Part Five

CONTROL AND
OPTIMIZATION TECHNIQUES

The association o f a number o f refining and/or petrochemical units is often

termed an industrial complex. The interactions among the differentparts making
up the whole are so diverse and o f such a variety of types that the name is well
suited to the reality. The complexity can be seen even in a simple processing unit,
such as distillation or catalytic cracking.
In order to control this complexity, the operator relies on all the measurement
and control techniques available, which may sometimes seem highly sophisti-
cated. Nevertheless, given the size o f equipment and the capacity o f the units,
nothing must be left to chance for reasons o f both safety and profitability.
Chapter 10 presents the measurement and control techniques commonly used
in refineries, while Chapter I I deals with certain aspects o f energy optimization
in a refining or petrochemical complex.
 10
Control and Monitoring

Jacques Ludaescher

10.1 Introduction

Refining is a processing industry that implements a variety of processes such

as those described in the preceding volumes. Refinery processing must be car-
ried out in the most reliable and automatic way possible. There must be a min-
imum number of shutdowns, constant reproducible quality and only the
strictly necessary manpower.
These objectives can be reached thanks t o the use of process control tech-
niques that allow all the physicochemical parameters @ressure, temperature,
flow rate, level, etc.) to be measured. Then by comparison with the target Val-
ues, corrective action is worked out that is consistent with the process being
controlled.
The introduction of industrial data processing permits advanced optimized
control of refining procedures on the basis of sophisticated strategies. It also
allows linkage with production management and maintenance work.
The following sections will give only a condensed description of the major
techniques used in controlling refining units. Readers seeking a more in-depth
study of certain aspects can consult the reference section at the end of the
chapter.

1C.L Basic Principles of Industrial Measurements

Measurements in an industrial context must be able to quantify, compare and

check, duplicate, recopy and repeat all the physical and chemical values
needed to monitor processing.
618 ChaDter 10 CONTROL AND MONITORING

In industry two types of measurement are used:

Industrial measurement, which is essentially utilitarian and character-
ized by its effectiveness and must meet operation and safety needs.
Accounting measurement, which is identical t o the first type, but gov-
erned by methods and modes of expressing results and requires means
of verification and control.
A number of terms are used to define the type of value measured, with the
most common being the following:
Absolute: when the measurement does not depend on an arbitrarily cho-
sen reference (for example absolute temperature, absolute pressure).
Analog: when the value can vary continuously (movement of a needle on a
graduated dial for example).
Rating: the maximum capacity of a measurement instrument.
Precision class: defines the maximum absolute error of a measurement
instrument.
Discrete: describes a value that varies by step.
Scale: a graduation used t o compare the values representing variables. The
scale must be given by the two extreme values expressed in standard Val-
ues of the variable under consideration.
Error: deviations between the result of a measurement and the theoretical
value of the measurement, classified according t o two criteria:
- expression: absolute or relative,
- causes: random or systematic.
Absolute error: deviation between the result and reality, expressed in
the same unit as the result.
Relative error: quotient of the absolute error by the result, generally
expressed in %.
Random error: error that is not governed by any law or error on a sin-
gle result.
Systematic error: error due to a known and recognizable cause.
Standard: a value that serves as a reference to determine units and to cal-
ibrate measurement instruments.
Reliability: the aptitude of an instrument to fulfill the function it was built
for.
Repeatability: the quality of a measurement instrument whose measure-
ment uncertainties are small and leave little room for random errors.
Fineness: the quality of an instrument that does not disturb the quality or
state of the variable being measured.
Hysteresis: the sensitivity of an instrument in detecting a change in the
value being measured.
Accuracy: the quality of an instrument that leaves little room for system-
atic errors, the measured value is therefore very close to the true value.
 Chapter 10. CONTROL AND MONITORING 619

Linearity: the quality of an instrument whose indication varies proportion-

ally to the variations in the variable being measured.
Resolution: the aptitude of an instrument to perceive small differences.
Accuracy: the quality of an instrument that is exact and repeatable, gener-
ally expressed in % of the maximum value of the adjustment range.
Unit: a quantity defined arbitrarily or according to known rules, used to
establish comparisons. The official system is the international system of
units (So.

10.2.1 Pressure
The concept of pressure is generally known as an elementary force exerted by
a fluid on a unit of surface area. There are different types of pressure:
Static pressure: the pressure that a motionless fluid exerts on the parts of
the vessel containing it.
The steady state flow of an ideal fluid (zero viscosity) in the field of gravity
is governed by the Bernoulli theorem. This theorem is expressed by the fol-
lowing equation:
p.
2
'+ pgz + P = Constant
with:
p fluid density (kg/m3)
u fluid velocity (m/s)
z height related to a reference plane (m)
g acceleration of gravity (= 9.81 m/s2)
P pressure (Pa)
Static pressure (P,) corresponds to the fluid at rest (u = 0), (Fig. 10.1A):
P, = P + pgz
Dynamic pressure: pressure (Pd) due to velocity, whose existence is
demonstrated by opposing an obstacle to the movement (Fig. 10.1B):
P U2
pd= 2
Total pressure: total pressure (PJ is the sum of the two previous pressures
(Fig. 10.1C):
Pt = P, + Pd
In a real fluid the forces due to viscosity are not zero and oppose the slip-
ping of fluid layers in relation to one another. They are dependent on velocity
and tend toward zero with it.
The Bernoulli equation is not applicable for viscous flow. The total pres-
sure is not constant along a stream line, it decreases. There is said to be a
pressure drop along the flow.
620 Chapter 10 CONTROL AND MONITORING

‘1‘
To pressure gauge

‘1‘
To pressure gauge

Figure
A. Static pressure probe (A frer “Capteurs industriels’y. B. Measurement of
dynamic pressure. C Total pressure probe.

Atmospheric pressure: this pressure is variable depending on the location

and the atmospheric conditions. It is the pressure measured by a barometer
and is approximately 760 mm of mercury, i.e. 1013 mbar.
Absolute pressure: this is the pressure measured in relation to an absolute
vacuum.
Gauge pressure: this is the pressure measured in relation to atmospheric
pressure. It is the pressure normally measured in an industrial facility. Vacuum
pressure (or negative pressure) is a relative pressure below atmospheric level.
Differential pressure: this is the difference between two pressures
expressed in identical units.
Pressure unit
In the SI system the unit of pressure is the pascal (pa) which corresponds
to the pressure exerted by a force of 1 newton uniformly distributed on a sur-
face area of 1 m2. The pascal is mainly employed for very low pressures. Two
multiples of the pascal are commonly used:
the bar for medium and high pressures, 1 bar = lo5 Pa,
the mbar for low pressures, 1 mbar = lo2 Pa.
 Chapter 10 CONTROL AND MONITORING 621

C D

--
7 t
P
P

Figure
10.2 Principles of pressure measurement (After “Capteurs industriels’y
A. Capsule. B. Bellows. C. Bourdon tube. D. Diaphragm.

