Chemical Technology

PROCESS CONTROL BLOCK III

1) Introduction and control theory

A chemical process is a set of chemical and/or physical operations.
ordered to transform raw materials into products
different finals. A chemical system (process) is composed of
subsystems, called unit operations.
The control of chemical processes is a specialty of Automation that
It deals with the selection and application of techniques for operation
safe and efficient processing plants.
The objective of a chemical plant is to transform a raw material into
products. This transformation must be safe, economical, respectful
with the environment and with the specified quality. To carry out
this, the teams that make up the plant must operate correctly, but
since during the operation the plant is subject to disturbances, it is done
it is necessary to control the process by installing a set of
measuring and process handling instruments, which
they constitute the Control System.

Therefore, in this section, the following sections will be examined:

1. Manual operation of a process. Need for control

2. Automatic operation of a process
3. Feedback control system
4. Feed Forward Control Systems
[Link] transform tables
6. System Dynamics

1. Manual operation of a process. Need for control

The processes are of a dynamic nature, so they will change.
continuously and if the appropriate actions are not taken, the
process variables will not meet the design variables.

page 1
Control is necessary because disturbances exist. A disturbance is
an external variable that destabilizes the process.
Let's analyze the following figure:

We have a free outlet tank that is fed with a flow rate.i, that not
it is controlled (wild inflow) and exits the tank with a flow rate
q0The controlled variable is the height of the tank (h). Through a
Meter, the operator looks at the tank level and compares it with the Set-Point.
setpoint. Calculate the error, and if it is not 0, apply a
corrective measure, which in this case will be to open or close the valve. This is what it
we call it a manual control system. In this case, the manipulated variable
it is the output flow, since the valve acts as the final element of
control, as it is not possible to control the input flow.

Therefore, we can define:

Controlled variable: It is the variable that must be maintained or

control that should not deviate from the desired value.
Control point: It is the value that we want the variable to have.
controlled.
Manipulated variable: It is the variable that is used to maintain
the controlled variable at the control point.
Perturbation: It is any variable that causes the variable of
control deviates from the checkpoint.

page 2
 Set-Point: It is the value that we want to
keep the variable controlled.

Control actions:

Measurement (M) of the controlled variable.

Verification or comparison (C) of the measured value with the value
what I wish the variable (Set-Point) to have.
Decision (D): Operator's brain. Controller. Error calculation
Act (A): On the final control element to correct the error.

Block diagram:

Manual operation or open loop

2. Automatic operation of a process

Control is only possible when the engineers provide the equipment.
Required during the design process: Equipment and instrumentation.
The process is carried out from control rooms using systems of
distributed control.
The operation and control of a process require:
Understand its dynamic functioning.
Maintain your variables at the desired values despite the
disturbances.

page 3
The actions for an automatic control system are the same as
for a manual control system, the only thing that changes is the
procedure.

Measurement of the process variable to be controlled, by means of a

primary measurement instrument or sensor. Therefore, it changes the
vision of the operator manipulating the valve, by a sensor, which also
it will be much more accurate.
Transmission of the measured signal from the process to the controller
through a transmitter. The perception of the operator looking changes
the level, to a controller that will perform its calculations in an algorithm of
control.
Comparison of the Set-Point with the process variable, to calculate the
deviation or error, and the automatic controller decides based on a
control algorithm. It is replaced by the operator's brain.
Corrective action: Automatic control valve. A signal is changed.
electric or pneumatic with the force exerted by the operator.

Differences between open loop and closed loop

Open loop: The controller does not receive information from the variable, it is
to say, there is no feedback. The controller does not provide any
relative function of how to keep the controlled variable at the point of
control.

Toaster

White Bread Electrical power Toasted bread

WHEEL OF PROCESS
LEVEL FOR TOASTER
TOSTAR

The value of the output has no influence on the input. That the
the result does not come out as we expected (a little toasted or very toasted) it is not
reason to suppose that if it has it, so if we wanted to roast more
the bread, we should repeat the process from the beginning. The same
We can apply the methodology to a washing machine.

page 4
 Closed loop: The controller acts constantly until there is none
error. It can be manual or automatic.
This is the case, for example, of a stove that heats a room.
It is considered a loop or closed loop, because if it suffers any disturbance,
the system can provide more or less heat, depending on what is
require at that moment.

