07/03/2020

Introduction to process control

DINAMICA E CONTROLLO DEI PROCESSI CHIMICI

A.A. 2019-2020
Corso di Laurea Magistrale in Ingegneria di Processo e dei Materiali

ANDREA MIO, PhD

MolBNL @ DIA-UniTS; www.molbnl.it
[email protected] - 0405583440

17

DEFINITION OF PROCESS

ENERGY

PRODUCTS
RAW MATERIALS
BYPRODUCTS

The process represents the set of transformations ENERGY

converting raw materials into products (and byproducts):
▰ Usually, energy is exchanged with the surroundings
▰ Sometimes the process involves energy
exchange/transformation only
Often the term «process» is used to indicate both the processing operation and the processing equipment.
We will also assign the same meaning to the words «process» and «system».

18

1
 07/03/2020

MOTIVATION OF PROCESS CONTROL

▰ Safety first
▻ A process unit must operate far from conditions potentially harmful for:
 The operators’ lives
 The surrounding environment (on both short and long terms)
 Equipment integrity

▰ The profit motive

▻ Meeting final product specifications
 In «chemical» products, quality specifications are often related to such properties as concentration, viscosity, melt index, …
 In other industrial sectors (e.g., food, pharma, …) product quality is defined by other properties, such as color, roughness,
crispiness, hardness, …
▻ Minimizing waste production
▻ Minimizing environmental impact
▻ Minimizing energy and/or raw materials use
▻ Maximizing overall production rate

19

THE CONTROL SYSTEM

▰ A process control system must perform two basic tasks

▻ Monitoring the state of certain variables (i.e. controlled variables, CVs) that can provide
indication of the current state of the system
▻ Introducing changes in appropriate variables (i.e. manipulated variables, MVs) in order to
improve the operation of the process

▰ Notice for Italian-speaking students

▻ The word «controllo» must be intended with an active meaning
 It does not simply refer to observation and supervision
 Rather, it means intervention, action
▻ For these reasons, the best italian translation for «control» would be «regolazione»

20

2
 07/03/2020

AN EXAMPLE: TEMPERATURE CONTROL

▰ Shell&Tube heat exchanger

▻ Objective: heating the process fluid from 𝑇 𝑡 to 𝑇 𝑡 = 𝑇 by
condensing a utility stream (steam) within the exchanger
▰ Manual control:
▻ If 𝑇 𝑡 < 𝑇 slightly open the steam valve to transfer more energy
to the process fluid
▻ If 𝑇 𝑡 > 𝑇 slightly close the steam valve to transfer less energy
▰ Problems!
▻ An operator is needed to continuously observe 𝑇 𝑡 , so as to achieve
prompt intervention
▻ Different operators would act with different rationales on the steam
valve, making the exchanger operation inconsistent
▻ Several operators would be required, because in a true process there
might be hundreds of variables to be controlled simultaneously

21

AN EXAMPLE: TEMPERATURE CONTROL

▰ One may think to automate the series of operations involved in the manual
control of the exchanger
▰ Automatic control (closed loop):
▻ A sensor (e.g., a thermocouple) measures the variable of interest (output; CV)
▻ The sensor is physically connected to a transmitter (TT), which receives
the signal and converts it so as to make it «strong» enough to be
transmitted to the controller
▻ The controller (TC) compares the signal to a reference one (set point, SP)
 Depending on the result, it decides if, and by how much, the valve opening should be
changed
 It sends the corresponding signal to the valve

▰ This control strategy is called feedback control or «controllo in retroazione»

22

3
 07/03/2020

THE CONTROL LOOP

OPEN LOOP CLOSED LOOP

▰ The basic elements of a control loop are:

▻ Sensor-transmitter: the plant state is monitored ▰ For the controller to be effective,
 Sometimes called «primary» and «secondary» measurement elements
the action must have an effect on
▻ Controller: a decision is taken on the corrective action to be taken
the measured variable
 The brain of the control system
▻ Final control element: an action is taken
 Usually a valve
 Other examples: variable speed pump, electric motor, electric resistance, …

23

THE JARGON OF PROCESS DYNAMICS AND CONTROL

▰ Inputs: variable that independently «stimulate» a system; they can induce changes in the internal process conditions
(states)
▻ Manipulated variables (u, m, MVs): at the disposal of the control system
▻ Disturbances (d, DVs): it is not possible to act on them
▰ Outputs: measurements (y), through which information on the internal of the system is obtained (e.g., temperature, level,
viscosity, refraction index, …)
▰ States: minimum set of variables (x) that are necessary to completely describe the internal condition of a system (e.g.,
composition, holdup, enthalpy, …)
INPUTS OUTPUTS

Measured disturbance

Manipulated Variables Measured Controlled Variables

Manipulated Variables Unmeasured Controlled variables

Unmeasured disturbance

24

4
 07/03/2020

THE TWO TYPICAL CONTROL PROBLEMS

▰ Regulatory control
▻ The control system objective is to cancel the effect of disturbances,
so as to keep the output to the constant value specified as the
controller set-point
▻ Disturbance rejection

