Control at the Process Unit Level

General Requirements
1 Safety
1. Safety. It is imperative that industrial plants operate safely
so as to promote the well-being of people and equipment
12

within the pplant and in the nearbyy communities. Thus,, plant

p
C pter 1

safety is always the most important control objective.

2. Environmental Regulations. Industrial plants must comply
Chap

with environmental regulations concerning the discharge of

gases, liquids, and solids beyond the plant boundaries.
3. Product Specifications and Production Rate. In order to be
profitable, a plant
p p must make products
p that meet
specifications concerning product quality and production
rate.

4. Economic Plant Operation. It is an economic reality that the

plant
l t operation
ti over long
l periods
i d off time
ti mustt be
b profitable.
fit bl
Thus, the control objectives must be consistent with the
economic objectives.
5. Stable Plant Operation. The control system should facilitate
12

smooth stable plant operation without excessive oscillation in

smooth,
C pter 1

key process variables. Thus, it is desirable to have smooth,

p set-point
rapid p changes
g and rapid p recoveryy from plant
p
Chap

disturbances such as changes in feed composition.

2
 Steps in Control System Design

After the control objectives have been formulated, the control

system can be designed
designed. The design procedure consists of three
main steps:
12
C pter 1

1. Select controlled, manipulated, and measured variables.

2. Choose the control strategy (multiloop control vs.
Chap

multivariable control) and the control structure

(e.g., pairing of controlled and manipulated variables).
3. Specify controller settings.

Control Strategies
Multiloop Control:
Each output variable is controlled using a single input
variable.
12

Multivariable Control:
C pter 1

Each output variable is controlled using more than one

input variable.
variable
Chap

4
 Degrees of Freedom for Process Control
The important concept of degrees of freedom was introduced in
Section 2.3, in connection with process modeling.
The degrees of freedom NF is the number or process variables
that must be specified in order to be able to determine the
12
C pter 1

remaining
i i process variables.
i bl
If a dynamic model of the process is available, NF can be
Chap

determined from a relation that was introduced in Chapter 2,

N F NV N E (13 1)
(13-1)

where NV is the total number of process variables,

variables and NE is the
number of independent equations.

For process control applications, it is very important to determine

the maximum number of process variables that can be
independently controlled, that is, to determine the control degrees
of freedom, NFC:
Definition. The control degrees of freedom, NFC, is the
number of process variables (e.g., temperatures, levels,
12
C pter 1

flow rates, compositions) that can be independently

controlled.
Chap

In order to make a clear distinction between NF and NFC, we will

refer to NF as the model degrees of freedom and NFC as the
control degrees of freedom.
Note that NF and NFC are related byy the followingg equation,
q ,
N F N FC N D (13-2)

where ND is the number of disturbance variables (i.e., input

variables that cannot be manipulated.) 6
 General Rule. For many practical control problems, the
control
t l degrees
d off freedom
f d NFC is
i equall to
t the
th number
b off
independent material and energy streams that can be
manipulated.
12

Example
C pter 1

Determine NF and NFC for the steam-heated, stirred-tank system

Chap

modeled
d l d by
b Eqs.
E 2-50
2 50 2-52
2 52 iin Chapter
Ch t 2. 2 Assume
A that
th t only
l the
th
steam pressure Ps can be manipulated.

Solution
In order to calculate NF from Eq. 13
13-1,
1, we need to determine NV
and NE. The dynamic model in Eqs. 2-50 2-52 contains three
equations (NE = 3) and six process variables (NV = 6): Ts, Ps, w, Ti,
T, and Tw. Thus, NF = 6 3 = 3.
7
12
C pter 1
Chap

Two examples where all three process streams cannot

be manipulated independently.

8
 Stirred-Tank Heating Process
12
C pter 1
Chap

Stirred-tank
Stirred tank heating process with constant holdup, V.