Industrially the principles of pressure measurement are as follows

(Fig. 10.2):
Measurement of the physical deformation applied to a pressure element
such a s capsule, bellows, Bourdon tube or diaphragm. The deformations,
movements or elongations are generally proportional to the pressure.
They are amplified by a mechanical system and transmitted to a pointer
and/or an electric or pneumatic amplifying and conversion device.
Measurement of the deformation of a wire strain gauge inserted in a
Wheatstone bridge assembly. Two types of gauges are used: metallic and
semi-conductor. The second type are now the most common since they
offer a number of advantages over conventional metallic gauges. They
have a high gauge factor (from 50 to 270), practically no hysteresis, great
precision and good stability over time.
Capacitive measurement on a ceramic element.

10.2.2 Flow Rate

The flow rate of a fluid in pipe is the amount of the fluid that goes through a
flow cross-section during the unit of time. The amount can be expressed either
by the volume (Qv) or the mass (Q,).
As a result:
Q,=PQv
622 Chapter 10 CONTROL AND MONITORING

The velocity provided by the differential pressure is generally used.

Differential pressure is obtained by inserting a flow element (orifice plate, noz-
zle, etc.) in the pipe of cross-section S. The volume flow rate is given by:
Q,, = KSU
with K being the calibration coefficient.
The velocity u is a function of the differential pressure h, as indicated
below:

with the differential pressure h expressed in millimeters of water column.

In the pipe a static pressure (P,) prevails which can be higher or lower than
the atmospheric pressure, along with a total pressure (Pt) equal to P, + h. With
respect to the circulation direction of the fluid in the pipe, the upstream pres-
sure indicated (+) is higher than the downstream pressure noted (-). The dif-
ference h measures the pressure drop in the channel of the differential pres-
sure sensor. Figure 10.3 summarizes the different definitions and the way they
are measured, along with the generally accepted set up for a flow-rate mea-
surement.
In an initial approximation, liquids can be considered as incompressible,
i.e. their volume depends only slightly on the pressure. This is not true for
gases and vapors whose density is proportional to the pressure (gas likened
to an ideal gas). It can therefore be considered that at constant temperature
the mass flow rate of a liquid is proportional to its volume flow rate. For a gas
there is proportionality only at constant pressure and temperature. There are
a number of formulas allowing a volume flow rate to be converted into a mass
flow rate and vice versa, depending on the different pressure and temperature
conditions and the nature of the fluid, etc.

-_ _
Figure
10.3 Measuring differential pressure.
 Chapter 10 CONTROL AND MONITORING 623

The instantaneous (volume or mass) flow rate of a liquid or a gas is mea-

sured by flow meters. Most flow meters indicate the volume that has flowed
per unit of time. These are volumetric flow meters. In some applications (cal-
culating the heat energy of a fuel for example) it is necessary t o determine the
mass flow rate, which can be calculated from the volume flow rate or mea-
sured directly.
The total flow rate is the mass or volume of fluid that has flowed during a
specified time. In order to measure a total flow rate, a positive displacement
flow meter is used. An instantaneous flow rate measurement can also be inte-
grated. The main principles of flow rate measurement are discussed below.
1. Flow meters with a differential pressure element (Fig. 10.4), based on
the Bernoulli law relating the flow rate and the pressure drop resulting from a
change in pipe cross-section. These devices can be used only when the flow is
turbulent, the differential pressure generated is then proportional to the
square of the average fluid velocity, with the pipe cross-section being known.
The main flow elements are as follows:
The orifice (Fig. 10.4A) is the simplest, least expensive and takes up the
least amount of space, but must be used on clean fluids and when there
is no limitation on pressure drop.
The nozzle (Fig. 10.4B) is an orifice that is streamlined so that it follows
the fluid stream better. The pressure drop is thus lower.
The venturi tube (Fig. 10.4C) is the device with the best pressure drop,
wear and fouling performance. It provides good precision, but is costly
and takes up a lot of space.

A B

___ I
Figure
10.4 Flow meter with a differential pressure element (After “Capteurs industriels’y.
A. Orifice. B. Nozzle. C. Venturi.
624 ChaDler 10 AND MONITORING
CONTROL

2. Pilot tubes. Pitot and Annubar tubes, consist of a probe that is inserted
in the pipe. The precision obtained is approximately 1%of the real value. They
are used on large dimension pipe and when the allowable pressure drop is low.
3. Area meters with a rotating float placed in a vertical conical tube. The
float is supported by the differential pressure generated by the annular restric-
tion between the float and the tube. The float moves up or down until the
annular restriction creates just enough pressure drop to balance the weight of
the float. The float position in front of the tube graduation, or a mechanical or
magnetic coupling, allows the measurement to be read.
4. Vortex flow meters are based on a natural phenomenon generated when
fluid flow encounters a non-streamlined obstacle, thereby forming eddies that
are carried downstream. The frequency of eddy emission is proportional to
the flow velocity. It is measured by means of a sensor that is sensitive to oscil-
latory variations in pressure or velocity.
5. Electromagnetic flow meters are based on Faraday's law. The magnetic
field is generated by two inductor coils placed on either side of the same pip-
ing diameter. The fluid itself is the conductor. The electromotive force gener-
ated is proportional to the average liquid velocity and therefore to the volume
flow rate of the liquid.
6. Coriolis mass flow meters are based on Newton's law of gravitation (law
of gravitation, i.e. force = mass x acceleration, F = ma). The device is com-
posed of two measurement tubes excited by an electromagnetic device that
causes them to vibrate in phase opposition on their resonance frequency, like
a tuning fork. The degree of twisting in the tube is proportional to the mass
flow rate. Electromagnetic sensors located on each side of the tube induce
sinusoidal tensions that allow the tube positions to be measured at these two
points. Given the set up, the specific gravity measurement is also available and
the measurement is temperature compensated.
7. Turbine or propeller velocity meters can be classified into two cate-
gories: axial turbine and vertical turbine meters. In both cases, the fluid flow
causes a turbine or propeller to rotate and the rotation speed is proportional
to the flow rate.
8. Positive displacement meter measurement involves a well defined vol-
ume of fluid. There are different types depending on the type of test body
used: meters with vanes, pistons and oval wheels. Transmission between the
test body and the indicator is generally mechanical with coupling to a meter
or a pulse generator allowing remote transmission.

