Process Control

Advanced Topics

2. Process Dynamics

Alexandria University - Chemical Engineering Department

 2. Process Dynamics
• Process dynamics refers to the time trajectory of a variable in response
to a change in an input to the process.

• All of us have an inherent appreciation of process dynamics in the sense

that the e ect of a cause takes time to manifest itself. It thus takes 20
minutes for a pot of rice to cook over a ame, 5-10 minutes for the water
in the geyser to heat up su ciently, years and years of dedicated
practice to become an adept musician (or a good engineer, for that
matter!) and so on so forth.

• In each of these examples, a change in the causal variable ( ame,

electric heating or dedicated practice) results in a change over time in
the e ected variable (degree of “cookedness” of rice, geyser water
temperature or a musician’s virtuosity).

• Process dynamics deals with the systematic characterization of the time

response of the e ected variable to a change in the causal variable. In
process control parlance, the causal variable is referred to as an input
variable and the e ected variable 2is referred to as an output variable.

ff
ff
ff
ff
ffi
fl
fl
 2.1. Standard Input Changes
• To systematically characterize the transient response of an output
to a change in the input, the input change is usually standardized
to a step change, a pulse change or an impulse change. These
standard input changes are depicted in the gure. A step change
in the input, the simplest input change pattern, is mainly used in
this course to characterize the process dynamics.
Step Pulse

A A
u(t) u(t)

0 0

ta t
Impulse tb t
ta

u(t) Area =1

t
3



fi
 2. 1. Standard Input Changes

Step Pulse
A

u(t) u(t)

0
0

ta t ta tb t

Impulse
u(t)

t 4



 2.2. Basic Response Types

• The variety of transient responses can be characterized as an
appropriate combination of one or more basic response types

• Transient responses correspond to solution of ordinary di erential

equations

2
d y (t )
2 dy (t )
Eg.
τ 2
+ 2ζτ + y (t ) = K p u (t ) Second order
dt dt

Response of the first order

process for unit step change

dy (t )
τp + y (t ) = K p u (t )
dt

5
ff
2.2. Basic Response Types

Response of the first order dy (t )

process for unit step change τp + y (t ) = K p u (t )
6 dt
 2.2. Basic Response Types

dy (t )
Response of the first order τp + y (t ) = K p u (t )
process for unit step dt
change

= Amount of time counted from the moment the variable starts to

respond that it takes the process variable to reach 63.2% of its
total change (for single capacitance
7
process)
𝜏
 2.2. Basic Response Types
It is important to realize that the gain, K, relates only steady-state
values, that is, how much a change in the input variable a ects the
output variable. Therefore, the gain is a steady-state characteristic
of the process. The gain does not tell us anything about the
dynamics of the process, that is, how fast changes occur.
To describe the dynamics of the process, the following two terms are
needed: the time constant ( ) and the dead time ( or to).

to = nite amount of time between the change in input variable and

when the output variable starts to respond

= Amount of time counted from the moment the variable starts to

respond that it takes the process variable to reach 63.2% of its total
change (for a single-capacitance process)
8
𝜏
fi
𝜏
𝜽
ff
2.2. Basic Response Types
How many time constants are required for the process to line out?
Some may consider the process to be lined out when it is within 5%
of its equilibrium value. For them, the answer is three time
constants. Most engineers would agree that the process is lined out
when it is within 1% of its equilibrium state, which requires ve time
constants. Thus, the lineout time is not a precise number, but it is
between three and ve time constants.

9
fi
fi
2.2. Basic Response Types
Higher Order Response: Second Order Response
2
d y (t )
2 dy (t )
τ 2
+ 2ζτ + y (t ) = K p u (t )
dt dt

10
2.2. Basic Response Types
Higher Order Response:

Connect rst order lags in series Sluggish S-shaped response

11
fi
2.2. Basic Response Types
Higher Order Response:

12
2.2. Basic Response Types
Higher Order Response:

13
2.2. Basic Response Types
Higher Order Response:

14
2.2. Basic Response Types
Higher Order Response:

Every time a capacitance is encountered, it slows the dynamics,

leading to longer time constant ( ) and dead time (to or ) of the
process.

To summarize, multicapacitance, or higher-order processes, are

most often encountered. The reason for this is that processes
are usually formed by single capacitances in series.

15
𝜏
𝜽
 2.2. Basic Response Types
Higher Order Response:

Output response may be oscillatory.

