Papadopoulos
PID Controller
Tuning Using
the Magnitude
Optimum
Criterion
Konstantinos G. Papadopoulos
123
Konstantinos G. Papadopoulos
ATDD
ABB Industries
Turgi, Aargau
Switzerland
ISBN 9783319072623
DOI 10.1007/9783319072630
ISBN 9783319072630
(eBook)
To my mother Maria,
To the memory of my father Georgios,
(19371999)
To Eirini,
To the memory of my advisor Nikolaos
Margaris.
(19492013)
Acknowledgments
viii
Acknowledgments
Thessaloniki, Greece
Zrich, Switzerland
July 2014
Konstantinos G. Papadopoulos
Synopsis
This book introduces a systematic controller design strategy for typeI, typeII and
typeIII linear singleinput singleoutput closed loop control systems regardless of
the process complexity. The main advantage of typeI, typeII, and typeIII loops
is their ability to track fast reference signals since their output variable achieves
zero steady state position, velocity, and acceleration error at the presence of step,
ramp, and parabolic reference input signals, respectively. Since such kind of loops
are often met in many industry applications (electrical and chemical engineering)
the proposed control law is of PID, and therefore fast and quick integration of the
proposed approach can be achieved on a real time application platform.
The development of the proposed theory lies in the wellknown Magnitude
Optimum criterion, takes place in the frequency domain, and is carried out into
two directions. The first direction of the proposed approach deals with the direct
tuning of PID regulators and the second direction deals with the wellknown term
automatic tuning of PID regulators.1
For the direct tuning of PID regulators and further to the control laws proof, a
general transfer function of the process model is involved and based on the type of
the control loop to be designed (typeI, typeII, typeIII), the three parameters of
the PID controller are explicitly determined in terms of closed form expressions.
These closed form expressions involve all process modeled parameters and can be
applied for the control of any SISO process regardless of its complexity.
Therefore, if system identification techniques are followed for the determination of
the transfer function of the plant, the proposed PID controller parameters can be
directly calculated.
Once this step is complete, a new approach to a common problem met in many
real world applications is presented, which is associated with the automatic tuning
of PID regulators. Note that for this problem, given little information about the
ix
Synopsis
Synopsis
xi
regarding the tuning of the PID current control loop of a grid connected converter
through the proposed method is presented.
In similar fashion with Chap. 3, Chap. 4 presents the application of the Magnitude Optimum principle to typeII control loops, commonly known within the
academic and industrial literature as Symmetrical Optimum tuning. Again, the
conventional PID tuning via the Magnitude Optimum criterion for typeII control
loops is presented so that advantages and disadvantages of the current state of the
art are made clear to the reader. To cope with the remarked drawbacks, a revised
PID control law is presented for typeII control loops, introducing again an explicit
solution for the controller parameters. Once more a comparison section for several
benchmark processes follows, both for the conventional and the revised method.
The chapter closes with a practical example of a typeII control loop related to the
control of actual DC link voltage in an AC/DC/AC converter arrangement often
met in the field of electric motor drives.
In Chap. 5 the design of a PID typeIII control loop is presented for first time
over the literature. To achieve this, a similar to the conventional typeII PID design
procedure is introduced which leads effortlessly to the development of the optimal
PID control law for typeIII control loops regardless of the process complexity.
Again, a comparison between the conventional and the revised control law is
performed for several benchmark process models. The chapter closes with the
extension of the conventional PID typeIII tuning to the design of typeIV, typeV
and finally typep control loops.
In Chap. 6, the revised control law is presented for digital control loops3 and
therefore the sampling period of the controller Ts is introduced within the explicit
solution. Comparison results are presented for analog and digital design focusing
on the effect of the sampling period on the control loops performance both in the
time and frequency domain.
Part III consists of Chaps. 7 and 8. In Chap. 7, the proposed automatic tuning
method for typeI control loops is presented. The same principle is extended in
cases where the process contains conjugate complex poles. The application of the
proposed method requires (1) an open loop experiment for initializing the algorithm and (2) access to the output of the process and not to its states. Simulation
examples between the explicit solution and the proposed method justify the
potential of the current approach.
In Chap. 8, the contribution of the proposed theory is summarized and directions to control engineers are given so that the explicit solution and the automatic
tuning algorithm are integrated within a real time application platform.
Finally, all proofs of the revised PID control law for typeI, typeII, typeIII
control loops (analog and digital design) are summarized in Appendices A, B and
C. In Appendix A the principle of the Magnitude Optimum criterion is presented
xii
Synopsis
and certain optimization conditions are extracted, which serve as the basis for the
development of both the optimal analog and digital control law. Appendices B and
C present the proof of the analog and digital control law (typeI, typeII, typeIII),
respectively.
Contents
Part I
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 Target of the Proposed Theory . . . . . . . . . . . . . . . .
1.3 State of the ArtThe Magnitude Optimum Criterion .
1.3.1 TypeI Control Loops . . . . . . . . . . . . . . . . .
1.3.2 TypeII Control Loops . . . . . . . . . . . . . . . . .
1.3.3 TypeIII Control Loops . . . . . . . . . . . . . . . .
1.4 Automatic Tuning of PID Controllers . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3
3
5
6
6
7
8
8
9
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
11
11
14
15
18
19
21
23
26
27
31
32
Part II
3
.
.
.
.
.
.
.
.
.
.
33
xiii
xiv
Contents
3.2.1
3.2.2
I Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Preservation of the Shape of the Step
and Frequency Response . . . . . . . . . . . . . . . . . . .
3.2.3 PI Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.4 PID Control . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.5 Drawbacks of the Conventional Tuning Method . . .
3.2.6 Why PID Control? . . . . . . . . . . . . . . . . . . . . . . .
3.3 Revised PID Tuning Via the Magnitude Optimum Criterion
3.4 Performance Comparison Between Conventional
and Revised PID Tuning. . . . . . . . . . . . . . . . . . . . . . . . .
3.4.1 Plant with One and Two Dominant Time Constants.
3.4.2 Plant with Five Dominant Time Constants . . . . . . .
3.4.3 A Pure Time Delay Process . . . . . . . . . . . . . . . . .
3.4.4 A Nonminimum Phase Process . . . . . . . . . . . . . . .
3.4.5 A Process with Large Zeros . . . . . . . . . . . . . . . . .
3.4.6 Comments on PoleZero Cancellation . . . . . . . . . .
3.4.7 Comments on Disturbances Rejection . . . . . . . . . .
3.4.8 Rejection of Output Disturbances . . . . . . . . . . . . .
3.4.9 Rejection of Input Disturbances. . . . . . . . . . . . . . .
3.4.10 Robustness to Model Uncertainties . . . . . . . . . . . .
3.5 Performance Comparison Between Revised PID Tuning
and Other Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.1 Internal Model Control. . . . . . . . . . . . . . . . . . . . .
3.5.2 ZieglerNichols Step Response Method . . . . . . . . .
3.5.3 Simulation Results. . . . . . . . . . . . . . . . . . . . . . . .
3.6 Explicit Tuning of PID Controllers Applied
to Grid Converters . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.1 Simplified Control Model and Parameters. . . . . . . .
3.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
....
34
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
35
37
38
41
41
42
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
46
46
47
49
51
51
53
55
57
60
62
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
66
67
70
71
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
76
77
82
83
..
..
85
85
.
.
.
.
.
.
.
.
.
.
.
.
88
88
90
90
94
94
..
..
98
98
Contents
xv
.
.
.
.
.
.
.
.
.
.
.
.
101
103
106
106
.......
.......
110
111
.......
.......
.......
111
114
114
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
117
117
119
119
123
127
133
135
146
154
158
159
........
........
161
161
........
........
165
171
........
........
174
179
........
181
........
........
........
188
196
196
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
xvi
Contents
Part III
199
199
202
203
204
204
205
206
210
212
212
215
..
216
..
..
218
221
..
224
..
225
.
.
.
.
.
.
.
.
228
234
236
240
......
243
......
......
......
243
247
247
249
253
269
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
293
.
.
.
.
.
.
.
.
.
.
.
.
Author Biography
xvii
Notations
ap
av
aa
Gs
e
Gs
di s
do s
es
ess
Ts
Ffp s
Fol s
Cs
Cex s
CZOH s
I
kh
kp
no s
nr s
PI
PID
rs
Sno s
xix
xx
Ss
Si s
Su s
STG s
STkh s
Ti ; ti
TRc ; tRc
TRp ; tRp
Tp ; tp
Tz ; tz
Td ; td
trt
Ts ; fs
tss
us
ys
yf s
f
T; s
Notations
, while all
Sensitivity of the closed loop control system, (Ss dys
o s
other inputs within the control loop are set to zero)
, while all other inputs
Input sensitivity transfer function, (Ss dys
i s
within the control loop are set to zero)
Sensitivity of the command signal at the presence of output
disturbance do s
Sensitivity of the closed loop control system in the presence of plants
model variations
Sensitivity of the closed loop control system in the presence of
variations of the feedback path
Time constant (and normalized time constant) of the integral action in
the PID controller
Time constant (and normalized time constant) of the controllers
parasitic dynamics
Parasitic time constant (and normalized time constant) of the plant
Time constant (and normalized time constant) within the plant transfer
function (corresponds to poles of the process)
Time constant (and normalized time constant) within the plant transfer
function (corresponds to zeros of the process)
Time delay constant (and normalized time delay constant) of the
process Gs
Rise time of the step response
Sampling time, sampling frequency
Settling time of the step response
Command signal (output of the control action of Cs)
Output of the closed loop control system
Output of the open loop transfer function Fol s
Damping ratio of a plant with conjugate complex poles
Time constant (and normalized time constant) of a plant with
conjugate complex poles
Part I
Chapter 1
Overview
Abstract Since this book is dedicated to the definition of a general theory of tuning
of the PID controller using the Magnitude Optimum criterion, a brief retrospect
relevant to the evolution of the PID control and the Magnitude Optimum criterion is
presented in this chapter. The strong effectiveness of the PID controller along with
the simplicity of the criterions principle justifies its strong application within many
industry applications till date. By presenting concrete examples from the industry,
the scope of the chapter is also to argue and justify why the functionality of tuning
the PID controller via the Magnitude Optimum criterion has a long history along
with still a much promising future.
1.1 Introduction
The proposed theory presented in this book copes with the design of the PID controller in singleinput singleoutput control loops given also the fact that the complete
knowledge of the process is often unknown, see [13, 20].
This unawareness of the processs behavior is owed to the fact that exact measurement of the states of the process itself is sometimes unfeasible. The reason for this
issue is either the nature of the states within the process, or the lack of proper equipment able for accurate measurement. The unawareness of the process in this book
is called unmodelled dynamics, which as shown in the sequel, plays an important
role in the whole control loops performance. To this end, the proposed theory does
not follow the classical line of wellestablished classical theories, which are often
based on known process models. Examples of such approaches are (1) the linear
state feedback control law, (2) the measures of a control loops performance such
as ITAE (integral timeweighted absolute error), ISE (integral squared error), IAE
(integral absolute error), and (3) the root locus analysis.
In contrast to the aforementioned control design principles, an interesting method
for designing control loops was proposed in the early 80s by Zames, entitled
Feedback and optimal sensitivity, see [40] or commonly known as H design
control principles, see also [10, 14, 15, 30]. Let it be recalled that the goal of this
principle is to design a closedloop control system such that the maximum magnitude
Springer International Publishing Switzerland 2015
K.G. Papadopoulos, PID Controller Tuning Using the Magnitude Optimum Criterion,
DOI 10.1007/9783319072630_1
1 Overview
in any direction and at any frequency of the control loops frequency response is minimized.
Tracking the literature further back in the past, it is found that the idea for minimizing the maximum magnitude of the frequency response in a control loop stems
from the results found in the master thesis of Sartorius, which was published in
Stuttgart in 1945, see [29]. Later on and in 1954, these results were also published
by Oldenbourg and Sartorius in [23] where in this work, the authors concentrate on
applying the proposed principle for the design of typeI1 control loops. This kind of
control design is commonly known within the German literature as Betragsoptimum,
or BO or Magnitude Optimum, see [11].
The design of typeI control loops motivated Kessler in 1955 to apply the aforementioned principle to the design of typeII control loops.2 This method is commonly
known in the German literature as Symmetrische Optimum, Symmetrical Optimum, or SO, see [17, 18] and again is dedicated to the design of typeII control
loops by minimizing the maximum magnitude of the closed loops frequency response at any frequency. For this reason, since the principle for designing a control
loop, either typeI or typeII according to the aforementioned citations, is common,
we refer to the proposed method in this book as the principle of Magnitude Optimum
criterion.
A common feature between the principle of the Magnitude Optimum criterion
and the H control methods is the minimization of the maximum magnitude in the
frequency response of the closedloop transfer function. However, an H control
method is looking for a controller, the order of which is most often not a constraint
in the problem formulation of a strong mathematical optimization procedure, since
the basic goal of this principle is to stick to the basic requirement, which is the
minimization of the maximum magnitude of the control loops frequency response.
For this reason, many are the times where the resulting controller coming out of these
methods is of much high order, the realtime implementation of which, is often under
discussion.
In contrast to the high order of the controller that comes out of an H control
method, the PID controller (only three terms) proves to be simple and effective among
various control schemes proposed in the literature, see [1]. The aforementioned
industrial applications reported in [1] raise automatically the big question, see [2]:
Why does the PID controller stand so vigorously over the various more complex
control methods that have been reported over the literature? What does the order
of an industrial controller have to be? Of course, the answer to this question is not
straightforward, since every industry application introduces its own requirements
and specifications, which of course make the problem more complex.
However, since the three big requirements in any realtime environment are (1)
effectiveness and efficiency, (2) simplicity of implementation, (3) and cost, this book
1 At this point it has to be mentioned that typeI control loops are those loops that are able to track
step reference signals with zero steady state position error, see Sect. 2.5.
2 TypeII control loops, are loops that are able to track step and ramp reference signals with zero
steady state position and velocity error, see Sect. 2.5.
1.1 Introduction
concentrates on the PID control solution. According to the authors opinion, the
effectiveness and simplicity of the PID controller along with the attractive property
of the Magnitude Optimum criterion comprise a downtoearth recipe for acceptable
and satisfactory results in a wide variety of industry applications.
The higher the type of a control loop is, the faster reference signals can track.
1 Overview
that the automatic tuning of the PID controller is a practical problem, which often
comes up on the table of control engineers within a realtime environment.
8. The MO technique may lead to poor attenuation of load disturbances. It was shown
that disturbance rejection can be significantly improved by using a twodegreesof
freedom controller structure [36].
In our opinion, all the above remarks need to be revised for three reasons.
1. First, as it will be proved in the sequel, the conventional design4 procedure via the
Magnitude Optimum criterion for PID type controllers, restricts the controllers
zeros to be tuned only with real zeros leading finally to poor tuning. This approach
does not take into account the fact that the optimal values for the PID controllers
zeros may be conjugate complex, which might result in more robust tuning than
the principle of polezero cancellation.
2. Second, for determining the PID controllers zeros, exact polezero cancellation
has to be achieved between the processs poles and the controllers zeros [3]. This
approach disregards all other plant parameters for the optimal control law and as
a result, the PID parameter tuning is poor and suboptimal.