If the feed temperature Ti and mass flow rate w are considered to

be disturbance variables, ND = 2 and thus NFC = 1 from
Eq. (13-2).
It would be reasonable to use this single degree of freedom to
control temperature T by manipulating steam pressure, Ps.
12
C pter 1

E
Example
l
The blending system in Fig. 13.3 has a bypass stream that allows a
Chap

fraction f of inlet stream w2 to bypass the stirred tank. It is

proposed that product composition x be controlled by adjusting f
via
i the
th control
t l valve.
l Analyze
A l the
th feasibility
f ibilit off this
thi control
t l
scheme by considering its steady-state and dynamic
characteristics.
characteristics
In your analysis, assume that x1 is the principal disturbance and
that x2, w1, and w2 are constant
constant. Variations in the volume of liquid
in the tank can be neglected because w2 << w1.
10
 12
C pter 1
Chap

Blending system with bypass line

line.

11

Solution
The dynamic characteristics of the proposed control scheme are
quite favorable because the product composition x responds
rapidly to a change in the bypass flow rate.
In order to evaluate the steady-state characteristics, consider a
12

component
p balance over the entire system:
y
C pter 1

w1 x1 w2 x2 wx (13-3)
Solving for the controlled variable gives,
gives
Chap

w1x1 w2 x2
x ((13-4))
w

Thus x depends on the value of the disturbance variable x1 and

four constants (w1, w2, x2, and w).
But it does not depend on the bypass function
function, f.
f

12
 Thus, it is not possible to compensate for sustained disturbances
in x1 byy adjusting
j g ff.
For this reason, the proposed control scheme is not feasible.
Because f does not appear in (13-4), the steady-state gain
between x and f is zero. Thus, although the bypass flow rate can
12

be adjusted,
adjusted it does not provide a control degree of freedom.
freedom
C pter 1

However, if w2 could also be adjusted, then manipulating both f

Chap

andd w2 could
ld produce
d excellent
ll t control
t l off the
th product
d t
composition.

13

Effect of Feedback Control

Next we consider the effect of feedback control on the control
degrees of freedom.
In general, adding a feedback controller (e.g., PI or PID) assigns
a control degree of freedom because a manipulated variable is
12
C pter 1

adjusted
dj t d by
b the
th controller.
t ll
However, if the controller set point is continually adjusted by a
Chap

higher-level (or supervisory) control system, then neither NF nor

NFC change.
To illustrate this point, consider the feedback control law for a
standard PI controller:

14
 1 t
u t u K c e t e d ((10-5))
1 0
w e e e(t) = ysp(t) y(t) and
where d ysp iss thee set
se po
point.. We consider
co s de two
wo
cases:
12
C pter 1

Case 1. The
h set point is constant, or only
l adjusted
d d manually
ll on an
infrequent basis.
Chap

For this situation, ysp is considered to be a parameter instead of a

variable.
Introduction of the control law adds one equation but no new
variables because u and y are already
y included in the process
p
model.
Thus, NE increases by one, NV is unchanged, and Eqs. 10
10-1
1 and
10-2 indicate that NF and NFC decrease by one.
15

Case 2. The set point is adjusted frequently by a higher level

controller.
The set point is now considered to be a variable. Consequently,
the introduction of the control law adds one new equation and
one new variable,
i bl ysp.
Equations 13-1 and 13-2 indicate that NF and NFC do not change.
12
C pter 1

The importance of this conclusion will be more apparent when

p 16.
cascade control is considered in Chapter
Chap

Selection of Controlled Variables

G id li 1.
Guideline 1
All variables that are not self-regulating must be controlled.

Guideline 2.
Choose output variables that must be kept within equipment and
operating constraints (e.g., temperatures, pressures, and
compositions). 16
 12
C pter 1
Chap

General representation of a control problem.

problem

17

Guideline 3.
Select output variables that are a direct measure of product
quality (e.g., composition, refractive index) or that strongly affect
it (e.g., temperature or pressure).
12
C pter 1

Guideline 4.
Choose output variables that seriously interact with other
Chap

controlled variables.

Guideline 5.
Choose output variables that have favorable dynamic and static
characteristics.

18
 Selection of Manipulated Variables
Guideline 6.
Select inputs that have large effects on controlled variables.
12

Guideline 7.
7
C pter 1

Choose inputs that rapidly affect the controlled variables.