Oscillations before reaching final steady state

Sustained oscillations Unstable oscillatory response

16
2.2. Basic Response Types
Pure Integrator (non self-regulating):

The most common example of a pure integrator is the response of

the tank level to change in the inlet / outlet feed rate. Unless the inlet
and outlet ows are perfectly equal, the tank level is either rising or
falling in direct proportion to the mismatch in the ows.

The level in a tank is thus non-self regulating with respect to the

connected ows. A controller must be used to stabilize all such non-
self regulating process variables. 17
fl
fl
fl
2.2. Basic Response Types
Inverse Response:

A very common example of the inverse response is the response of

the liquid level in a boiler to a change in the heating duty. As the
heating duty is increased, the vapour volume entrapped in the liquid
increases causing the liquid interface level to rise initially.

Over longer duration, the level of course reduces since more liquid is
being vaporized. As will be seen later, dead time and inverse
response can create control di culties.

18
ffi
2.2. Basic Response Types
Pure Dead Time:

19
 2.2. Basic Response Types
Unstable Response:

Y(t)

Oscillatory non-oscillatory

t t
Some systems may be inherently unstable. Unstable transient
responses are shown in the gure . The unstable response may be
non-oscillatory or oscillatory as in the gure. Reactor temperature
runaway is an example of an unstable process.

A control system must be used to stabilize an inherently unstable

system.
20
fi
fi
 2.3. Combination of Basic Response Types
Many transient responses are well represented by combining basic
response types

First Order + Dead Time

21
2.3. Combination of Basic Response Types

Inverse Response

22
 2.3. Combination of Basic Response Types
Inverse response can be represented by the parallel combination of two
rst order lags. One of the lags has a small gain and a small time constant
(ie a fast response) while the other lag has a gain of larger magnitude and
opposite sign with a much larger time constant (i.e. a slow response in the
opposite direction).

23
fi
Example:

A simple red heater is shown in the gure. It has no automatic

controls in place and the minimum of instrumentation – a
temperature indicator (TI) and a fuel control valve. The aim is to
ultimately commission a temperature controller which will use the
temperature as its process variable (PV) and the fuel valve position
as its manipulated variable (MV).

PV

MV

24
fi
fi
 Example:

PV
MV

The gure above shows the e ect of manually increasing the

opening of the valve. While the temperature clearly rises as the
valve is opened, the temperature trend is somewhat di erent
from that of the valve. We use a number of parameters to
quantify these di erences.
25
fi
ff
ff
ff
Example:

PV
MV

The test was begun with the process steady then su cient time
was given for the process to reach a new steady state. We observe
that the steady state change in temperature is di erent from that of
the valve. This di erence is quanti ed by the steady state process
gain and is de ned by the expression

26
fi
ff
fi
ff
ffi
Example:

PV
MV

Process gain, occasionally also called process sensitivity, is

given the symbol Kp. If we are designing controls to be installed
in the DCS, as opposed to a computer‐based MPC, Kp should
generally have no dimensions. This is because the DCS works
internally with measurements represented as fractions (or
percentages) of instrument range.

27
Example:

PV
MV

28
Example:

Instrument ranges are de ned when the system is rst con gured
and generally remain constant.

However, it is often overlooked that the process gain changes if

an instrument is later reranged and, if that instrument is either a
PV or MV of a controller, then the controller should be re‐tuned
to retain the same performance.
29
fi
fi
fi
Example:

The other di erences, in the gure, between the trends of

temperature and valve position are to do with timing. We can see
that the temperature begins moving some time after the valve is
opened.

This delay is known as the process deadtime; until we develop a

better de nition, it is the time di erence between the change in MV
and the rst perceptible change in PV. It is usually given the symbol
θ. Deadtime is caused by transport delays.
30
fi
fi
ff
fi
ff
Example:

PV

MV

Indeed, in some texts, it is described as transport lag or distance

velocity lag. In our example the prime cause of the delay is the
time it takes for the heated uid to move from the rebox to the
temperature instrument. We describe later how deadtime can
also be introduced by the instrumentation. Clearly the value of θ
must be positive but otherwise there is no constraint on its value.
Many processes will exhibit virtually no delay; there are some
where the delay can be measured in hours or even in days.
31
fl
fi
Example:

PV

MV

Finally, the shape of the temperature trend is very di erent from that
of the valve position. This is caused by the ‘inertia’ or capacitance of
the system to store mass or energy.

