TECHNOLOGY REPORT

Flow Measurement
Spring 2020
 www.controlglobal.com

TABLE OF CONTENTS

Going with the flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Advances in inline blending offer fresh opportunities to optimize profits.

Slurry flow measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Even ordinary process fluids can present problems due to suspended solids.

Knowing the best is the best . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Understand the details to be sure you’re specifying the right flowmeter.

When to use positioners in flow control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

A positioner on an actuator that’s too slow for the process can be worse than none.

Custody transfer flow measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Why can a DP flowmeter be used for gas but not liquid?

AD INDEX
ACROMAG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Krohne . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Smart Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Endress+Hauser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Flow Measurement, Spring 2020 2

 www.controlglobal.com

Going with the flow

Advances in inline blending offer fresh opportunities to optimize profits.

by John Rezabek

T
he last gasoline blender upgrade was a decade or more in the past, and once
again claimed millions in benefits thanks to controlling the blend so precisely that
giveaway was next to nothing. Giveaway, simply put, is shipping a fuel that ex-
ceeds the specifications. By blending gasoline precisely on spec, a refinery can minimize
the proportion of valued and less abundant (e.g., high octane) components, and maximize
the amount of cheap components (i.e., butane), which are extremely profitable to sell as
gasoline. Millions are claimed for shaving the spec a tenth one way or the other.

Many of us who ascribe a price or a premium for quality might be dismayed that so much
effort and attention is paid to delivering the very least the specification allows. But blending
a fuel or synthesizing a commodity chemical differs from, say, manufacturing a car. There
are few secrets and the products are virtually interchangeable. In a commodity business, the
enterprise wins through volume, reliability and efficiency.

Many of the largest fuel producers blend directly to a pipeline, which in turn ends up in
terminals where the petrol is combined with ethanol and other additives before shipping
to retail outlets. Keeping this substantial stream of precisely blended fuel on specification
requires the same of the instruments providing flow control and online analysis. While an
off-spec tank at a terminal can sometimes be “doctored” to correct an octane or vapor
pressure deficiency, it’s still a substantial disruption to the quickly consumed inventory.

Flow Measurement, Spring 2020 3

 www.controlglobal.com

For decades, the turbine flowmeter has turndown. Innately digital (measuring fre-
been the staple of gasoline and similar quency and phase shift resulting from the
blending operations. It’s a positive-dis- Coriolis effect, the Coriolis meter requires
placement, volumetric meter that relies on no straight run, and diagnostics can imme-
a spinning turbine immersed in the flowing diately detect two-phase flow or blockage.
stream. The turbine meter is innately digi-
tal—the raw signal is the frequency gener- The Coriolis phase shift is directly propor-
ated as a transducer (e.g., magnetic pickup) tional to mass flow, and the characteristic
effectively counts the blades spinning by. frequency of the meter tubes can repeat-
The frequency is directly proportional to ably measure density to three or even
volumetric flow. But such simplicity is com- four decimal places. For the optimum
plicated by its mechanical design—bearings accuracy and turndown, a line-size re-
immersed in a flowing stream will eventually duction is not uncommon, and dual-tube
wear out and fail, so meters are routinely meters will further reduce the effective
pulled or exchanged for freshly calibrated line size as the fluid passes through the
sensors. Straight runs are required to ensure meter body—hence the pressure loss can
a symmetrical flow and limit turbulence, and be higher and might be a limiting factor.
the streams must be filtered to prevent en- Different face-to-face dimensions will
trained particulates and other flotsam from most likely require some piping modifica-
corrupting the precision-engineered vanes tions; and improper pipe alignment caus-
and bearings. Each stream must be reliably ing twisting or bending torque on the
single-phase (all gas or all liquid) and the flow tube should also be avoided.
flowing temperature (and pressure for gas
applications) must be monitored to account In recent years, Coriolis specialist Micro
for the density of the stream. Motion has introduced a feature they call
Smart Meter Verification. Especially useful
While it sometimes seems like refinery in a blending application, it allows valida-
cultures cling to past practices, the more tion of a sensor in the line without having to
modern Coriolis flowmeter can help ensure remove it for recertification.
a blender upgrade’s million-dollar return-
on-investment is realized. Particularly Especially in large line sizes, Coriolis me-
when integrated over fieldbus, Ethernet, or ters are far from the least expensive solu-
equivalent high-speed and time-sensitive tion. But their capabilities and features in a
digital network, Coriolis meters equal or digitally-integrated blender have potential
exceed the turbine meter for accuracy and to justify the premium.

