I

nstrumentation is critical to un-

derstanding and troubleshooting
all processes. Very few engineers
specialize in this field, and many
learn about instrumentation through
experience, myth and rumor. A good
understanding of the various types of
instrumentation used on columns is a
valuable tool for engineers when evalu-
ating column performance, starting up
new towers or troubleshooting any type
of problem. This article gives an over-
view of the common types of instru-
ments used for pressure, differential
pressure, level, temperature and flow.
A discussion of their accuracy, common
installation problems and troubleshoot-
ing examples are also included.
The purpose of this article is to pro-
vide some basic information regarding
the common types of instrumentation
found on distillation towers so that
process engineers and designers can
do their jobs more effectively.
Introduction
Anyone trying to complete a simple
mass balance around a column under-
stands that process data contain some
error. Closing a mass balance within
10% using plant data is usually consid-
ered very good. Generally, some values
must be thrown out when matching a
model to plant data. Understanding
which measured plant data is likely
to be most accurate is invaluable in
making good decisions about a model
of the plant, column performance and
future designs.
The following is a real case and a
telling example of how little the aver-
age chemical engineer may understand
about instrumentation. A process engi-
neer with over 20 years of experience
was doing a material balance around a
distillation tower, illustrated in Figure
1. Based on the material balance, the
engineer concluded that the bottoms
flowrate must be in error and wrote
a work order to have the flowmeter
recalibrated. The instrument group
disagreed heartily. By the end of this
article, the reader will understand the
instrument groups response.

Pressure
There are three common types of pres-
sure transmitters: flush-mounted
diaphragm transmitters, remote-seal
diaphragm transmitters and impulse-
line transmitters. All use a flexible
disk, or diaphragm, as the measuring
element. The deflection of the flexible
disk is measured to infer pressure.
The diaphragm can be made of many
different materials of construction,
but the disk is thin and there is little
tolerance for corrosion. Coating of the
diaphragm leads to error in the mea-
surement. The instrument accuracy
of all three types of pressure trans-
mitters is similar, usually 0.1% of the
span, or calibrated range.
Flush-mounted diaphragms
These pressure transmitters are com-
mon in low-temperature services, such
as in scrubbers and storage tanks.
The process diaphragm, an integral
part of the transmitter, is mounted
on a nozzle directly on the vessel, and
the transmitter is mounted directly
on the nozzle.
Remote-seal diaphragm
Used in higher temperature service
when the electronics must be mounted
away from the process, a flush-mounted
diaphragm is installed on a nozzle at
the process vessel. A capillary tube
filled with hydraulic fluid connects the
flush-mounted diaphragm to a second
diaphragm, which is located at the re-
motely mounted pressure transmitter.
The hydraulic fluid must be appropri-
ate for the process temperature and
pressure. Hydraulic fluid leaks will
lead to errors in measurement. Cali-
bration is complex because the head
from the hydraulic fluid must be con-
sidered. The calibration changes if the
transmitter is moved, the relative po-
sition of the diaphragms changes or if
the hydraulic fluid is changed.
Impulse-line
Impulse-line pressure transmitters
can either be purged or non-purged.
Feature report
48 ChemiCal engineering www.Che.Com marCh 2008
Feature report
Ruth R. Sands
DuPont Engineering
Research & Technology
Column Instrumentation
Basics
'5
'5
'5
&SSPS
'5
Figure 1. Which flowmeter is the most
accurate? What is the source of error in
the material balance?
Figure 2. Flush-mounted diaphragm
pressure transmitters are common in
low-temperature services
An understanding of instrumentation is valuable in
evaluating and troubleshooting column performance
Purged impulse-line pressure trans-
mitters measure purge-fluid pressure
to infer the process pressure. Most
commonly, the purge fluid is nitro-
gen, but it can also be air or other
clean fluids. The purge fluid is added
to an impulse line of tubing to detect
pressure at the desired point in the
process. The purge fluid enters the
process and must be compatible with
it. Check valves are required to en-
sure that process material does not
back up into the purge-fluid header.
The system must be designed so that
the pressure drop through the im-
pulse line is negligible. A pressure
transmitter measures the purge-fluid
pressure with a diaphragm to infer
the process pressure.
Non-purged, impulse-line
Rather than a purge fluid, this type
of pressure transmitter uses process
fluid. Usually, this style is chosen
when the process is non-fouling or
it is undesirable to add inerts to the
process. One example is a situation
where emissions from an overhead
condenser vent must be minimized.
An impulse line is connected from
the desired measurement point in
the process to a pressure transmitter,
which measures the process pressure
at the remote point. The system must
be designed so that the pressure drop
through the impulse line is negligible.