10.2.3 Level
Liquid sensors and detectors can be classified into three main types depend-
ing on the method employed:
Liquid level measurement and monitoring based on hydrostatic proper-
ties of liquids (hydrostatic pressure, buoyancy). A pneumatic or electric
translation is obtained by detecting a position, a movement, a force or a
pressure.
 Chapter 10. CONTROL AND MONITORING 625

Measurement utilizing the electric properties of liquids (electric con-

ductivity and capacitance). They directly translate the level into an elec-
tric signal.
Measurement using radiation to measure levels without any contact with
the product.

a. Hydrostatic Level Measurements

Displacer measurement: the displacer is subjected to a force by the liquid
depending on the depth of immersion. The level measurement is actually the
measurement of the buoyancy exerted on the displacer. It is accomplished by
means of a torque tube. This principle is used for measurement ranges of less
than 2 meters and for interface measurements of two liquids with different spe-
cific gravities.
Float measurement: this is the oldest device for measuring and especially
for detecting a liquid level. It uses buoyancy directly with a solid float that
detects the position of the liquid level. The float is in direct contact with the
liquid, so the construction materials must be appropriate. The system is not
well suited to viscous liquids or liquids with variable specific gravities.
Gauge measurement: the level measurement is deduced from a measure-
ment of the pressure exerted by the hydrostatic pressure on the measuring
element. It must be corrected taking into account the static pressure prevail-
ing when the liquid is in a closed pressurized tank. The most common method
employs effective or differential pressure sensors. This technique can be
adapted to almost all liquids by using separators or indirect measurement by
bubbling if necessary.
Detecting pin measurement: the principle of detecting pin devices con-
sists in measuring the distance separating the measuring device from the sur-
face of the product. They are used in measuring the level of variable specific
gravity liquids and are also suited for use with a solid product. The moving
detecting pin generates pulses at regular intervals, which are counted and
thereby give the measured level.

b. Electric Level Measurements

Capacitive measurement: this type is suited to measuring and detecting the
level in all types of conductive or non-conductive products. It works with a
probe submerged in the tank. For non-conductive products (petroleum, lube
oil, hydrocarbons) the probe consists of a metallic rod insulated from the tank
and forming a capacitor with it. When the probe is emerged, the dielectric is
the ambient air (dielectric constant equal to 1). In the presence of a non-con-
ductive product the capacitance increases because the products have a dielec-
tric constant greater than 1.
Conductive measurement: this type is suited only for conductive prod-
ucts. It allows high, low or intermediate level thresholds to be detected. When
the liquid touches an electrode, it grounds a low voltage alternating current
circuit. The grounding element is either the metallic tank or another electrode
when the tank is not metallic.
626 ChaDler 10 CONTROL AND I\/IONITORING

c. Radiation Level Measuvemenfs

Radio-isotope measurement: this equipment uses a radioactive cobalt source
(Co60) with a lifetime of approximately 5 years or a cesium source ( C S ~with~ ~ )
a lifetime of 32 years. The principle is based on the difference in gamma ray
absorption when the rays go through the air in the tank or through the prod-
uct. The method can be applied only to non-radioactive products whose den-
sity is very different from that of air. Since there is no contact with the prod-
uct being measured, the method is mainly utilized for large grain size solids, at
high temperature or under high pressure.
Ultrasound or radar measurement: this equipment is increasingly used,
especially for storage tanks containing raw, intermediate or finished products.
In ultrasound level measurement, the probe emits a train of pulses that are
reflected at the liquid surface. They are then detected by the probe, which is
both transmitter and receiver. The level distance is directly proportional to
the pulse train two-way travel time.

10.2.4 Temperature
The two basic principles used for measuring temperature are:
electric measurement by thermocouples, metallic resistances, semi-con-
ductors and quartz elements as primary measuring devices,
mechanical measurement with dilatation of a solid or expansion of a
fluid (gas or liquid).
Thermocouples: a thermocouple consists of two junctions or welds con-
necting two different metals or alloys. One of the junctions, placed at the mea-
surement point, is identified as the hot junction and the other one is called the
cold junction. The temperature of the cold junction is known and serves as a
reference point. The working principle of this device is based on the Seebeck
effect and uses the electromotive force generated by the temperature differ-
ence between the two junctions. Although the electromotive force is low it is
measurable and depends on the nature of the materials.
Metallic resistance elements: this measurement principle uses the varia-
tion in electric resistance of a metallic conductor with temperature. Although
copper or nickel have advantageous characteristics, platinum is the most com-
monly used due t o the repeatability of its response and because probes of the
same model are absolutely interchangeable. Thermometric resistances pro-
vide an important advantage over thermocouples: the measurement depends
only on the temperature of the sensitive element. This makes measurements
simpler and more accurate, but the response time is longer.
Solid dilatation thermometer:this method is based on the linear dilatation
of solids with temperature. Bimetallic strip thermometers use this dilatation
that causes the strip to bend. Many grades of metallic strip are available and
selection is governed by the difference in dilatation, which should be as high
 Chanter 10 CONTROL AND MONITORING 627

as possible, and the type of environment. The sensitivity of a bimetallic ele-

ment is estimated by its Villarceau coefficient or by its specific deflection.
Liquid or gas expansion thermometer: these thermometers use the high
expansion coefficients of fluids. There are three types of thermometers: liquid
expansion, gas expansion and vapor pressure.
For the liquid type, temperature causes the liquid volume to expand.
These sensors feature good linearity and identical sensitivity over the
whole measurement scale.
For the gas type, temperature causes a variation in volume or pressure
of the deformable device. The advantages are similar to those of liquid
expansion thermometers.
Vapor pressure thermometers are based on the fact that the pressure of
a saturating vapor depends on the temperature. They are not linear.

10.2.5 Miscellaneous
Other measurements are required in order to monitor and operate refinery
equipment, such as mechanical measurements used for example to monitor
rotating machines (compressors, pumps, turbines, etc.):
Measurement of acceleration corresponds to a relationship between a
force and a mass (Fig. 10.5A). Accelerometers can be classified according
to the phenomena analyzed, and in particular their levels of acceleration
and their frequency ranges. Measurement of accelerations, vibrations or
impacts (Fig. 10.5B) can be involved depending on the case.
Measurement of motion uses the principles of electric measurement of
the resistive, inductive or capacitive type. The choice is made according
t o the values of motion being measured.

A
-
I
Spring
B

Vibration

Acceleration
Damping device Seismic mass
L
Duration

-
Figure
10.5 A. Principle of an acceleration sensor (After “Capteurs industriels ’7.
B. Acceleration curves.
628 Chapter 10 CONTROL AND MONITORING

Measurement of velocity or tachymeter: there are two types of sensors

depending on whether a linear or an angular velocity is measured. For lin-
ear velocities, wire sensors or differential transformer sensors are used.
For angular velocities, pulse tachymeters are found, which give a digital
output signal using inductive, Hall effect, capacitive or photoelectric
technologies along with analog output tachymeters which use direct or
alternating current electromagnetic technologies.
Position detectors are chiefly used to detect an open or closed position
(on-off valve), or a phenomenon that is absent or present (flame for
example). The two technologies employed can be either mechanical or
without any mechanical connection of the inductive, capacitive or pho-
toelectric type. Inductive or capacitive proximity detectors are the most
commonly used today.

10.3 Industrial Instrumentation Equipment

10.3.1 Sensors
The sensor is a device to receive information, the first step in any sequence of
measurements or data acquisition, in any pilot-operated, control, regulation,
monitoring or safety system.
The component parts of a sensor are as follows (Fig. 10.6):
the primary element (cell),
the transducer element (or sensitive element),
the housing,
the conditioning module (usually electronic, sometimes pneumatic).