Flow Measurement, Spring 2020 4

Because We
Know I/O
You Get Personalized Support-
Guaranteed
You can trust Acromag for all your signal conditioning and
remote I/O solutions. You can be confident that with our
more than 60 years I/O experience that you will get
professional personalized support – guaranteed.

Products Designed for

Dependable Value
Acromag’s signal conditioning line features more than 100
transmitters, isolators, alarms, and computation modules.
If networked I/O is required, Acromag offers analog and
discrete I/O modules for Ethernet, Modbus, and Profibus.

Visit Acromag.com/IOSolutions
TO SEE WHAT’S NEW
877-295-7066 Remote I/O Solutions You Can Depend On.
 www.controlglobal.com

Slurry flow measurement

Even ordinary process fluids can present problems due to suspended solids.

by Ian Verhappen

F
low measurement is important, especially when it’s a custody flowmeter that’s also
the cash register. Obviously, there are many standards related to accurate flow
measurement, including upstream and downstream straight-run rules to assure a
nice, clean profile through the meter.

But very few process streams are 100% single-phase, with most having either small bubbles
or low levels of solids. These are normally suspended close to homogenously, so the fluid
still behaves as if it were single-phase, therefore not affecting the measurement. Once they
start having a noticeable number of solids—less than 200 microns primarily by accident,
non-settling slurry with specific gravity less than 1.05, and less than 5% solids by weight—
streams are considered a “light slurry.” Most of the time, until the second phase constitutes
a significant percent of total fluid fraction, the fluid can be treated as a single phase.

Low levels of solids may be (and likely are) present in your process, even though you don’t
think so. If your flowmeter is downstream of a catalyst bed or some other process vessel
with solids, chances are there will be some carryover, particularly during startup. Similarly,
almost every well will contain some level of entrained solids, which normally continue with
the liquid stream until they’re removed in dedicated vessels. However, during different
stages of plant operation, such as startup and shutdown, when the stream is at other than
steady-state, even low levels of solids can have an impact.

Flow Measurement, Spring 2020 6

 www.controlglobal.com

Suspended solids at low levels that will not However, in mining, “real” slurries with
affect a single-phase measurement (other targets of close to 40% solids are of-
than perhaps higher wear) may not always ten used in hydro transport technology
stay suspended, especially if the flow is where the solids are mixed with water
stopped. Without the fluid velocity to keep and transported via pipeline relatively
them in suspension, they will settle to the long (tens of miles) distances. One vari-
bottom of the pipe. This may affect a wide able that needs to be controlled with
range of flowmeters, especially if they have hydro transport is the solids ratio, to
some part of the sensor in this position. prevent settling while still minimizing the
Therefore, if you suspect solids may occur, amount of water being recycled.
it’s always a good idea to rotate your me-
ter or impulse line by one bolt hole to keep Oil sands mining takes this concept one
sensitive parts of the instrument away from step further by going from two-phase,
settling materials. solid/liquid flow to four-phase flows of
vapor(air), bitumen, water and solids (4-
In many cases with a light slurry, some me- in. to micron size), where it’s necessary to
ters will no longer work and the application maintain both the solids content as well as
will require special meters to handle the the air content in the line. The good news
solids. Typical flowmeters used for slurry is that with the help of industry and oil
streams are ideally non-intrusive or venturi sands researchers, this has been done reli-
because they have less sudden change in ably for many years. The how is a story for
the flow path. When using a venturi meter another day, though I did enjoy working in
to measure a slurry that’s severe due to this challenging environment.
either the type or size of solids, the meter
internals are often coated with a hardened Like many other field sensor applications,
material, which adds roughness at the sacri- flow may seem relatively straightforward.
fice of accuracy, or is fitted with a hardened However, like many other parts of our in-
cast lining that can be replaced when the dustry, there are also many ways for the
accuracy falls below what’s acceptable. For- process to be “not as it appears,” and as
tunately, these meters are rarely used for a result, the opportunity arises to excel or
custody transfer, but more for plant balance fail miserably. Small details, such as low
and simple regulatory control. solids concentrations, can have a signifi-
cant impact on your measurement—even
The above examples apply for light or though the stream may not be consid-
even medium (5-20% solids) slurries. ered a slurry.