The system designer must consider
the safety implications of an impulse-
line failure. The consequence of releas-
ing hazardous material from a tubing
failure may warrant the selection of a
different type of pressure transmitter.
Adequate freeze protection on the im-
pulse lines is also important to obtain
accurate measurements.
Example 1. A good example of a prob-
lem with impulse-line pressure trans-
mitters can be found in Kisters Distil-
lation Troubleshooting [2]. Case Study
25.3 (p. 354), contributed by Dave
Simpson of Koch-Glitsch U.K., de-
scribes three redundant impulse-line
pressure transmitters used to mea-
sure column head pressure. Following
a tray retrofit, operating difficulties
eventually led to suspicion of the head
pressure readings. The impulse lines
and pressure transmitters had been
moved during the turnaround. The
transmitters had been moved below
the pressure taps on the vessel. Con-
densate filled the impulse lines and
caused a false high reading. Relocat-
ing the transmitters to the original
location above the nozzles solved the
problem by allowing condensate to
drain back into the tower.
Transmitters in vacuum service
Pressure transmitters in vacuum ser-
vice are generally the most problem-
atic, leading to greater inaccuracy in
the measured value. Damage to the
diaphragm can occur from exceeding
the maximum pressure rating of the
instrument. Often, this happens on
startup, or it can happen when per-
forming a pressure test of the vessel.
The diaphragm deflects permanently
and introduces error.
Calibration of vacuum pressure
transmitters is more difficult for in-
strument mechanics. The operating
range must be clearly defined; for ex-
ample, is the range 100-mm Hg vac-
uum, 100-mm Hg absolute, or 650-mm
Hg absolute? Using different measure-
ment scales in the same plant is con-
fusing, and it can make it very hard
for mechanics to calibrate the pressure
transmitters accurately.
Another issue is measuring the
relief pressure. The system designer
must consider the instrument ranges
available and the accuracy of the
measurement for the operating range
versus the relief pressure range. It is
good practice to install a second pres-
sure transmitter on vacuum towers to
measure the relief pressure.
Example 2. An excellent example of
calibration problems is illustrated in
vacuum service in Reference [2]. Case
Study 25.1 (p. 348), contributed by
Dr. G. X. Chen of Fractionation Re-
search, Inc., describes several years
of troubleshooting a steam-jet system
in an attempt to achieve 16-mm Hg
absolute head pressure on a tower. It
was eventually determined that the
calibration of the top pressure trans-
mitter was wrong, and they had been
pulling deeper vacuum than they
thought. The top pressure transmitter
was calibrated using the local airport
barometric pressure, which was nor-
malized to sea-level pressure and was
off by 28-mm Hg.
Differential pressure
Differential pressure can be measured
either with a differential pressure (dP)
meter or by subtracting two pressure
ChemiCal engineering www.Che.Com marCh 2008 49
DEFINITIONS
Instrumentation range
The instrumentation range, the scale over which the instrument
is capable of measuring, is built into the device by the manu-
facturer. The purchaser defines the desired measured range,
and the vendor should provide a device that is appropriate for
the application.
Calibrated range
The calibrated range is the scale over which the instrument is
set to measure at the plant. It is a subset of the instrument range.
The calibration has a zero and a span. The zero is the minimum
reading, while the span is the width of the calibrated range.
The calibrated range will simply be referred to as the range at
a plant site.
Instrument accuracy
Accuracy
Error
Scale of Measurement
v100%
The instrument accuracy is published by the manufacturer in the
product documentation, which is easily obtained on-line. A few
examples of how accuracy can be expressed are:
Best-in-class performance with 0.025% accuracy
0.10% reference accuracy
0.065% of span
These examples refer to the ideal instrument accuracy, which is
only the accuracy of the measuring device itself. The total ac-
curacy, on the other hand, includes the instrument accuracy plus
all other factors that contribute to error in the measured reading
as compared to the actual value. These other factors can include
digital to analog conversions, density errors, piping configura-
tions, calibration errors, vibration errors, plugging and more.
Turndown ratio
The ratio of the maximum to minimum accurate value is an impor-
tant factor in considering the total accuracy of a measured value.
Turndown ratio
imum accurate value
imu
=
max
min mm accurate value
For example, an instrument with 100:1 turndown and 0100-
psi instrument range would have the stated instrument accuracy
down to 1 psi. Below 1 psi, the instrument might read, but it will
have greater inaccuracy.
measurements. Subtracting two pres-
sure readings is not always accurate
enough to obtain a meaningful mea-
surement, so it is important to consider
the span of the anticipated measured
readings. If the dP is a substantial frac-
tion of the top pressure, then it is okay
to subtract the readings of two pres-
sure transmitters. However, if the dP
is a small fraction of the top pressure,
then it will be within the instrument
error of the pressure transmitter.