Figure
10.6 Component parts of a sensor (After “Capteursindustriels”).
 Chapter 10. CONTROL AND MONITORING 629

The primary element is a mechanical device that reacts to the variable

being measured. Its function is to convert this variable into another measur-
able physical variable. This variable is the primary element reaction.
The transducer element is a sensitive element connected to the primary
element. It translates primary element reactions into a variable constituting
the output signal.
The housing is a mechanical element to protect, hold and attach the sen-
sor.
The electronic conditioning module has the following functions depend-
ing on the case:
electric power supply to the sensor (if needed),
shaping and amplifying the output signal,
filtering, correcting, processing the signal,
converting the signal (analog/digital, voltage/frequency, etc.),
adjusting the output signal for remote transmission.
This electronic module is called a transmitter and may or may not be incor-
porated in the sensor.

10.3.1.1 Pressure Instruments

A distinction must be made between absolute, relative and differential pres-

sure sensors.
Absolute pressure sensor: one of the faces (face -) of the primary element
comprises the wall of a chamber where a vacuum prevails. The other face
(face +) can be either in contact with the fluid whose pressure is being mea-
sured or connected to a diaphragm. The test body is always deformed in the
same direction from its initial position at rest.
Relative pressure sensor: the primary element face previously subjected
to the vacuum is left at atmospheric pressure. Precautions (selection of mate-
rials) are required if use in a difficult environment (a moist, salty or dusty, etc.
atmosphere) is anticipated.
Differential pressure sensor: this device can be used only to measure the
differences in pressure between some fluid on the (face +) side and a dry neu-
tral gas on the (face -) side. The sensor is said to be unidirectional if the pri-
mary element is deformed only in one direction from its position at rest and
bidirectional if it can be deformed indifferently in either direction of the posi-
tion at rest. Liquid-liquid bidirectional sensors have a specific design. If nec-
essary, they must be able to measure a very low differential pressure while
withstanding very high line pressures (or static pressures). They must be pro-
tected against accidental connection of a single side or non-synchronized pres-
surization of the two sides.
 Accuracy and safety
Linearization discrepancy lower than
0.1% at the end of adjusted scale
Rangeability of 20/1 I
Q
Temperature effect lower than 0.1%
from -40°C to +85"C
Long-term stability of 0.1% per year
22
Zero setting adjustable between 3
-100% and +loo% of the nominal P
value b
5
P
0
System integration z
Centralized display and command a
Direct connection to a PC
Via standard networks such as
s
0
PROFIBUS, MODBUS or FIP

FIP
MODBUS
*.
I PROFIBUS

Figure
10.7 Component parts of a sensor (Endress + Hauser documentation).
 Chapter 10 CONTROL AND MONITORING 631

A pressure sensor (Fig. 10.7) is selected essentially on the basis of the fol-
lowing points:
type of measurement element, according to the measurement scale;
cell and primary element construction materials according to the con-
tacting fluid;
type of signal: electric or pneumatic;
type of conditioning according to the environment (explosion proof,
intrinsic safety, corrosion);
accessories, indicators, isolating diaphragms, connections, capillaries;
communication protocol.

10.3.1.2 Flow Rate Instruments

There are two principles of flow rate measurement sensors: mechanical and
static.
A flow rate sensor should be selected taking the following main criteria into
account:

the characteristics of the measurement (type of signal, measurement

range, accuracy, etc.);
the type of fluid (nature of the fluid, impurities transported, viscosity,
flow regime, pressure, temperature, compressibility, etc.);
the constraints of the installation (piping diameter, operating position,
floor space, pressure drop generated, maintenance, etc.).
In the refining industry, whenever measurement conditions allow, static
flow meters with a differential pressure element are selected a s they can be
used on liquid, gas or vapor.
The major selection criteria for a differential pressure element are given in
the table below.

I Criteria Orifice Nozzle Venturi

Accuracy Good Average Good

Pressure drop High Medium Low
cost Low Medium High
Wear Average Slight Very slight
Installation Easy Average Cumbersom1
Loaded fluid Not recommended Fairly good Recommend

Other static flow rate measurement sensors are available (Fig. 10.Q with
much more complicated selection criteria, summed up a s follows.
632 Chapter 10 CONTROL AND MONITORING

Fluid criteria
Measurement principle
Nonconducting
Vapor
liquid liquid

Electromagnetic
(Fig. 10.8D) Yes - -
Coriolis (Fig. 10.8F) Yes Yes Yes
Thermal (Fig. 10.8E) - - Yes
Vortex Yes Yes -
Ultrasound
(Fig. 10.8G and H) Yes Yes -

Other selection criteria should also be taken into account, especially:

Electro-
Criteria Coriolis Thermal Vortex Ultrasound
magnetic
I

Measurement Wide Wide Narrow Wide Wide

range
Accuracy Good Good Good Good Good
Pressure Zero Low Low High Zero
drop
Pression High Medium High Medium Medium
Temperature High Medium Medium Medium Medium
Response Good Good Average Average Rapid
time
cost High High Average Average High
Main Loaded Mass Variable Wide mea: High
advantage fluid flow rate pressure and urement flow rate
temperature range

Three types of mechanical flow meters are available:

Flow meters with a rotating float (Fig. lO.SA), mainly used for local mea-
surements and offering the advantage of low pressure drop and relatively
low cost.
Positive displacement meters (Fig. lO.SB), chiefly used for measurements
in commercial transactions despite severe drawbacks (high pressure
drop, great sensitivity t o impurities and long response time).
Turbine meters (Fig. lO.SC), used especially for measurements with wide
variations in flow rates and for commercial transactions. Since they are
lower cost and withstand high temperature and pressure well, they
should be given priority over positive displacement meters despite their
lesser degree of accuracy.
 A

\7/ \ flow 1

D
2 3
Operating cycle
4

coil
5

E
-m- h
\
Electrodes

Upstream temperature power Downstream temperature

sensor T1 sensor T2

Ill1 HIOI A Tube

~ Heating coil

Power in Amplifier

AT detection bridge --I

ti H Transmitter Receiving
element element

hid
Flow d i r e c t i o q . ’.. - ..
. ..

Figure
10.8 Flow rate measurement sensors (After “Capteursindustriels’7.
A. Principle o f the rotative float meter. B. Principle o f the positive displacement
634 Chapter 70 CONTROL AND MONITORING

10.3.1.3 Level Instruments

Three types of level measurement sensors are found in refineries: hydrostatic,
electric and radiation sensors.
Selection of a level sensor must take the following main criteria into
account:
the characteristics of the measurement (type of signal, measurement
range, accuracy, etc.);
the type of fluid being measured (clean or loaded fluid, unchanging or
variable specific gravity, pressure, temperature, pulverulent or granulous
products, etc.);
the installation constraints (accessibility for reading, measurement tap
location, etc.).
Hydrostatic level measurements are the most frequently used. This is
because of the wide variety of sensors suitable for most common require-
ments and their wide and accurate measurement range (Fig. 10.9). Electric
measurements are also widely utilized. They are generally static sensors that
withstand pressure and temperature well, but are especially employed for con-
tact measurements. Level measurements by radiation are ordinarily limited to
applications that can not use the previous two types, chiefly because of their
high price and mediocre accuracy.
Selection of a level sensor must take into account the type of measurement
and the medium being measured