Flow Measurement, Spring 2020 7

Flow rates
up to 169,000 lbs/h
OPTIMASS 2400 – World’s highest capacity Coriolis mass flowmeter
for liquids and gases
• Multiple straight tube in sizes DN100…400/4…16”
• Entrained Gas Management (EGMTM): reliable indication
of gas entrainments, maintaining operation at all times
• Highest safety factor with optional pressure rating 180 bar/2,610 psi,
PED approved secondary containment up to 150 bar/2,175 psi
• NACE compliant with Duplex or Super Duplex as wetted material
• ATEX, IECEx, CSA, FM, NEPSI, CEPAL, MID, OIML

us.krohne.com/optimass2400

products solutions services

 www.controlglobal.com

Knowing the best is the best

Understand the details to be sure you’re specifying the right flowmeter.

by Greg McMillan

Greg: We often forget that the performance of equipment is in the care of the automation
system that is the window into the process and the means of affecting the process. It starts
with the measurement. You can only do as well as what the measurement tells you. De-
spite the process being in the critical care of the measurement, we may obsess over project
budgets, losing sight of the bigger picture. The instrumentation is typically less than 10% of
the cost of the equipment and piping, yet we see projects cut corners from misguided goals
based purely on hardware cost.

Here, we look at flow measurement with the insight that the process is defined and manipu-
lated by the mass flows as seen on the process flow diagram (PFD). While volumetric flows
may be useful, mass flows define the material balance for changes in production rates and
operating conditions, plus the process gains derived in Appendix F of my ISA books (e.g.,
101 Tips for a Successful Automation Career).

There are many mass flow measurements, but most depend on adding temperature and
pressure measurements, and assume a constant composition, often not realized by the user
or in the process. A composition of each stream being exactly what’s shown on the PFD is a
rare occurrence. The fact that processes aren’t perfect is realty, and is why process control
systems are so valuable.

Flow Measurement, Spring 2020 9

 www.controlglobal.com

The Coriolis meter offers a true mass flow While the Coriolis flowmeter is far less sus-
measurement with extraordinary accuracy ceptible to problems, it’s useful to explore
and rangeability, particularly for liquid flow possible effects to gain confidence and justifi-
measurement, independent of chemical cation for future applications and to alert the
concentration, density and viscosity. The user to the special cases where this extraor-
accuracy can be as good as 0.05% of rate dinary performance is threatened. “Knowing
for liquid flow with a rangeability of 200:1. the best is the best” is my goal. I’m impressed
Additionally, the density measurement how technical advances seen in Smart Meter
offered is as good 0.1 kg/m3 for measure- Verification (SMV) for smart meters can help
ments between 300 and 5,000 kg/m3. For us in this path. The following conversation is
streams where a single component has offered to gain insights on SMV capability and
a density that’s quite different than the achieving the best by Tom O’Banion, director
solvent or other component, the density of chemical industry innovation for Emerson
provides a fast, reliable measurement of Automation Solutions.
component concentration.
Tom, what are effects of corrosion and
True mass flow measurements are criti- erosion?
cally important for control of many other
types of process equipment besides reac- Tom: Although rare in practice, corrosion
tors, such as crystallizers, evaporators and causes a thinning of the tube, which in the
distillation columns. The Coriolis flowme- short term usually causes the meter to over-
ter provides this capability up front, and register (mass flow) and under-register
retains this capability, holding its calibra- (density). The thinner tube twists a bit more
tion despite many adverse situations. (so the flow reading is now biased high),
There are no upstream or downstream and the frequency increases (so the meter
straight run requirements and basically thinks the fluid density has gone down).
no maintenance requirements. While the SMV can “see” this happening before the
upfront cost is higher, the lifecycle cost sensor tube integrity is affected. The goal
is typically less, particularly when you is to get that meter out of service before a
consider the cost of maintaining impulse (very rare) tube breach could occur. Ero-
lines. If you consider the impact on pro- sion also thins the sensor tubes with virtu-
cess efficiency, the extraordinary accuracy ally identical effects as corrosion. In both
and rangeability, and a true mass flow examples, it’s difficult to predict how uni-
measurement, translates into an incredible form the corrosion or erosion is, so there’s
return on investment. no 1:1 correlation to accuracy.

