manuIacturer should provide the recommended Reynolds number range Ior its
Ilowmeter design. A minimum Reynolds number is calculated by using the minimum
expected Iluid Ilow and density while using the maximum expected viscosity. A
maximum Reynolds number is calculated using the maximum expected Iluid Ilow and
density while using the minimum expected viscosity
Accuracy and Repeatability: What accuracy and repeatability are needed? Is
repeatability more important? II so, the plant will have more leeway in selecting a
less-expensive Ilow technology such as an insertion meter.
In general, a Iull bore in-line meter is more accurate than an insertion meter,
especially in situations in which velocity proIile distortions are present. However, Ior
larger pipes, an insertion meter is less expensive, and the lower accuracy may still be
suIIicient Ior the application at hand. Remember, in some applications, it is
repeatability that is the critical Iactor, not necessarily accuracy. Both insertion meters
and Iull-bore meters can provide excellent repeatability.
II necessary, in-line Ilow conditioners should be used Ior liquid swirling and/or Ilow
proIile instability that occurs aIter pipe bends; otherwise, the manuIacturers
recommended upstream and downstream piping requirements should be Iollowed
strictly.
II it is possible, plants should pick a point in the piping system that ensures major
disturbances are at least Iive pipe diameters downstream instead oI upstream. For
example, a valve should be placed downstream oI the Ilowmeter instead oI upstream
whenever possible. The same is true Ior pipe bends, expanders or reducers.
Positioning Within the Pipe: For insertion meters, it is important Ior plants to Iind
out whether the manuIacturer recommends center-line positioning or an oII-center
installation where the velocity point corresponds to the average velocity. Positioning
Ior best accuracy will change, based on the Reynolds number. In some cases, more
than one insertion sensor can be used at diIIerent depths to provide a more complete
picture. Sensor arrays that stretch across a large pipe are available Ior just this
purpose. Remember, even iI the Ilow proIile isn't completely stable, it is possible
with a little work and patienceto use an insertion meter in a large pipe and attain
very stable repeatability and an indication oI the average pipe Ilow.


keferences
1 olearmer 20012002 caLalog vernon Pllls ll 2000
2 aLrlck u and lardo S lndusLrlal rocess onLrol SysLems uelmar ubllshers
n? 1997
3 arr L A lndusLrlal onLrol Pandbook 2nd ed 8uLLerworLhPelnemann
Lngland 1993
4 Mlller 8W llow MeasuremenL Lnglneerlng Pandbook 3rd ed McCrawPlll
n? 1996
3 MofL 8 Applled lluld Mechanlcs 3Lh ed renLlce Pall n! 2000
6 8enard l Pandbook of lluld llowmeLerlng 1sL ed 1he 1rade 1echnlcal ress
LlmlLed Lngland 1988
7 8elf u MaLchlng Lhe flowmeLer Lo Lhe [ob llow onLrol vol lll no 3 1997
WlLLer ubllshlng orp n!
8 Swearlngen hooslng Lhe 8esL flowmeLer hemlcal Lnglneerlng !uly 1999
hemlcal Week ubllshlng n?
9 Swearlngen hooslng Lhe 8lghL llowmeLer hemlcal Lnglneerlng !anuary
2001 hemlcal Week ubllshlng n?

10 Pammond M ls a Lurblne flowmeLer rlghL for your appllcaLlon? llow onLrol vol
lv no 4 1998 WlLLer ubllshlng orp n!
11 Lsplna ulLrasonlc clampon flowmeLers Pave Lhey flnally arrlved? llow
onLrol !anuary 1997 WlLLer ubllshlng orp n!
12 Ldelhauser M A comparlson of conLlnuous wave uoppler vs pulsed uoppler
proflllng 1echnology llow onLrol lebruary 1999 WlLLer ubllshlng orp n!
13 !ohnson u llLLlng Lhe lnsLrumenL lnLo Lhe process llow onLrol Aprll 1999
WlLLer ubllshlng orp n!
14 Sllverberg Poney l shrunk Lhe llowmeLer! hemlcal Lnglneerlng !anuary
1997 hemlcal Week ubllshlng n?
13 Clnesl u and Crege C llowmeLers A performance revlew hemlcal
Lnglneerlng !une 1987 hemlcal Week ubllshlng n?
16 leener u Applylng llow ondlLlonlng onLrol Lnglneerlng lebruary 1999
ahners MA
17 kosla L and MuLsakls M new lnplpe flow condlLloner cuLs fluld problems
hemlcal Lnglneerlng SepLember 1992 hemlcal Week ubllshlng n?
18 8oyes W Pow Lo choose an lnserLlon flowmeLer llow onLrol lebruary 2000
WlLLer ubllshlng orp n!