3. Third, the conventional design procedure via the Magnitude Optimum criterion
has been tested only on a limited class of simple process models [37] and not on
benchmark processes.
Industrial examples of typeI control loops are found in the field of electric motor
drives and grid connected converters. Specifically in grid connected converters, a
typeI control loop is met in vector controlled AC/DC power converters where there
is an inner loop responsible for regulating the current and an outer loop responsible
for regulating the DC link voltage to be utilized by another DC/AC electric motor
drive. In this case, the inner current control loop is of typeI since in its openloop
transfer function there exists only one integrator coming from the PID control action,
see Sects. 2.5 and 3.6.
1 Overview
In addition, as far as the motor connected drive is concerned, a typeII control loop
is the speed control loop in vector controlled or direct torque controlled drives. In this
case, one integrator comes from the speed PI control action and another integrator
comes from the inertia ( s1J ) of the shaft of the motor, the speed of which is controlled.
Apart from the definition of the conventional PID tuning principle via the Symmetrical Optimum criterion for the control of integrating processes, no other work
has been reported regarding the tuning of the PID controller through the Magnitude Optimum criterion. However, the problem for controlling such processes has
been approached by many researchers after incorporating the Smith predictor, see
[6, 21, 31, 32, 39, 41], the Internal Model Control (IMC) principle, see [22] or other
optimization methods [16, 27].
The problem of automatic tuning of regulators can be seen in cases where the
derivation of the process of the model is almost impossible. This may happen due to
the nature of the states of the process or due to the lack of proper measuring equipment
for identifying a model of the process, see [4]. Given these problem restrictions, an
automatically tuned controller has basically to satisfy the requirements listed below.
1. An automatic controller tuning procedure has to decide the proper type of control
action (P control action, PI control action or PID). There are many applications
where the question arises whether the D term is to be added or omitted.
2. An automatic controller tuning procedure has to end up in such controller parameters so that robust performance is achieved by the process in terms of reference
tracking and output disturbance rejection.
3. An automatic controller tuning procedure must have the ability to retune the
controllers parameters in cases where the plant dynamics change in such a way
that finally make the initial controller tuning unacceptable.
As mentioned in Sect. 1.1, the Magnitude Optimum criterion is again adopted for
developing such a technique, see Part III of this book. For developing the proposed
method, it is assumed that access to the output and not to the plants states is possible5
an open loop experiment from the process itself is available, which serves to initialize
the proposed algorithm.
References
1. Ang KH, Chong G, Li Y (2005) PID control system analysis, design, and technology. IEEE
Trans Control Syst Technol 13(4):559576
2. strm KJ, Hgglund T (2001) The future of PID control. Control Eng Pract 9(11):11631175
3. strm KJ, Hgglund T (1995) PID controllers: theory, design and tuning, 2nd edn. Instrument
Society of America, North Carolina
4. strm KJ, Wittenmark B (1973) On self tuning regulators. Automatica 9(2):185199
5. strm KJ, Hgglund T, Hang CC, Ho WK (1993) Automatic tuning and adaptation for PID
controllersa survey. Control Eng Pract 1(4):699714
6. strm KJ, Hang CC, Lim BC (1994) A new Smith predictor for controlling a process with
an integrator and long dead time. IEEE Trans Autom Control 39(2):343345
7. Buxbaum A (1967) Berechnung von regelkreisen der antriebstechnik. Frankfurt am Main,
AEGTelefunken AG, Berlin
8. Buxbaum A, Schierau K, Straughen A (1990) Design of control systems for DC drives. Springer,
Berlin
9. Courtiol B, Landau ID (1975) High speed adaptation system for controlled electrical drives.
Automatica 11(2):119127
10. Doyle CJ, Francis BA, Tannenbaum AR (2009) Feedback control theory. Dover Publications,
New York
11. Fllinger O (1994) Regelungstechnik. Hthig, Heidelberg
12. Frhr F, Orttenburger F (1982) Introduction to electronic control engineering. Siemens, Berlin
13. Gunter S (2003) Respect the unstable. IEEE Control Syst Mag 23(4):1225
14. Helton JW, Merino O (1998) Classical control using H methods. SIAM: Society for Industrial
and Applied Mathematics, Philadelphia
15. Ho MT (2003) Synthesis of H PID controllers: a parametric approach. Automatica
39(6):10691075
16. Isaksson AJ, Graebe SF (1999) Analytical PID parameter expressions for higher order system.
Automatica 35(6):11211130
17. Kessler C (1955) UG ber die Vorausberechnung optimal abgestimmter regelkreise teil III. Die
optimale einstellung des reglers nach dem betragsoptimum. Regelungstechnik 3:4049
5
10
1 Overview
18. Kessler C (1958) Das Symmetrische Optimum. Regelungstechnik, pp 395400 and 432426
19. Loron L (1997) Tuning of PID controllers by the nonsymmetrical optimum method. Automatica 33(1):103107
20. Margaris NI (2003) Lectures in applied automatic control (in Greek), 1st edn. Tziolas
21. Matauek MR, Micic AD (1999) On the modified Smith predictor for controlling a process
with an integrator and long deadtime. IEEE Trans Autom Control 44(8):16031606
22. Morari M, Zafiriou E (1989) Robust process control, 1st edn. PrenticeHall, New Jersey
23. Oldenbourg RC, Sartorius H (1954) A uniform approach to the optimum adjustment of control
loops. Trans ASME 76:12651279
24. Papadopoulos KG, Margaris NI (2012) Extending the symmetrical optimum criterion to the
design of PID typep control loops. J Process Control 12(1):1125
25. Papadopoulos KG, Papastefanaki EN, Margaris NI (2011) Optimal tuning of PID controllers
for typeIII control loops. In: 19th Mediterranean conference on control & automation (MED).
IEEE, Corfu, Greece, pp 12951300
26. Papadopoulos KG, Papastefanaki EN, Margaris NI (2012) Automatic tuning of PID typeIII
control loops via the symmetrical optimum criterion. In: International conference on industrial
technology, (ICIT). IEEE, Athens, Greece, pp 881886
27. Poulin E, Pomerleau A (1999) PI settings for integrating processes based on ultimate cycle
information. IEEE Trans Control Syst Technol 7(4):509511
28. Preitl S, Precup RE (1999) An extension of tuning relation after symmetrical optimum method
for PI and PID controllers. Automatica 35(10):17311736
29. Sartorius H (1945) Die zweckmssige festlegung der frei whlbaren regelungskonstanten.
Master thesis, Technische Hochscule, Stuttgart, Germany
30. Skogestad S, Postlethwaite I (2005) Multivariable feedback control: analysis and design. Wiley,
New York
31. Smith OJM (1959) Closed control of loops with deadtime. Chem Eng Sci 53:217219
32. Stojic MR, Matijevic MS, Draganovic LS (2001) A robust Smith predictor modified by internal
models for integrating process with dead time. IEEE Trans Autom Control 46(8):12931298
33. Umland WJ, Safiuddin M (1990) Magnitude and symmetric optimum criterion for the design
of linear control systems: what is it and how does it compare with the others? IEEE Trans Ind
Appl 26(3):489497
34. Voda AA, Landau ID (1995) A method for the autocalibration of PID controllers. Automatica
31(1):4153
35. Vrancic D (1997) Design of antiwindup and bumpless transfer protection. PhD thesis, University of Ljubljana, Faculty of Electrical Engineering, Ljubljana
36. Vrancic D, Peng YSS (1999) A new PID controller tuning method based on multiple integrations. Control Eng Pract 7(5):623633
37. Vrancic D, Strmcnik S (1999) Practical guidelines for tuning PID controllers by using MOMI
method. In: International symposium on industrial electronics, IEEE, vol 3, pp 11301134
38. Washburn DC (1967) Optimization of feedback control loops. Westinghouse Industrial Control
Seminars
39. Watanabe K, Ito M (1981) A process model control for linear systems with delay. IEEE Trans
Autom Control 26(6):12611268
40. Zames G (1981) Feedback and optimal sensitivity: model reference transformations, multiplicative seminorms, and approximate inverses. IEEE Trans Autom Control 26(2):301320
41. Zhang W, Sun Y, Xu X (1998) Two degreeoffreedom Smith predictor for processes with time
delay. Automatica 34(10):12791282
Chapter 2
Abstract In this chapter, fundamental definitions and terminology are given to the
reader regarding the closedloop control system. The analysis of the control loop
takes place in the frequency domain and, therefore all necessary transfer functions of
the control loop are presented in Sect. 2.2. The important aspect of internal stability
of a control loop is presented in Sect. 2.3, whereas in Sect. 2.4 the property of robustness in a control loop is analyzed. In Sect. 2.5, a clear definition of the type of the
control loop is given, since in Part II, the proposed theory is dedicated to the design
of typeI, typeII, and typeIII, typep control loops. Last but not least, in Sect. 2.6,
the definitions of sensitivity and complementary sensitivity functions are presented
so that the tradeoff feature in terms of controller performance that these two functions introduce is made clear to the reader. Finally, in Sect. 2.7, the principle of the
Magnitude Optimum criterion is presented and certain optimization conditions are
proved that comprise the basic tool for all control laws proof throughout this book.
These optimization conditions serve to maintain the magnitude of the closedloop
frequency response equal to the unity in the widest possible frequency range as the
Magnitude Optimum criterion implies. In the same section, the Magnitude Optimum
criterion is proved to be considered as a practical aspect of the H design control
principle.
11
12
d( t )
u(t)
controlled
process
y(t)
x(t)
I.e., rise of temperature during a motors operation, aging of materials after a certain time (for
example, copper conductors in a squirrelcage induction motor).
2 Negative feedback is present to the water clock invented by Ktesibios (Greek inventor and mathematician in Alexandreia, 285222 BC) and in the steam engine governor patented by James Watt
in 1788.
13
d( t )
comparator
r( t )
reference
filter
r ( t ) +
e( t )
controller
u(t)
actuator
controlled
process
y(t)
x(t)
h( t )
measuring
equipment
Fig. 2.2 General form of a closedloop control system with negative feedback. The path that
connects the reference input r (t) with the output control loop y(t) is called forward path. The path
that connects the measuring equipment with the states and the output of the process is the feedback
path
The path connecting the measuring equipment with the states and the output of the
process is called feedback path. The logic for the existence of these two paths is as
follows:
All information h(t) that can be potentially accessed in the process either from
the output y(t) or the states x(t), is collected by the measuring equipment and is
transferred to the input of the comparator. This information h(t), is then compared
with a reference signal r (t) that describes the desired behavior of the process. This
comparison takes place within the comparator unit, the output of which is the error
e(t) = r (t) h(t). This error enters the controller unit, passes through the actuator,
and finally enters the input u(t) of the process. The goal of u(t) is to make the output
of the process y(t) track perfectly the reference signal r (t).
With respect to the above, it becomes apparent that the aforementioned goal has
to be achieved by the controller unit, which basically, given the error e(t) and the
presence of any disturbance d(t) entering the plant, tries to calculate the proper u(t)
command signal such that the output y(t) tracks perfectly the reference signal r (t). In
principle and as previously mentioned, both perfect disturbance rejection and perfect
tracking of the reference at the same time cannot take place. Therefore, the design
of a control action has to take always into account this compromise and deliver this
command signal to the plant, which satisfies certain constraints according to the
application.
In many industry applications, a control engineer sets as a first priority to design
such a control unit able for fast suppression of disturbances. The reason for this is due
to the nature of these signals which often enter the control loop suddenly and without
any prediction. Thus, tracking of the reference signal is set as a second priority for
the control actions design, since r (t) does not change frequently while its value is
known a priori before setting the control loop into operation.
Finally, once disturbance rejection is achieved for improving reference tracking,
many are the times when the reference signal is filtered to avoid high overshoot at
step changes of r (t), see Fig. 2.2. This control scheme is also known in the literature
as a two degree of freedom controller (2DoF).
14
n r ( s)
+
r ( s)
e( s)
controller
C ( s)
(2.1)
y(s)
= kp C(s)G(s),
e(s)
(2.2)
di ( s )
u ( s)
+
do ( s)
y f ( s)
S
Ffp (s)
y(s)
=
,
r (s)
1 + Fol (s)
+
kp
kh
G ( s)
y ( s)
+
+
n o ( s)
Fig. 2.3 Block diagram of the closedloop control system. G(s) is the plant transfer function, C(s)
is the controller transfer function, r (s) is the reference signal, y(s) is the output of the control loop,
yf (s) is the output signal after kh , do (s) and di (s) are the output and input disturbance signals,
respectively, and n r (s) and n o (s) are the noise signals at the reference input and process output,
respectively. kp stands for the plants dc gain and kh is the feedback path
3
In the case of electric motor drives, for example, kp stands for the proportional gain introduced by
the power electronics circuit, which finally applies the command signal u(t) to the plant which in
this case is the electric motor. For voltage source inverters, the command signal u(t) is voltage. In
the sequel it is explained that the gain introduced by the actuator has to be linear and proportional
so that the command signal of the controller remains unaltered. In the specific case of electric motor
drives, the power part introduces also a time delay with time constant Td which corresponds to
the time the controller decides the command u(t), until the time it is finally applied by the power
electronic circuit. Therefore in this case, the model of the actuator is given as kp esTd
15
yf (s)
= kh kp C(s)G(s),
e(s)
(2.3)
ydo (s)
1
=
,
do (s)
1 + Fol (s)
(2.4)
which expresses the variation of the output ydo (s), in the presence of output disturbance do (s).
Input sensitivity function
Si (s) =
kp G(s)
ydi (s)
=
= kp G(s)S(s),
di (s)
1 + Fol (s)
(2.5)
which expresses the variation of the output ydi (s), in the presence of input disturbance
di (s).
Control (command) signal sensitivity function
Su (s) =
kh C(s)
u(s)
=
= kh C(s)S(s),
do (s)
1 + Fol (s)
(2.6)
which expresses the variation of the command signal u(s) of the controller in the
presence of output disturbance do (s).
In general, if we consider that all inputs of the control loop are acting at the same
time, then after applying the theorem of superposition among (2.1), (2.4), and (2.5),
it becomes apparent that the output of the control loop is determined as
y(s) = T (s)[r (s) + n r (s) kh n o (s)] + S(s)[do (s) + kp G(s)di (s)].
(2.7)
16
A classic reference that remains modern till date, is the paper by Gunter Stein,
Respect the Unstable, which describes accurately the importance of stability in
modern control systems, see [2].
Definition 1 Any closedloop control system is said to be internally stable if for any
bounded signal entering the control loop, all other generated responses (states, output)
remain bounded.
Definition 2 A linear timeinvariant system (LTI) is said to be internally stable, if
and only if, every transfer function from whichever input to whichever output within
the control loop is stable. In other words, every transfer function from whichever
input to whichever output within the control loop must introduce poles only in the
lefthalf plane (LHP).
From the control loop structure presented in Fig. 2.3, it is seen that the difference
between the reference signal r (s) and the output of the control loop y(s) is expressed
by the error signal e(s), because e(s) = r (s) y(s). Since r (s) is bounded and
r (s) = e(s) + y(s), for checking the internal stability of the control loop, it is
sufficient to track either the response of the output signal y(s) or the error signal
e(s). Assuming a stable controller design of C(s) it is apparent that u(s) is also
stable, since u(s) = C(s)e(s). As a result, for checking the internal stability of the
control loop, it is again sufficient to track either the response of the output signal
y(s) or the controllers command signal u(s) in the presence of the bounded signal
r (s).