Flow Measurement, Spring 2020 10

 www.controlglobal.com

Greg: What is the effect of coatings? gas, whether it’s well mixed, bubbles or
actual gas slugs, cause more adverse ef-
Tom: Tube coating in general does not af- fects on flow and density accuracy—the gas
fect mass flow very much, but it does affect and liquid decouple from the tube vibration,
the density reading, and thus, concentra- causing larger errors due to a velocity of
tion measurement and volume flow (Coriolis sound effect.
volume flow = mass/density). Coatings tend
to be denser than the fluid, but it varies. Improved two-phase flow measurement
It’s more difficult to detect, but our new ability is a big subject in industry, and so
SMV Professional has a coating detection our research continues. We have a few
feature. If the coating is uneven, then it’s software options that help mitigate the
relatively easy to see. If it’s highly uniform, effect. Advanced Phase Measurement is
then we also examine other variables from one prominent one. It reads the last “good”
the meter (drive gain, density trend) to fer- values for density, and uses that to “steer”
ret out what’s happening. through the chaotic and periodic periods of
substantial gas bubbles.
Greg: What is the effect of freezing?
Greg: What is the effect of liquid droplets in
Tom: Freezing (water) will cause swelling of gases?
the tubes, and thus, an increase in stiffness.
This will cause the flow reading to under- Tom: Liquid or “wet” gas has a somewhat
register (tubes twist less for a given amount similar effect as gas bubbles in a liquid
of flow because they’re stiffer). In addition, stream. The liquid tends to stream down
the effect on density is difficult to predict, the tubes, dampening some of the vibra-
but accuracy will be degraded. It should be tion, causing density to over-read. Flow and
noted that, in most cases, the meter is per- density readings may become erratic and
manently damaged. less accurate.

Greg: What is the effect of gas bubbles in Greg: What’s the effect of poor installation?
liquids?
Tom: In the early days of Coriolis (pre-
Tom: We’ve done a lot of research on two- 1990s), installation was more critical, as
phase flow over the past 15-plus years. It’s some designs were sensitive to pipe mount,
a complex fluid interaction. Small amounts stress and cross-talk. These have largely
of gas and good mixing cause only small been eliminated. The most common field
accuracy degradation. Larger amounts of issue I’ve seen recently involves the meter

Flow Measurement, Spring 2020 11

 www.controlglobal.com

not staying full of liquid, or trapping bub- Greg: Can we put Coriolis meters in series
bles, or trapping liquids in a gas application. and parallel?
It’s a piping design challenge, usually easily
rectified. Occasionally we see a meter under Tom: Absolutely! Two applications are
large axial or radial forces, and the meter safety instrumented systems (SIS) and large
zero is shifted. flow-rate applications. Large fiscal transfer

TOP 10 THINGS YOU DON’T WANT TO HEAR ABOUT FLOWMETERS

(10) The stream composition, density and flow must be exactly what are on the PFD.
(9) Meter sizing was so easy; we just went with line size meters everywhere.
(8) We put flowmeters at the high point in the line to make sure they are empty for maintenance.
(7) We put flowmeters in the pump suction to reduce the required pressure rating.
(6) We standardized on the same materials of construction for all applications.
(5) My impulse is to use impulse lines.
(4) What is a Coriolis meter?
(3) If we get a reading during water batching, we are all set.
(2) Who needs turndown and ratio flow control?
(1) We must minimize instrumentation costs.

Flow Measurement, Spring 2020 12

 www.controlglobal.com

stations often rely on multiple large Coriolis Tom: Yes, our meters are often used in
meters in parallel. I’ve seen applications in regulated industries such as food & bever-
30-in. crude oil lines require four or more age and life sciences, where master meters
large Coriolis meters. They’re known for are installed on portable carts and moved
working well in high-accuracy, custody- around plants, calibrating other meters.
transfer conditions. These masters are available with full ISO-
17025 NIST traceability. On a side note,
Greg: Could we temporarily insert a Corio- customer needs still include a clamp-on Co-
lis meter to verify the calibration of other riolis, to measure mass flow and hopefully
meters, or to find the real density for their fluid density. It’s still a “gleam in our eye,”
calibration? but very challenging to do.