For example, a column at a plant runs
at 30 psia top pressure. The expected
dP is 2-in. H
2
O over a few trays. The in-
strument error for a 050 psi pressure
transmitter is 1.4-in. H
2
O. The mea-
surement is within the accuracy of the
pressure transmitters, and a dP meter
is the appropriate meter to obtain an
accurate measurement. The downside
of dP meters is that very long impulse
lines are required on tall towers.
LeveL
Level and flow are the hardest basic
things to measure on a distillation
tower. Kister reports that tower base
level and reboiler return problems
rank second in the top ten tower mal-
functions, citing that Half of the case
studies reported were liquid levels ris-
ing above the reboiler return inlet or
the bottom gas feed. Faulty level mea-
surement or control tops the causes
of these high levels...Results in tower
flooding, instability, and poor separa-
tion...Vapor slugging through the liq-
uid also caused tray or packing uplift
and damage. (Reference 2, p. 145)
One of the main reasons for faulty
level indications is that dP me-
ters are the most common type of
level instrument, and an accurate
density is required to convert the dP
reading to a level reading. In many
cases, froth in the liquid level de-
creases the actual density and causes
faulty readings. Changes in composi-
tion or the introduction of a different
process feed with a different density
are cited several times as reasons for
level measurement problems. Plug-
ging of impulse lines and equipment
arrangements that make accurate
readings impossible are also very
common problems.
Differential pressure transmitters
are the most common type of level
transmitter. The accuracy of the in-
strument is quite good, at 0.1% of
span (calibrated range). Any type of
dP meter can be used: flush-mounted
diaphragms, remote-seal diaphragms,
purged impulse-line, or non-purged
impulse-line pressure transmitters.
The level measurement is dependent
on the density of the fluid:
P
height of liquid
R
l
ft ,
An accurate density is required for
calibration. Changes in composition or
the introduction of a process feed with
a different density will cause errone-
ous readings. Level transmitters suf-
fer from the same problems that occur
in pressure transmitters. Hydraulic
fluid leaks, compatibility of the hy-
draulic fluid, damage to diaphragms,
and plugging or freezing of impulse
lines are just a few of the problems
that can be encountered with dP level
transmitters.
Example 1. A column in a high-tem-
perature, fouling service began to ex-
perience high pressure drop, and the
plant engineers were concerned that
they were flooding the column. Calcu-
lations showed that the tower should
not be flooding if the trays were not
damaged. Downcomer flooding was a
possibility if the cartridge trays had
become dislodged and reduced the
downcomer clearance. The tower was
taken down, and internal inspection
revealed no damage to the internals.
It was determined that a false low
level caused the bottoms flow control-
ler to close. This raised the level in the
tower above the reboiler return line
and above the lower column pressure
tap. The column dP meter was reading
the height of liquid above the lower-
column pressure tap. Consultation
with the instrument manufacturer
revealed that the remote, seal hydrau-
lic fluid was not appropriate for the
high temperature of the process. The
hydraulic fluid was boiling in the cap-
illary tubes and had deformed the dia-
phragm, which was also coated from
the fouling service. The level transmit-
ter was switched to a periodic, purged
impulse-line dP meter. An automated
high-flow nitrogen purge prevents ac-
cumulation of the solids in the impulse
lines and is done once per shift. Logic
was added to the control loop to main-
tain the previous level reading during
the short nitrogen purge, a method
that has eliminated the problem with
the level.
Example 2. Another common ex-
ample of a level transmitter failure
is based on the fact that equipment
is designed in such a way that an ac-
curate level reading can never be ob-
tained. Though this may be surpris-
Feature report
50 ChemiCal engineering www.Che.Com marCh 2008
LT
ReboiIer
return
Figure 3. (left) Remote-seal
diaphragm pressure transmitters
are used in high-temperature
service
Figure 4. (above) Location of
reboiler return nozzle does not
allow for accurate level reading
Figure 6. Non-contact radar level
transmitters generate waves that are
reflected from the surface of the level
back to the transmitter
Figure 5. Nuclear level transmit-
ters are non-contact devices
ing, it is mentioned in Ref. [3], Case
Study 8.4 (p. 149), as illustrated in
Figure 4. A column that was being
retrofitted was originally designed
so that the reboiler return was intro-
duced directly between the two liq-
uid-level taps. The level in the tower
could never be accurately measured,
and it was modified on the retrofit to
rectify this situation.
Nuclear level transmitters
Common in polymer, slurry and
highly corrosive or fouling services,
these instruments work by placing a
radioactive source on one side of the
vessel and a detector on the other side.