The heater coil will comprise a large mass of steel. Burning more fuel
will cause the temperature in the rebox to rise quickly and hence
raise the temperature of the external surface of the steel. But it will
take longer for this to have an impact on the internal surface of the
steel in contact with the uid. Similarly the coil will contain a large
quantity of uid and it will take time for the bulk temperature to
increase.
32
fl
fl
fi
ff
 Example:

PV

MV

The eld instrumentation can add to the lag. For example, the
temperature is likely to be a thermocouple located in a steel
thermowell.

The thermowell may have thick walls which cause a lag in the
detection of an increase in temperature. Lag is quite di erent from
deadtime.
Lag does not delay the start of the change in PV. Without deadtime
the PV will begin changing immediately but, because of lag, takes
time to reach a new steady state. We normally use the symbol τ to
represent lag.
33
fi
ff
Example:

The DCS will also be a source of deadtime, on average equal to

half the controller sampling period – more usually known as the
scan interval (ts).

For example, if a measurement is scanned every two seconds,

there will be a delay of up two seconds in detecting a change.
While this is usually insigni cant compared to any delay in the
process, it is a factor in the design of controllers operating on
processes with very fast dynamics – such as compressors.

The delay can be increased further by the resolution (also called

quantisation) of the eld instrumentation. Resolution is the least
interval between two adjacent discrete values that can be
distinguished from one another. Imagine that this is 0.1% of range
and that the measurement is ramped up 10% of range over an
hour. The instrumentation will not report a change until it has
exceeded 0.1%; this will incur additional deadtime of 36 seconds.
34
fi
fi
Example:
Again, only when the process dynamics are extremely fast, do we
have to concern ourselves about the variable delay this can cause.
A much larger source of deadtime is discontinuous measurement.

This is common for many types of on‐stream analysers, such as

chromatographs, which take a sample, analyse it and report the
result some minutes later. Added to this are delays which might
occur in transporting the sample to the analyser and any time
required by the analyser preparing to receive the next sample. Such
delays are often comparable to the process dynamics and need to
be taken account of in controller design.

When trying to characterise the shape of the PV trend we also have

to consider the order (n) of the process. While, in theory, processes
can have very high orders, in practice, we can usually assume that
they are rst order. However, there are occasions where this
assumption can cause problems, so it is important to understand
how to recognise this situation.
35
fi
 2.4. PROCESS AND IMPORTANCE OF PROCESS CHARACTERISTICS
It is important at this time to describe what a process is from a control's
point of view. To do this, consider the heat exchanger shown again in the
gure. The controller’s job is to control the process. In the example at
hand, the controller is to control the outlet temperature. However, realize
that the controller only receives the signal from the transmitter.

It is through the transmitter that the

controller “sees” the controlled
variable. Thus, as far as the
controller is concerned, the MV
controlled variable (PV or c(t)) is
the transmitter output.
The controller only looks at the
process through the transmitter. PV
The relation between the
transmitter output and the
process variable is given by
the transmitter calibration.
36
fi
 2.4. PROCESS AND IMPORTANCE OF PROCESS CHARACTERISTICS

In this example the controller is to manipulate the steam valve

position to maintain the controlled variable at the set point. Realize,
however, that the way the controller manipulates the valve position is
by changing its signal to the valve (or transducer).

Thus the controller does not

manipulate the valve position
MV
directly; it only manipulates its
output signal.

PV
Thus, as far as the
controller is concerned,
the manipulated variable
is its own output.
37
 2.4. PROCESS AND IMPORTANCE OF PROCESS CHARACTERISTICS

If the controller is to control the process, we can therefore de ne the

process as anything between the controller’s output and the signal
the controller receives. Referring to the gure, the process is anything
within the area delineated by the curve.
The process includes the I/P
transducer, valve, heat
exchanger with associated MV
piping, sensor, and transmitter.

PV

That is, the process is

everything except the
controller.
38
fi
fi
2.4. PROCESS AND IMPORTANCE OF PROCESS CHARACTERISTICS

MV Controlled Variable

PV
Manipulated
Variable

MV

PV

39
 2.4. PROCESS AND IMPORTANCE OF PROCESS CHARACTERISTICS
Controlled Variable
MV
The diagram clearly
shows that the process
output is the transmitter Manipulated
PV
Variable
output and the process
input is provided by the
controller output.
Note that we refer to the output
of the transmitter as c(t) to
stress the fact that this signal is MV

the real controlled variable; the

unit of c(t) is %TO (transmitter
output).