The same investigation has to take place also for the affect of the disturbance
signals d(s) which enter the control loop either on the input di (s) or the output do (s)
of the process. Therefore, it is necessary to investigate the effect of the signals di (s)
or do (s) on the response of u(s), since both di (s) and do (s) are bounded.
For investigating the way how signals y(s), u(s) are affected in the presence of
the reference signal r (s) and disturbance d(s) (di (s) or di (s)), the internal stability
matrix of (2.8) is introduced
y(s)
T (s) kp G(s)S(s) r (s)
.
(2.8)
=
di (s)
u(s)
C(s)S(s) kh T (s)
From (2.8), it is concluded that internal stability for the control loop of Fig. 2.3 is
guaranteed only if each one of the transfer functions T (s), Si (s), Su (s) is stable.
For the definition of T (s), Si (s), Su (s) see accordingly (2.1), (2.5), and (2.6). After
algebraic manipulation of (2.8), it is seen that
y(s) = r (s)T (s) + di (s)kp G(s)S(s)
(2.9)
which is valid if we set in the general expression of y(s) (2.7), n r (s) = 0, n o (s) = 0,
and do (s) = 0. Moreover, from (2.8) it is seen that
u(s) = r (s)C(s)S(s) kh di (s)T (s)
(2.10)
17
which is also valid. If di (s) = 0 and assuming then that r (s) is the only active input
in the control loop, it is necessary to prove that u(s) = r (s)C(s)S(s). This is proved
from Fig. 2.3, since if di = do = n r (s) = n o (s) = 0 then e(s) = r (s)kh kp G(s)u(s)
u(s)
= r (s) kh kp G(s)u(s),
C(s)
(2.11)
(2.12)
(2.13)
(2.14)
or
or
or finally
1
From (2.14) and along with (2.4) it is apparent that u(s) S(s)
= r (s)C(s) or finally
u(s) = r (s)C(s)S(s).
(2.15)
In a similar fashion, it can be proved that u(s) = di (s)kh T (s) assuming all other
inputs within the control loop are set to zero.
From Fig. 2.3 it is obvious that
u(s) + di (s) =
or
u(s) +
u(s)
kp kh C(s)G(s)
u(s)
+ di (s) = 0.
kp kh C(s)G(s)
(2.16)
(2.17)
1
u(s)
kh
1 + kp kh C(s)G(s)
kp C(s)G(s)
+ di (s) = 0
(2.18)
1
+ di (s) = 0,
kh T (s)
(2.19)
which is equal to
u(s) = kh di (s)T (s).
(2.20)
18
2.4 Robustness
Robust performance is of primary importance when designing a control law. In
other words, it is related to the ability of the controller to deliver the necessary
command signal to the plant, which both makes the plant achieve perfect tracking
of the reference along with satisfactory disturbance rejection and regardless of the
changes that might take place within the process during its operation.
For measuring robustness, the functions of sensitivity and complementary sensitivity are introduced. The sensitivity function for two functions F, S is given as
SGF (s) =
G dF
dF/F
=
dG/G
F dG
(2.21)
see [3]. By applying the aforementioned definition to the sensitivity of the closedloop transfer function T with respect to changes in the transfer function of the process
G see Fig. 2.3, results in
SGT (s) =
1
1
G dT
=
=
= S(s).
T dG
1 + kp kh C(s)G(s)
1 + Fol (s)
(2.22)
kp kh C(s)G(s)
Fol (s)
kh dT
=
.
=
T dkh
1 + kp kh C(s)G(s)
1 + Fol (s)
(2.23)
If the magnitude of the openloop transfer function Fol (s) is fairly high compared
to unity (Fol (s) 1) then (2.22) and (2.23) are transformed into
SGT (s) =
G dT
1,
T dG
(2.24)
SkTh (s) =
kh dT
1.
T dkh
(2.25)
and
Equation (2.24) reveals that possible changes on the model G of the process do not
affect seriously the behavior of the closedloop transfer function T and therefore
of the closedloop control system. Moreover, from (2.25), it is concluded that any
variation that takes place in the feedback path kh , is transferred directly and without
any change to the output of the closedloop control system T .
With respect to the above, it is apparent that the sensitivity of the units located
in the forward path of the closedloop control system is directly transmitted to the
feedback path. As a result, when designing a closedloop control system, extra care
must be taken by the control engineer for the sensitivity of the feedback path. After
2.4 Robustness
19
(2.26)
Note at this point that (2.26) is the fundamental equation that connects the sensitivity
S with the transfer function of the closedloop control system T , via the feedback
path kh . In case of unity feedback systems kh = 1, (2.26) is rewritten as follows:
T (s) + S(s) = 1,
(2.27)
bm s m + bm1 s m1 + + b1 s + b0
an s n + an1 s n1 + + a1 s + a0
(2.28)
an s n + + cm s m + + c1 s + c0
r (s)
an s n + an1 s n1 + + a1 s + a0
(2.29)
If r (s) =
1
s
an s n + + c2 s 2 + c1 s + c0
r (s).
an s n + an1 s n1 + + a1 s + a0
then
e() = lim
s0
c0
a0
(2.30)
(2.31)
y(s) 4
do (s)
S(s) stands for the sensitivity of the closedloop control system and is defined as S(s) =
when r (s) = n r (s) = di (s) = n r (s) = 0.
4
y(s)
do (s)
20
T (s) =
S(s) = s
s m bm + + s 2 b2 + sb1 + a0
,
s n an + + s 2 a2 + sa1 + a0
(2.32)
an s n1 + an1 s n2 + + (am bm )s m1
+ (am1 bm1 )s m2 + s(a2 b2 ) + a1 b1
,
s n an + s n1 an1 + + s 2 a2 + sa1 + a0
(2.33)
respectively. If (2.32) and (2.33) hold by the closedloop control system is said to be
of typeI. In a similar fashion, if r (s) = s12 then the velocity error is equal to
e() = lim
s0
an s n + + cm s m + + c1 s + c0
an s n + an1 s n1 + + a1 s + a0
1
s
(2.34)
s m bm + s m1 bm1 + + sa1 + a0
,
s n an + s n1 an1 + + sa1 + a0
an s n2 bm s m2 bm1 s m3 + + a2 b2
.
s n an + s n1 an1 + + s 2 a2 + sa1 + a0
(2.36)
(2.37)
an s n p + an1 s n1 p bm s m p
bm1 s m1 p + a p b p
n
s an + s n1 an1 + + s 2 a2 + sa1 + a0
(2.38)
In gridconnected power converters and when vector control is followed for regulating the DC
link voltage to be utilized by the motor connected converter, there is one inner loop for regulating
the current of the power converter and one outer loop for regulating its DC link voltage. In this
case, the inner current control loop is of typeI, since in its openloop transfer function there exists
only one integrator coming from the current PI control action, whereas the outer control loop is of
typeII, since the openloop transfer function introduces two integrators, one coming from the DC
1
link voltage PI control action and another coming from the capacitor bank path ( sC
). A case of
typeII control loop in the field of electric motor drives is the speed control loop in vectorcontrolled
or direct torquecontrolled drives. In this case, one integrator comes from the speed PI control action
and another integrator comes from the inertia ( s1J ) of the shaft of the motor the speed of which is
controlled.
and
T (s) =
21
bm s m + + a p s p + a p1 s p1 + + a1 s + a0
,
an s n + + a p s p + a p1 s p1 + + a1 s + a0
(2.39)
respectively. Also, one could argue according, to (2.38), that type p control loops
are characterized by the order of zeros at s = 0 in the sensitivity function S(s),
see (2.33), (2.37) and (2.38). In a similar fashion, the type of the control loop is
automatically defined by the closedloop transfer function T (s) when observing the
terms of s j ( j = 0 . . . p1) both in the numerator and the denominators polynomial.
(2.40)
(2.41)
(2.42)
(2.43)
holds by within the closedloop control system of Fig. 2.3. From (2.43), it is apparent
that if sensitivity S is large enough, any disturbance signal (do (s) or di (s)) entering
the control loop is amplified, and as a result the output of the control loop y(s) can
hardly track the reference signal r (s). To this end, the main problem which a control
engineer faces when designing an output feedback control loop, is that in such a
system, it is impossible to have perfect tracking of the reference signal r (s) along
with optimal disturbance and noise rejection at the same time.
Looking further on this statement, one can claim that the aforementioned conclusion is not 100 % correct, if we consider the frequency spectrum of both the noise
and disturbance signals that enter the control loop. Often in many realtime applications, the reference signal r (s) along with disturbances do (s) (i.e., load disturbance
in electric motor drives operation) that appear at the output of the process, are signals
of low frequency. By contrast, noise signals come basically by measuring equipment
and most of the time contain highfrequency components.
22
Taking into account these facts, it becomes apparent that if the magnitude of T
remains equal to unity in the widest possible frequency range, then complementary
sensitivity is low enough, see Fig. 2.4 and therefore lowfrequency disturbances are
not amplified by low sensitivity S in the lowfrequency region. As a result, satisfactory
tracking of the reference can be achieved while disturbances are suppressed.
On the other hand, since noise signals appear in the higher frequency region, they
cannot be amplified by the low complementary sensitivity T since it is close to zero
in the highfrequency region; see Fig. 2.4. Finally, no disturbances can be amplified
by the high magnitude of the complementary sensitivity S, since they do not exist in
this highfrequency region.
However, it has to be pointed out that a high magnitude of T does not necessarily
mean that the magnitude of S is low, since the relation (2.40) between T, S is a relation
between vectors, see also Fig. 2.5. The aforementioned statement is true only in the
Fig. 2.5 Geometric
interpretation of
T (s) + S(s) = 1 in the
complex plane
23
case where the angle cl of T is very low. As a result, it becomes apparent that
optimal disturbance rejection along with perfect reference tracking can be achieved
only when
(2.44)
T ( j) 10 .
Since practically this kind of design cannot be achieved, control engineers have to
design control loops such that the frequency response of the closedloop control
system does not exhibit any resonance all over the low and highfrequency regions.
(2.45)
bm s m + bm1 s m1 + + b2 s 2 + b1 s + b0
an s n + an1 s n1 + + a2 s 2 + a1 s + a0
(n m) .
(2.46)
n
T ( j)n d .
= min lim
(2.47)
(2.48)
24
(a)
(b)
 T ( j )
 T ( j ) 3
 T ( j ) 2
 T max 
max
 T ( j )
Fig. 2.6 a Frequency response of T ( j) with resonant peak Tmax . b Frequency response of
T ( j)n for various values of parameter n, n = 1, 2, . . .
In a similar way, for calculating the surface of T ( j)n , we can rewrite according
to
n
.
(2.49)
E = Tmax
For calculating the surface of T ( j)n , we can also write
E = T ( j)n d.
(2.50)
Note that (2.49) is equal to (2.50) in case where n becomes sufficiently high and
the term becomes sufficiently small. Strictly speaking, (2.49) is equal to (2.50)
when n and 0. For this reason, after taking the lim of both (2.49) and
(2.50) when n , we can rewrite
n
T ( j)n d = lim (Tmax
).
lim
(2.51)
n
n
T ( j)n d = lim Tmax .
lim
n
(2.52)
= 1 .
(2.53)
25
n
T ( j)n d = lim (Tmax ) .
lim
(2.54)
n
T ( j)n d = min lim (Tmax ) .
= min lim
(2.55)
For this, we have to invent a systematic approach that satisfies the condition
H = min(Tmax ).
(2.56)
(a)
(2.57)
(b)
 T ( j )
 T ( j )
 Tmax 
n=1
T
Fig. 2.7 a Frequency response of T ( j)n for various values of parameter n and when n .
b Desired frequency response of T ( j) after minimization of any resonant peak at any resonance
frequency
26
(2.58)
(2.59)
D( j)2 + A8 16 + A7 14 + A6 12 + A5 10
(2.60)
+ A4 + A3 + A2 + A1 + A0 ,
(2.61)
(2.62)
(2.63)
(2.64)
2.8 Summary
In this chapter, preliminary definitions of the operation of closedloop control systems
were presented in Sect. 2.1. It was shown how the problem of perfect reference
tracking is in conflict with any kind of disturbance entering the control loop from
the outer world. To justify this statement, in Sect. 2.2, the closedloop control system
was presented in a more concrete mathematical modeling by the frequency domain
approach. With respect to this approach, basic transfer functions of the control loop
were presented, which serve as proof of the proposed PID control law, which follows
in the next chapters for any typeI, typeII, and typeIII control loops.
Given the aforementioned necessary definitions regarding the transfer functions
involved within a closedloop control system, in Sects. 2.3 and 2.4 the important
aspect of internal stability and robustness in a control loop were covered. In Sect. 2.5,
a mathematical approach was presented relevant to the type of feedback control loop.
This section aims at giving the reader quick hints on how to easily identify, given the
2.8 Summary
27
closedloop transfer function of a control system, its exact type (typeI, typeII, and
typeIII).
In Sect. 2.6 it was shown why it is important to keep the magnitude of the closedloop control system equal to unity in the widest possible frequency range (T ( j)),
since under certain circumstances this principle leads to satisfactory disturbance
rejection both at the input and output of the process. This section is also the connecting ring to the principle of the Magnitude Optimum criterion which is finally
presented in Sect. 2.7. The principle of the Magnitude Optimum criterion is considered as a practical aspect of the H and is used to deploy the proposed PID control
laws presented in the following chapters. Finally, certain optimization conditions are
derived in Sect. 2.7 which serve as the basis for the explicit definition of the PID
control action irrespective of the process complexity.
