THE COPPERBELT UNIVERSITY

School of mines and minerals sciences

CE 560 / MT 580
Process control
Lecture 04
Design stages of a control system
(feedback, feedforward & inferential control)

C. Botha (Mr.)
Contacts: Chemical Engineering Department
Email: [email protected]
Alt. Email: [email protected]

January, 2024
 Stage 1: Define control objectives

o Coming up with a control system involves a number of stages.

o First stage is to establish the control objectives for a control system.

o Asking appropriate questions during this stage can help to establish suitable
objectives.

o E.g., Specifically what is it about a given process that should be controlled?

o Answer(s) to such a question determines the control objectives.

o The answers and hence control objectives may have to do with one or more of
the process control activities or inceptives discussed.
 Design elements of a control system: control objectives

o E.g., ensure stability, supress external disturbance effect, etc. or a combination of

the motives.

o However, objectives based on functions are general and should be refined.

o Initially, control objectives are described qualitatively.

 Can be refined to be more specific at later stages, i.e., quantitative.

 Usually, control objectives are described in terms of the output variables.

o For the stirred heater tank (refer to earlier notes):

o Control objective was stated qualitatively as maintaining operating temperature

and liquid volume at their desired values.
 Design elements of a control system: control objectives

o Quantitatively, these objectives can simply be defined as:

 Maintaining temperature 𝑇 at 𝑇𝑠 and liquid level ℎ(𝑉) at ℎ𝑠 (𝑉𝑠 ).

o For the CSTR, the control objective can be defined qualitatively as:

 To ensure stability of the unstable mid-state operating point.

 A more useful control objective would be a quantitative translation of the

above.

 E.g., the temperature of the unstable state P2 may not deviate by more than
5% from its nominal value.

 The nominal value being the desired steady state value of the temperature.
 Stage 2: Select measurements

o Once control objectives are defined, the means to monitor a process must be
established.

o I.e., the measurements

o This can be done by measuring appropriate variables, input (s) or output (s).

o Again, suitable question(s) may assist in identifying appropriate measurements.

o E.g., what variable(s) must be measured to properly monitor the process?

o Naturally, monitoring the variables that directly represent the control objectives is
done whenever possible.

o Such measurements are called primary measurements.

 Stage 2: Select measurements cont’d…

o In some cases, variables directly representing the control objectives may not be
measurable.

o That is, they are classified as unmeasured variables.

o In that case, other variables related to the unmeasured variables can be identified
and measured.

o Such variables/measurements are called secondary variables/measurements.

o A mathematical relationship or correlation between unmeasured variables and

secondary measurements can then be established.
𝑢𝑛𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑 𝑜𝑢𝑡𝑝𝑢𝑡 = 𝑓 𝑠𝑒𝑐𝑜𝑛𝑑𝑎𝑟𝑦 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑛𝑡𝑠 4.1
 Stage 2: Select measurements cont’d…

o Once the secondary measurements are available, unmeasured variables can be

determined using Eq. 4.1.

o Exact form of Eq. 4.1 may be determined using semi-empirical, experimental and/or
theoretical approaches.

o Consider the distillation process in Fig. 4.1.

o Process separates hexane and pentane into two product streams.

o I.e., pentane (distillate) and hexane (bottoms).

o Control objective: Maintain composition of the distillate near 95 mole% pentane.

o Disturbance: changes in feed composition.

 Stage 2: Select measurements cont’d…

Set point controller

95 mole % pentane
Composition
analyser

Feed

Hexane

(a)
(b)

Fig.4.1 Distillate composition control: (a) feedback control and (b) feedforward control scheme
 Stage 2: Select measurements cont’d…

o The stated control objective can be achieved by:

 Using a composition analyser to measure pentane amount in distillate. Fig. 4.1

(a).

 Reflux ratio can then be manipulated to keep distillate at the desired

composition (95 mole%).

 This is a feedback configuration (more on this later).

o Alternatively, feed composition can be measured and the reflux ratio adjusted
(feedforward configuration)

o The two control configuration considered depends on the composition analyser.

 Stage 2: Select measurements cont’d…

o Such analysers are typically expensive to acquire and may have poor reliability.

