Industrial Automation and Control (PE-EC802B), ECE-8th Semester

Module-3
Advanced multiloop control techniques:
We have discussed about the different aspects of PID controllers, their tuning, implementation
etc. In all the cases the control action considered was feedback type; i.e. the output was fed back
and compared with the set point and the error was fed to the controller. The output from the
controller was used to control the manipulating variable. However there are several cases, where
apart from the feedback action few other control structures are incorporated in order to satisfy
certain requirements. There are a number of more complex control techniques that should also be
considered, which can provide improved and economic process control. In these control
configuration one manipulated variable and more than one measurement, or one measurement
and more than one manipulated variable. In such cases, control systems with multiple loops are
to be employed. Systems with multiple loops improve the controllability of process. These
schemes are classified as ‘advanced classical’ multiloop control techniques and are widely used
in industry.

1
In order to illustrate the effect of feedforward control, let us consider the heat exchange process
shown in Fig.1. The cold water comes from a tank and flows to the heat exchanger. The flow rate
of cold water can be considered as a disturbance. The change in input flow line may occur due to
the change in water level in the tank. Suppose, the feedforward line is not connected, and the
controller acts as a feedback control only. If the water inlet flow rate increases, the temperature
of the outlet hot water flow will decrease. This will be sensed by the temperature sensor that will
compare with the set point temperature and the temperature controller will send signal to open
the control valve to allow more steam at the steam inlet. The whole operation is a time
consuming and as a result the response of the controller due to the disturbance (inlet water flow
rate) is normally slow. But if we measure the change in inlet flow rate by a flowmeter and feed
this information to the controller, the controller can immediately take the correcting action
anticipating the change in outlet temperature. This will improve the speed of response. Thus
feedforward action, in addition to the feedback control improves the performance of the system,
but provided, the disturbance is measurable.

2
Let us now draw the block diagram of the overall control operation of the system shown in Fig.
1. The block diagram representation is shown in Fig. 2.

3
4
5
Design of feedforward controller

6
7
8
9
10
11
12
13
14
Cascade control as multiloop control

Consider the heat exchange process shown in Fig. 1.Steam is used to heat the water in the heat
exchanger. Feedback temperature controller is used to compare the water outlet temperature with
the set point and control the steam flow rate by opening or closing the control valve. However
due to the change in upstream steam pressure (Pi), the steam flow rate may change, though the
control valve is at the same position. This will affect the amount of heat exchanged and the
temperature at the water outlet. It will take some time to detect the change in temperature and
take subsequent corrective action. In a cascade control, this problem can be overcome by
measuring the disturbance (change in flowrate in steam due to upstream pressure variation and a
corrective action is taken to maintain constant flowrate of steam. There is an additional controller
(flow controller) whose set point is decided by the temperature controller. The schematic
arrangement of cascade control is shown in Fig.2. its block diagram is shown in Fig.3. Clearly,
there are two control loops – outer and inner, and two controllers. The set point of the inner loop
controller is decided by the outer loop primary controller.

15
16
17
18
Example of cascade control

19
Tuning of cascade control

The correct sequence of operations while tuning the controllers that work in cascade is:

1. Set the primary controller to manual

2. Tune the secondary controller
3. Put the secondary controller in automatic mode
4. Tune the primary controller

Further analysis of feedforward and cascade control

One of the disadvantages to using feedback control is that a disturbance must propagate through
the process before it is detected and action is taken to correct it. This type of control is sufficient
for processes in which some deviation from the set point is acceptable. However, there are
certain processes where this set point deviation must be minimized. Feedforward control can
accomplish this because it corrects and/or minimizes disturbances before they enter the process
[3,4]. A typical feedforward control system is shown in Figure 6.5.

Figure 6.5 Feedforward control system.

In its simplest form, a feedforward controller merely proportions the corrective action to the size
of the disturbance. In other words, the control equation is merely a gain based on steady state,
that is, mass or energy balance at steady state. This does not take into account any of the process
dynamics of the system. If there is a difference, or lag, in the speed of the process response to the
control action when compared to that of the disturbance, it may be necessary to introduce some
dynamic compensation into the control equation. The dynamic compensation correctly times the
control action and response thus giving increased accuracy in the feedforward control.

20
In general, the feedforward dynamic elements will not be physically realizable. In other words,
they cannot be implemented exactly. For instance, if the process disturbance measurement
contains dead time, or lag, the feedforward dynamic compensation would have to be a predictor,
which of course is impossible unless an exact and very fast dynamic model of the process is
available. In practice, the feedforward dynamic elements are approximated by a lead lag network
that is adjusted to yield as much disturbance rejection as possible over as wide a range as
possible.