References
1. Ang KH, Chong G, Li Y (2005) PID control system analysis, design, and technology. IEEE
Trans Control Syst Technol 13(4):559576
2. Gunter S (2003) Respect the unstable. IEEE Control Syst Mag 23(4):1225
3. Horowitz I (1963) Synthesis of feedback systems. Academic Press, London
4. Margaris NI (2003) Lectures in applied automatic control (in Greek), 1st edn. Tziolas, Greece
5. Middleton RH (1991) Tradeoffs in linear control system design. Automatica 27(2):281292
6. Morari M, Zafiriou E (1989) Robust process control, 1st edn. PrenticeHall, New Jersey
7. Petridis V (2001) Automatic control systems, part B (in Greek), 2nd edn. Ziti, Greece
8. Voda AA, Landau ID (1995) A method for the autocalibration of PID controllers. Automatica
31(1):4153
9. Voronov AA (1985) Basic principles of automatic control theoryspecial linear and nonlinear
systems. MIR Publishers, Moscow
10. Zames G, Francis BA (1983) Feedback, minimax sensitivity, and optimal robustness. IEEE
Trans Autom Control 28(5):585600
Part II
Chapter 3
Abstract In this chapter, the tuning of the PID controller via the Magnitude
Optimum criterion for typeI control loops is presented. Initially, the revision of
the conventional Magnitude Optimum design criterion for tuning the PID type controllers parameters is presented in Sect. 3.2, which serves as a basis for the reader
to understand the current state of the art, see Sects. 3.2.13.2.4. This revision reveals
three fundamental drawbacks, which are summarized in Sect. 3.2.5 and prove to
restrict the PID controllers optimal tuning in terms of robustness and disturbance
rejection at the output of the plant. Sorting out these drawbacks in the beginning,
one can argue that: (1) with the conventional PID tuning and for determining the
PID controllers zeros, exact polezero cancellation has to be achieved between the
processes poles and the controllers zeros. (2) To this end, the conventional PID tuning via the Magnitude Optimum criterion restricts the controllers zeros to be tuned
only with real zeros. (3) Last but not the least, the conventional design procedure via
the Magnitude Optimum criterion has been tested only to a limited class of simple
process models. To overcome the aforementioned drawbacks, a revised PID type
control law is then proposed in Sect. 3.3. For the development of the control law a
general transfer function process model is employed in the frequency domain. The
final control law consists of analytical expressions that involve all modeled process
parameters. The resulting control law can be applied directly to any linear single
input single output stable process regardless of its complexity. A summary of the
explicit solution is presented in Sect. 3.3 and the analytical proof of the control law
is presented in Appendix B.1. For evaluating the proposed theory, an extensive simulation test batch between the conventional and the revised PID tuning is performed
in Sect. 3.4 for various benchmark processes. Throughout this evaluation, the validity of several literature comments related to the Magnitude Optimum criterion is
discussed in Sects. 3.4.6 and 3.4.7. Finally, it is shown that the performance of the
proposed control law compared to the conventional PID design procedure achieves
satisfactory results both in the time and the frequency domain, in terms of robustness
and disturbance rejection.
31
32
3.1 Introduction
The principle of the Magnitude Optimum criterion, introduced by Sartorius and
Oldenbourg [21, 22] is based on the idea of designing a controller, which renders the
magnitude of the closed loop frequency response as close as possible to unity, in the
widest possible frequency range. Oldenbourg and Sartorius applied the Magnitude
Optimum criterion in typeI systems with stable real poles. In succession, Kessler
suggested the Symmetrical Optimum criterion [15, 16]. The name of this criterion
comes from the symmetry exhibited by the open loop frequency response. In reality,
the Symmetrical Optimum criterion is not something different, but the application
of the Magnitude Optimum criterion to typeII control systems.
The design of control systems both with the Magnitude Optimum and Symmetrical
Optimum criteria of OldenbourgSartorius and Kessler, respectively, presents at least
two important advantages according to [2, 24]: firstly they do not require the complete
plant model see [24] and secondly, the setpoint response of the closed loop system
is excellent [2]. Further to these statements, the Symmetrical Optimum criterion is
more known because of its successful application in the control of electric motor
drives, see [6, 9, 17, 25]. However, excluding the German bibliography [8, 10, 11,
18], the Magnitude Optimum criterion is rarely referred today. In addition to this,
the limited impact of both the Magnitude and Symmetrical design criteria is stressed
in [24], and it might be owed to the negative comments that occasionally have been
stated in the literature. Some of these comments are presented in Sect. 1.3.1 all of
which, need, in our opinion, to be revised for three reasons.
1. Firstly, as it is proved in the sequel, the conventional design1 procedure via the
Magnitude Optimum criterion for PID type controllers, restricts the controllers
zeros to be tuned only with real zeros leading finally to poor tuning. This approach,
does not take into account the fact that the optimal values for the PID controllers
zeros may be conjugate complex, which might result to more robust tuning than
the principle of polezero cancellation.
2. Secondly, for determining the PID controllers zeros, exact polezero cancellation
has to be achieved between the processs poles and the controllers zeros [2]. This
approach disregards all other plant parameters for the optimal control law and as
a result, the PID parameter tuning is poor and suboptimal.
3. Thirdly, the conventional design procedure via the Magnitude Optimum criterion
has been tested only to a limited class of simple process models [26, 27], and not
to benchmark processes as it is carried out in Sect. 3.4 of this chapter.
Based on the above and for the sake of a clear presentation of the proposed theory, this
chapter is organized as follows. In Sect. 3.2, the conventional tuning method of PID
controllers via the Magnitude Optimum criterion is presented, so that all drawbacks
are made clear, see Sect. 3.2.5.
3.1 Introduction
33
Taking into account the aforementioned drawbacks, in Sect. 3.3, the revised PID
type control law is developed. For the control laws proof, a general transfer function
of the process model is employed consisting of n poles m zeros plus time delay
d. In Sect. 3.4, the conventional and the revised PID control law are compared via
simulation examples for benchmark processes met in many industry applications.
The comparison focuses on the performance of the control law in terms of disturbance rejection and reference tracking. Finally, after the verification of the proposed
control law, the validity of the several negative comments toward the Magnitude
Optimum criterion presented in this section, is investigated in Sects. 3.4.6, 3.4.7
and 3.4.10.
(3.1)
for which Tp1 > Tp2 > > Tpn is also considered. Note that kp stands for the
plants DC gain at steady state. Supposing that no information about the real process
is available, it is conceived as a first order one [2, 9, 13], defined by the approximation
n r (s)
r (s)
+
+
e(s)
controller
C (s)
di ( s )
u (s)
+
y f (s)
S
do (s)
kp
kh
G (s)
+
+
y (s)
+
+
n o (s)
Fig. 3.1 Block diagram of the closedloop control system. G(s) is the plant transfer function, C(s)
is the controller transfer function, r (s) is the reference signal, do (s) and di (s) are the output and
input disturbance signals, respectively, and n r (s), n o (s) are the noise signals at the reference input
and process output, respectively. kp stands for the plants DC gain and kh is the feedback path
34
=
G(s)
where Tp =
n
i=1 Tpi
1
,
1 + sTp
(3.2)
3.2.1 I Control
When the information about the plant is limited, the control that can consciously be
applied is limited to integral control action, so that the system exhibits at least zero
steady state position error. By applying integral action given by
C(s) =
1
sTiI 1 + sTc
(3.3)
to the approximate plant (3.2), the resulting closed loop transfer function takes the
form of
kp
sTiI (1 + sTc )(1 + sTp ) + kh kp
kp
2
,
s TiI T + sTiI + kp kh
(s) =
T
(3.4)
for which
Tc Tp 0 and T = Tc + Tp
(3.5)
has been considered. Note that Tc stands for the controllers unmodeled dynamics,
which are involved between the output of the controller and the input signal to the
plant.
According to the conventional design via the Magnitude Optimum criterion, the
integration time constant TiI of the controller and the parameter kh in the feedback
path are determined so that the amplitude of the closed loop transfer function T (s)
( j) 1 in the wider possible frequency range. The
is forced equal to one T
magnitude of (3.4) is given by
( j)
T
kp2
.
Ti2I T2 4 + TiI 2kp kh T TiI 2 + kp2 kh2
(3.6)
(3.7)
35
1
,
4T4 4 + 1
(3.8)
2T2 s 2
1
.
+ 2T s + 1
(3.9)
1
.
2s 2 + 2s + 1
(3.10)
2
Tn
1
s n+1
j=1 Tp j
+ 2s 2 + 2s
+1
+ +
2
T2
n
3
(3.11)
where s = sT . Comparing (3.9) with (3.11), it becomes apparent that in the
approximate design, the terms of order higher than s 2 are being neglected in the
denominator polynomial.
However, these terms have negligible effect on the dynamic behavior of the control
loop, because their coefficients are small (they are divided by a power of the closed
loop sum time constant T of higher order). Therefore, the two systems exhibit
almost the same dynamic behavior. The accuracy of the approximation depends on
the distribution of the plant time constants Tp j , j = 1, 2, . . . , n. In cases where ratio
T
= Tp1 0, the accuracy is especially satisfactory both in the time and frequency
domain, Fig. 3.2.
Figure 3.2a presents the step response of the exact and approximate closed loop
system to the reference input r (s) and to the output disturbance do (s), for two extreme
distributions of the plant time constants ( = 0.3 and = 0.9). The coincidence of
the two responses is especially satisfactory, despite the fact that the determination of
36
(a)
y r ( )
= T p 1 / T
= t/ T
(b)
S ( ju )
 F( ju )
= 0.9
= 0.3
= T p 1 / T
u = T
parameters Ti I and kh was based on a rough plant model. Figure 3.2b, presents the
closed loop transfer and output sensitivity frequency responses of the exact and the
approximate systems, for the two extreme distributions of the plant time constants
( = 0.3 and = 0.9).
With respect to the above analysis, it is concluded that by using a rough model of
the plant and applying only integral control through the conventional design method
via the Magnitude Optimum criterion, a closed loop system with satisfactory response
results. The features of these response are listed below.
Mean rise time tr = 4.40T (4.7T for 0.9 and 4.1T for = 0.3).
Mean settling time tss = 7.86T (8.40T for 0.9 and 7.32T for = 0.3).
Mean overshoot 4.47 % (4.32 % for 0.9 and 4.62 % for = 0.3).
Gain margin m = 205 db.
Phase margin m = 65.27 .
37
3.2.3 PI Control
In cases now where the dominant time constant Tp1 of the plant (conventional design
method via the Magnitude Optimum criterion) is evaluated, an approximate transfer
function of (3.1) is defined by
=
G(s)
1
,
(1 + sTp1 )(1 + sT1p )
(3.12)
where
T1p =
Tpi
(3.13)
i=2
stands for the parasitic time constant of the plant. Since the plant has a dominant
time constant, by imposing PI control through the controller
C(s) =
1 + sTn
,
sTiPI (1 + sTc )
(3.14)
kp (1 + sTn )
.
sTiPI (1 + sTp1 )(1 + sT1 ) + kh kp (1 + sTn )
(3.15)
Note again that for the derivation of (3.15), Tc T1p and T1 = T1p + Tc =
T Tp1 has been set.
According to the conventional Magnitude Optimum criterion design, for determining the zero Tn of the PI controller, polezero cancellation between the processs
dominant time constant Tp1 and the controllers zero Tn has to take place. To this end
Tn = Tp1
(3.16)
(3.17)
kp
.
s 2 TiPI T1 + sTiPI + kh kp
(3.18)
(s) =
T
which yields
(s) =
T
38
kh = 1
(3.19)
is set. Note that T1 = T Tp1 and therefore the final PI control law is equal to
kh = 1,
Tn = Tp1 ,
TiPI = 2kp kh T1
= 2kp kh (T Tp1 )
= 2kp kh (T Tn ).
(3.20)
(3.21)
(3.22)
(3.23)
Let it be noted that for the derivation of the control law, (3.20)(3.22), exact polezero cancellation has been assumed (conventional design method via the Magnitude
Optimum criterion) (3.21). Substituting (3.20)(3.22) into (3.15) results in
(s) =
T
2T21 s 2
1
.
+ 2T1 s + 1
(3.24)
1
.
2s 2 + 2s + 1
(3.25)
Comparing (3.25) with (3.10), it is concluded that with the application of PI control
via the conventional design of the Magnitude Optimum criterion, a closed loop system
with time and frequency response of the same shape results.
However, the response of (3.25) is faster than of (3.10), because the time scale
(T1 < T ) is smaller. In other words, the compensation of the dominant time
constant Tp1 has left the shape of the system time and frequency responses unaltered
and produced only a change both in the time and frequency scale, respectively.
1
,
(1 + sTp1 )(1 + sTp2 )(1 + sT2p )
(3.26)
where
T2p =
n
i=3
Tpi
(3.27)
39
represents the parasitic time constant of the plant. Since the plant has now two
dominant time constants, the PID control law defined by
C (s) =
(1 + sTn ) (1 + sTv )
sTiPID 1 + sTc
(3.28)
is imposed to (3.28). Assuming that Tc T2 p and T2 = T2 p + Tc = T
Tp1 Tp2 , the transfer function of the closed loop control system is equal to
kp (1 + sTn )(1 + sTv )
.
sTiPID (1 + sTp1 )(1 + sTp2 )(1 + sT2 ) + kh kp (1 + sTn )(1 + sTv )
(3.29)
According to the conventional Magnitude Optimum criterion design, for determining
the zeros Tn , Tv of the PID controller, polezero cancellation between the processs
dominant time constants Tp1 , Tp2 and the controller zero Tn , Tv has to take place.
To this end
(s) =
T
Tn = Tp1
(3.30)
Tv = Tp2
(3.31)
(3.32)
kp
.
s 2 TiPID T2 + sTiPID + kh kp
(3.33)
(s) =
T
or
(s) =
T
(3.34)
Note that T2 = T Tp1 Tp2 , and therefore the final PID control law is equal to
kh = 1,
Tn = Tp1 ,
Tv = Tp2 ,
TiPID = 2kp kh T2
= 2kp kh (T Tp1 Tp2 )
= 2kp kh (T Tn Tv ).
(3.35)
(3.36)
(3.37)
(3.38)
(3.39)
After substituting the control law given by (3.35)(3.38) into (3.29), results in
40
(s) =
T
2T22 s 2
1
.
+ 2T2 s + 1
(3.40)
1
2s 2
+ 2s + 1
(3.41)
Comparing (3.41) with (3.25) and (3.10) it is concluded that with the application of
PID control, a closed loop system with time and frequency responses of the same
shape results, but with even smaller time scale (T2 < T1 < T ) and consequently
even faster (Fig. 3.3).
Fig. 3.3 a Step response.
b Frequency response.
Comparison study of the step
and frequency response of the
closed loop control system
defined by (3.10), (3.25) and
(3.41), respectively. The
is
approximate process G(s)
controlled by I, PI, PID
control action through the
conventional tuning
(a)
y r ( )
Icontrol
PIcontrol
y o ( )
PID control
= t/ T
(b)
S ( ju )
 T ( ju )
PID control
PI control
I control
u = T
41
kp
.
(1 + sTp1 )(1 + sTp2 ) (1 + sTpn2 ) (1 + sTpn )
(3.42)
sn
s = s
Tp j = sT
j=1
(3.44)
42
G(s ) =
kp
n
3 1
i=1 Tp j + + s T 3
j=k=l=1 Tp j Tpk Tpl
2 1 n
+ s T2
j=k=1 Tp j Tpk + s + 1
s n T1n
(3.45)
From (3.45) it is apparent that the higher order terms of s are divided by Tn where
n = 0, 1, 2 . . . At this point, let it be noted that in principle, the value of the sum
time constant T is relatively high.
With respect to the above, it can be concluded that higher order type systems,
can under certain circumstances, be approximated by low order systems. The error
of this approximation lies in the distribution of the time constants of the process
itself. Obviously, the worst case takes place in case the time constants of the plant
are equally distributed. For example, in case of a process with five equal dominant
time constants given by
G(s) =
kp
(1 + s)
1 + 5s
+ 10s 2
kp
+ 10s 3 + 5s 4 + s 5
(3.46)
kp
1 5
5 4
10 3 10 2
s +
s +
s + s + s + 1
3125
625
125
25
(3.47)
or
G(s ) =
kp
3.2 104 s 5 + 8 103 s 4 + 8 102 s 3 + 4 101 s 2 + s + 1
(3.48)
which can be easily controlled by a PID controller tuning according to the method
described in Sect. 3.2.