Type of level :ontinuow measurement Detection

measurement Liquids Solids Liquids Solids
-
Capacitive Yes Yes Yes Yes
Conductive Yes - - -
Hydrostatic Yes - Yes -
Radio-isotope Yes Yes Yes Yes
Ultrasound Yes Yes - -
Microwave Yes Yes Yes Yes
Mechanical or electric detector Yes Yes Yes -
Vibrating strips - - - Yes

10.3.1.4 Temperature Instruments

Measurements for remote transmission are almost exclusively made by ther-
mometric elements with a thermocouple or a metallic resistance (Fig. 10.10).
Local temperature measurements mainly use solid dilatation thermometers.
 Chapter 10 CONTROL AND MONITORING 635

Torque tube
Disolacer arm

Displacer
Chamber Flow outlet

mber assembly
(source: Fisher Controls S.A.)
a) Vertical assembly
(source: Hartman and Braun)

Flow outlet

6 Pulleys

( 5 5Counterweight
~

Counterweight

GI

Figure

A and ’ Level measurement sensors (After “Capteursindustriels’y).

A. Displacer sensors. B. Float level detector.
636 Chapter 10. CONTROL AND MONITORING

Glass
tube

Magnetic
index

Measuring rule

Gauge
D r7Y

\
Flow rate
controller

Bubble
tube

E c
r
a

a) Relative b) Differential pressure

pressure sensor
sensor
1. Float
2. Float axis o
3. Nozzle
h 4. Flange
5. Housing
6. Plate
7. Housing screw
c) Immersion probe 8. Connecting stuffing box

1 Figure I
Level measurement sensors (After “Capteursindustriels’9.
C. Magnetic level. D. Principle of bubble tube measurement.
E. Principle of pressure sensor measurements. G. Principle of float level
measurement.
 Chapter 10. CONTROL AND MONITORING 637

When the dial is far from the measurement point, liquid or gas expansion ther-
mometers with capillaries are employed.
The measurement ranges in each category are significant, but require
selection of an element suited to the measurement within each measurement
category. The temperature ranges covered are as follows:
Thermocouple: from -270 to 2 700"C,
Metallic resistance: from -260 to 1400"C,
Solid expansion: from -60 to 500°C,
Liquid expansion: from -55 to 600°C,
Gas expansion: from -100 to 600"C,
Vapor pressure: from 0 to 400°C.

10.3.1.5 Construction and Protection

The component parts of an industrial sensor are the primary element, the
transducer element and the electronic conditioning module, all of which are
incorporated inside a housing.
When the primary element is in direct contact with the medium being mea-
sured, its construction must allow communication and isolation between the
equipment and the sensor. It also has to facilitate maintenance operations
while keeping the industrial installation in operation.
The housing must provide effective protection against:
external aggression such as influxes of dust and small diameter solid par-
ticles,
mechanical impacts.
Degrees of protection refer t o the 1EC standard which defines them by the
characteristic letters IP and three numbers identifying the degree of:
protection against solid particles,
water tightness,
protection against impacts.
For sensors installed in potentially explosive media, the sensor housing
must be built in compliance with the IEC standard with the following protec-
tion features:
explosion-proof shell (IEC 79.1), which works by confining any explosion
inside the housing;
enhanced safety shell (IEC 79.7), that can be used when the elements
installed inside the housing can not produce any sparks;
intrinsic safety shell (IEC 79.3-79.11), which works by limiting the energy
to below thresholds where an explosion could occur.
638 Chapter 10 CONTROL AND MONITORING

A B
Wire A
Cold weld Measu-
rement
device couple housing device

Reference junction
(cold weld)

C
Wire A
- copper
F

Wire B ” -copper
Reference junction
(cold weld)
-
D

Measurement
device with

Wire B lcompensation I

1. Sensitive element
2. Electric wire insulators
3. Protection well
4. Fastening system
5. Guard tube
6. Connecting head
7. Compensation cable (or copper wire)
8. Transmitter (optional)

Temperature measurement sensors (After “Capteursindustriels’3).

A. Conventional thermocouple assembly. B. Thermocouple and compensa-
tion housing assembly. C. Thermocouple and compensation cable assembly
D. Assembly with cold junction electric compensation. E. Flat temperature
measurement elements. E Thermocouple assembly.
 Chapter 10 CONTROL AND MONITORING 639

10.3.1.6 Installation
Industrial sensor installation must comply with the following requirements:
Sensors installed in line must have flange or screw-type connections that
are compatible with piping and equipment construction materials, espe-
cially with respect to the temperature and pressure of the medium being
measured. In-line installation of instruments must allow continuous oper-
ation even when the instruments are placed off stream by means of iso-
lating valves placed upstream and downstream and addition of a by-pass.
When sensors are installed near impulse connections, these latter must
theoretically be located below the sensor for measurements on gas and
above it for measurements on liquid or steam.
In all cases installation must protect the sensor from any effects that
could be detrimental to measurement accuracy and quality. It also must
facilitate maintenance operations.

10.3.2 Remote Transmission

Centralized operation of refining units requires remote transmission of mea-
surement and control data. Several transmission modes are available:
pneumatic analog transmission (practically no more used),
electric transmission using three types of signals: high and low level ana-
log, digital or logic.

10.3.2.1 Pneumatic Analog Transmission

The conventional pneumatic signal from 0.2 to 1 bar (3 to 15 psi) varies pro-
portionally with the calibration measurement of the pneumatic sensor. This
type of signal is transmitted by a nozzle-flapper system associated with an
amplifying pneumatic relay. The power supply air has a standard pressure of
1.4 bar (20 psi). Today pneumatic measurements are mainly used for local indi-
cations. The main application of this signal is in electro-pneumatic valve posi-
tioners and converters. These change an electronic analog signal (4-20 mA)
into a pneumatic signal that can be employed to change the position of pneu-
matic servomotors associated with control valves for example.

10.3.2.2 Electric Analog Transmission

Two types of analog signals can be used:
“low level” analog signals for remote transmission, generated by temper-
ature sensors (thermocouple or resistance probe), but also by a variety
of particular sensors of the flow meter turbine type (frequency varia-
tion), radioactive level sensors, capacitive sensors, mechanical vibration
or motion measurement sensors, etc.;
640 ChaDter 10 CONTROL AND MONITORING

“high level” electric signals, which are originally usually low level signals
amplified inside the sensor t o generate a conventional signal of the
4-20 mA type.
Analog signals are related to the measurement by a continuous, sometimes
linear, law that reflects the variation in the physical phenomena being mea-
sured. They may be of all types (voltage, current or load) and can be trans-
mitted directly or modulated. Direct transmission is the simplest mode. It
makes an amplitude variation proportional to the signal correspond t o each
variation in the measurement. Direct transmission is generally over short dis-
tances. Transmission by modulation of a carrier wave consists in modifying
one of the characteristics of a high frequency wave, called a “carrier”, by the
measurement signal. The carrier wave acts as a support for the signal all
throughout transmission. It is usually a sine wave whose amplitude, frequency
or phase is modulated. Frequency modulation is the most widespread method.