When feedforward control is used equations are needed to calculate the amount of the
manipulated variable needed in order to compensate for the disturbance. This sounds simple
enough; however, the equations must incorporate an understanding of the exact effect of the
disturbances on the process variable. Therefore, one disadvantage of feedforward control is that
the controllers often require sophisticated calculations as even steady models can be nonlinear
and thus need more technical and engineering expertise in their implementation.

Another disadvantage of feedforward control is that all of the possible disturbances and their
effects on the process must be precisely known. If unexpected disturbances enter the process
when only feedforward control is used, no corrective action is taken and the errors will build up
in the system. If all the disturbances were measurable and their effects on the process precisely
known, a feedback control system for regulatory purposes would not be needed. However, such
complete and error-free knowledge is never available, so feedforward is generally combined with
feedback, as illustrated in Figure 6.6. The intent of this union is that the feedforward mitigates
most of the effects of the principal disturbances and the feedback loops provide residual control
and set point tracking.

Figure 6.6 Feedforward/feedback control system

21
Consider the following example of the feedforward control of a heat exchanger with cascaded
feedback trim control (shown in Figure 6.7). The addition of feedback and cascade control serves
to eliminate offset due to modelling inaccuracies and other non-measured disturbances.

Figure 6.7 Combined feedforward and cascade control of a heat exchanger.

At steady state an overall heat balance can be written for the process as shown in Equations 6.1,
6.2 and 6.3:

(6.1)

(6.2)

Or

(6.3)

where λ is the enthalpy transferred by the steam condensing to form condensate (kJ/kg) and Cp is
the heat capacity of the process fluid (kJ/kg K).

22
In this example, the inlet flow of liquid, W, and the temperature, T1, are measured to determine
the amount of steam required as per Equation 6.3. The desired outlet temperature, T2, is the set
point into the feedforward controller. The feedback temperature controller on the liquid stream
measures T2 to adjust for any disturbances that are not corrected by the feedforward controller.

Typical response curves for a load upset would appear as shown in Figure 6.8. Included in this
figure for comparison is the response curve for feedback control only (with a PID controller) on
the same process.

Figure 6.8 Typical responses of the heat exchanger.

The response of the outlet temperature, T2, for the base case of the feedback control shows the
type of improvement in control that can be achieved with even a simple steady-state feedforward
controller. The lead/lag dynamic compensation shows further improvement over the steady-state
feedforward control.

Common industrial examples of feedforward control in addition to heat exchanger control as

shown above include control of fired heaters, chemical reactors and distillation columns. Other
examples include control of biological process systems, such as fermentors and activated sludge
processes and so on. An industrial illustration of when feedforward control was combined with

23
feedback to control dissolved oxygen addition to a municipal wastewater treatment process
resulted in significant savings in air blower energy consumption for the wastewater treatment
plant . Figure 6.9 illustrates the successful control scheme.

Figure 6.9 Feedforward/feedback control of a wastewater denitrification process.

24
Ratio control as multiloop control

Ratio control is a special type of feedforward control where the disturbance is measured and the ratio
of the process output and the disturbance is held constant. It is mostly used to control the ratio of
flow rates of two streams. Flow rates of both the stream are measured, but only one of them is
controlled. There can be many examples of application of ratio control. Few examples are:
1. fuel-air ratio control in burners,
2. control of ratio of two reactants entering a reactor at a desired ratio,
3. maintaining the ratio of two blended streams constant in order to maintain the composition of
the blend at the desired value.

There can be two schemes for achieving ratio control. The first scheme is shown in Fig. 5. In this
configuration the ratio of flow rates of two streams is measured and compared with the desired
ratio. The error is fed to the controller and the controller output is used to control the flow rate of
stream B.

The second possible scheme for ratio control is shown in Fig. 6. Suppose the flow rate of fluid B
has to be maintained at a constant fraction of flow rate of fluid A, irrespective of variation of
flow rate of A (qA). In this scheme the flow rate of fluid A is multiplied with the desired ratio (set
externally) that gives the desired flow rate of fluid B. This is compared with the actual flow rate
of fluid B and fed to the controller that operates the control valve.

25
Suppose that the above scheme (Fig. 6) is used for controlling the fuel-air ratio in a burner where
airflow rate (fluid B) is controlled. But the desired ratio is also dependent on the temperature of
the air. So an auxiliary measurement is needed to measure the temperature of air and set the
desired ratio. Such a scheme is shown in Fig. 7.

26
Example

We have mentioned earlier that ratio control is a special type of feedforward control. We shall
elaborate this point in the following example [4]. Effects of control actions in feedback only
mode, feedforward only mode and feedforward-feedback mode will also be elaborated in the
context of control of a blending process.