G(s) =
s m m + s m1 m1 + + s 2 2 + s1 + 1 sTd
e
s n1 n1 + + s 3 3 + s 2 2 + s1 + 1
43
(3.49)
where n 1 > m. The proposed PIDtype controller is given by the flexible form
C(s) =
1 + s X + s2Y
sTi (1 + sTpn )
(3.50)
allowing its zeros to become conjugate complex. Time constant Tpn stands for the
unmodeled controller dynamics coming from the controllers implementation.
According to Fig. 3.1, the closed loop transfer function T (s) is given by
T (s) =
kp C(s)G p (s)
N (s)
N (s)
=
=
.
1 + kh kp C(s)G p (s)
D(s)
D1 (s) + kh N (s)
(3.51)
m
(s i i ),
(3.52)
i=0
n
(s j j )
(3.53)
j=0
m
(s i z i )
(3.54)
i=0
D1 (s ) = s ti es d
n
(s j r j )
(3.55)
j=0
respectively. The corresponding normalized terms involved in the control loop are
given by
x=
ri =
i
c1i
X
,
c1
y=
Y
Ti
Td
, ti = , d =
,
c1
c1
c12
, i = 1, 2, . . . , n, z j =
j
j
c1
, j = 1, 2, . . . , m.
The normalized time delay constant d is substituted with the all pole series approximation
44
es d =
1 k k
k! s d
= 1 + sd +
1 2 2
2! s d
1 3 3
3! s d
1 4 4
4! s d
1 5 5
5! s d
k=0
(3.56)
By substituting (3.50) into (3.55), D1 s becomes
k
i
D1 s =
(ti s q(i1) ), q0 = 1,
(3.57)
i=1
where
qk =
k
1 i
r(ki)
d , k = 0, 1, 2, . . . , n, r0 = 1
i!
(3.58)
i=0
or
1
r1 + d
q0
1 2
q1
r2 + r1 d + d
2!
q2
1 2
1 3
r3 + r2 d + d r1 + d
q3
2!
3!
=
.
q4
1 2
1 3
1 4
d
d
d
r
+
r
d
+
r
+
r
+
4
3
2
1
q5
2!
3!
4!
1
1
1
1
r + r d + d 2r + d 3r + d 4r + d 5
..
4
3
2
1
5
.
2!
3!
4!
5!
..
.
(3.59)
n
s i kp (z (i) + z (i1) x + z (i2) y) ,
(3.60)
i=0
k
j
s ti q( j1) + kp kh z ( j) + z ( j1) x + z ( j2) y
D s =
(3.61)
j=0
N (s )
D(s )
=
k
j=0
i
s kp z (i) + z (i1) x + z (i2) y
.
i=0
j
s ti q( j1) + kp kh z ( j) + z ( j1) x + z ( j2) y
n
(3.62)
45
The problem to be solved now for determining the optimal control law is as follows: given measured the parameters of the process kp , z i , q j , calculate controller
parameters ti , x, y, kh as a function of kp , z i , q j . For doing this, the principle of
the Magnitude Optimum criterion is adopted, which is presented in Appendix B.1.
There, a general closed loop transfer function is formulated the magnitude of which
is forced to be equal to the unity in the widest possible frequency range, T ( j) 1.
Once this is completed, a set of optimization conditions2 are derived, which comprise
the basis for proving the proposed optimal control law.
In Appendix B.1 the optimal control law is proved to be equal to
1
ti
2kp kh (q1 z 1 )
1 2kp kh 0
x = 0 1 a12
b1
0 1
a22
b2
y
(3.63)
b11
q1 z 1
,
(q1 z 1 ) q1 (q2 z 2 )
2
q1 2q2 (q1 z 1 ) + q1 z 2 q2 z 1 + q3 z 3
=
(q1 z 1 ) q1 (q2 z 2 )
a22 =
(3.64)
(3.65)
q1 z 2 q2 z 1 + q3 z 3
q22 2q1 q3 q2 z 2 + q1 z 3 + q3 z 1 + q4 z 4
(3.66)
Q0 Q1 + Q2
Q3
(3.67)
and
b22 =
and
Q 0 = q22 2q1 q3 + 2q4
(3.68)
Q 1 = q1 z 1
Q 2 = q2 z 3 q3 z 2 q1 z 4 + q4 z 1 q5 + z 5
(3.69)
(3.70)
Q 3 = q22 2q1 q3 q2 z 2 + q1 z 3 + q3 z 1 + q4 z 4 .
(3.71)
Finally, the corresponding I, PI control law can be easily derived in Table 3.1. It
is necessary to mention that the new integrators time constant is equal to ti =
2kp kh (q1 z 1 x) or finally
2
These optimization conditions are between the numerator and the denominator of the closed loop
transfer function.
46
1
1
2kp (r1 + d z 1 )
2kp (r1 + d z 1 x)
PID
2kp (r1 + d z 1 x)
b11
b22
b11
a11 + a22
n
m
Ti = 2kp kh Tpi + Td
Tzi X .
i=1
i=1
"
(3.72)
"
Another conclusion which is derived from (3.72) is that the integrators time constant
is equal to the sum o the poles of Fol (s) minus the sum of zeros of Fol (s). As a result,
necessary condition for the control loop to be controllable is Ti or
n
Tpi + Td >
i=1
Tzi X.
(3.73)
i=1
j=1 (1 + a
j1 s )
(3.74a)
G 2 (s ) =
(1 + s )2 4j=2 (1 + a j s )
47
(3.74b)
where a = 0.1. For controlling G 1 , the resulting PI controllers via the conventional
and the revised design procedure are given by,
1 + s
,
0.42s (1 + s tsc )
1 + 1.0035s
.
Crev (s ) =
0.415s (1 + s tsc )
Ccon (s ) =
(3.75a)
(3.75b)
Ccon (s ) =
(3.76a)
(3.76b)
(3.76c)
From (3.75a), (3.75b), (3.76a), (3.76b) and Fig. 3.4 it is apparent that the two control
loops, both for PI and PID control law exhibit almost the same behavior regarding
reference tracking and output disturbance rejection.
Let it be noted that the revised PID control law has led to a PID controller consisting
of conjugate complex zeros with a very close to zero imaginary part. In both cases
(PI, PID control) a step disturbance is applied in the input di (s) and the output do (s)
of the process. Disturbance rejection remains the same for both tuning methods
(conventional and revised).
1
(1 + s )5
(3.77)
Ccon (s ) =
(3.78a)
(3.78b)
48
(a)
conventional
y ( )
revised
di ( ) = 0.5r ( ) do ( ) = 0.5r ( )
PI control
= t/ T p1
(b)
revised
di ( ) = 0.5r ( )
y ( )
do ( ) = 0.5r ( )
conventional
PID control
= t/ T p1
1 + s (1.7 + 0.57i) 1 + s (1.7 0.57i)
.
=
3.34s (1 + s tsc )
(3.78c)
From Fig. 3.5a it is apparent that disturbance rejection has been improved since
tss = 110 +21.9 to tss = 110 +10.6 (51.6 % decrease) when the PID controller
is tuned via the revised method.
Robustness of the control loop has been increased, since in the frequency domain,
Srev ( ju) < Scon ( ju) holds by in the lower frequency region, Fig. 3.5b. The cost
of this improvement is paid in the overshoot of the output where there has been an
increase from 4.65 to 8.07 %. Once more, the revised PID type controller involves
conjugate complex zeros in its transfer function.
49
(a)
PID control
do ( ) = 0.5r ( )
t ss = 21.9
ovs = 4.65%
di ( ) = 0.5r ( )
conventional
y ( )
S(s) = dy(s)
. Control of a
o (s)
process with five dominant
time constants defined by
(3.77). Comparison between
the conventional and the
revised Magnitude Optimum
criterion
revised
ovs = 8.07%
t ss = 10.6
y ( )
= t/ T p1
(b)
 T ( ju )
S ( ju )
revised
conventional
u = T p1
1
(1 + s )5
e4s .
(3.79)
The conventional and the revised PID tuning via the Magnitude Optimum criterion
has led to
Ccon (s ) =
(1 + s )(1 + s )
,
14.2s (1 + s tsc )
(3.80a)
50
(a)
PID control
do ( )= 0.5r ( )
ovs = 6.23%
revised
y ( )
S(s) = dy(s)
. Control of a
o (s)
process with long time delay
defined by (3.79).
Comparison between the
conventional and the revised
Magnitude Optimum criterion
di ( )= 0.5r ( )
conventional
ovs = 4.24%
= t/ T p1
(b)
 T ( ju )
S ( ju )
revised
conventional
u = T p1
1 + 5.08s + 9.22s 2
8.02s (1 + s tsc )
1 + s (2.5 + 1.66i) 1 + s (2.5 1.66i)
=
8.02s (1 + s tsc )
Crev (s ) =
(3.80b)
(3.80c)
respectively. The revised PID controller involves conjugate complex zeros while
disturbance rejection has been improved (tss = 180 + 44.6 180 + 23.4 )
up to (47.5 % decrease) compared to the standard design, Fig. 3.6a. Let it be noted
that Trev ( ju) > Tcon ( ju) holds for a wider band in the lower frequency region as
well, Fig. 3.6b.
51
(3.81)
The respective PID controllers via the conventional and the revised method are
defined by
(1 + s )(1 + s )
,
9.4s (1 + s tsc )
1 + 3.77s + 4.04s 2
Crev (s ) =
5.85s (1 + s tsc )
1 + s (1.88 + 0.7i) 1 + s (1.88 0.7i)
=
.
5.85s (1 + s tsc )
Ccon (s ) =
(3.82a)
(3.82b)
(3.82c)
The resulting step and frequency responses in terms of disturbance rejection show an
improvement of up to 50.6 % as far as tss is concerned, Fig. 3.7a. Robustness of the
control loop has also been improved since output sensitivity Srev ( ju) < Scon ( ju)
holds by, in the lower frequency region, Fig. 3.7b.
(3.83)
In that case, there is a loss of controllability both for the revised PI and PID type
control law, Fig. 3.8. This is due to the fact that the integral gain becomes negative
since large zeros are involved in the process (3.83). This is justified by taking into
account that the revised definition of the integral gain is given by
Ti = 2kp kh
Tpi + Td
i=1
Tzi X ,
i=1
i=1
(3.84)
52
(a)
ovs = 6.08%
ovs = 4.14%
y r ( )
conventional
PIDcontrol
revised
y o ( )
t ss = 14.7
t ss = 29.8
= t/ T p1
(b)
 T ( ju )
S ( ju )
revised
conventional
u = T p1
More specifically in the case of PI control the integral gain is ti = 3.4286 and in
the case of PID control law the integral gain is ti = 2.7379. Note that only I control
leads to a stable but still oscillatory control loop, Fig. 3.8. In order to overcome that
obstacle, PI and PID control are turned into PIlag and PIDlag, respectively, by
adding a lag time constant tx in the initial PI, PID controller so that Ti becomes
positive again, Ti > 0. By choosing a lag time constant tx = 5 (Tx = 5Tp1 ) results in
CrevPI (s ) =
1
1 + sx
,
6.5714s (1 + s tsc ) (1 + 5s )
CrevPID (s ) =
1
1 + s x + s2 y
.
7.2621s (1 + s tsc ) (1 + 5s )
(3.85)
53
I control
d o ( )= 0.5r ( )
d i ( )= 0.5r ( )
revised tuning
= t / T p1
n
i=1
Tpi + Td + Tx
Tzi X
(3.86)
i=1
while the optimal solutions for X, Y remain the same. The resulting step responses
are shown in Fig. 3.8. In conclusion, the revised design procedure can overcome the
obstacle of large zeros in a process by turning the PI or PID control law into PI or
PIDlag respectively.
G (s) = 5
(3.87)
G s = 5
tpj
Tpi
Tp1 ,
i = 1, . . . , 5, ti = TTpi and x =
1
= a ( j1) , j = 1, . . . , 5 into G s we obtain
where t pi =
X
Tp1 .
(3.88)
Substituting again by
54
1
1 + sx
.
, C s =
( j1)
s ti 1 + s t p 6
j=1 1 + s a
G s = 5
(3.89)
If a < 0.3 the resulting process consists of a relatively large time constant whereas
if a > 0.8 the process consists of relatively equivalent dominant time constants.
The optimal PI control law proved in Sect. 3.3 results in x = b11 . Since the class of
processes a does not contain any zeros, z i = 0, i = 1, 2, . . . , m, it is concluded
from (3.63) that,
x=
(3.90)
q12 q2
where the qi coefficients are defined in (3.59). Rolling back the qi coefficients,
results in
1
1 i
d = r1 + r0 d,
r(1i)
q1 =
i!
(3.91)
2
1 i
1
d = r(2) + r1 d + d 2 r0 ,
r(1i)
i!
2!
(3.92)
3
1 i
1
1
d = r3 + r2 d + d 2 r1 + d 3 r0 .
r(3i)
i!
2!
3!
(3.93)
i=0
q2 =
i=0
q3 =
i=0
(3.94)
r12 r2
6
6
t pi , r2 =
t pi t p j , r3 =
i=1
i= j=1
t pi t p j t pk .
i= j=k=1
From Fig. 3.9 it is evident that the revised design method via the Magnitude Optimum
criterion, also drives the optimal PI controller parameter x to polezero cancellation
[2] only in case when the process contains one dominant time constant, Sect. 3.4.1.
The same result can be proved also for the PID controller, Sect. 3.4.1.
Hence, in cases where the process contains only one or two dominant time constants the revised PI, PID control law leads to polezero cancellation, respectively.
In any other case, neither the PI nor the PID controller tuning through the revised
55
t sc = 0.1
t sc = 0.01
1
y(s)
=
do (s)
1 + kp kh C(s)G(s)
(3.95)
y(s)
= kp G(s)So (s)
di (s)
(3.96)
u(s)
= kh So (s)C(s).
do (s)
(3.97)
N ydo (s)
ydo (s)
=
,
do (s)
D ydo (s)
(3.98)
56
respectively, where
N ydo (s) = 2s
Tpi Tn Tv
i=1
%
D ydo (s) = 2s
n
$
(1 + sTpi ),
(3.99)
i=1
Tpi Tn Tv
i=1
n
$
&
(1 + sTpi ) + 1
i=3
(3.100)
Additionally, we have
Si (s) =
N ydi (s)
ydi (s)
=
di (s)
D ydi (s)
(3.101)
where
N ydi (s) = 2kp s
Tpi Tn Tv ,
(3.102)
i=1
%
D ydi (s) = 2s
Tpi Tn Tv
i=1
n
$
&
(1 + sTpi ) + 1
i=3
(3.103)
Finally,
Su (s) =
Nu do (s)
u do (s)
=
do (s)
Du do (s)
n
$
(3.104)
(1 + sTpi )
(3.105)
i=1
%
Du do (s) = kp 2s
Tpi Tn Tv
i=1
&
n
$
(1 + sTpi ) + 1
i=3
(3.106)
According to [2], polezero cancellation may lead to poor rejection of load and
input disturbances [12], if the compensated modes are excited by disturbances, espe
57
cially if they are slow compared to the dominant closedloop poles. K.J. strm and
T. Hgglund discovered the above drawbacks of the polezero cancellation by examining the tuning method of [13] and extended their conclusion to other methods such
as the Internal Model Control [20], and the Magnitude Optimum design criterion [21,
22]. K.J. strm and T. Hgglund attribute the poor rejection of load disturbance on
the loss of the system controllability for the specific modes.