▰ Servo control
▻ The objective is to make the output track a set-point trajectory
(hence, the set-point is time-varying)
▻ Set-point tracking

25

EXAMPLE: A MIXING PROCESS

▰ Objective:
mixing the two inlet streams to obtain an output stream of
assigned composition 𝑥 (weight fraction of A) 1 2

▰ Assumptions:
▻ Constant volume
▻ Constant mass flow 𝑤
▻ Variable composition 𝑥
▻ Stream (2) is pure A
▻ Stream (2) flow 𝑤 can be manipulated

▰ Controlled variable (CV)? 𝑥

▰ Disturbance (DV)? 𝑥
▰ If the nominal value of 𝑥 is 𝑥̅ , which nominal 𝑤 is
▰ Manipulated variable (MV)? 𝑤
required to obtain the desired concentration 𝑥 in the
outlet stream?

26

5
 07/03/2020

EXAMPLE: A MIXING PROCESS

▰ Overall mass balance (nominal conditions)

𝑤 +𝑤 =𝑤
▰ Material balance on species A (nominal conditions)
1 2
𝑤 𝑥̅ + 𝑤 𝑥̅ = 𝑤 𝑥̅

𝑥 − 𝑥̅
⇒ 𝑤 =𝑤
1−𝑥

This is the design equation at the nominal steady state

If the inlet concentration 𝑥 changes with time, how can one

ensure that the outlet concentration 𝑥 is always kept at the
desired value 𝑥 ?

27

EXAMPLE: A MIXING PROCESS

1 2

▰ Method 1: Measure 𝑥 and adjust 𝑤 𝑥 − 𝑥̅

The rationale 𝑤 =𝑤
▻ 1−𝑥
 If 𝑥 is too large, decrease 𝑤 and vice versa
▻ This strategy can be implemented manually or
automatically
 Manual control vs automatic control
▻ How much to decrease (or increase) 𝑤 ?
 An idea: the more 𝑥 differs from 𝑥 , the more FEEDBACK
one should change 𝑤 CONTROL
 i.e., the 𝑤 variation is proportional to the
distance of 𝑥(𝑡) from the set point

▰ Control law ⇒ 𝑤 𝑡 = 𝑤 + 𝐾 [𝑥 −𝑥 𝑡 ] 𝐾𝑐: controller gain or «guadagno»

[𝑥 − 𝑥 𝑡 ]: error

28

6
 07/03/2020

EXAMPLE: A MIXING PROCESS

1 2

▰ Method 2: Measure 𝑥 and adjust 𝑤 𝑥 − 𝑥̅

The rationale 𝑤 =𝑤
▻ 1−𝑥
 If 𝑥 is greater than the nominal value, decrease
𝑤 and vice versa
▻ How much to decrease (or increase) 𝑤 ?
 An idea: using the steady state design equation
under unsteady conditions

FEEDFORWARD
▰ Control Law CONTROL
𝑥 − 𝑥 (𝑡)
⇒ 𝑤 𝑡 =𝑤
1−𝑥

Is the control law truly effective?

29

EXAMPLE: A MIXING PROCESS

▰ Method 3: Measure 𝑥 and 𝑥, and adjust 𝑤

1 2

▰ Method 4: Use a larger mixing tank

Method Measured variable Manipuated variable Category

1 𝑥 𝑤 Feedback (FB)

2 𝑥 𝑤 Feedforward (FF)

3 𝑥, 𝑥 𝑤 FF+FB

4 - - Design change

30

7
 07/03/2020

Block Diagram of FEEDBACK CONTROL

▰ The P&I diagram shows the physical connections

between the control system elements
▰ The block diagram shows the logical connections and
the information flux between the control system
elements

▰ Each block can be

described by a differential
or an algebraic equation

How to obtain these equations?

31

SISO and MIMO systems

▰ Outputs (to be controlled)
▰ SISO system (single input, single output) ▻ 𝑥 ;𝑥
▻ A single manipulated variable (input) and a single ▻ P
controlled variable (output) exist ▻ ℎ ;ℎ
▰ Inputs (to be manipulated)
▰ MIMO system (multiple input, multiple output) ▻ D; B
▻ Manipulated variables and controlled variables are both > 1 ▻ R
▻ 𝑄 ;𝑄

5x5 system ▰ Decentralized control (multiloop control)

▻ 5 control loops of the 1x1 type

▰ Centralized control (multivariable control)

▻ Each MV is manipulated depending
on the values of all the CVs
▻ A dynamic model of the process is
needed

32

8
 07/03/2020

Instrumentation symbols (see SEMD Appendix D, p.487)

T I C - 101
As the 1° letter: As the 2° or 3° or 4° letter:
Letter
measured or controlled variable function or modifier succeeding
1° letter letters Loop number
A Analyzer (composition) Alarm
C - Controller
F Flow rate -
G User’s choice -
H High
I Current Indicator
L Level Low
P Pressure -
R - Registered
S Speed Switch
T Temperature Transmitter

33