The amount of radiation reaching the
detector depends on how much mate-
rial is inside the vessel. A strip source
and strip detector are more accurate
than a single source, strip detector. A
sketch of a single source, strip detector
is shown in Figure 5. The advantage
of nuclear level transmitters is that
they are non-contact devices, making
them ideal for services where the pro-
cess fluid would coat or damage other
types of level instruments.
Nuclear level transmitters are more
expensive than other level devices.
They also require permits and a radia-
tion safety officer, so they are often only
used as a last resort. The instrument
accuracy is generally 1% of span. The
total accuracy depends on how well the
system was understood by the designer
and installer. The thickness of the ves-
sel walls and any other metal protru-
sions in the measuring range, such as
baffles, must be taken into account in
the calibration, along with the correct
rate of decay of the source. Build-up of
solids in the measuring range will also
result in error.
Radar level
transmitters
This type of level transmit-
ter has been used in the
chemical processing in-
dustries (CPI) for the last
30 years. They demonstrate high ac-
curacy on oil tankers and have been
used frequently in storage-tank appli-
cations. Radar level transmitters are
now being applied to distillation tow-
ers but are still more commonly found
on auxiliary equipment, like reflux
tanks. There are contact and non-con-
tact types of radar level instruments.
A non-contact, radar level transmit-
ter generates an electromagnetic wave
from above the level being measured.
The wave hits the surface of the level
and is partially reflected to the instru-
ment. The distance to the surface is
calculated by measuring the time of
flight, which is the time it takes for the
reflected signal to reach the transmit-
ter. Some things that cause inaccuracy
with non-contact radar are: size of the
cone, heaving foaming, turbulence,
deposits on the antenna, and varying
dielectric constants caused by changes
in composition or service. The instru-
ment accuracy is reported as 5 mm.
Contact radar sends an electromag-
netic pulse down a wire to the vapor-
liquid interface. A sudden change in
the dielectric constant between the
vapor and the liquid causes some of
the signal to be reflected to the trans-
mitter. The time of flight of the re-
flected signal determines the level.
Guided wave radar can be used for
services where the dielectric constant
changes, but is not a good fit for fouling
services. A bridle (Figure 7), is used on
distillation towers to reduce turbulence
and foaming and therefore increases
the accuracy of the measurement. In-
strument accuracy is 0.1% of span.
Example 3. A reflux tank on a batch
distillation tower had a non-contact
radar level transmitter. The tower
stepped through a series of water
washes, solvent washes, and process
cuts. The reflux-tank level transmit-
ter gave false high readings during the
solvent wash cycle, which used tolu-
ene. The reflux pumps would always
gas off during this part of the process.
The dielectric constants of the various
fluids in the reflux tank, of which tolu-
ene had the lowest dielectric constant,
varied ten times during the cycle, af-
fecting the height of liquid able to be
measured. Larger antennas focus the
signal more and give greater signal
strength. As the dielectric constant de-
creases, a larger antenna is required to
measure the same height of fluid. The
level transmitter used in this service
was not appropriate for all measured
fluids and could not accurately mea-
sure the liquid level when the reflux
drum was inventoried with toluene.
TemPeraTure
There are two common types of tem-
perature transmitters in distillation
service thermocouples and Resis-
tive Temperature Devices (RTDs).
Both are installed in thermowells.
Thermocouples. The most popular
temperature transmitter, thermocou-
ples, consist of two wires of dissimilar
metals connected at one end. An elec-
tric potential is generated when there
ChemiCal engineering www.Che.Com marCh 2008 51
Figure 9. For
vapor or gas applica-
tions, orifice flowme-
ters require tempera-
ture and pressure
compensation
Figure 8. Resis-
tive temperature de-
tectors respond to a
temperature change
with a change in
resistance
Figure 7. Guided wave radar level
transmitter on a distillation
tower level [5]
is a temperature delta between the
joined end and the reference junction.
Type J thermocouples, made of iron
and Constantine, are commonly used
in the CPI for measuring tempera-
tures under 1,000C.
RTDs
The second most-common type of tem-
perature transmitter, RTDs consist of
a metal wire or fiber that responds to
a temperature change by changing
its resistance. Though RTDs are less
rugged than thermocouples, they are
also more accurate. Typically, they are
made of platinum. The instrument
accuracy of thermcouples and RTDs
is very good in both. However, ther-
mocouples have a higher error than
RTDs. The total accuracy of a thermo-
couple is 12C. There is greater error
due to calibration errors and cold-ref-
erence junction error.
It is important to note that, with
temperature transmitters, there is
a lag in the dynamic response to
changes in process temperatures. All
temperature measurements have a
slow response, because the mass of
the thermowell must change in tem-
perature before the thermocouple or
RTD can see the change. The lag time
will depend on the thickness of the
thermowell and on the installation.