Raise Your Plant's Performance

ABOUT THE EVENT
Industrial with a ca pital “I”: Smart Industry
Base Camp is exclusively focused on the digital
Our sister publication, Smart Industry, hosts Baseof Camp,
transformation the
industrial organizations,

only industry event designed to deliver a step-by-step action

including manufacturing and raw materials
processing as well as energy and water
plan to take your organization’s transformation
infrastructureto the next
and industrial machine builders.

level. Attend Base Camp in March to: Personalized curriculum: Our content agenda
includes industry-specific Wayfinding Sessions

• Assess your organization’s digital maturity

designed tofrom a your own personalized
help you plan
curriculum. You'll identify which use cases hold
range of perspectives the most potential for your organization's
strategic priorities, then learn how to affect
• Align people, processes and technologies to change,
organizational effect demonstrate ROI, and
scale successful pilots into enterprise roll-outs.
organizational change
Unrivaled line-up of industry practitioners: Our

• Prioritize those opportunities that provide quick

speakers are and
from leading manufacturing,
processing and consulting organizations. They
scaleable ROI know what works—and what doesn’t.

Networking, networking, networking: From our

popular speed networking session to more
2020
Digital Manufacturing Tours
traditional receptions, meals and breaks,
attendees will have numerous opportunities to

Networking with Experts

explore new relationships with peers and solution
providers.

MARCH 30-APRIL 1 | CHICAGO, IL Keynotes

Convenient location:
&AExhibits
quick trip downtown and
a 5-minute shuttle from the airport, the Loews
Chicago O’Hare offers a big city vibe, newly

REGISTER TODAY FOR JUST $750! renovated rooms and award-winning onsite
restaurants.

Use code PUB at event.smartindustry.com

Flow Measurement, Spring 2020 13

 We understand you need insightful process
information to help you run your plant e�ciently.

MEASURED VALUE
+ ADDED VALUE You make confident decisions backed by process data and
a complete portfolio of services and solutions to support you.

Heartbeat Technology takes the pulse of your measurement to save costs:

• Permanent diagnostics of processes and devices guarantee

an economical maintenance and safe plant operations.
• Verification of devices happens without process interruptions.
It ensures high plant availability.
• Monitoring of all information enables predictive maintenance
and process optimizations.

Do you want to learn more?

www.us.endress.com/heartbeat-technology
 www.controlglobal.com

When to use positioners

in flow control
A positioner on an actuator that’s too slow for the process can be
worse than none.

by Béla Lipták

Q: I stumbled on www.ControlGlobal.com while searching for information about PID control

of flow. This question may have been answered, but the couple of threads I found didn’t
seem helpful specifically for my problem. I’m controlling gas flow with a flow sensor and an
electrically actuated (servo) valve positioner. The positioner is slow, taking up to 90 sec-
onds for full travel.

I’m using a microcontroller to read the flow signal (F) and call out a control output of c(t) =
Kp + Ki*[ {e(t)* t}] + Kd*( F/ t). The error is e(t) = Fsp - F.

The slow positioner has been causing constant position adjustment and overshoots. I’m able
to get decent control only with a huge averaging of the input flow signal (over almost a
two-minute time scale). I guess we can live with it, but is there a better way?
Murthy Tata
[email protected]

A: This is a very basic and very important question. I’m glad you brought it up because, in
my experience, the use and selection of positioners is the cause of the most frequent con-
trol problems. (The other error seen most often is usually caused by valve response.)

The answer to your question applies to all cascade systems. The slave controller in all