10.3.2.3 Digital Transmission

The digital signal is generated by the sensor and directly coded in digital form
in the sensor itself. The signal can be absolute when transmission is in paral-
lel, with the signal coded digitally on several bits transmitted in parallel on
several wires. Otherwise, it can be incremental when transmission is in series,
with the signal coded digitally and the code bits transmitted sequentially on a
single wire. In this second case, the signal is made up of a series of electric
pulses that must be counted to find the measured value. Digital sensors are
called coders or counters depending on the case.
Digital transmission exhibits great immunity to noise. Additionally, it is
compatible with computerized processing systems, which are increasingly
numerous and diverse in their construction and applications. These are called
high speed digital communications networks and use a protocol that allows
data transmission management (Fig. 10.1 1).
The data often come from analog sensors connected by hard wiring to ana-
log/digital conversion modules, for utilization in a group of processing units
and monitoring interfaces, of the digital control and supervision system type
(DCS). The present-day trend is to place a group of compatible sensors in a
“cluster” on local buses with direct connection to sensors so a s to minimize
hard wiring.

10.3.2.4 Logic Transmission

The signal has only two states or possible values and is generated by sensors
of the on-off type. The transition between the two states corresponds t o pass-
ing a threshold or a preset measurement value.

10.3.2.5 Transmission by Smart Sensors

A sensor is said to be smart when it incorporates a microprocessor that allows
it to carry out a set of functions unavailable in a traditional sensor. It can
 Chapter 10 CONTROL AND MONITORING 641

I & ' 8 * , 1

ZA gateway
for MODBUS
PROFIBUS
FIP

Summary of advantages
0 Interactive calibration without
applying external pressure,
with visual support.
0 Parameterization by means of
a portable terminal with
HART@or INTENSOR@proto-
col, wherever the location.
0 Centralized display and super-
vision with a computer thanks
to system link.
0 Integration in a field network
such as PROFIBUS or FIP
with the corresponding ZA
gateway.

I Figure
10.11 Architecture of networked sensors (Endress + Hauser documentation].
642 Chapter 10 CONTROL AND MONITORING

linearize measurements, compensate them for temperature and permit better

accuracy by the use of a digital measurement sensor. It can combine several
functions allowing for example a volume flow rate measurement to be con-
verted into a mass flow rate measurement with temperature and specific grav-
ity compensation. It can also accomplish recalibration and self-testing,
thereby simplifying maintenance operations (Fig. 10.11).
These functions require the use of a configurator that may have several
aspects, from a simple hand-held device to complete integration in a commu-
nication system incorporating a PC or a digital control and supervision system
(DCS). These configuring tools use the hard wiring support of the electric ana-
log signal (4-20 mA) with a digital signal superimposed on it. This communica-
tion line employs a communication protocol of the “Hart” type, which allows
remote reading of calibration parameters. It also permits manual or automatic
modification by an applicative program.
The use of a standard protocol means that different instruments can be
connected on the same communication line, provided they have been previ-
ously identified by an address that allows the description of data transmitted
in series. This type of sensor is increasingly utilized and is becoming standard,
due to its cost which is today approximately the same as that of a conventional
sensor.

Spring adjusting screw -

Lower spring support ring

Servomotor stem

Servomotor stem guide

Travel indicator plate

-
Figure
10.12 Cutaway view of a pneumatic servomotor.
 Chapter 10 CONTROL AND MONITORING 643

10.3.3 Control Valves

The control valve is a regulating device used to vary the cross-sectional area
for a fluid in piping by manual or automatic command. The cross-section may
vary from open to totally closed. The valve includes two distinct parts plus a
number of accessories required by the various valve operations:
The actuator (Fig. 10.12) is the device that receives a positioning signal
issued by manual command or by a controller and converts it into a
force. The value of the force is such that it can modify the internal adjust-
ment from one specific position to another.
The body (Fig. 10.13)is the part of the control valve where the fluid can
be stopped, or on the contrary circulate more or less freely. It contains
all the fluid regulating elements.

Valve stem

Stuffing box flange nuts 1

1 /
,- Arch niit

Stuffing box flange

\
Stuffing box seal ring c

Packing distance piece

4
Stuffing box packing

Cap

Body seal --.

ail

Figure
10.13 Cutaway view of a valve body (Masoneilan documentation).
~
644 Chaoter 10 CONTROL AND MONITORING

The accessories installed on a control valve are numerous and complete

or enhance the intrinsic valve functions (air regulator, valve positioner,
solenoid valve, noise reduction device, etc.).

10.3.3.1 Types of Valves

Valves with straight bodies are the most common in refining. They can be two-
or three-way, with one or two seats and feature several internal technologies.
They are screwed or flanged on to the piping. There are several types:
Globe valves (Fig. 10.14A, B, C and D) can have a single or double seat
and plugs of a variety of shapes (parabolic, V port, linear or equal per-
centage, throttle, needle, on-off, etc.) depending on the characteristics of
the process control and the desired flow rate.
Cage valves (Fig. 10.14E) comprise cylindrical internals provided with
ports whose shape allows the same flow rate law to be reproduced as
with globe valves. A solid piston moves in the cylinder and closes off a
varying number of apertures.
Multi-stage valve internals are similar to those of the cage valve, except
that there are a large number of baffles or labyrinths that the fluid passes
through. This reduces the upstream pressure of the fluid being controlled
as many times as there are baffles.
Self-regulating valves are identical in construction to globe valves, but
the power fluid is the one being controlled. These valves work as escape
or overflow valves.
Eccentric rotative valves are simplified in construction, consisting of a
body mounted between two piping flanges. In the body a plug in the shape of
a disk or a ball pivots around an axis that is perpendicular to the piping,
according to the required control adjustments.
Angle valves (Fig. 10.14F) have a body at a 90" angle between the fluid inlet
(generally via the valve seat) and the outlet which is horizontal. The internals
can be of the same type as those of the straight valves mentioned above.
Butterfly valves have a circular body inserted between two flanges. A flat
or streamlined disk is installed in the body and mounted on axes provided
with ball bearings or sintered packing. The flow rate is controlled by rotation
of the disk, limited to 60" for a flat one and 90"for a streamlined one.
Ball valves are similar in construction to eccentric swing-disk valves,
except that the internals consist of a solid ball when the valve is used as an
isolating device. The ball can also have a V-shaped port which allows a control
profile to be achieved.