The thermocouple and RTD must be
touching the tip of the thermowell for
best performance. If there is an air
gap between the thermowell and the
measuring device, the heat-transfer
resistance of the air will add substan-
tially to the lag time, which is also
why temperature transmitters work
better in liquid service. The response
time for temperature transmitters
in liquid service is between 110 s,
whereas the response time for tem-
perature transmitters in vapor service
is about 30 s. Heat-transfer paste is a
thermally conductive silicone grease;
it has been used with success in some
plants to improve the response time of
temperature transmitters.
Example. The plant in this example
experienced a temperature lag prob-
lem. A thermocouple near the bottom
of a large tower controlled the steam
to the reboiler. The temperature con-
trol point had a 10-min delayed re-
sponse to changes in steam flowrate.
The rest of the column responded to
the change in boilup in about 3 min.
The lag in the control point caused
cycling of the steam flowrate and cre-
ated an unstable control loop. The
cause was determined to be a thermo-
couple that was too short for its ther-
mowell. Normally, thermocouples are
spring-loaded to ensure that the tip
is touching the end of the thermow-
ell, but the instrument mechanics had
installed a thermocouple of the wrong
length because they lacked the proper
replacement part. The poor heat
transfer through the air gap between
the end of the thermocouple and the
thermowell caused the delay in tem-
perature response. Replacing the in-
stalled thermocouple with one of the
proper length fixed the problem.
FLow
There are many different types of
flowmeters. Here, the types commonly
used in plants will be discussed: orifice
plates, vortex shedding meters, mag-
netic flowmeters and mass flowmeters.
Orifice plates
Orifice plates are the most common
type of industrial flowmeter. They are
inexpensive, but they also have the
greatest error of all the common types
of flowmeters. Orifice plates measure
volumetric flowrate according to the
following equation:
Q C
P
v

$
R
1
2
Q is the volumetric flowrate, C is a con-
stant, P is the pressure drop across
the orifice, and is the fluid density.
To obtain an accurate flowrate, an ac-
curate fluid density must be known.
Temperature and pressure compen-
sation are required for vapor or gas
applications and may be required
for some liquids. Figure 9 shows the
equipment arrangement for an orifice
flowmeter with temperature and pres-
sure compensation.
Typical turndown for orifice plates is
10:1. Below 10% of span, the measure-
ment is extremely erroneous because
the volumetric flowrate is proportional
to the square root of the P. At 10% of
span, the meter is only measuring 1%
of the P span (Figure 10).
Multiple meters can be used to
overcome the turndown ratio when
high accuracy is required over the
entire span. This is often worth the
effort when measuring the flowrate
of raw materials or final products. At
one plant, three orifice plates in par-
allel were used to measure the plant-
boundary steam flowrate due to the
large span and the accuracy required
at the low end of the range. This re-
sulted in a very complicated system.
There are many common problems
that lead to error in the orifice plate
measurement, including inaccurate
density, impulse-line problems, erosion
of the orifice plate, and an inadequate
number of pipe diameters upstream
and downstream of the orifice plate.
An accurate density is required to
obtain an accurate flowrate. In a plant
that has a process feed that varies
from as low as 12% to as high as 30%
water, the density changes signifi-
cantly, and therefore an orifice meter
will not provide an accurate reading
without density compensation.
Impulse line problems include
plugging, freezing due to loss of elec-
tric heat tracing, and leaking. Con-
densate filling the impulse lines in
vapor/gas service and gas bubbles
in the impulse lines in liquid service
are also commonly cited. Figure 11
shows a pipe just upstream of an ori-
fice that was in clean water service
for two years. There was a filter just
upstream of this section of pipe. The
Feature report
52 ChemiCal engineering www.Che.Com marCh 2008
Orifice PIate Square Root Function
0
0 10 20 30 40 50 60 70 80 90 100
10
20
30
40
50
60
70
80
90
100
F
I
o
w
r
a
t
e
,

g
p
m
BeIow 10% of scaIe, error is high for
fIow sensor and dp measurement
Orifice pIate pressure drop (in H2O)
Figure 10. Volumetric flowrate is proportional to
the square root of the P, causing high error at less
than 10% of span
Figure 11. Due to impulse
line problems, this "clean" ser-
vice did not meet standards
impulse lines to the orifice plate flow-
meter were completely plugged. This
section of pipe was removed and a
Teflon-lined magnetic flowmeter was
installed instead.
Orifice plates can erode, especially
in vapor service with some entrained
liquid. This is common in steam ser-
vice, and orifice plates should be
checked every three years for wear.
Orifice plates generally need 20
pipe diameters upstream and 10 pipe
diameters downstream of the orifice
plate for the velocity profile to fully
develop for predictable pressure-drop
measurement. This requirement var-
ies with the orifice type and the piping
arrangement. This is rarely achieved
in a plant, which introduces error in
the measurement.