Flow Measurement, Spring 2020 15

 www.controlglobal.com

cascade control loops must be faster than A pneumatic positioner increases the ac-
the master controller. No exception! This tuator speed or thrust by increasing the
is kind of self-evident from everyday life. actuator pressure and/or air flow volume,
After all, how could, say, a pilot do his/her and can modify the valve characteristics
job if the second pilot was still busy carry- through the use of mechanical links and
ing out the previous order when the next cams or by electronic function generators.
was given? While these positioner capabilities are very
important, some of these capabilities can
People often don’t realize that a positioner also be obtained or approximated with
is a slave controller, with all its advantages other accessories.
and limitations. It can both help and hurt
the performance of the loop. Therefore, I’ll The positioner/actuator combination is
give a brief discussion of the subject of po- the secondary control loop in a cascade
sitioners before giving my advice concern- system, receiving its setpoint from the
ing your specific question. master controller. For a cascade secondary
to improve control, it must respond more
The purpose of a positioner is to improve quickly than the primary loop. The ideal
the accuracy and response of control situation would be if the time constant of
valves. This means it will help to have the the secondary was one-tenth (open-loop
valve position more closely approach the speed of response 10 times as fast) of that
position commanded by the controller. A of the primary (but certainly less than half).
positioner can reduce the effects of many No process control response time can be
dynamic variations. These include changes faster than the slowest element in the con-
in packing friction due to dirt, corrosion, trol loop, therefore, the performance of a
lubrication or lack of lubrication; variations cascade system improves with fast sec-
in the dynamic forces of the process; sloppy ondary response.
linkages (causing dead band); and nonlin-
earities in the valve actuator. It’s clear that poor valve response reduces
the quality of control much more than one
The dead band of a good valve/actua- would expect. If that’s the case, the control-
tor is up to 5%. Large plug valves and ball ler tuning must be modified (gain reduced,
valves with less than perfect linkages and integral lengthened) to avoid oscillation due
inadequate actuators may be far worse. A to dead band and response delay. In typical
better positioner with the proper actuator real applications, controller tuning is con-
can often reduce the dead band to less than servative and avoids any hint of oscillation,
0.5% of stroke. and therefore response is even worse than

Flow Measurement, Spring 2020 16

 www.controlglobal.com

Time response of a fast process (0.03 sec. time constant) Time response of a slow process (100 sec. time constant)

Controller settings Controller settings

PB = 45% (gain = 2.2) PB = 5% (gain = 20)
I = 20 rep./min. I = 0.5 rep./min.
Response without a positioner Response without a positioner

PB = 500% (gain = 0.2) PB = 1% (gain = 100)

I = 20 rep./min. I = 1.0 rep./min.
Response with a positioner
Response with a positioner

Setpoint change (5%) Setpoint change (5%)

WHEN POSITIONERS DON’T HELP

Figure 1: In slow processes (right), a positioner always improves loop performance, but in fast pro-
cesses (left), particularly if the controller is pneumatic and the process is faster than the positioner, it
can hurt loop performance..

it it could be. (Operators usually object to useful to understand the cascade nature
cycling processes.) of positioning, as described above. So,
turning to your specific question, assum-
Figure 1 illustrates a fast flow process, ing that your valve is linear and you have
where the loop without a positioner can no derivative in your controller, you can
be tuned more tightly (for higher gain and disconnect the positioner. Or if the posi-
more repeats/minute). Such a loop re- tioner is an integral part of the actuator,
sponds better without a poor positioner. It insert a lag into the controller output to
might also be noted that after a new state make its time constant at least double
is reached, the positioned installation gives that of the positioner.
better and noisier control because of in-
creased speed of response. Another option is to use gap control, where
as long as the flow is inside the gap, the
Electric actuators are used where air isn’t controller stays inactive. Today’s micro-
available and where their typically slow op- processor controllers provide broad tun-
erating speed is acceptable. There are some ing adjustment ranges, so you can set the
specialized electric valve actuators that can controller gain much lower (portional band
provide very high position precision; others wider) and leave the integral (reset) fast to
can provide reasonably fast response. match the time constant of your relatively
fast flow process.
Sorry for taking so long to come to an Béla Lipták
answer to your question, but I believe it’s [email protected]

Flow Measurement, Spring 2020 17

 www.controlglobal.com

A: I have few follow-up questions before such as air dampers in boilers, but are not
I can guide you for a reasonably stable suitable for gas or liquid flow control.
tuning. Which software PID block are you
using? Is it PID_IND or PID_ISA block? Or If your actuator is more like a large air
are you doing your own logic? Have you damper rather than a control valve, then
done a step test? I’ve tuned liquid flow you’d probably benefit from using a
control, and we don’t use D, but for gas, gap-action control algorithm or an error-
you’ll need to use D gain. squared control algorithm. These algo-
rithms have a characteristic of taking little
Another problem I’ve seen is the range for action when the measured value is close to
the transmitter. What is the inherent dead- the setpoint. Both algorithms are common
band of your positioner? Discuss it with the variants of the PID algorithm you seem to
manufacturer of the positioner, and ensure be using, and are commonly found in the
that the setup/configuration is appropriate algorithms of most DCSs and even most
for your application. PLCs used for closed-loop control.
Hiten A. Dalal PE
PMP Automation Engineer Most process control textbooks also dis-
[email protected] cuss these variations of PID control. The
gap-action controller is simple: establish
A: Yes, there is a better way to do flow con- a zone around the setpoint and when the
trol, but it would require that you replace the error is within this gap, don’t change the
slow-acting electric control valve actuator controller output. The error-squared con-
with a pneumatic or a direct-acting, electric troller is more complicated, and you can
solenoid-type control valve actuator. What find a discussion at this website: https://
you describe is not a solenoid actuator but pdfs.semanticscholar.org/921e/c98836d-
a reversible electric motor driving the con- da49347d1d99f6a751b548b475283.pdf.
trol valve position through a screw-thread Richard H. Caro
mechanical linkage. These have been in use CEO, CMC Associates, ISA Life Fellow
for many years to control very large objects, [email protected]