In (3.98), along with (3.99), and (3.100) it is observed that indeed, there is a polezero cancellation for the compensated time constants and the loss of controllability
is possibly justified. On the contrary, as observed in (3.101), in the case of input
disturbances a polezero cancellation does not occur. Therefore, in this case the
loss of controllability is not justified. For the verification of the correctness of this
belief, let us examine the sensitivity functions of the closed loop system, by imposing
disturbances of the form (3.107),
di (s) = do (s) =
Tp j
1 + sTp j
j = 1, 2, . . . , n.
(3.107)
Tp j
N ydo
N1do (s)
=
D ydo (1 + sTp j )
D1do (s)
(3.108)
where
N1do (s) = 2sTp j
Tpi Tn Tv
i=1
D1do (s) = 2s
% n
$
&
(1 + sTpi )
(3.109)
i=1
&
&
% n
$
1 + sTpi + 1 1 + sTp j
(3.110)
Tpi Tn Tv
i=1
while
u do (s) =
i=1
Tp j
Nu do
N2do (s)
=
Du do (1 + sTp j )
D2do (s)
(3.111)
and
N2do (s) = Tp j (1 + sTp1 )(1 + sTp2 )
n
$
i=1
(1 + sTpi )
(3.112)
58
Fig. 3.10 a ydo ( ) response
to output disturbances. b u( )
response to output
disturbances. The disturbance
excites a canceled slow mode
(a)
T p
do = e
y do ( )
= t/ T
(b)
T p
do = e
u ( )
= t/ T
n
n
k [2s
Tpi Tn Tv
(1 + sTpi ) + 1]
p
D2do (s) =
(1 + sTp j ). (3.113)
i=1
i=3
59
(a)
T p
do = e
y do ( )
= t/ T
(b)
T p
do = e
u ( )
= t/ T
As observed in Figs. 3.12b, and 3.13b if we violate the optimal control law through
incorrect tuning of the integration time constant, the control input u( ) is kept almost
constant and consequently the system appears as uncontrollable. Let it be noted that
this behavior appears independently of the excitation of a compensated or uncompensated mode. These results seem to agree with the statement of K.J. strm and
T. Hgglund, that the attenuation of load disturbance is improved considerably by
reducing the integral time of the controller [2].
60
(a)
T
T p
do = e
Ti = 2k p k h T
Ti = 2k p k h ( T T p1 )
y do ( )
Ti = 2k p k h ( T T p1 T p2 )
= t/ T
(b)
T
T p
do = e
Ti = 2k p k h ( T T p1 )
Ti = 2k p k h T
Ti = 2k p k h ( T T p1 T p2 )
u ( )
= t/ T
N ydi
Tp j
D ydi (1 + sTp j )
N3di (s)
D3di (s)
(3.114)
where
N3di (s) = 2kp Tp j (
n
i=1
Tpi Tn Tv )s
(3.115)
61
(a)
T
T p
do = e
Ti = 2k p k h T
Ti = 2k p k h ( T T p1 )
y do ( )
Ti = 2k p k h ( T T p1 T p2 )
= t/ T
(b)
T p
do = e
Ti = 2k p k h T
Ti = 2k p k h ( T T p1 T p2 )
u ( )
Ti = 2k p k h ( T T p1 )
= t/ T
%
D3di (s) = 2s
n
i=1
Tpi Tn Tv
n
$
&
(1 + sTpi ) + 1
i=3
(3.116)
As shown in Fig. 3.14a, b the rejection of the input disturbance is not poor, whether
the disturbance excites the compensated dominant time constant Tp1 , or the uncompensated time constant Tp3 respectively.
On the contrary, poor attenuation of the disturbance occurs again, when the integration time constant has not been properly tuned, as shown in Fig. 3.15a, b. Again the
attenuation of load disturbances is improved considerably by reducing the integral
time of the controller [2]. From the analysis presented in that section, it is evident
that the polezero cancellation does not lead to poor disturbances rejection. On the
contrary, poor disturbance rejection is caused by incorrect tuning of the integration
62
Fig. 3.14 a The disturbance
excites a canceled slow mode.
b The disturbance excites an
uncanceled fast mode.
Response to input disturbance
(a)
do = e
T
Tp
3
T i = 2k p k h T
y do ( )
T i = 2k p k h ( T T p 1 )
y do ( )
T i = 2k p k h ( T T p 1 T p 2 )
= t / T
(b)
di = e
T
Tp
3
y di ( )
= t / T
time constant of the PID controller and not by the polezero cancellation tuning
method.
63
(a)
di = e
T
Tp
1
T i = 2k p k h T
T i = 2k p k h ( T T p 1 )
T i = 2k p k h ( T T p 1 T p 2 )
y di ( )
= t / T
(b)
di = e
T
Tp
3
T i = 2k p k h T
T i = 2k p k h ( T T p 1 )
T i = 2k p k h ( T T p 1 T p2 )
y ( )
y di ( )
= t / T
C(s) =
(3.117)
n
where Tn 0 , Tv0 , T20 = i=1
Tpi Tn 0 Tv0 , kp0 , kh0 , are the nominal values of
the system parameters, see (3.35)(3.38). Using (3.117), the closed loop transfer
function (3.29) obtains the form,
T (s) = 
(3.118)
64
2T22 s 2
0
1
.
+ 2T20 s + 1 + a
(3.119)
2T22 s 2
0
1+b
.
+ 2T20 s + 1 + b
(3.120)
Figure 3.16, shows the step responses of the nominal (b = 0) and modified system.
It is apparent that variations of parameter kp cause variations on the overshoot, but
for variation up to 20 %, the settling time remains practically unchanged. Moreover,
from Fig. 3.16 it becomes obvious that a variation less than 10 % in kp does not have
a significant effect on the system response.
Therefore, system response cannot be considered unacceptable while modifications to parameter kp take place. Let it be noted that in vector controlled induction
motor drives and when carrier based modulation methods are adopted, kp stands for
the pulse width modulator gain when the modulator is modulating in its linear region.
Variations of kp can take place in cases when modulation enters the so called nonlinear region.
a3
s3
sb1 + b0
+ a2 s 2 + a1 s + a0
(3.121)
65
(a)
k p < k p0
k p = k p0
y ( )
k p > k p0
= 0.1
= t/ T 2
(b)
k p < k p0
k p = k p0
y ( )
k p > k p0
= 0.1
= t/ T 2
a3 = 2(1 + c)
a1 = 2 +
Tp10
T20
Tp10
T20
%
T32 ,
0
a2 = 2 1 +
T20 , b1 =
Tp10
T20
Tp10
T20
&
(1 + c) T22
T20 .
From the step responses of the nominal (c = 0) and modified system, presented
in Fig. 3.17a, it is concluded that the variation of the dominant time constant Tp1
manifests with a variation of the overshoot and mainly of the settling time. Moreover,
as shown in Fig. 3.17b, variations of the dominant time constant Tp1 less than 10 %
have no effect on the system response.
66
Fig. 3.17 a Dominant time
constant Tp1 varies up to
20 %. b Dominant time
constant Tp1 varies up to
30 %. Effect of variation of
parameter Tp1 affect the rise
and settling time of the
optimal closed loop control
system
(a)
T p1 > T p1
T p1 = T p1
y ( )
T p1 < T p1
c = 0.2
= t / T 2
(b)
T p1 = T p1
T p1 > T p1
y ( )
T p1 < T p1
c = 0.3
= t / T 2
3.5 Performance Comparison Between Revised PID Tuning and Other Methods
n r ( s)
r ( s)
e( s)
+
q( s, f )
do ( s)
p( s)
controller
u ( s)
67
kp
G ( s)
+

p( s)
y f ( s)
y ( s)
+
+
n o ( s)
Fig. 3.18 The internal model control (IMC) principle. G(s) is the plant transfer function, C(s) is
the controller transfer function, r (s) is the reference signal, do (s) and di (s) are the output and input
disturbance signals respectively and n r (s), n o (s) are the noise signals at the reference input and
process output respectively. kp stands for the plants DC gain and kh is the feedback path.
p (s) is
the approximated model of kp G(s) coming out of an open loop experiment, measurements etc
methods are used for regulating the same process and three curves are presented in
each figure, in Sect. 3.5.3.
q(s)
u(s)
=
.
e(s)
1 q(s)
p (s)
(3.122)
y(s)
p(s)C(s)
p(s)q(s)
.
=
=
r (s)
1 + p(s)C(s)
1 + q(s) ( p(s)
p (s))
(3.123)
68
r (s)
p(s)
controller
e(s)
u (s)
q(s, f )
+
G (s)
kp
C (s)=
p(s)
do (s)
+
+
y (s)
u (s)
e(s)
y f (s)
+
+
n o (s)
1
y(s)
=
.
do (s)
1 + p(s)C(s)
(3.124)
C(s)
u(s)
=
= So (s)C(s).
do (s)
1 + p(s)C(s)
(3.125)
p(s)q(s)
r (s),
1 + ( p(s)
p (s)) q(s)
(3.126)
y(s) =
1 p(s)q(s)
do (s).
1 + ( p(s)
p (s)) q(s)
(3.127)
and
According to [5, 20], goal of the ideal control action is to make the output y(s) of
the control loop track perfectly its reference signal r (s) and suppress perfectly
output disturbances. Those two goals can be interpreted mathematically by
T (s) =
So (s) =
y(s)
=1
r (s)
(3.128)
y(s)
= 0.
do (s)
(3.129)
3.5 Performance Comparison Between Revised PID Tuning and Other Methods
69
(3.130)
(3.131)
(3.132)
1
esTd
1 + sTp1
(3.133)
1
the inverse kp G(s)
transfer function is given by
1
1 + sTp1 sTd
=
e .
kp G(s)
kp
(3.134)
1 + sTp1 1
kp
1 + sf
(3.135)
for which parameter f is chosen such that modeling errors in the approximated
model are corrected. In the general case where the real and approximated plant
transfer function are defined by
N (s) sTd
e
D(s)
(3.136)
= N (s) esTd
G(s)
D (s)
(3.137)
G(s) =
and
respectively, the proposed controller according to [5, 20] depends on the characteristics of polynomial N (s). In this case, the controllers transfer function is given
by
70
kp
L
a
A
B
G: step response
t
CIMC (s) =
1
D (s)
.
N (s) (1 + f s)r
(3.138)
Parameter r is named with the term relative order and is equal to order of D (s)
minus the order of N (s).
(3.139)
3.5 Performance Comparison Between Revised PID Tuning and Other Methods
71
where b is equal to
b = y1 t1
(3.140)
From (3.139), the values of a, L can be easily calculated. These two values are used
for determining the PID controller parameters according to Table 3.2.
(3.141)
(1 + s )5
1
.
(1 + s )(1 + 0.9s )(1 + 0.88s )(1 + 0.7s )(1 + 0.68s )
(3.142)
In this case, the PID controller regarding the revised Magnitude Optimum method
and the ZieglerNichols step response method are tuned based on (3.142) whereas
the resulting control law is applied to the real process (3.141).
Table 3.2 PID tuning
formulas based on the
ZieglerNichols step response
method
Controller
I
PI
PID
kh
Ti
Td
1
1
1
1
a
0.9
a
1.2
a
3L
2L
L
2
72
Fig. 3.21 a Step response of
the control loop. Input and
output disturbance di (s) and
do (s) are applied at t = 100
and t = 200 respectively.
b Response of the command
signal at the presence of input
disturbance di ( ). Control of
a process with five dominant
time constants defined by
(3.141). PID control action:
comparison between the
revised Magnitude Optimum
criterion, the Internal Model
Control principle (IMC,
f = 1.5) and the
ZieglerNichols step response
method. All PID controllers
are tuned based on the
approximate model defined
by (3.142)
(a)
do ( )= r ( )
di ( )= r ( )
ZieglerNichols
IMC
Magnitude Optimum
y ( )
PID control
= t/ T p1
(b)
di ( )= r ( )
PID control
IMC
Magnitude Optimum
u ( )
ZieglerNichols
= t/ T p1
From Fig. 3.21a it is apparent that the ZieglerNichols step response method
leads to an oscillatory step response with an undesired overshoot of 60 %, which is
resulted by the aggressive command signal as shown in Fig. 3.21b. On the contrary,
the same method shows the minimum peak value regarding the input disturbance
rejection di (s) and almost the same settling time compared with the other two methods.
The IMC tuning leads to a satisfactory overshoot of (4 %) compared with the
revised PID tuning method via the Magnitude Optimum criterion (21 %) and as far
as the step response is concerned. Finally, output disturbance rejection is considered
acceptable only in the case of IMC, since its settling time and undershoot exhibit the
minimum values with respect to the other two methods, see Fig. 3.22.
By selecting a different time constant f = 0.5 in the IMC tuning principle,
the step response of the control loop becomes faster and so does input and output
disturbance rejection. In this case the control loop in the case of internal model control
outperforms the revised PID tuning method via the Magnitude Optimum criterion,
see Fig. 3.22.
3.5 Performance Comparison Between Revised PID Tuning and Other Methods
Fig. 3.22 a Step response of
the control loop. Input and
output disturbance di (s) and
do (s) are applied at t = 100
and t = 200 respectively.
b Response of the command
signal at the presence of input
disturbance di ( ). Control of
a process with five dominant
time constants defined by
(3.141). PID control action:
comparison between the
revised Magnitude Optimum
criterion,1 the Internal Model
Control principle (IMC,
f = 0.5) and the
ZieglerNichols step response
method. All PID controllers
are tuned based on the
approximate model defined
by (3.142)
73
(a)
di ( ) = r ( )
do ( ) = r ( )
y ( )
PID control
= t/ T p1
(b)
PID control
di ( ) = r ( )
Magnitude Optimum
u ( )
IMC
= t/ T p1
1
(1 + s )5
e4s ,
(3.143)
1
(1 + s )(1 + 0.95s )(1 + 0.8s )(1 + 0.75s )(1 + 0.7s )
e3.5s . (3.144)
In this case, the ZieglerNichols step response method gives an unstable response,
and this is why it is not depicted either in Figs. 3.23 and 3.24. In Fig. 3.23a, b the
filter time constant has been chosen equal to f = 3 and slow disturbance rejection is
74
Fig. 3.23 a Step response of
the control loop. Input and
output disturbance di (s) and
do (s) are applied at t = 100
and t = 200 , respectively.
b Response of the command
signal at the presence of input
disturbance di ( ). Control of
a process with long time
delay defined by (3.143). PID
control action: comparison
between the revised
Magnitude Optimum
criterion, the Internal Model
Control principle (IMC,
f = 3) and the
ZieglerNichols step response
method. All PID controllers
are tuned based on the
approximate model defined
by (3.144)
(a)
Magnitude Optimum
PID control
IMC
y ( )
di ( ) = r ( )
do ( ) = r ( )
= t/ T p1
(b)
PID control
di ( ) = r ( )
IMC
u ( )
Magnitude Optimum
= t/ T p1
observed both in the input di ( ) and the output do ( ) of the control loop compared
to the Magnitude Optimum PID tuning.