PV
We refer to the signal from the
controller as m(t) to stress the
fact that this signal is the real
manipulated variable; the unit of
m(t) is %CO (controller output).
40
 2.4. PROCESS AND IMPORTANCE OF PROCESS CHARACTERISTICS
Now that we have de ned the process to be controlled, it is
necessary to explain why it is important to understand the terms that
describe its characteristics.

The control response depends on the tuning of the controller. The

optimum tunings depend on the process to be controlled. As we well
know, every process is di erent, and consequently, to tune the
controller, the process characteristics must rst be obtained.
What we do is to adapt the controller to the process.

It is important to realize that

once a process is built and MV

installed, it is not easy to change

it. That is, the process is not very
exible. All the exibility resides PV

in the controller since it is very

easy to change its tunings.

41
fl
fl
fi
ff
fi
 2.5. Process Nonlinearities.
The numerical value of K, , and to depend on the process operating
conditions. Processes where these numerical values are constant
over the entire operating range, known as linear processes, occur
very infrequently. Most often, processes are nonlinear. In these
processes the numerical values of K, , and to vary with operating
conditions. Nonlinear processes are the norm.

The gure shows a simple example of a nonlinear process. A

horizontal tank with dished ends is shown with two di erent heights,
h1 and h2.

42
fi
𝜏
𝜏
ff
2.5. Process Nonlinearities.
Because the cross section of the tank at h1 is less than at h2, the
level at h1 will respond faster to changes in inlet, or outlet, ow
than the level at h2. That is, the dynamics of the process at h1 are
faster than at h2.
A detailed analysis of the process shows that the gain depends on
the square root of the pressure drop across the valve. This
pressure drop depends on the liquid head in the tank. Thus the
numerical value of the gain will vary as the liquid head in the tank
varies.

43

fl
2.5. Process Nonlinearities.
The tank process is mainly nonlinear because of the shape of the
tank. Most processes are nonlinear, however, because of their
physical–chemical characteristics.

To mention a few, consider the relation between the temperature

and the rate of reaction (exponential, the Arrhenius expression);
between the temperature and the vapor pressure (another
exponential, the Antoine expression); between ow through a pipe
and the heat transfer coe cients; and nally, the pH.

The nonlinear characteristics of processes are most important

from a process control point of view. As we have already
discussed, the controller is always adapted to the process. Thus,
if the process characteristics change with operating conditions,
the controller tunings should also change, to maintain control
performance.

44
ffi
fi
fl
 2.5. OBTAINING PROCESS CHARACTERISTICS
FROM PROCESS DATA
In this section we learn how to obtain the process characteristics,
K, and (or to) from process data for self-regulating processes.
We have already learned that most processes are self-regulating
and of higher order, with a general transfer function as

Higher-order processes can be approximated by a second-order-

plus-dead-time (SOPDT) transfer function,

What happens in practice, though, is that there is no easy, reliable,

and consistent method to approximate a higher-order process by
this type of transfer function.
45
𝜏
𝜽
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
What it is usually done, therefore, is to approximate a higher-order
system by a rst-order- plus-dead-time (FOPDT) transfer function,

Thus we approximate
higher-order
processes by a low-
order-plus-dead-time
model.

46
fi
 2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA

What it is usually done, therefore, is to approximate a higher-order

system by a rst-order- plus-dead-time (FOPDT) transfer function,

Thus we approximate higher-order processes by a low-order-plus-

dead-time model.
To use a concrete example, consider the heat exchanger shown in
the gure. Assume that the temperature transmitter has a
calibration of 100 to 250°C. To obtain the necessary process data,
the following steps are used:

1. Set the controller to manual mode.

E ectively, the controller is removed.

2. Make a step change in the controller output.

3. Record the process variable.

47
ff
fi
fi
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA

What we just constructed is called the Process reaction curve -

It is the simplest and most often used method. It gives nice visual
interpretation as well.

48
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA

Process reaction curve

Temperature, C

Time, min

t0.283 ΔO t0.632 ΔO
49
 2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Process reaction curve
The method consists in obtaining
two data points from the response

Temperature, C
curve (process reaction curve).