Flow Measurement, Spring 2020 18

 www.controlglobal.com

Custody transfer
flow measurement
Why can a DP flowmeter be used for gas but not liquid?

by Béla Lipták

Q: I’m interested to know why orifice differential pressure (DP) flowmeters aren’t used in
liquid metering systems (for custody transfer purposes), whereas they’re widely used in gas
metering systems? What makes an orifice flowmeter a viable, cost-effective choice for gas
metering systems only?
A. Rashimi
[email protected]

A: The short answer is that orifices are used for both, but because custody transfer (fiscal
metering) is such an important topic (see Chapter 2.20 in volume one of my handbook), I
will give a more detailed answer.

Flowmeter selection can be based on government regulations, industry or national stan-

dards, and contractual agreements, and can also be subject to the approval of such organi-
zations as API, AGA and ISO.

The acceptable uncertainty in the quantity of transferred liquid or gas determines me-
ter selection. The uncertainty is the sum of the errors of all components of the metering
system. In case of volumetric flowmeters, this includes errors by flow, pressure, tempera-
ture, density, composition sensors, their A/D converters, and in calculating the amount
of energy (not mass, but energy) transferred. The hydrocarbon industry claims that cus-

Flow Measurement, Spring 2020 19

 www.controlglobal.com

TABLE I: FLOWMETERS USED IN CUSTODY TRANSFER

Features Orifice and (Venturi) Coriolis Rotary PD liquid (gas) Turbine dual helical Ultrasonic, multipath
Chapter(s) 2.21-2.34 2.16 2.24, 2.25 2.31 2.32
Accuracy at max. flow 1% (0.25%) 0.15% 0.2% (1%-2%) Liq: 0.25%, Gas: 0.5% Liq: 0.25%, Gas: 0.5%
Accuracy at min. flow 2% (0.5%) 1% 0.1% (2%) Liq: 0.5%, Gas: 1.0% Liq: 0.5%, Gas: 1.0%
Rangeability 3:1-4:1 Up to 100:1 ~ 15:1 Liq: 10:1, Gas: >20:1 >20:1
Reynolds (RE) limitations >10,000 (>100,000) Debated Insensitive Insensitive < 2,000 and >8,000
Size range 0.5-24 in. (1-120 in.) 1.0 mm to 16 in. 1-18 Liq: 1-20, Gas: 2-12 in. Liq: 2- >12, Gas: 2- >42 in.
Straight run up/downstream 20/5 (5/0-3) None None 15-20/5 20/5
Pressure Drop High (low) ~ 10 psid High ~ 5 psid Low
Installation Critical Not critical Not critical Important Important
High because High because
Maintenance High (high) Low Low
moving parts moving parts
Mass flow or multphase No Yes No No No
Moving parts No Vibration Yes Yes No
Bidirectional Some w/2DP Yes No Some Yes

tody transfer operates at an uncertainty the transferred quantity without flowme-

of ±0.25% on liquid and ±1.0% or better on ters, just by measuring the level change of
gas service, but I consider these numbers liquids or pressure change of gases.
overly optimistic.
A typical liquid custody transfer skid
Table I lists flowmeters that can be used includes multiple flowmeters (master and
to measure hydrocarbon liquids and gases. operating meters), flow computers and
The table also gives the chapter numbers meter provers. For pipe sizes below 42-in.
where each is described in my handbook, diameter (1.07 m), onsite provers can be
their accuracies (if they’re correctly sized, used and API requires prover accuracy
installed and maintained), and other main to be 0.02%. The meter prover volume is
features. Table I doesn’t list their first costs calibrated against Seraphin cans, whose
because, in larger transactions, the cost dif- precise volume is traceable to NIST. Reca-
ferences between meters are small in com- librations should be performed frequently,
parison to the cost of measurement errors. typically before, during and after the
batch transfer.
For example, when oil costs $60/barrel and Béla Lipták
we’re unloading a 500,000-barrel tanker, [email protected]
each 0.1% uncertainity corresponds to
$50,000. In cases of smaller quantities, me- A: Orifice meters are still widely used for
ter cost differences can be considered, and liquid measurements and have been for
if accuracy is not critical, one can determine many years. The orifice meter accuracy is