The instrument accuracy of orifice
plates ranges from 0.752% of the
measured volumetric flowrate. Vari-
ous problems are encountered with
orifice plate installations, and they
have the highest error of all flowme-
ters. Orifice plates are, however, quite
sensitive to a variety of error-induc-
ing conditions. Precision in the bore
calculations, the quality of the instal-
lation, and the condition of the plate
itself determine total performance.
Installation factors include tap loca-
tion and condition, condition of the
process pipe, adequacy of straight
pipe runs, gasket interference, mis-
alignment of pipe and orifice bores,
and lead line design. Other adverse
conditions include the dulling of the
sharp edge or nicks caused by corro-
sion or erosion, warpage of the plate
due to water hammer and dirt, and
grease or secondary phase deposits
on either orifice surface. Any of the
above conditions can change the ori-
fice discharge coefficient by as much
as 10%. In combination, these prob-
lems can be even more worrisome and
the net effect unpredictable. There-
fore, under average operating condi-
tions, a typical orifice installation can
be expected to have an overall inac-
curacy in the range of 2 to 5% AR (ac-
tual reading) [6].
Vortex shedding meters
Vortex shedding meters contain a
bluff body, or a shedder bar, that cre-
ates vortices downstream of the object
when a fluid flows past it. The meters
utilize the principle that the frequency
of vortex generation is proportional
to the velocity of the fluid. The whis-
tling sound that wind makes blowing
through tree branches demonstrates
the same phenomenon.
The fluids density and viscosity are
used to set a k factor, which is used
to calculate the fluid velocity from
the frequency measurement. The fre-
quency, or vibration, sensor can either
be internal or external to the shedder
bar. The velocity of the fluid is con-
verted to a mass flowrate using the
fluid density. Therefore, accurate fluid
density is important for accurate mea-
surements. Vortex meters work well
both in liquid and gas service. They
are commonly used in steam service
because they can handle high tem-
peratures. They are available in many
different materials of construction and
can be used in corrosive service.
Vortex meters have lower pressure
drop and higher accuracy than orifice
plates. A minimum Reynolds num-
ber (Re
min
) is required to achieve the
manufacturers stated accuracy. Vortex
meters exhibit non-linear operation
as they transition from turbulent to
laminar flow. Typical accuracy above
the Re
min
is 0.651.5% of the actual
reading. In general, the meter size
must be smaller than the piping size
to stay above the Re
min
throughout
the desired span. The requirements
for straight runs of pipe upstream and
downstream of the meter vary, but
both are usually longer than for orifice
plates. In general, 30 pipe diameters
are required upstream and 15 pipe
diameters downstream. The upstream
and downstream piping must be the
same size pipe as the meter.
There are only a few problems com-
monly encountered with vortex me-
ters. Older models may be sensitive to
building vibrations, but newer models
have overcome this issue. If the shed-
der bar becomes coated or fouled, the
internal vibration sensor will cease to
work. This can be avoided by using an
external vibration sensor. The most
common issue is failing to meet the
Re
min
requirements over the desired
span. At one plant, every vortex meter
was line-sized, which means it was
the same size as the surrounding pip-
ing. The flow went into the laminar
region in the desired measured range
in every case. The flow read zero when
it transitions to laminar, making the
meters useless.
Example. Another good example of
failing to meet the Re
min
require-
ments over the desired span hap-
pened on a project where a tower that
had been out of service for some time
was recommissioned. The distillate
flowrate was substantially lower than
the original tower design and was in
the laminar flow region over the en-
tire operating range. The distillate
flow was a major control point on the
tower, but the vortex meter could not
read the flowrate. The control strat-
egy had to be changed to work around
this issue until an appropriate meter
could be installed.
Magnetic flowmeters
Faradays law states that the voltage
induced across any conductor as it
moves at right angles through a mag-
ChemiCal engineering www.Che.Com marCh 2008 53
Figure 12. Vortex meters contain a shedder bar that creates vortices down-
stream when fluid flows past it (left). Depending on the application and pipe size,
vortex shedding meters are available in a range of sizes and shapes (right)
Figure 13. The magnetic flowmeter prin-
ciple states that the voltage induced across
a conductor as it moves at right angles
through a magnetic field is proportional to its
velocity.
netic field is proportional to the
velocity of that conductor. This
is the principle used to measure
velocity in magnetic flowmeters,
which are commonly referenced
to as mag meters (Figure 13).