3. Feedback control system

The output variable is measured, it is sent to the controller, and the operation is carried out.
verification, and corrective action is taken only when there is an error.

Disadvantage: In order for the control system to function, it has to be produced.

an error.

page 5
4. Feed Forward control systems
No prior knowledge of the process is required. The output variable does not
it is measured, and the controller does not receive information about it.

In this case, the control system measures the variations in input or

perturbations. Through a mathematical model, it calculates what the
effect on the system and take a corrective measure in advance.
So in this method, we do not know what the controlled variable is.

page 6
The drawback of this method is that another disturbance may arise that
has not been taken into account and destabilizes the whole system.

Feedback control combined with feedforward loop

page 7
5. Laplace Transform Tables
With the Laplace transform, it converts into the mathematical model of
a system, in algebraic equations in the s domain. We will obtain the
output variable as a function of s.
Then it is inversely transformed to obtain the desired solution in
the mastery of time.
Ec. Differential -> Laplace Transform -> Algebraic Solution -> Inverse Laplace Transform ->
Sun.
For a step function, the Laplace transform = k/s
One of the most important properties of the transform is the Theorem
of the Final Value. This theorem allows predicting the behavior in
steady state in the time domain, without having to perform the
inverse transform, determining the value of the output variable.

limf(t)=lims·f(s)
→
∞
t s→0

6. System dynamics

a. First order systems or delays

They are called first order elements or first order delay elements,
to those with transfer function (TF) is of the type:

page 8
 K
G=
ts+1
Being:
Profit
Time constant (s)

Both the gain and the time constant are characteristics

common to all first-order systems.

The transfer function is given in the Laplace domain and not in

the time. The transfer function is now defined as:
The quotient between the Laplace transform of the output function C
(t) and the Laplace transform of the input function R(t).

The order of an element, process, or system can be identified if

we know its transfer function or its mathematical model that
represents the dynamic behavior of the output variable against
a change (step jump) in the input variable.

Within first-order systems, systems can be specified

of order 0, where the output response is equal to the input.
input variable changes in a stepwise jump, the variable of
output as well. As can be the case with a digital thermometer, where you
give the instant value. It is placed under the arm, the thermometer changes to
temperature, and the device gives you the value instantly.
Let's assume we have a step jump change,
for example, a valve that opens, a thermometer of
mercury, a variable level tank with free discharge to the atmosphere…
It is about finding out how a change in influences the output variable.
step jump form in the input variable.
Mercury thermometer
Input variable

page 9
 Change in step form (u)

If the temperature at t=0 is 20º, when t>0, the temperature will be:

T = 20 + ∆ T

What will happen with the output variable? It is mathematically demonstrated

that the graph is like this, because it takes longer than a digital thermometer.

This phenomenon is called Self-Regulating System, because it is based on a

steady state, and after going through a transient state, ends in another
steady state.

What information can we obtain from this graph?

a. Stability input output
A system is input-output stable when a bounded input,
it corresponds to a limited output.

b. Time constant ( ) τ
Time required for the output variable to reach 63.2%
its final value.
We can also define the time constant as the time that
the system's output variable would take time to reach its final value if

page 10
 its rate of change was constant and equal to the speed of
initial response, that is always equal to the slope of the line
tangent to the curve (graphical representation of the response) at t=0.

c. Profit (K)
What the output variable gains regarding a change in the variable
as an appetizer.

d. Settle time
Time required to reach 95% of the final value

Systems with the same gain (K) and different time constant
(τ )

The final value is the same.

They have different response speeds.
The smaller the time constant, the less time it takes to
to reach the final value, which means it takes longer to stabilize.
System 1 is the fastest and 3 is the fastest, due to the
slope inclination. The steeper, the faster.