These two points are the time it

takes the process to reach 63.2% of
the total change in output, or
t0.632ΔO, and the time it takes the Time, min

process to reach 28.3% of the total t0.283 ΔO t0.632 ΔO

change in output, or t0.283ΔO; these
two points are shown in the Fig.

Controller Output, %
Time zero is the time when the step
change in controller output occurs.

50 Time, min
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Process reaction curve

With these two data points, and to are

Temperature, C
obtained from the following equations:

Time, min
The units of and to are the same time t0.283 ΔO t0.632 ΔO
units as those used by the control
system. Now that and to have been

Controller Output, %
evaluated, we proceed to evaluate K.
Following the de nition of gain,

51 Time, min
𝜏
fi
𝜏
𝜏
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Process reaction curve

Temperature, C
This gain says that at the present
operating condition, a change of
1%CO results in a change of 1.2°C
in outlet process temperature. This Time, min
gain correctly describes the sensitivity t0.283 ΔO t0.632 ΔO
of the outlet process temperature to a
change in controller output.

However, this gain is only a partial

process gain and not the total
process gain.

52
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Process reaction curve

Temperature, C
However, this gain is only a partial
process gain and not the total process
gain.

The total process gain is the one

Time, min
that says how much the process
output, c(t) in %TO, changes per t0.283 ΔO t0.632 ΔO
change in process input, m(t) in %CO

53
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Process reaction curve

Temperature, C
That is, the process output is given by
the transmitter output and it is not the
temperature. Therefore, we are
interested in how much the transmitter
output changes per change in
controller output, or
Time, min

t0.283 ΔO t0.632 ΔO

54
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Process reaction curve

Temperature, C
The change in transmitter output is
calculated as follows: Time, min

t0.283 ΔO t0.632 ΔO

or, in general,

55
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA

Temperature, C
Therefore, the total process gain for Time, min
this example is
t0.283 ΔO t0.632 ΔO

If the process variable had been recorded in percent of transmitter output,

there would be no need for any extra calculation.
We can now write the transfer function for
this process as

56
 2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA

Temperature, C
This transfer function describes the
relation between the transmitter output Time, min

and the controller output. t0.283 ΔO t0.632 ΔO

If a transfer function describing the relation between the transmitter

output and any other process input (other than the controller output)
is desired, the same procedure is then followed to evaluate K, and
(or to).

In this case the units of the K will be di erent than before; that is,
they will not be %TO/%CO. The units will depend on the units of the
particular input.
57
𝜽
ff
𝜏
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA

58
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods
The previously described graphical method had two major
limitations:

• a FOPDT model

• and a perfect step input.

Statistical model identi cation methods provide more exible

approaches to identi cation that relax these limits to model
structure and experimental design.

In addition, the statistical method uses all data and not just a few
points from the response, which should provide better parameter
estimates from noisy process data.

A simple version of statistical model tting is presented here to

introduce the concept and provide another useful identi cation
method.
59
fi
fi
fi
fl
fi
 2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods

1. Introduce a perturbation (or sequence of perturbations)

in the input variable.
There is no restriction on the shape of the perturbation, but the
e ect on the output must be large enough to enable a model
to be identi ed.



2. Collect input and output response data. It is not
necessary that the process regain steady state at the end of
the experiment.



3. Calculate the model parameters as described in the
subsequent slides.

60
ff
fi
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods
Beside linearized forms of the FOPDT equation, non-linear
regression could be performed directly on the response curve of
an FOPDT driven by a step change by creating an objective
function of the form:
n
n
n
n=1

and then directly iterating on the values of , , and even y’∞

Where t is the time and y’n is deviation from the initial

steady-state value.

61
𝜏
𝜽
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods
n
n
n
n=1

The form of the response curve is:

Where y* is the value at the initial steady state.

62
 2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods

The gure shows part of a Microsoft Excel spreadsheet with

synthetic data corresponding to an ideal second order response
de ned by 1 and 2 subjected to a unit step disturbance.
63
fi
fi
𝜏
𝜏
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods

The function in column B calculates the second order response

using t, 1 and 2 .

64
𝜏
𝜏
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods

In column D, the predicted value of y is calculated from the formula

Using the initial estimates of K, td and , with y* equal to zero.