If the filter time constant f is reduced from f = 3 to f = 1 the step response and
disturbance rejection become faster in the case of IMC tuning, see Fig. 3.24. In this
case, the IMC tuning principle outperforms the PID tuning via the revised method
since it exhibits almost the same settling time regarding disturbance rejection, but
with less undershoot, Fig. 3.24a.
(3.145)
3.5 Performance Comparison Between Revised PID Tuning and Other Methods
Fig. 3.24 a Step response of
the control loop. Input and
output disturbance di (s) and
do (s) are applied at t = 100
and t = 200 , respectively.
b Response of the command
signal at the presence of input
disturbance di ( ). Control of
a process with long time
delay defined by (3.143). PID
control action: comparison
between the revised
Magnitude Optimum
criterion, the Internal Model
Control principle (IMC,
f = 1) and the
ZieglerNichols step response
method. All PID controllers
are tuned based on the
approximate model defined
by (3.144)
75
(a)
di ( ) = r ( )
do ( ) = r ( )
Magnitude Optimum
y ( )
IMC
PID control
= t/ T p1
(b)
di ( ) = r ( )
PID control
IMC
u ( )
Magnitude Optimum
= t/ T p1
(3.146)
In this case the PID controller via for the ZieglerNichols and the Magnitude
Optimum criterion is tuned via the (3.146) whereas and the resulting control law is
applied to (3.145).
The ZieglerNichols step response tuning method leads to an unstable control
loop and therefore is not depicted in Figs. 3.25 and 3.26.
76
Fig. 3.25 a Step response of
the control loop. Input and
output disturbance di (s) and
do (s) are applied at t = 100
and t = 200 respectively.
b Response of the command
signal at the presence of input
disturbance di ( ). Control of
a nonminimum phase process
defined by (3.146). PID
control action: comparison
between the revised
Magnitude Optimum
criterion, the Internal Model
Control principle (IMC,
f = 1) and the
ZieglerNichols step response
method. All PID controllers
are tuned based on the
approximate model defined
by (3.146)
(a)
do ( ) = r ( )
di ( ) = r ( )
Magnitude Optimum
IMC
y ( )
PID control
= t/ T p1
(b)
di ( ) = r ( )
PID control
Magnitude Optimum
IMC
u ( )
= t/ T p1
77
(a)
di ( ) = r ( )
do ( ) = r ( )
IMC
Magnitude Optimum
y ( )
PID control
= t/ T p1
(b)
u ( )
IMC
Magnitude Optimum
do ( ) = r ( )
PID control
= t/ T p1
Fig. 3.26 a Step response of the control loop. Input and output disturbance di (s) and do (s) are
applied at t = 100 and t = 200 , respectively. b Response of the command signal at the presence
of input disturbance di ( ). Control of a nonminimum phase process defined by (3.145). PID control
action: comparison between the revised Magnitude Optimum criterion, the Internal Model Control
principle (IMC, f = 2) and the ZieglerNichols step response method. All PID controllers are
tuned based on the approximate model defined by (3.146)
1 + sTc
,
sT2
(3.147)
78
Ideal source
Line impedance
Transformer model
I f eed
I ar
RM
CDC
LM
V 50Hz
DC LinkIload
V net
AC/DC
V ar
V DC
Description
Three phase voltage measured at PCCa
Pulsation of the network
Three phase voltage at the grid converter
Three phase current to grid converter
DC link voltage
Feeding current from grid converter
DC link capacitor current
Load current
V
Hz
V
A
V
A
A
A
of common coupling
Description
F
H
H
H
DC link capacitor
Leakage inductance of the transformer
Leakage resistance of the transformer
Magnetizing inductance of the transformer
Magnetizing resistance of the transformer
Equivalent line inductance of the network
Equivalent line resistance of the network
CI =
1 + sTLR
,
sT1
(3.148)
1
,
1 + sTF
(3.149)
CIF =
stand for the model of the current PI controller, where T1 , TLR are the current controllers integrator time constant, the current controllers zero to be determined and
TF stands for the unmodeled controller dynamics. Let it be noted that the analysis
of the control loop takes place in the d q reference frame. From the output of the
V DC
re f
+
CV
DC
+

voltage controller
VD
current controller
CI
ID
CI
Vnet
net
+
+
Iq
act
79
GM
kp
[ T +1 ]
dq
Iload
Iar
GT
V ar
GC
V DC
I f eed
Plant
I AR
interface
Iar
current controller, the modulation index MAR and modulation angle are constructed
out of the expressions
/
VD2
VQ2
and atan
VD
VQ
.
(3.150)
Note that VD , VQ is the output of the current PI controller within the d and q path,
respectively.
The modulator itself is modeled by first order process kp G M model, where kp
V
0
stands for the modulators gain in terms of fundamental amplitude (kp = Vout
) and
in0
Tm is the time delay introduced from the time the controller decides the command
until the final voltage is applied by the power part of the inverter.
GM =
1
.
1 + sTm
(3.151)
1
,
R + s L
(3.152)
and the capacitor bank path within the DC link of the inverter
GC =
1
sCDC
(3.153)
respectively. The DClink voltage controller provides the current reference to the
grid current controller which itself provides a reference to the modulator through the
modulation index Mar . The load current Iload is the main perturbation of the system
and a power feedforward current Ipff can be provided to the current controller for
enhancing its dynamics.
Although the description of the synchronization to the grid through a dedicated
PLL [19] is not the scope of this section, it cannot be ignored since it provides
the reference for the vector controller, Fig. 3.28. The grid voltages and currents are
described in the synchronous reference frame computed by the wellknown Park
transformation. Only the active part of the vector control is considered in the scheme
80
Q axis
N
I re f
V ar
Iqre f
Vq
V qnet
jX re f
V
V
V
V
A
A
A
A
A
A
D axis
V net
V dnet
Vd
Idre f
Description
Modulation index of grid converter AC voltage
Voltage control value (active part)
Voltage control value (reactive part)
Grid voltage measurement (active part)
Grid voltage measurement (reactive part)
Grid current measurement (active part)
Grid current measurement (reactive part)
Current feedforward from load drive
Grid current reference (active part)
Grid current reference (reactive part)
DC link voltage reference
depicted on Fig. 3.28, but the reactive part is also controlled by a reference set often
to zero or to a nonzero value when a reactive power controller is active (Fig. 3.29).
The cross coupling due to the use of a synchronous reference frame needed in the
current controller [6, 7], is also not depicted. The signals entering and generated by
the controller are summarized in Table 3.5.
Considering the optimal tuning of the vector control parameters in the synchronous
reference frame, the crosscoupling gain is a value corresponding to the estimation
of the equivalent inductor value on the grid side of the power converter [23].
A more accurate way of decoupling this effect has been presented in [4]. Instead
of crosscoupling the current components through proportional gains as in classic
current controllers, one should crosscouple the error signals through integrators as
depicted in Fig. 3.30. The parameters of the crosscoupling integrators are identical
to the PI controllers parameters and can therefore be tuned using the optimal control
action described in Sect. 3.3.
Ip f f
Idre f
Id
81
Vdnet
1+ sTLR
sT1
Vdctrl
N TLR
sT1
N TLR
sT1
Iqre f
+

1+ sTLR
sT1
Iq
Vqctrl
Vdnet
The proposed method is applied to the inner control loop (current controller)
since the analysis within this chapter is dedicated to typeI control loops.3 The outer
control loop is of typeII (two integrators in the open loop transfer function) and is
out of the scope of this chapter. For measuring the DC gain of the process, an open
loop experiment from Mar to Iar is carried out. In this case a good estimation of R
is acquired since
G(s) = G M G T =
1
1
=
(1 + sTm ) (R + s L )
R (1 + sTm ) (1 + s LR )
(3.154)
(3.155)
where Tk = LR .
The current controller step response is evaluated alone since the optimal tuning
of the DClink controller is considered in Chap. 4. As illustrated in Fig. 3.31a, the
controller is first submitted to a reference step of the current Iref , then to a perturbation
of the net voltage Vnet and of the load current Iload . One can see that the system is
completely decoupled from the voltage perturbations, this is possible only with the
use of an accurate PLL that is synchronizing the system to the grid. The behavior of
the feeding current Ifeed shows the accuracy of the proposed method.
3
Loops that track step reference signals with zero steady state error.
82
Fig. 3.31 a Reference and
load step response of the
system. b Detailed view on
step response of degraded
systems. Step response of the
system
(a)
do ( ) = 0.7r ( )
ovs = 4.47%
PI control
(b)
t (sec)
ovs = 4.47%
PI control
t (sec)
One can read on Fig. 3.31b an overshoot of some 5 % and a 7 ms response. If one
modify the time constant TM of the modulator by a factor a = 20 %, the effect on
the response time is obviously increasing its overshoot or its time response, showing
that the proposed tuning gives a satisfactory behavior to to the current control loop.
3.7 Summary
In Sect. 3.2 the conventional PID controller tuning via the Magnitude Optimum criterion was presented. It was shown that controller parameters
3.7 Summary
83
References
1. strm KJ (1995) Model uncertainty and robust control. Tech. rep., Department of Automatic
Control, Lund University, Lund, Sweden
2. strm KJ, Hagglund T (1995) PID controllers: theory, design and tuning, 2nd edn. Instrument
Society of America
3. strm KJ, Hagglund T (2004) Revisiting the ZieglerNichols step response method for PID
control. J Process Control 14(6):635650
4. Bahrani B, Kenzelmann S, Rufer A (2011) MultivariablePIbased current control of voltage source converters with superior axis decoupling capability. IEEE Trans Ind Electron
58(7):30163026
5. Brosilow C, Joseph B (2002) Techniques of modelbased control, 1st edn. PrenticeHall, New
Jersey
6. Bhler H (1979) lectronique de reglage et de commande. Dunod, Paris
7. Bhler HR (1997) Reglage des systemes delectronique de puissance, vol 1, 2 and 3, Theorie,
1st edn. PPUR: Presses Polytechniques et Universitaires romandes
8. Buxbaum A, Schierau K, Straughen A (1990) Design of control systemsfor DC drives. Springer,
Berlin
9. Courtiol B, Landau ID (1975) High speed adaptation system for controlled electrical drives.
Automatica 11(2):119127
10. Fllinger O (1994) Regelungstechnik. Hthig, Heidelberg
11. Frhr F, Orttenburger F (1982) Introduction to electronic control engineering. Siemens, Berlin
12. Goodwin GC, Graebe SF, Salgado ME (2001) Control system design. Prentice Hall, New Jersey
13. Haalman A (1965) Adjusting controllers for a dead time process. Control Engineering Practice,
pp 7173
84
14. Habetler TG (1993) A space vectorbased rectifier regulator for AC/DC/AC converters. IEEE
Trans Power Electron 8(1):3036
15. Kessler C (1955) UG ber die Vorausberechnung optimal abgestimmter regelkreise teil III. Die
optimale einstellung des reglers nach dem betragsoptimum. Regelungstechnik 3:4049
16. Kessler C (1958) Das symmetrische optimum. Regelungstechnik, pp 395400 and 432426
17. Loron L (1997) Tuning of PID controllers by the nonsymmetrical optimum method. Automatica 33(1):103107
18. Lutz H, Wendt W (1998) Taschenbuch der regelungstechnik, 1st edn. Frankfurt am Main:
Verlag, Harri Deutsch
19. Mohan N, Undeland TM, Robbind WF (1989) Power electronics: converters, applications and
design, 1st edn. Wiley, New York
20. Morari M, Zafiriou E (1989) Robust process control, 1st edn. PrenticeHall, New Jersey
21. Oldenbourg RC, Sartorius H (1954) A uniform approach to the optimum adjustment of control
loops. Trans ASME 76:12651279
22. Sartorius H (1945) Die zweckmssige festlegung der frei whlbaren regelungskonstanten.
Master thesis, Technische Hochscule, Stuttgart, Germany
23. Schauder C, Mehta H (1993) Vector analysis and control of advanced static VAr compensators.
IEE Proc Gener, Transm Distrib 140(4):299306
24. Umland WJ, Safiuddin M (1990) Magnitude and symmetric optimum criterion for the design
of linear controlsystems: what is it and how does it compare with the others? IEEE Trans Ind
Appl 26(3):489497
25. Voda AA, Landau ID (1995) A method for the autocalibration of PID controllers. Automatica
31(1):4153
26. Vrancic D, Strmcnik S (1999) Practical guidelines for tuning PID controllers by using MOMI
method. In: International symposium on industrial electronics, IEEE, vol 3, pp 11301134
27. Vrancic D, Kristiansson B, Strmcnik S (2004) Reduced MO tuning method for PID controllers.
In: 5th Asian control conference, IEEE, vol 1, pp 460465
Chapter 4
Abstract In this chapter, the explicit solution for tuning the PID controller
parameters in the presence of integrating process is presented. The presence of one
integrator coming from the plant along with one integrator coming from the PIDtype
control action results in a typeII control loop according to Sect. 2.5. The proposed
control law is developed again in the frequency domain and lies in the principle of
the symmetrical optimum criterion which, strictly speaking, is the application of the
Magnitude Optimum criterion in typeII control loops. Therefore, the desired control
action requires again that the magnitude of the closedloop transfer function is equal
to the unity in the widest possible frequency range. For the proof of the control law,
a general transfer function process model is adopted consisting of n poles, m zeros
plus unknown time delay d. The final solution determines explicitly the P, I, and
D parameters as a function of all time constants involved within the control loop
and irrespective of the process complexity. The potential of the proposed method
is tested both (1) on benchmark process models (integrating process with dominant
time constants, integrating nonminimum phase process, integrating process with
long time delay). The proposed control action is tested also for the control of the
actual DC link voltage in an AC/DC grid connected converter. In all cases, an extensive comparison test is presented between the conventional current stateoftheart
PID tuning and the proposed control law, justifying the potential of the proposed
method.
4.1 Introduction
In the literature, the demanding problem of controlling integrating processes has
driven many researchers at employing or modifying wellestablished control techniques [17, 22], such as the Smith predictor, see [1, 7, 14, 18, 21] and the internal
model control (IMC) principle [20]. More specifically, in [1, 14], an extension of
Watanabes Smith predictor is proposed where for its tuning an accurate estimation
of the input disturbance is required [1]. The proposed method involves adjustable
tuning and not explicit solution for the controllers parameters, whereas in [7, 21],
85
86
the proposed modified Smith predictor restricts its focus on controlling integrating
processes with long dead time.
The control loops where integrating processes are involved are also called in the
literature as typeII control loops [11]. The basic advantage of such control loops is the
ability the output variable of the control loop exhibits, to track perfectly step and ramp
reference signals with zero steady state position and velocity error respectively. On
a theoretical basis and if frequency domain modeling is followed for the controllers
design, typeII control loops are also characterized by the presence of two pure
integrators within the openloop transfer function.
On a practical basis, industrial examples of this case arise frequently in the area of
AC/DC/AC power converters and drive systems in principle. Representative industry
applications where the aforementioned AC/DC/AC configuration is met is (1) a
wind energy conversion system [8], (2) a shaft generator system [6, 9], and (3) an
AC/DC/AC arrangement operating in motoring mode.1
Taking the case of a shaft generator system as an example, the AC/DC/AC configuration has to operate often in island network mode so that the vessels efficiency
is improved.2 Island network mode means that the grid side converter has to deliver
to the grid3 the required AC signal of certain amplitude and certain frequency, given
constant DC link between the two converters. In this case, constant DC link is guaranteed by the outer DC link voltage control loop of the shaft side converter which
takes the energy from the shaft generator.