Mag flowmeters measure the
volumetric flowrate of conduc-
tive liquids. Fluids like pure
organics or deionized water do
not have a high enough conduc-
tivity for a mag meter. An ac-
curate density is required to convert
the volumetric flowrate to a mass
flowrate. The meters are line-sized,
but they have a minimum and maxi-
mum velocity to achieve the stated
instrument accuracy. A smaller line
size may be necessary to achieve the
velocity requirements throughout
the desired span. The instrument
accuracy is quite good, generally
at 0.5% of the actual reading. The
error is very high below the mini-
mum velocity. Turndown for newer
mag meters is 30:1, but older models
will be closer to 10:1.
Mag meters do not have a lot of op-
erating problems. They must be liq-
uid-full to get an accurate reading and
are often placed in vertical piping to
achieve this. They rarely plug as they
can be specified with Teflon liners and
are often used in slurry service. Mag
flowmeters are more expensive to
install because they usually require
110-V power.
Mass flowmeters
Mass flowmeters use the Coriolis effect
to measure mass flowrate and density.
A very small oscillating force is applied
to the meters flowtube, perpendicular
to the direction of the flowing fluid. The
oscillations cause Coriolis forces in the
fluid, which deform or twist the flow-
tube. Sensors at the inlet and outlet of
the flowtube measure the change in the
geometry of the flowtube, which is used
to calculate the mass flowrate. The os-
cillation frequency is used to measure
the fluids density. The temperature of
the fluid is measured to compensate for
thermal influences and can be chosen
as an output of the meter.
The original mass meters were U-
tubes, but several different shapes
are now available, including straight
tubes as shown in Figure 14. Mass
flowmeters have the highest ac-
curacy of all the different types of
flowmeters, usually 0.10.4% of the
actual reading. The measurement is
independent of the fluids physical
properties, making mass flowmeters
unique in that most flowmeters re-
quire the fluid density as an input.
Mass flowmeters are insensitive to
upstream and downstream pipe con-
figurations. Practical turndown is
100:1, although the manufacturers
claim 1,000:1. The density measure-
ment is not as accurate as a density
meter. Mass flowmeters are gener-
ally very reliable and only require
periodic calibration to zero them.
Mass flowmeters are on the expen-
sive end to purchase and to install.
They require 110-V power. Pressure
drop can sometimes be an issue, and
the meters are only available in line
sizes up to 6 in. Coating of the inside
of the flowtube will result in higher
pressure drop and can result in loss of
range and accuracy if the tube is re-
stricted. Wear and corrosion can result
in a gradual change of the mechanical
characteristics of the tube, resulting
in error. Zero stability was an issue
with older meters but this problem
has been solved in newer units.
Example 1. The reflux flowrate on
a final product column was an im-
portant measurement, and the reli-
ability of the existing flowmeter was
questioned. Product literature for
mass flowmeters promised high ac-
curacy and low pressure drop. The
plant-area engineer coordinated a
small project to replace the existing
orifice plate flowmeter with a mass
flowmeter. Column performance was
very poor after startup. The new
meter had to be bypassed to operate
the column normally. The overhead
condenser was gravity drained, and
the new mass flowmeter had enough
additional pressure drop to force the
liquid level into the condenser tubes
and restrict rates an expensive les-
son for a new engineer.
Example 2. Another tower had a
mass flowmeter installed on the bot-
toms flow, which was pumped but
not cooled. The mass flowmeter al-
ways had erratic readings and was
never believed. A closer examination
of the system revealed enough pres-
sure drop through the mass flowme-
ter to result in flashing in the flow-
tube. The two-phase flow caused the
erratic readings.

Epilogue
With a knowledge of the basics of
column instrumentation, the ques-
tion posed in the introduction should
seem trivial. Our experienced en-
gineer had concluded that the bot-
toms flowrate of the column had to
be erroneous, but the instrument
group had disagreed. The flowmeter
in question was a mass flowmeter in
relatively clean and non-corrosive
service. The other three flowmeters
on the column were orifice plates
and are known to have a myriad of
problems that introduce error.
Summary
Some basic knowledge of instrumen-
tation can be a very valuable trouble-
shooting and design tool. Gauging
whether an instrument installation
will ever give accurate readings or
whether it is an expensive spool piece
is useful in itself. Being able to assess
the relative accuracy of two measure-
ments will help determine from which
data to draw conclusions. Knowledge
of common instrument problems can
help in troubleshooting.