Systems with different gain (K) and the same time constant
(τ )

page 11
They have the same response speed.
Different accuracy.
For the same increase in the input value, the final value of the
the output will be different in systems with different gain.
The response speed is the same, they reach the final value at the same time.
but the speed of change is different.

b. Higher-order dynamic systems

They are those described by differential equations of order
superior to 1. They are also known as second-order elements of
delay, since the final value is not obtained immediately.
In process control, second-order systems are formed by
two capacitors in series, for example:
System formed by two tanks in cascade. Between the flow that exits
from the first and the one that enters the second there is no resistance whatsoever, it is
a non-interactive system. In the event that there was resistance, it would be
interactive.
Thermometer element in its housing has heat capacity.
different.
There are other systems, called 'intrinsically second'
"order" like the mercury manometer or Bourdon tube.
Differences between first-order and second-order systems:
First-order systems are defined by a parameter, the
time constant. Second-order systems are characterized by
two parameters:
Natural frequency in rad/s.
Damping coefficient. This parameter will depend on the
system response.
The response of the second-order system can be approximated by that of
first-order plus a delay.
Higher order elements than 2: Multicapacitive, dead time.
For second-order elements, the transfer function will be of the
type:

page 12
 K
G= 2
1 2δ
2
s+ s+1
w wn

Profit

w n : Undamped natural frequency δ Coefficient of

damping

Developing the quadratic equation:

Different solutions can be obtained depending on the value of δ :

Yes δ 1 to 2 real roots (Overdamped, does not oscillate)

The greater the damping coefficient, the less tendency it will have to oscillate.
If δ = 1 → 2 real roots (Critically damped)
If 0< δ 2 complex roots (Underdamped system, oscillates)
Yes δ = 0 -> Stability limit. The oscillations remain in the
time, constant amplitude.
If δ Unstable. Oscillations without damping.

Typical response of an underdamped system

The advantage of these systems compared to over-damped ones is that

Although its response oscillates, it reaches the steady value sooner.

It is normal to encounter over-damped processes, but when

a controller is introduced in the process and the control loop is closed
it says that the controller interacts with the process and the set of
the system behaves like a underdamped system.

page 13
Peak time (Tp): It is defined as the time required for the
response to a step input, reach the first maximum.
Overpeak (Mp): It is the value of the first oscillation. It depends only on
δ .
Settling time (Tss): Time required for the output
reached a certain percentage of the final value and remain within
from said margin.

2) Instrumentation
We have the following process: A shell-and-tube heat exchanger,
where an incoming fuel current arrives through pipes, and
what we intend to cool with cooling water that enters through
case, and we want to store it in a storage tank.

P&ID Diagram (Process and Instrumentation Diagram)

Measuring and regulating instruments represented by circles and letters
connected by connection lines. They are the most used:

Temperature transmitter
TIC: Temperature controller
Converter
Valve: Final control element

page 14
The instruments that we can find in a P&ID are:
Indicators
Transmitters
Registrars
Converters
Drivers
Actuators (Valve)
Transducers
Connected by transmission lines: Pneumatic, electric or
digital

1st LETTER: Variable 2nd LETTER: It can 3rd LETTER: Function

measure or qualify for of the instrument
related first

Analysis D: Differential I: Indicator

D: Density Relationship R: Register

V: Voltage Security C: Control

Flow Integration Transmitter

Current V: Valve

Power Y: Calculation

L: Level High

Humidity Low

Pressure

Speed

e
m
T
eu
ra
tp
e
r

V: Viscosity

Weight

Z: Position
 The most well-known are: LC (Level Controller), TC (Temperature Controller)
temperature
Next, we will detail the most important aspects:

Transmitters

Sensor: Primary element sensitive to a physical property related

with the variable that one wants to measure.
Transmitter: A system connected to the sensor that converts, conditions and
It normalizes its signal to transmit it over a distance.
Indicator: Combines a sensor and an analog measurement system or
digital.

Pneumatic signal: 3-15 psi

Electrical signal: 4-20 mA

Actuators
Final control elements. They modify the manipulated variable of the process.
according to the controller's signal.
The most common are the valves:

Valves: Device that allows to vary the flow that passes through a
conducting by modifying the pressure drop in it through a
variable closure
They can be: manual closure, retention, safety, On/Off and
regulation.

I/P Converter
Convert the electrical signal (4-20 mA) to pneumatic signal (3-15 psi)

3) PID Controllers

page 16
 Brain of the control circuit.