Consider a blending system where two streams are blended; one is uncontrolled wild stream that
acts as disturbance and another is controlled and acts as the manipulated variable. The ratio of
these two streams affect the quality of the output of the process.

First consider a simple feedback system shown in Fig. 8(a). The composition of the output
stream is unmeasured and fed to the composition controller that controls the flow rate of the
manipulated stream. But the performance of this system may not be satisfactory, since any
variation of the wild stream will cause the change in composition of the output, that will be felt
after certain time. As a result the performance of the feedback controller is sluggish.

27
On the other hand the performance of the blending process can be made considerably faster if he
incorporate the ratio control scheme as shown in Fig. 8(b). In this case, the disturbance (flow rate
of wild stream) is measured and corrective action is taken immediately by controlling
manipulated stream so as to maintain a constant ratio between them. It is evident that this the
control action is feedforward only, since the composition of the output stream is not unmeasured.

But the above scheme also suffers from certain limitations, since the composition variation due
to other disturbances is not taken into account. So though faster, if may not yield the desired
performance. The performance may be further improved by choosing a feedforward-feedback
control scheme (Fig. 8(c)) where the composition of the output stream is measured and the
composition controller sets the desired ratio to be maintained.

28
Further analysis of Ratio Control

Ratio control involves keeping the ratio between two variables fixed [2,4], as illustrated in Figure
6.9. Typically, these two variables, y and yw, are flow rates, where yw is the wild or
uncontrolled flow rate and y is the manipulated or controlled flow rate. The wild flow rate is
measured, and the controlled flow rate is then adjusted to maintain a fixed ratio between the two.

Ratio control can be considered a form of feedforward control. This is obviously true since in
ratio control the process variable is measured upstream of the process, as is the case in
feedforward control (Figure 6.4). Take, for example, a reactor with two liquid feed streams.
Ratio control would ensure that these streams were being stoichiometrically fed to the reactor by
measuring the flow of one stream and adjusting the other accordingly. The product stream would
be of no real use in determining whether the stoichiometric ratio was met.

There are two methods by which the ratio between the two variables can be fixed when only one
stream is being manipulated. The first is shown in Figure 6.10.

Figure 6.10 Typical ratio control system.

In this first scheme, both flows are measured and divided to obtain the actual ratio. This is then
compared to the set point and the flow of y is adjusted based on the difference. The set point to
the ratio controller is the desired ratio.

In the scheme shown in Figure 6.11, the flow of the wild stream, yw, is measured and then
multiplied by the desired ratio. The output from the multiplier is the set point for the controller,
29
which compares it to the measured flow and adjusts the flow of y accordingly. In this scenario,
the desired ratio is a constant variable in the multiplier and if a new value for the ratio is needed,
it must be set in the multiplier.

Figure 6.11 Typical ratio control system.

The ratio control configuration shown in Figure 6.10 will not have a steady loop gain because the
ratio calculation itself is in the loop. The loop in Figure 6.10 may become nonlinear, making the
control configuration in Figure 6.11 a more reliable model, since its loop gain is constant.

A common example of ratio control is the case of an adsorption column where a fixed ratio of
V/L is desired. The wild flow rate is the vapour feed to the column, V, while the controlled flow
rate is the liquid flow rate, L. The ratio control seeks to maintain constant absorption factors in
the column by keeping a constant V/L profile.

30
Advanced Control Systems

Selective Control Systems

These are control systems that involve one manipulated variable and several controlled outputs.
Since with one manipulated variable we can control only one output, the selective control
systems transfer control action from one controlled output to another according to need.

Override Control

During the normal operation of a plant or during its startup or shutdown it is possible that
dangerous situations may arise which may lead to destruction of equipment and operating
personnel. In such cases it is necessary to change from the normal control action and attempt to
prevent a process variable from exceeding an allowable upper or lower limit. This can be
achieved through the use of special types of switches:

• HSS (high selector switch):- whenever a variable should not exceed an upper limit.

• LSS (low selector switch); to prevent a process variable from exceeding a lower limit.

31
32
Further study of override control

Override Control (Auto Selectors)

Frequently a situation is encountered where two or more variables must not be allowed to exceed
specified limits for reasons of economy, efficiency or safety. If the number of controlled
variables is greater than the number of manipulated variables, a selection must be made for
control purposes (single input/single output). A selector is used to accomplish this. Selectors are
available in both electronic and pneumatic versions. The only difference between selectors is the
number of inputs a particular hardware implementation may be able to accommodate. In this
section specific examples of such selectors will be discussed. It must be kept in mind that these
are only a few examples of such auto selectors [4].