65
𝜏
 2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods

Manually changing the values of the unknown parameters (K, td and

) will change the shape and position of the green least-squares
tted curve. We can keep trying with di erent values until we obtain
a good t.
66
fi
𝜏
fi
ff
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods

It would be better if we minimize the sum of squared deviations

using Solver add-in.
67
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods

Solver will optimize the parameters values to obtain the best t

that corresponds to the least squared deviations by using an
optimization algorithm. 68

fi
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Statistical Methods

Focusing on the initial stage of the response, we can see the

approximation of a FOPDT model tted to a 2nd order response
that does not have a dead time. 69
fi
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Process reaction curve vs. Statistical Methods

Feature Process reaction curve Statistical method

Input magnitude Signal/noise > 5 Can be much smaller
Experiment duration Reach steady state Steady state not required
Input change shape Nearly perfect step Arbitrary, sufficient “information”
required
Model structure First order with dead time General linear dynamic model
Accuracy with Poor with significant disturbance Poor with significant disturbance
unmeasured
disturbances
Diagnostics Plot prediction vs data Plot residuals
Calculations simple Requires spreadsheet or other
computer program

70
2.5. OBTAINING PROCESS CHARACTERISTICS FROM PROCESS DATA
Typical Loop Dynamics

71
2.6. CONSIDERATIONS WHEN PERFORMING PROCESS TESTING
55
Data can be smoothed visually
or using a numerical method

output variable, degrees C

51
input variable, %

47

43

55 39

45
0 10 20 30 40

time
72
2.6. CONSIDERATIONS WHEN PERFORMING PROCESS TESTING
45 15

35 11

input variable in deviation (% open)

output variable in deviation (K)

25 7

15 3

5 -1

-5 -5
0 10 20 30 40
time (min)

Input should be close to a perfect step;

this was basis of equations. If not, cannot
use data for process reaction curve.
73
2.6. CONSIDERATIONS WHEN PERFORMING PROCESS TESTING

45 15

35 11

output variable, degrees C

input variable, % open
25 7

15 3

5 -1

-5 -5
0 10 20 30 40
time

The output must be “moved” enough. Rule of thumb

for process reaction curve:
74
Signal/noise > 5

2.6. CONSIDERATIONS WHEN PERFORMING PROCESS TESTING

Output did not return

45 close to the10initial
value, although input
35
returned to6initial
value

output variable, degrees C

input variable, % open

25 2

15 -2

5 -6

-5 -10
0 20 40 60 80
time

This is a good experimental design; it checks for

disturbances
2.6. CONSIDERATIONS WHEN PERFORMING PROCESS TESTING

Obvious
nonlinearity
2.7. QUESTIONS WHEN PERFORMING PROCESS TESTING

The following questions must be answered when performing process

testing.

1. In what direction should the controller output be moved?

Safety is the most important consideration. You always want to

go in a safe direction.

2. By how much (%) should the controller output be moved?

Move by the smallest amount that gives you a good (readable)

answer. There are two reasons for this: (1) if you move the
process far from its present steady state, this makes the
operating personnel nervous; and (2) you want to obtain the
characteristics close to the operating point, because much away
from it, nonlinearities may start having an e ect.
77
ff
2.7. QUESTIONS WHEN PERFORMING PROCESS TESTING

3. How many tests should be performed?

We want to have repeatability in the results; therefore, we

could say that we should have as many tests as possible to
ensure repeatability. However, many tests are not realistic
either. To start, perhaps two tests, in di erent directions, are
enough. Once the numerical values of the characteristics are
obtained, they can be com- pared, and if not similar, more
tests may be justi ed. If they are similar, an average can then
be calculated.

4. What about noise?

Noise is a fact in many processes. Once a recording is obtained,

an average process curve can be drawn freehand to obtain an
average curve. This in itself is a way to lter the noise.
78
fi
ff
fi
2.7. QUESTIONS WHEN PERFORMING PROCESS TESTING

5. What if an upset enters the process while it is being

tested?

The purpose is to learn how the manipulated variable a ects

the controlled variable. If a disturbance enters the process
while it is being tested, the results will be due to both inputs
(manipulated variable and disturbance). It is very di cult to
deconvolute the e ects. This disturbance may be the reason
why two tests may provide much di erent results. The test
should be done under the most possible steady-state
conditions.

79
ff
ff
ffi
ff
End of Chapter 2