He is the one who makes the decision: He compares the signal (the signals are normalized,
normally from 4-20 mA) that comes from the transmitter with the Set-Point and
calculate the error. Send the appropriate signal to the actuator, which normally
It is a control valve.
This controller can apply a single action, two or three actions:

Proportional controller
PI: Proportional and Integral Controller
PD: Proportional and Derivative Controller
PID: Proportional, Integral, and Derivative Controller.

To calculate the error, a control algorithm is needed, which comes

given by the following equation, the controller equation:

m=m0 ± K c·e+
( 1
T i∫
it is+T d
of
dt )
Where:

m0 BIAS (Constant value at the output of the controller in the 1st state
stationary

K c Controller gain (Proportional term)

T i Integral time constant (RESET, minutes or seconds)

T d Derivative time constant (RATE)

Difference between DIRECT ACTION/INVERSE ACTION

Let's suppose we have a process in which we intend for the level
do not exceed:

page 17
 a. If the amount of fluid is above the Set-Point, the controller
You must close the control valve to prevent more liquid from passing through.
We have an air valve to open (AA) so it must be
reduce the output of the controller. To make this decision the
the controller must be in reverse action. Some manufacturers
design that action as a decrement, that is, when there is a
increase in the signal entering the controller, then it presents
a decrease in the signal coming out of it.

b. If the fluid quantity is above the Set-Point, the controller

you should open the control valve to try to let the liquid out of the
tank. We have an air valve to open (AA) so that
must increase the output of the controller. To make this decision
The controller must be in direct action. Some manufacturers
design the action as an increase, that is, when there is a
increase in the signal entering the controller, then it presents
an increase in the signal coming out of it.

PROPORTIONAL CONTROLLER
It is a PID controller in which only the proportional action takes place.

TI = 0

Td = ∞ m=m 0± K c· e

With a proportional controller, we do not manage to eliminate the error. For

To manage to eliminate it, the gain of the controller will need to be increased.

If the gain is increased too much, above the value of 'gain.

"last", the process can become oscillatory, even unstable.

p. 18
Proportional band
B.P = 100/Kc
The corrective action is greater the higher Kc is and the lower BP is.

A 100% BP means that the controlled variable must change by 100%

for m to vary by 100%.

2. INTEGRAL CONTROLLER
With comprehensive action, it is possible to eliminate the error.

m=m0 ± K c.e+
( 1
T i∫
e dt
)
The controller output will continue to change until the error is 0.
Ti: Integration or readjustment time (minutes/repetition), that is, time
how long the integral action takes to repeat the proportional action if the error is
constant.
As Ti decreases, the integral action and the aggressiveness of the action increase.
It can't be lowered too much either, as it can be done.
unstable.

Reset wind-up
Delay in correction when the integral term exceeds the limits
superiors or inferiors of the control signal.
Anti wind-up systems, or integral term saturation systems, prevent this.
phenomenon.

3. DERIVATIVE CONTROLLER
This controller anticipates where the process is headed, observing the
error change.
Since what it does is derive the error, it only changes if the error is different from
zero.
The derivative action is known as the speed of derivation or preactivation.
Td: Time it takes for the derivative action to give the same signal as the
proportional action.
The derivative action is not always a good control action, as it does not
It is useful for processes that change rapidly.

page 19
Noisy process signals provoke inappropriate actions, it doesn't matter the
derivative action.

PID (NON-INTERACTIVE)

IDEAL
It is not physically achievable.
Very sensitive to noise
REAL
Physically realizable
Incorporate a filter in the derivative term

PID (SERIES/PARALLEL OR INTERACTIVE)

Used in analog or loop regulators

Typical response of a Feedback control System

In this graph, it can be observed how if a variable is not controlled, it

has suffered a disturbance, it goes out of control, and that the PID controller is the
what best controls the variable.

PID regulators solve most problems well.

monovariable control in slow processes (flow, pressure, velocity...)
PI: Indicated in most cases.
P: Processes with integrator or where an error is not important
null stationary.
- In processes with high delay: Smith predictor.
PID is not suitable in cases of difficult dynamics, or with specifications
demanding
Big delays
Instability

page 20
 Inverse response
Minimum variance
Big dead times
PID tuning methods

Error testing methods

Experimental based methods
Analytical methods based on models
Methods of automatic tuning

4) TYPICAL CONTROL LINKS

page 21