The two basic building blocks for selector systems are the high and the low selector. The high
selector, shown in Figure 6.12, will pass the highest value of the multiple inputs to the output
signal, ignoring all other inputs.

Figure 6.12 High selector.

Figure 6.12 High selector

The low selector, shown in Figure 6.13, will choose the lowest of inputs to pass through as the
output while ignoring all other inputs.

Figure 6.13 Low selector.

33
By using combinations of these basic building blocks it is possible to build other types of
selectors, such as a median value selection, shown in Figure 6.14. The selector output for a
median value selector is a signal that falls between the highest and lowest input.

Figure 6.14 Median value selector.

Let us investigate some typical applications of these selectors in four areas:

1. Protection of equipment.
2. Auctioneering (choosing from several signals).
3. Redundant instrumentation (used commonly with process analytical equipment).
4. Artificial measurements (establishing artificial limits).
6.4.1 Protection of Equipment
To illustrate how selectors can be used to protect equipment, examine the pump system shown in
Figure 6.15. The pump system demonstrates a situation where there are multiple measurements,
multiple controllers and only one manipulated variable that can provide the following protection:

• Surge protection: when Pin drops below a certain minimum value, close the valve.

• High temperature: when the temperature of the pump exceeds a certain maximum
temperature, close the valve.

• Excessive downstream pressure: when Po exceeds a certain maximum pressure, close the
valve. (Assume Po > P shut off).

34
Figure 6.15 Protection of equipment – pump.

Surge Protection

As Pin begins to drop, the output, m1, will also decrease (note Increase/Increase action on
pressure controller). The output, m1, will be selected by the first and second low selector and
will be passed through as the manipulated variable causing the valve to close.

High Temperature and Excessive Downstream Pressure

If either the pump temperature or outlet pressure begins to increase, both outputs m2 and m3
begin to decrease (note Increase/Decrease action on both of these controllers). The smallest value
will be chosen and passed through to manipulate the valve. In general, the smallest output from
either of the controllers will always be operating the valve.

6.4.2 Auctioneering

The objective of auctioneering is to protect against the highest temperature sensed by one of
many temperature transmitters. In the example shown in Figure 6.16, the control equipment
consists of one controller, four transmitters and one FCE. The highest temperature will be
selected by the high selectors and will be used as the measurement for controlling the fuel to the
oven.

35
Figure 6.16 Auctioneering – temperature control in an oven.

6.4.3 Redundant Instrumentation

For an exothermic reactor (shown in Figure 6.17) too much catalyst can prove disastrous. By
implementing a fail-safe scheme that consists of two composition transmitters that are analysers
and a high selector, the highest reading from the analysers will be utilized by the composition
controller to control catalyst flow. The following actions will occur in the event of catastrophic
failure of the analysers:

1. Downscale failure of analyser – If one analyser fails to zero, the other will be selected to
control catalyst flow and production will not be interrupted.

2. Upscale failure of analyser – If one analyser fails to full scale it will get selected and the
catalyst flow will be stopped. Production is stopped and a possible hazardous situation is
avoided.

36
Figure 6.17 Redundant instrumentation – reactor.

An alternate scheme, shown in Figure 6.18, implements analysers with a medium selector that
will keep the process operating regardless of the failure mode of one of the analysers. The
measurement variable to the controller will always be the median transmitter output. If one of the
analysers fails, either upscale or downscale, the selector will still choose the median value.

Figure 6.18 Redundant instrument – reactor/median selector.

In summary, the amount or quantity of redundancy depends on the importance of the process unit
(reactor, distillation column, etc.). This is because the higher the quantity of redundancy the
higher the cost (capital/operating) becomes and, therefore, the economics must be justified.

6.4.4 Artificial Measurements

Some processes require certain operating constraints to be set. These are referred to as artificial
measurements. These operating constraints can be set through the use of selectors. For example,
consider a distillation column whose feed versus steam characteristic is shown in Figure 6.19.

37
Figure 6.19 Feed/steam characteristic of a distillation column.

Instead of operating the steam versus feed flow on a straight line, operating constraints are set.
The operating constraints require a minimum steam rate of 10%, even if the feed rate drops to
zero. This sets the low limit of the steam flow. Furthermore, at maximum feed rate the steam rate
should not exceed a maximum flow of 90%, the high limit. These constraints can be
implemented as shown in Figure 6.20.

Figure 6.20 Artificial constraints.