10.3.3.2 Selecting a Valve

A control valve is selected with due consideration given to the different ser-
vice conditions such as the ones listed below.
 Chapter 10 CONTROL AND MONITORING 645

A B

C D

E F

-
Figure
10.14 Differenttypes o f control valves.
A. Single seat, reversed plug. B. Single seat, standard plug. C. Double seat,
646 Chanter 10 CONTROL AND MONITORING

control or on-off function,

control coefficient and characteristic,
large or small pressure drop,
cavitation, vaporization, mixed fluid,
small or large valve dimension,
noise, etc.
Selecting a control valve always requires a good understanding of the prob-
lem to be solved and of the intrinsic control possibilities of each type of valve.
Globe valves are the most widely used type in refining and allow most com-
mon control operations. The characteristic of the most frequently used plug is
equal percentage (= %). Here the flow rate at constant pressure drop is related
to the stem travel by an exponential law. This means that a given increase in
stem travel causes a constant percentage of the flow rate variation compared
to the flow rate prior to the increase. The characteristic of the linear plug with
constant pressure drop allows the flow rate to be adjusted proportionally to
the stem travel. In other words, a given increase in stem travel causes a con-
stant flow rate variation, whatever the flow rate prior to the increase.
The plug characteristic can be appropriately selected on the basis of the
following criteria:
Linear plugs will be preferred when the pressure drop across the control
valve under all operating conditions is greater than 2/3 of the closed
valve differential pressure, as well as for flow rate and pressure control of
liquids and composition control.
Equal percentage (= %) plugs will be specified for all other cases except
for control valves used for the following services:
- level control,
- compressor anti-surge control,
- split range control,
- use of two valves in parallel to increase the control coefficient,
- manual control (HIC),
- control of the minimum pump protection flow rate,
- control by miniature control valves.
Cage valves are employed in exactly the same way as globe valves, but for
more severe services on clean fluids. They cost more than globe valves and as
a result their use has been limited for standard services.
Multi-stage valves are utilized with high pressure drop and when stability
and a high control coefficient are required. They cost so much that their use
is confined to particular applications such as by-passing a compressor or a
steam turbine.
Self-regulatingvalves are used for simple pressure reduction or overflow
services on utility fluids (air, steam, water, etc.).
 Chapter 10 CONTROL AND MONITORING 647

Eccentric swingdisk valves can be employed for all standard services of

fluid control with a low pressure drop and a limited control coefficient. They
cost much less than straight valves and so are preferred for many applications.
Butterfly valves are chiefly used to control low pressure gas flow rates
with a low pressure drop and when there is a large volume of fluid to control,
therefore requiring large size valves.
Angle valves can be selected when service conditions d o not allow the use
of a straight valve, to solve problems of noise or cavitation since the venturi
shape body causes lower internal pressure drops. They are also used on vis-
cous products.
Ball valves are chiefly employed in isolation a s on-off valves. Some valves
with specially designed balls are also used to control loaded and vaporization-
prone fluids.
Control valve selection must also take into account the type of valve action
in the event the power fluid is lost (closing or opening by lack of power fluid)
and the type of power supply available (compressed air, hydraulic network,
electricity).

10.3.3.3 Sizing
The flow rate coefficient, C,,, has become the universal characteristic of flow
rate measurement for a fluid flowing through a valve. By definition the coeffi-
cient C,, is the number of US gallons of water passing through a restricted
cross-section in a minute when the pressure drop across the restriction is
1 psi.
Precise sizing presupposes that specific service conditions are known.
Generally speaking, one of the service conditions must be estimated (for
example the pressure drop) and the flow rate coefficient C,, of the valve will
often have to be calculated on the basis of this estimate. Here common sense
and long experience in the industry are indispensable. Sizing errors usually
come from an accumulation of safety coefficients, which makes the calculated
C,, too large. Under these conditions the plug has to achieve control with an
overly small opening and this is unfavorable.
The main service conditions that have to be known or estimated most pre-
cisely are:
the flow rate of fluid being controlled at normal, maximum and minimum
values;
the normal, maximum and minimum pressure drop in the valve;
the fluid specific gravity.
The normal flow rate will be calculated or estimated. The maximum and
minimum flow rates will be determined versus the different operating cases in
the installation being controlled and the valve must be able to handle the
extreme flow rates between 10 and 90% of its stem travel. If only the normal
flow rate is known, the valve will have to handle an additional flow rate of 15
to 50%.
648 ChaDter 10 CONTROL AND MONITORING

The pressure drop in the valve is sometimes well defined: simple expan-
sion between two known pressures, overflow to the atmosphere or control
between practically constant levels. Lacking this information, if the pressure
drop in the rest of the control loop are known, a minimum of 1/3 of the pres-
sure drop value will be allocated to the control valve. If there is a centrifugal
pump with a high discharge head, the pump characteristic takes on great
importance. The pump discharge pressure will have to be identified at mini-
mum and maximum flow rates to assess the control valve pressure drops in
both operating situations. The comparison of these values will determine the
values of C, to be considered in selecting the valve. It may seem attractive to
reduce the power consumed in the control loop and in particular the pressure
drop at the control valve. However, it is impossible to obtain good results if the
valve can not absorb enough power to keep the controlled variable constant.
The fluid specific gravity is involved in calculating the flow rate only by its
square root. Consequently it is not absolutely necessary to determine fluid
specific gravity with great accuracy. If for example the specific gravity goes
from 0.8 to 0.9, the variation in flow rate will only be 5%.
The pressure drop value of a wide open valve is proportional to the flow
rate to the second power and to the specific gravity for liquids:
AP = K Q 2 d
with:
AP pressure drop (bar)
Q liquid flow rate (m3/h)
d specific gravity with respect to water
hence: AP
-= K = Constant
Q2d
The constant K therefore depends on the valve area and internal construc-
tion.
Control valves are sized on the basis of this general formula. There are
equations for flow rates expressed in volume or weight for:
a liquid in non-critical flow when AP < C, APs
a liquid in critical flow (cavitation or vaporization) when AP C, AP,with:

Ms=Pl-(0.96-0.28 g)fv
or simplifying, if P, < 0.5P1,then AP,= P, - P, with:
P, upstream pressure (bar abs)
P, critical pressure of the fluid (bar abs)
P, vapor pressure of the liquid at the upstream temperature (bar abs)
Af, pressure drop used for sizing with critical flow (bar)
C, (critical flow rate factor) a dimensionless coefficient representing the
critical pressure drop ratio of a valve for any fluid and is valid both
for liquid vaporizations and supersonic flows
 Chapter 10 CONTROL AND MONITORING 649