Get to know the instrumentation
Feature report
54 ChemiCal engineering www.Che.Com marCh 2008
A
B
A
B
A
B
1 2 3 No fIow FIow
Figure 14. Mass flowmeters use the Coriolis effect to infer mass flowrate from the mea-
surement of flowtube deflection
on your towers. Gather the manufac-
turers information so you can assess
the instrument accuracy. Keep in
mind that the manufacturers litera-
ture refers to the ideal instrument
accuracy, which is the accuracy of the
measuring device itself. There are
many other factors that contribute
to the accuracy of the reading that is
displayed on the DCS screen or in the
data historian. The total accuracy in-
cludes the instrument accuracy plus
all of the other things that contribute
to error in the measured reading as
compared to the actual value. Other
inaccuracies lie in digital to analog
conversions, density errors, piping
configurations, calibration errors, vi-
bration errors, and the list goes on
and on. Check the field installation
to see what types of problems your
meters will experience.
Get to know your mechanics and in-
strumentation experts at your plant.
Now that you know some of the lingo
of instrumentation, you can better
converse with your instrument engi-
neers and mechanics.
Acknowledgements
This paper is a compilation of in-
strumentation basics obtained from
the references listed below, of trou-
bleshooting experience from many
colleagues at DuPont, and of trou-
bleshooting examples from Henry
Kisters most recent book, Distilla-
tion Troubleshooting. Much of the
technical information and many of
the examples come from Nick Sands,
Process Control Leader for DuPont
Chemical Solutions Enterprise in
Deepwater, N.J. Nick has worked for
DuPont for 17 years and is a special-
ist in process control. In addition
to Nick, the following DuPont col-
leagues contributed their instrument
war stories, and the author is grate-
ful for their willingness to share their
experiences:
Jim England, DuPont Electronic
Technologies (Circleville, Ohio)
Charles Orrock, DuPont Advanced
Fibers Systems (Richmond, Va.)
Adrienne Ashley, DuPont Advanced
Fibers Systems (Richmond, Va.)
Joe Flowers, DuPont Engineering
Research & Technology (Wilming-
ton, Del.)
References
1. Gillum, Donald R., Industrial Pressure,
Level and Density Measurement. Resources
for Measurement and Control Series. ISA,
1995.
2. Kister, Henry Z., Distillation Troubleshoot-
ing, John Wiley & Sons, 2006.
3. Spitzer, David W., Industrial Flow Measure-
ment. Resources for Measurement and Con-
trol Series. ISA, 1990
4. Trevathan, V. L., editor. A Guide to the Auto-
mation Body of Knowledge. ISA, 2006.
5. emersonprocess.com/rosemount
6. omega.com
7. efunda.com
8. us.endress.com
9. spiraxsarco.com
A
B
A
B
A
B
1 2 3 No fIow FIow
Figure 15. With an under-
standing of the accuracies of
mass flowmeters and orifice
flowmeters, we revisit the
question Which flowmeter
is the most accurate?
Author
Ruth Sands is a senior con-
sulting engineer for DuPont
Engineering Research &
Technology (Heat, Mass &
Momentum Transfer Group,
1007 Market St., B8218,
Wilmington, DE 19898; Phone:
302-774-0016; Fax: 302-774-
2457; Email: ruth.r.sands@
usa.dupont.com). She has
specialized for the last nine
years in mass transfer unit
operations: distillation, extraction, absorption,
adsorption, and ion exchange. Her activities
include new designs and retrofits, pilot plant
testing, evaluation of flowsheet alternatives, and
troubleshooting. She has 17 years of experience
with DuPont, which includes assignments in
process engineering, manufacturing, and corpo-
rate recruiting. She holds a B.S.Ch.E. from West
Virginia University, is a registered professional
engineer in the state of Delaware, and is a mem-
ber of the FRI Executive Committee.
MI CROFI LTRATI ON
ULTRAFI LTRATI ON
NANOF I LT RAT I ON
WWW.MICRODYN-NADIR.COM
H
a
v
e

a

b
a
t
h

w
i
t
h

a

f
r
i
e
n
d
LEIPZIG 14. - 18. APRIL
STAND G9
I FAT MUNI CH
5. - 9. MAY
HALL A2
STAND 303
MICRODYN
TECHNOLOGIES INC
P. O. Box 98269
Raleigh, NC. 27624
Phone 001 - 919 - 341-5936
[email protected]
7>D"8:A

hjWbZg\ZYBZbWgVcZBdYjaZ
[dgW^dad\^XValVhiZlViZgigZVibZci
56x254_biocel_messe_usa_rz.indd 1 07.02.2008 15:36:39 Uhr
Circle 31 on p. 76 or go to
adlinks.che.com/7370-31
Mass
fIowmeter
Orifice
fIowmeter
Orifice
fIowmeter
Orifice
fIowmeter
Most IikeIy to be most accurate, even at
Iow fIowrates. Insensitive to piping con-
figuration. UnIess fIow is erratic due to
fIashing in Iine or the tube is corroded or
severeIy pIugged, the orifice pIates are
much more IikeIy to have higher error.
FT
FT
FT
FT