Chap

Guideline 8.
The manipulated variables should affect the controlled variables
directly rather than indirectly.

Guideline 9.
Avoid recycling of disturbances.
19

Selection of Measured Variables

Guideline 10.
Reliable, accurate measurements are essential for good control.
12

Guideline 11.
11
C pter 1

Select measurement points that have an adequate degree of

sensitivity.
sensitivity
Chap

G id li 12.
Guideline 12
Select measurement points that minimize time delays and time
constants

20
 Evaporator Control
12
C pter 1
Chap

Figure 13.5 Schematic diagram of an evaporator.

MVs: Ps, B and D DVs: xF and F

CVs: ??

21

Case (a): xB can be measured

12
C pter 1
Chap

22
 Case (b): ): xB cannot be measured
12
C pter 1
Chap

23

Distillation Column Control

( 55x5
(a 5 control
t l problem)
bl )
12
C pter 1
Chap

24
 Challenges for Distillation Control

1 There can be significant interaction between process

1.
variables.
13
C pter 1

2 Th
2. The column
l behavior
b h i can be b very nonlinear,
li especially
i ll for
f
high purity separations.
Chap

3. Distillation columns often have very slow dynamics.

4. Process constraints are important.

p

5. Product compositions are often not measured.

cf. Distillation Process Control Module (Appendix E)

25

Process Control Module:

Fired-Tube Furnace
12
C pter 1
Chap

The major gaseous combustion reactions in the furnace

f are:
3
CH 4 O2 CO 2H 2O
2
1
CO O2 CO2
2 26
 Process Control Module: Fired-Tube Furnace
Table 3.1
Key Process Variables for the PCM Furnace Module

Measured Output Variables Disturbance Variables

HC outlet temperature HC inlet temperature
12

Furnace temperature HC flow rate

C pter 1

Flue gas (exhaust gas) flow rate Inlet air temperature

O2 exit concentration FG temperature
FG purity (CH4
Chap

Manipulated Variables concentration)

Air flow rate
FG flow rate

27

Control Objectives
j for the Furnace

1. To heat the hydrocarbon stream to a desired exit

temperature
12

2. To avoid unsafe conditions due to the

C pter 1

interruption of fuel gas or hydrocarbon feed

p
3. To operate the furnace economicallyy byy
Chap

maintaining an optimum air-fuel ratio.

Safety Considerations?

g
Control Strategies?

28
 Catalytic Converters for Automobiles
Three-way catalytic converters (TWC) are designed to
reduce three types of harmful automobile emissions:
1. carbon monoxide (CO),
12

2. unburned hydrocarbons
y in the fuel ((HC))
C pter 1

3. nitrogen oxides (NOx).

Desired oxidation reactions (1000-1500 F
F, residence
Chap

times of ~ 0.05 s):

3n 1
2 +2 (
Cn H 2n+2 ) O 2 n CO 2 +(n+1) H 2O
2
1
CO O 2 CO 2
2
Desired reduction reaction:

2NO x x O 2 +N 2

29
12
C pter 1
Chap

TWC efficiency as a function of air-to-fuel ratio

(Guzzella 2008)
(Guzzella, 2008).

30
 TWC Control Strategy
12
C pter 1
Chap

31

Plasma Etching in Semiconductor Manufacturing

CDin

R i
Resist
in
Resist
BARC BARC
Oxide Oxide
12
C pter 1

Polysilicon Polysilicon
out
Gate oxide Gate oxide CDin
Chap

Silicon Silicon
Substrate Substrate

Pre-process Measurement Post-process Measurement

CDin
i CDout
(center, edge) Integrated Integrated (center, edge)
Etch Process
in Metrology Metrology out
(center) (center)

Figure 13.12 Inputs and outputs for polysilicon gate etch process in semiconductor manufacturing.
The measured inputs (CDin and in) in the incoming wafer can be used in feedforward control, while
the measured outputs (CDout and out) are used in feedback control. BARC is bottom anti-reflective
coating.

32