Flow Measurement, Spring 2020 20

 www.controlglobal.com

much affected by the details of installation, ments challenging estimated costs and ac-
and comments about poor accuracy can curacy. Support for this program went away
usually be explained by poor installation. as the costs of maintenance would be high
as the data changed.
The advantages and disadvantages of
the various meter technologies are well Good question, and we need to discuss
known, and the available technologies these things.
have changed. One example is how Coriolis Cullen Langford
meters have become more popular as the [email protected]
technology has matured and competition
has driven prices down. It’s clear that meter A: Orifice flow measurement is at best ±4%
selection is heavily affected by pipe size. I accurate. In spite of this, it was used for
find it hard to imagine a Coriolis meter in a custody transfer of liquids for many years
one-meter-diameter pipeline. And the cost? until better and more accurate instruments
became available. Today, the Coriolis flow-
National and international standards can meter is the standard for liquid flow cus-
affect decisions. Custody contracts may tody transfer due to its high accuracy and
well have statements limiting options. It often because it directly measures mass
can happen that non-technical people write flow instead of volumetric flow. It’s expen-
those contracts. sive, so sometimes if liquid is transferred
from tank to tank, the before and after tank
I’m prejudiced, but I have the impression level measurements are used for custody
that salespeople tend to suggest the more transfer of liquids. However, that requires
expensive choices in their catalogs. very accurate tank level measurement.

In the decision process, it’s common to un- Measuring the flow of gases with an orifice
derestimate the costs and details of installa- flowmeter with compensation for tempera-
tion for the various flowmeters. Accuracy is ture and pressure is typical for natural gas
expensive. transmission, but that still doesn’t make
it accurate. In most cases, the low value
I once developed a program to aid in flow- of natural gas makes it uneconomical to
meter selection. The user entered informa- spend extra money on more accurate gas
tion about the fluid and flows. The program flow measurement. For high-value gases,
then displayed a list of possible meter types it’s possible to use Coriolis flowmeters or a
with costs, accuracy and permanent pres- high-accuracy, positive-displacement me-
sure loss. This brought interesting com- ter. Sometimes, custody transfer of gases is

Flow Measurement, Spring 2020 21

 www.controlglobal.com

calculated from a change in pressure of the 1. During startup, while the flow stabilizes,
source of the gas, such as a cylinder. the error in measurment is generally un-
acceptable. The same happens when the
Orifice flow measurement depends on system is shut down.
Bernoulli’s law that relates pressure drop 2. Sizing an orifice plate to guarantee
thorough a sharp-edge orifice to volumetric custody transfer precision requirements
flow rate. The pressure drop is between the generally requires the orifice plate to be
upstream pressure (before the orifice) and designed and fabricated to extreme tol-
the pressure at the vena contracta formed erances that not every company can do.
by the increased velocity of the liquid or 3. The turndown ratio for liquid orifice
gas as it passes through the orifice. Unfor- plates would require users to have too
tunately, the location of the vena contracta many orifice plates available for when
varies with the flow rate, so there’s no prac- flow conditions change.
tical way to measure this pressure drop.
We do the best we can, and usually just To summarize, though custody transfer
measure the pressure drop at the flanges can be performed with orifice plates, it’s
that hold the orifice plate in place, and not recommended due to the fact that
depend on a correlation (approximation) signal instability can cause accounting
to estimate the pressure drop at the vena errors; design and fabrication of orifice
contracta, or just assume that the pressure plates is very expensive; and finally, if the
drop at the orifice is the same as that of the custody transfer is based on varying flow-
vena contracta. rates, then the need for additional orifice
Richard H. Caro, CEO, CMC Associates plates is increased.
[email protected] Alex (Alejandro) Varga
[email protected]
A: One of the biggest reasons why orifice
plates are not used for liquid custody me-
tering has to do with the following:

Flow Measurement, Spring 2020 22