38
If the feed flow is within the safe operating region, the signal from the multiplier will pass
through the high selector since it is higher than the low limit. It will also pass through the low
selector since it is lower than the high limit, and then act as the set point for the steam flow
controller. If the feed signal falls below the low limit or above the high limit the proper limit will
be selected and that limit will be a constant high or low signal to the steam flow controller. This
prevents the high and low limits from being exceeded.

Split Range Control

Split range control may be useful in processes where there is one controlled variable and there
are extra manipulated variables. Each of the extra manipulated variables must be able to affect
the controlled variable. An example of split range control is illustrated in the control of an
exothermic reactor, Figure 6.21, which will also be discussed in further detail in Chapter 7 when
we deal with reactor control more comprehensively.

Figure 6.21 Control scheme for an exothermic reactor.

The exothermic reactor in Figure 6.21 can be stabilized if the reaction temperature changes fairly
slowly when compared to the rate at which the jacket temperature changes. The idea of the split
range control loop shown in Figure 6.21 is that once the feedstock and catalyst are added, hot
water in the jacket is used to initiate the reaction. As the reaction temperature increases, the
controller output decreases, closing the hot water valve, which is air to open, and opening the
cold water valve, which is air to close. This is split range control.

It is important to note that there are several implementation issues when applying split range
control. One issue is the tuning of the system. When a different manipulated variable is selected
39
by the controller, the closed loop process dynamics may change. Therefore the controller tuning
should remain the same only if the process dynamics remain the same. For the example in Figure
6.21, this would mean that the hot water dynamics would be similar to the cold water dynamics.
If the closed loop dynamics are significantly different, the controller tuning should be
automatically changed, that is, a different set of tuning parameters should be used.

In practice there is a limit on how accurately a control valve can be adjusted. This has
implications for the number of manipulated variables that may be used in split range. Split range
control is normally limited to two or three manipulated variables for this reason. As a further
consequence of inaccurate valves a split range control system could cycle if there is a dead band
in which neither manipulated variable is adjusted. Valves in split range control systems are
therefore calibrated to have an overlap, for example, of 10% to prevent this from occurring.

Examples of split range control

This type of control is used, where there are several manipulated variables, but a single output
variable. The coordination among different manipulated variables is carried out by using Split
Range Control.

Fig. 8 shows an example of a typical split range control scheme. The steam discharges from
several boilers are combine at a steam header. Overall steam pressure at the header is to be
maintained constant through a pressure control loop. The command from the pressure controller
is used for controlling simultaneously the steam flow rates from the boilers in parallel. Clearly,
there is a single output variable (steam header pressure) while there are a number of
manipulating variables (discharge from different boilers).
40
In a split range control loop, output of the controller is split and sent to two or more control
valves. The splitter defines how each valve is sequenced as the controller output changes from 0
to 100%. In most split range applications, the controller adjusts the opening of one of the valves
when its output is in the range of 0 to 50% and the other valve when its output is in the range of
50% to 100%.

The principle of a split range control is illustrated in the following example:

In figure1, PIC-01 controls the pressure of the separator for liquid-vapor hydrocarbons, by mean
of a split range controller with the output signal split and sent to two pressure control valves PV-
A and PV-B. When pressure increases, the fluid shall be discharged to flare. When the pressure
decreases, Fuel gas is introduced to compensate the pressure of the separator.

41
The fuel gas valve (PV-B) needs to close in response to increasing of pressure of the separator,
while the flare valve (PV-A) will need to open when the pressure increases beyond setpoint.

- When the pressure increases beyond setpoint in range of with 0-50% controller output, PV-B
shall close from fully open to fully close.

- When the pressure increases beyond setpoint in range of with 50-100% controller output, PV-A
shall open from fully close to fully open.

In summary, valve actions by PIC-01 is as follow:

In this case, the service of both control valves is different, with respect to use of fuel gas and
flaring for pressure control. Another case of use of split range control loop is when one control
valve cannot be suitably designed to cover the complete operating range of the controller. In that
case, valve with a smaller Cv operated between 0-50% range and the other operates for 50-100%
range. For example, a pressure controller for accumulator drum in overhead of a stabilizer
column splits range to open 2 control valves. In the low range (0-50 % range in response to high
pressure of the stabilizer), the off gas is routed to a gas plant downstream, in the high range (50-
100% range in response to high high pressure of the stabilizer) the off gas goes to flare by
opening of valve B. The control valve actions is as follow:

42
Adaptive control (Programmed or scheduled and self adaptive control) and
internal mode control

43
Adaptive Schemes

◼ Gain scheduling

◼ Model-reference adaptive control

◼ Self-tuning regulator

44
Gain Scheduling

Model-Reference Adaptive Control

45
Self-Tuning Regulator

46
internal mode control

47