 I.e., fail often or give inaccurate readings.

o This would make it difficult to measure composition of either feed or distillate stream.

o However, it may be possible to measure easily and accurately the tray temperatures.

o This can be done reliably using thermocouples located on the respective trays (Fig.
4.2).

o Mathematical relationship can then be developed between the measured tray

temperatures and composition.
 Stage 2: Select measurements cont’d…

o Such a relationship can be developed from material and energy balances on each
tray.

o Balances are coupled with thermodynamic equilibrium relationships between liquid

and vapour on each tray.

o This results in a relationship for the distillate composition in terms of the tray
temperatures.

o A control mechanism will thus be based on the secondary measurement.

o I.e., tray temperatures to control the composition of the product. See Fig. 4.2.

o Such a control scheme is referred to as inferential control.

 Stage 2: Select measurements cont’d…

Controller uses the

estimate of
Set point
distillate
composition

Using T1, T2, T3

measurements, T1
T2 Distillate (95 mole %
computer estimates T3 pentane)
distillate composition

Feed

Bottoms
(Hexane)

Fig.4.2 Inferential control configuration for distillate composition

 Stage 3: Select manipulated variables

o Once the control objectives are established and measurements identified, next step is
to select manipulated variables.

o Such variables are related to the necessary changes required to control the output
variable.

o In a typical process, usually there is a number of variables that can be adjusted freely.

o Which variables are selected as manipulated is crucial as this affects the quality of the
control.

o Appropriate variables from the list of identified input/output variables must be selected
as manipulated variables.
 Stage 3: Select manipulated variables cont’d…

o In the case of the distillation, this can be the reflux ratio or steam flow rate for the
stirred tank heater.

o Similarly, for liquid level control, either inlet or outlet stream flow can be adjusted.

Stage 4: Select control configuration

o Suitable control configuration is required after establishing the manipulated

variables.

o A number of control strategies are available E.g., Feedback, Feedforward,

Inferential, cascade, ratio control etc.

o To be able to select the appropriate control configuration, these strategies together

with their suitability are discussed.
 Stage 4: Select control configuration –

o A control configuration refers to how information (measurement signal) flows from

the measurements to the available manipulated variables.

Feedback control

o In feedback control, the controlled variable (measured output) is measured first.

o This gives information about the state of a system i.e., how far the controlled variable
is from the set point.

o The measurement is then sent (fedback) or used to adjust the manipulated variable.

o Fig. 4.4. shows a feedback control configuration for the stirred tank heater.

o Note that, the disturbance variable is not measured in feedback control

Stage 4: Select control configuration – feedback control

Fi (ft3/min), Ti (oF)

Thermocouple
T
T h
- Q
Ts + F (ft3/min), T (oF)
ϵ

Controller Condensate

Fst (Ib/min)
Steam

Fig.4.3 Feedback control configurations for stirred tank heater.

 Stage 4: Select control configuration – feedback control

o In some cases, a number of control configurations are possible for feedback control.

o See Fig. 4.4 (a) and (b) for liquid level control.

o For the two alternative feedback control strategies in Fig. 4.4, the same
measurement information i.e., liquid level is used.

 In Fig. 4.4 (a), liquid level measurement is used to adjust the flow rate of the exit
stream.

 On the contrary, the measurement (liquid level) in Fig. 4.4 (b) is used to
manipulate flow rate of the feed stream.

o Which configuration is suitable depends on other factors.

 Select the control configuration: Feedback control strategy

Controller
3 o Differential pressure ϵ
Fi (ft /min), Ti ( F)= const.
(dp) cell hs
- + Set point
- h Fi (ft3/min), Ti (oF)= const. h
hs
Differential pressure
+ Set point (dp) cell
ϵ
T
T h
h Controller
Q
Q

Condensate F (ft3/min), T (oF)

3 o
Condensate F (ft /min), T ( F)
Fst (Ib/min)
Fst (Ib/min) Steam
Steam

(a) (b)

Fig.4.4 Feedback configurations for liquid level control (a) exit flow rate as the manipulated variable,
(b) feed flow rate as the manipulated variable