Cavitation is a detrimental phenomenon caused by part of the liquid vapor-

izing during the rapid acceleration when it passes between the seat and the
plug. Cavitation will occur whenever the static pressure reaches a value cor-
responding to the vapor pressure of the liquid at the flow temperature.
Cavitation must obviously be avoided, for example by increasing the down-
stream pressure, by selecting a valve type with a critical flow rate coefficient
as close as possible to 1, or in extreme cases by installing two similar valves
in series.
Other special calculation cases exist with high viscosity, laminar flow, the
presence of a diffuser or a two-phase flow. They will require the use of partic-
ular formulas or correction coefficients.
For control valves on gas or steam there are equations for:
non-critical flows when AP < 0.5 x C, x P, ,
critical flows when AP 3 0.5 x C, x PI.
There are particular formulas for calculating C, for gas flow rates in volume
or weight and for saturated or superheated steam in volume or weight flow
rates. Special calculation cases are needed with a diffuser, for the exiting veloc-
ity limitations and for the compressibility corrections. Depending on the type
of flow rate, when the valve flow has been determined, C, will be calculated
using one of the formulas available in the applicable standards that are also
found in supplier catalogues.
An important phenomenon to be considered when a control valve is being
selected is the noise it generates. Noise can be due to:
mechanical vibrations (plug resonance),
cavitation (noise caused by vapor bubbles bursting in the liquid flow),
aerodynamic turbulence, with critical flow at supersonic speed.
Environmental standards limit continuous noise levels to very low values,
so the noise generated by the valve must be calculated. If necessary, the selec-
tion is modified or the valve is provided with appropriate accessories to keep
the noise down to a compatible value which is generally 80 dB at 1 m.
A control valve must meet the control conditions dictated by process
needs. This concept is called rangeability and is the ratio R = (Qmax/Qmin)
The controllable flow rate is ordinarily confined to lift height varia-
Qcontrollable.
tions between 10 and 90% of the plug stem travel. When the control valve is
used for precise control, the concept of sensitivity, i.e. the ratio S = AQ/AL
(% of flow rate variation/lift variation), must also be verified.

10.3.4 Industrial Controllers

Whatever the technologies of industrial controllers, their main objectives are:
to control parameters (pressure, temperature, flow rate, level, etc.) to
preset values so that a process operates smoothly;
650 Chapter 10 CONTROL AND MONITORING

to fulfill a safety function by keeping these parameters within compatible

limits in the event of drift or malfunctions;
to play an economic role by keeping operating cycles regular, thereby
allowing stable, regular production of the manufactured products;
to eliminate human error by linking interactive control functions and
coordinating action t o correct the disturbances occurring in the process.
Whatever the controller, it processes and carries out the following func-
tions:
acquisition and processing of the measurement signal (M),
acquisition and/or generation of the set point signal (C),
determination of a correction signal (C/M),
generation of an output signal.
These functions are processed by a control module that carries out differ-
ent adjustment actions (set point, proportional, integral and derivative actions
(PID), high and low limits, manual output). At the same time the module is
linked to selectors (internal/external set point, direction of controller action,
automatic or manual operation) and to indicators (measurement, set point,
measurement/set point deviation, output, high and low limits). A simplified
representation of a controller is given in Figure 10.15.

Direct
Measurement
Automatic
output

Int. set point

Ext. set point Reverse

I 1
10.15 Block diagram of a controller.

10.3.4.1 Theory
Industrial controllers have three types of operating action that are indepen-
dent but ordinarily combined. A simplified description follows. The propor-
tional action controller (P) receives the measurement set point deviation
x = (M - C) and determines the output S given by the formula:
S=G(M-C)
 Chaoter 10 CONTROL AND MONITORING 651

C is the controller gain which is also expressed by the proportional band (BP)
expressed in per cent (%):
100 100
BP(OA)=- and G = -
G BP (A)
Hence a gain of 1 = 100%of BP and a gain of 2 = 50% of BP.
In a proportional and integral action controller (PI) the integral action is
expressed by the time T, (min or s) that is required for the output variation to
be equal to the input variation.
There are controllers with proportional + integral action in series where:

S = G ( M - C)

P
+-
+
“I
Ti
I
( M - C) dt

or in parallel where:
1
S = G ( M - C) + -j ( M - C) dt
Ti
The only difference in comparison with the in-series PI is that the gain does
not influence the integral action.
In a proportional and derivative action controller (PD) the derivative
action expresses a variation rate. By definition, the input data item at the oper-
ating point is the slope of the tangent to the curve of the variable versus time.
There are controllers with proportional and derivative action in series where:
d(M - C)
S = G ( M - C) + GT, dt
P + D
or in parallel where:
d(M- C)
s=G ( M - C) + Td dt

The only difference in comparison with an in-series PD is that the gain G

does not influence the derivative action.
There are numerous possibilities of associating the P, I and D modules. On
most present-day controllers these associations are available as standard fea-
tures as well as in-series, in-parallel or combination assemblies. Possibilities
also exist of having the derivative action on the error or on the measurement
(Fig. 10.16).
It is important to remember that:
Proportional action has the effect of stabilizing the variable but does not
correct the M/C deviations when the process is stabilized. Its use is gen-
eral and indispensable.
652 Chapter 10 CONTROL AND MONITORING

PID controller structures

~

Derivative on the error Derivative on the measurement

M

C S
+++
Ti+Td d(M-Cj C-
S=G,(-)(M-C)+
Ti

M.

+S
C-

+SO

C
M
&
-%s p+
d(M-C)
c -
(M-C)dt+G,Td -
dt +so

Figure
10.16 PID controller structure (After “Bouclesde r6gulations”)).

Integral action depends on time, eliminates the M/C deviation and surge.
Its use is necessary in controlling liquid pressure and flow rate, gas and
steam pressure, and sometimes for level control.
Derivative action reduces oscillation amplitude and period, accelerates
stabilization, allows proportional and integral action to be increased and
offsets the drawback of dead time. Its use is necessary especially in tem-
perature and pressure control of gas and steam and generally in control
with a significant lag effect.

10.3.4.2 Local Controllers

These controllers are directly in contact with the fluid being controlled. The
measurement element can be of different types:
torque tube for level controllers,
Bourdon tube for pressure controllers,
liquid or gas expansion element for temperature controllers,
differential pressure measurement element for flow rate controllers.
 Next Page

Chapter 10 CONTROL AND MONITORING 653

Set point generating modules are mechanical or pneumatic and the output
signal is pneumatic. There are also entirely mechanical local controllers.

10.3.4.3 Pneumatic Controllers

Pneumatic controllers for control room panels have almost totally been
phased out. The input and output signals are modulated air pressures gener-
ally on a range of 0.2 to 1 bar.

10.3.4.4 Electronic Controllers

These controllers use analog electronics, based on operational amplifiers.
They are being replaced by digital technology, but are still numerous in exist-
ing units.
The actions have value ranges that can be adjusted for:
the proportional band: from 3 to 400%,
integral action: from 0.05 to 40 min,
derivative action: from 0 to 5 min.

10.3.4.5 Digital Controllers

The basic difference in these controllers is the use of a micro-processor that
integrates all the control and computing functions. Moreover, the parameteri-
zation data are accessible via an integrated or connectable micro-computer.
The main access to control is as follows:
possibility of having several inputs;
choice of input signal (current, voltage, frequency, thermocouple, resis-
tance probe, etc.);
processing of the input signal (extracting the square root, filtering, lin-
earization, etc.);
scaling (value and format) of indicators;
choice of type of alarm either on the measurement or on the error;
precise display of data such as actions, limits, etc.;
choice of derivative mode on the measurement or on the error;
tracking set point: the set point follows the measurement in the manual
posit ion.
Digital technology also allows self-adjusting controllers to be designed
which calculate the PID actions at a given point. Another possibility is self-
adaptive controllers which calculate and adjust the parameters of their algo-
rithm according to process developments.
This type of controller connected on a communication bus is the con-
trol processing unit found on digital control and supervision systems of DCS
type.