Within the shaft side converter and from the control point of view, the resulting
control loop of the DC link voltage is proved to be of typeII, since one integrator comes from the capacitor bank of the DC link, whereas the other integrator is
coming from the PIDtype controller itself. Last but not least, in the case of the
motoring operation of an AC/DC/AC configuration system, the actual DC link voltage is regulated at a constant level by the grid connected converter which runs normally under a vector control scheme [2, 3]. In this case, there is again an inner
current control loop and an outer voltage control loop (for regulating the actual DC
link voltage) which afterwards is used by the motor side converter for driving the
machine.
Motivated by such practical industrial problems, the purpose of this chapter is
to provide control engineers with explicit tuning rules for the PID controller and
irrespective of the complexity of the integrating process so that robust performance
can be achieved by the output of a typeII control loop. To this end, development and
control engineers are provided with an explicit solution which
1. allows for accurate investigation of the robustness of the controller to possible
model uncertainties within the whole control loop;
1 In this case, a grid connected converter controls the DC Link which is then used by the motor side
converter which finally drives the motor.
2 In the case of the island network, auxiliary small diesel generators are completely switched off
since they are consuming expensive oil, and the energy is coming from the main diesel engine of
the ship which drives the propeller.
3 Grid of the vessel supporting the electrical load of the vessel.
4.1 Introduction
87
2. leads to reliable results before integrating finally the whole control law on a
realtime embedded system;
3. prevents onsite commissioning and service engineers from using heuristic tuning
rules which most of the times lead to poor performance of the drive itself, as far
as the field of power converters is concerned.
In order to develop the aforementioned control theory, the principle of the
Magnitude Optimum criterion is adopted [10, 16], see Appendix A.1. Oldenbourg
and Sartorius applied the Magnitude Optimum criterion in typeI systems and in
succession, Kessler suggested the symmetrical optimum criterion [4, 5, 19] which
in reality is the application of the Magnitude Optimum criterion to typeII control
systems. In this chapter, aim of the proposed theory is to revise thoroughly the current
stateoftheart in PID tuning via the symmetrical optimum criterion by pointing out
its drawbacks and improving them by
1. suggesting closedform expressions for the PID controllers parameters and
2. achieving robust and optimal performance of the control loop both in reference
tracking and disturbance rejection.
For the reasons above, and for the sake of a clear presentation of the proposed explicit
solution, the sections of this chapter are organized as follows. In Sect. 4.2, a short
presentation is given to the reader about the current stateoftheart relevant to the PID
tuning via the symmetrical optimum criterion. Its drawbacks are pointed out, which
are basically related to (1) the simple and poor process model used till date to adopt
the conventional PID tuning and (2) the polezero cancellation principle the current
stateoftheart method uses. Therefore, the proposed theory introduces a transfer
function of integrating behavior consisting of n poles, m zeros plus unknown time
delay d. Irrespective of the order of n, m, and d an explicit solution of the proposed
control law is presented within the same section without using the principle of polezero cancellation.
The proof of the control law lies in the wellknown Magnitude Optimum criterion
which is presented in the Appendix B.2. Therefore, in Sect. 4.4, we apply the theoretical modeling approach on five benchmark transfer function process models. The
proposed control law is also tested finally within the DC link voltage control path
on an AC/DC arrangement, see Sect. 4.5. The AC/DC configuration is presented on
system and closedloop control system level, and the control loop of actual DC link
is presented in the frequency domain. Controller performance is investigated in the
presence of output disturbances which in this case is the load current coming from
the inverter which in principle drives the electric motor.
In all examples, the proposed method is compared with the conventional stateoftheart PID tuning via the Magnitude Optimum criterion in terms of step and ramp
reference signals. Within this comparison, the output of the control loop along with
the command signal of the controller (control effort) are also measured. Results and
conclusions are summarized in Sect. 4.6.
88
4.2.1 I Control
Let us now consider the closedloop system of Fig. 4.1, where r(s), e(s), u(s), y(s),
do (s), and di (s) are the reference input, the control error, the input and output of
the plant, the output and the input disturbances, respectively. An integrating process
found in many industry applications can be defined by (4.1)
G(s) =
1
,
Tm s(1 + Tp1 s)(1 + Tp s)
(4.1)
where Tm is the integrators plant time constant, Tp1 the plants dominant time constant and Tp the process parasitic time constant [11]. Let it be noted that such type of
modeling is frequently used in vector controlled induction motor drives. More specifically, time constant Tm stands for the mechanical subsystem of the motor which is
the mechanism that involves the electromagnetic and load torque, the difference of
which, makes the shaft rotating.
n r (s)
r (s)
+
+
e(s)
controller
C (s)
di ( s )
u (s)
+
y f (s)
S
do (s)
kp
kh
G (s)
+
+
y (s)
+
+
n o (s)
Fig. 4.1 Block diagram of the closedloop control system. G(s) is the plant transfer function, C(s)
is the controller transfer function, r(s) is the reference signal, do (s) and di (s) are the output and
input disturbance signals, respectively, and nr (s), no (s) are the noise signals at the reference input
and process output, respectively. kp stands for the plants dc gain, and kh is the feedback path
89
Furthermore, time constant Tp1 is involved in the inner current control loop of
the electrical drive and represents the stator winding time constant. Finally, Tp
stands for the motors unmodeled dynamics. If vector control.4 is to be followed
(control of induction motor drives), kp stands for the pulse width modulators gain
(kPWM ) which is supposed to remain constant all over the whole operating range
(0 1p.u) regarding output frequency.5 Parameter kh is the feedback path of the
output measurement and as it is proved in the sequel, kh should satisfy condition
kh = 1.
Back to Fig. 4.1, for controlling (4.1), the PID controller defined by
C(s) =
(1 + Tn s)(1 + Tv s)
Ti s(1 + Tc s)
(4.2)
is adopted. For its tuning, the conventional symmetrical optimum design method
is employed. Time constant Tc stands for the controllers parasitic dynamics. If
Tn = Tv = 0, I control cannot be applied, because the closedloop transfer function
becomes unstable. This is justified as follows. If for controlling (4.1), I control of the
form
C(s) =
1
Ti s(1 + Tc s)
(4.3)
kp
2
Ti Tm s (1 + Tp1 s)(1 + T s) + kh kp
(4.4)
where
Tp Tc 0, and T = Tp + Tc .
(4.5)
kp
Ti Tm Tp1 T s4 + Ti Tm (Tp1 + T )s3 + Ti Tm s2 + kh kp
(4.6)
From (4.6), it is clear that T (s) is unstable since the term of s is missing.6
90
4.2.2 PI Control
In a similar fashion, if PI control of the form
C(s) =
1 + Tn s
Ti s(1 + Tc s)
(4.7)
is employed, then for determining controller parameter Tn via the conventional symmetrical optimum criterion, polezero cancellation must take place, Tn = Tp1 . Therefore, the dominant time constant Tp1 has to be evaluated and in that case, T (s) becomes
T (s) =
Ti Tm T
s4
kp
,
+ Ti Tm T s3 + Ti Tm s2 + kh kp
(4.8)
kp Tn s + kp
.
Ti Tm T s3 + Ti Tm s2 + kh kp Tn s + kh kp
(4.9)
(4.10)
2
+ kp kh Tn 2kp kh Ti Tp1 2 + kp2 kh2 .
(4.11)
(4.12)
91
Tn T
.
Tm
(4.13)
In similar fashion by setting the term of 2 equal to zero, see [11] results in
2
kp kh Tn2 = 2kp kh Ti Tm
Ti =
(4.14)
T2
1
kp kh n .
2
Tm
(4.15)
(4.16)
T2
.
Tm
(4.17)
(4.18)
4T
Tn
=
2
Ti 8kp kh T
Tm
kh
1
(4.19)
8T3 s
1 + 4T s
+ 8T2 s + 4T s + 1
(4.20)
1 + 4s
8s 3
+ 8s 2 + 4s + 1
(4.21)
92
(a)
43.4%
8.1%
yr( )
t rt = 3.1
= t/ T
(b)
without Cex (s)
S ( ju )
 T ( ju )
with Cex (s)
Mr
u=
Fig. 4.2 TypeII closedloop control system. a The effect of the two degrees of freedom controller
to the step response of the closedloop control system. Step response (solid black), filtered step
response (dotted black). b The effect of the two degree of freedom controller to the frequency
response of the closedloop control system
The respective step and frequency response of (4.20) are shown in Fig. 4.2a, b. From
there, it is clear that the step response of the closedloop control system exhibits an
undesired overshoot of 43.4 % in the time domain Fig. 4.2a, and a peak overshoot in
the frequency domain Fig. 4.2b.
This is also justified by the openloop frequency response Fig. 4.3 where the phase
margin in the crossover frequency
c =
1
2T
(4.22)
93
 Fol ( ju )
n = 7.46
n = 4.1
u c = 0.5 u c = 1
(u )
n = 4.1
m
n = 7.46
= 35
u=
is m 35 < 45 . Note also the symmetry of the critical frequencies 4T1 , T1
exhibited by Fol (j) where its slope is equal to 1/deg around the crossover frequency c = 2T1 , Fig. 4.3. The openloop transfer function is given by
Fol (s ) =
1 + 4s
.
8s2 (1 + s )
(4.23)
In order to overcome the obstacle of 43.4 % overshoot, the reference input is filtered
of the step
by adding an external controller Cex (s), Fig. 4.4. The great overshoot
response in (4.21) is owed to the zero of the transfer function, N s = 1 + 4s . This
can be removed by including that zero as a pole in the reference filter. In that, if an
r (s)
n r (s)
di ( s )
controller
r (s)
u (s)
+
+ e(s)
C (s)
Cex (s)
+
+
y f (s)
S
do (s)
kp
kh
G (s)
+
+
y (s)
+
+
n o (s)
Fig. 4.4 Two degrees of freedom controller. Controller Cex (s) filters the reference input so that
the undesired overshoot at the output y(s) is diminished. Controller Cex (s) affects the closedloop transfer function T (s) and not the output and input disturbance transfer functions So (s) =
y(s)
y(s)
do (s) , Si (s) = di (s)
94
r (s )
1
=
r(s )
1 + 4s
(4.24)
is chosen, the overshoot decreases from 43.4 to 8.1 %. Let it be noted that the rise
time increases from trt = 3.1T to trt = 6.6T . Such dynamics, can for sure be
improved by adding additional dynamics in the reference filter.
sm m + sm1 m1 + + s1 + 1 sTd
e
s(sn1 an1 + + s3 a3 + sa1 + 1)
(4.25)
1 + sX + s2 Y
sTi2 (1 + sTpn )
(4.26)
95
where parameter Tpn stands for the parasitic controllers time constant and is considered known from the controllers implementation. Note that the flexible form
of numerator Nc (s) = 1 + sX + s2 Y allows parameters X, Y to become complex
conjugate if needed. Purpose of the following analysis is to determine analytically
controller parameters as a function of all modeled time constants within the control
loop, X = f1 (i , aj , Td ), Y = f2 (i , aj , Td ), Ti = f3 (i , aj , Td ) and in contrast to the
conventional PID tuning see, [4, 5, 10, 16, 19], polezero cancellation does not take
place. According to (4.25) and (4.26), the product C(s)G(s) is defined by
j
(1 + sX + s2 Y ) m
j=0 (s j )
C(s)G(s) =
s2 Ti2 esTd ni=0 (si pi )
(4.27)
where
n
n1
(si pi ) = (1 + sTpn ) (sj aj ).
i=0
(4.28)
j=0
kp C(s)G(s)
Ffp (s)
=
1 + Fol (s)
1 + kp kh C(s)G(s)
(4.29)
where Ffp (s), Fol (s) stand for the forward path and the openloop transfer function
respectively. Along with the aid of (4.27), T (s) becomes equal to
T (s) =
s2 Ti2 esTd
n
kp (1 + sX + s2 Y )
i=0 (s
m
j=0 (s
i p ) + k k (1 + sX
i
p h
j )
j
+ s2 Y )
m
j=0 (s
j )
j
(4.30)
Y
Ti
X
, y = 2 , ti =
c1
c1
c1
j
Td
pi
, ri = i , i = 1, . . . , n, zj = j , j = 1, . . . , m.
c1
c1
c1
(4.31)
(4.32)
es d =
1
k=0
k!
sk d k
(4.33)
96
see [15]. Substituting the normalized parameters along with the approximation of
es d into (4.29) results in
j
kp (1 + s x + s2 y) m
j=0 s zj
m
T (s ) = 2 2 s d n i
2
j
s ti e
i=0 s ri + kp kh 1 + s x+s y
j=0 s zj
(4.34)
N(s )
N(s )
=
,
D1 (s ) + kh N(s )
D(s )
(4.35)
where
N(s ) = kp (1 + s x + s2 y)
m
(sj zj )
(4.36)
j=0
and
D1 (s ) = s2 ti2
7
1
k=0
k!
(sk )d k
n
(si ri ).
(4.37)
i=0
(4.38)
(4.39)
p
r
(s )(yzr2 + xzr1 + zr )
r=0
(4.40)
97
k
p
r
+ kh kp (s ) yzr2 + xzr1 + zr .
(4.41)
r=0
j=0
According to (4.35), (4.40) and (4.41), the resulting closedloop transfer function
is given by
p
r
r=0 (s )(yzr2 + xzr1 + zr )
p
2
(j+2) + k k
r
p h
j=0 (ti qj )(s )
r=0 (s ) yzr2 + xzr1
T (s ) = k
kp
+ zr
.
(4.42)
Since (4.42) is now written in the same form of (A.1), for determining the optimal
control law the optimization conditions proved in Appendix A.1 can now be used.
Eqs. (A.9)(A.12) are used for the derivation of the optimal control law. Therefore,
the problem to be solved is formulated as follows: given known the parameters of
the plant, calculate explicitly the PID control action x, y, ti . In Appendix B.2, the
proof of the optimal control law is presented which is proved to be equal to
kh = 1,
(4.43)
b1
c1
x+
=0
a1
a1
(4.44)
y = a2 x 2 + b2 x + c2
(4.45)
x2 +
ti2 =
1
1
kp kh (x 2 2y) + kp kh (z12 2z2 )
2
2
(4.46)
where
a1 = 2 q1 (q1 z1 ) q2 + z2
b1 = 4 q13 3q12 z1 + 2q1 z12 + q1 z2 + q2 z1 q3 + z3 2z1 z2
c1 =
(4.47)
(4.48)
2
2
2
q1 2q1 z1 + 2z2 z12 + 2z2 + 4q2
4q1 z1 + q1 2q2 z1 2z2
+4 q1 z3 + q3 z1 q4 z4 q2 z2
(4.49)
98
and
1
2
b2 = 2 (q1 z1 )
1
c2 = z12 + 2z2 + 4q2 4q1 z1 ,
2
a2 =
(4.50)
(4.51)
(4.52)
1
.
(1 + s )(1 + 0.2s )(1 + 0.1s )(1 + 0.1s )(1 + 0.05s )
(4.53)
For controlling (4.53) and after applying the revised PI tuning rules the controller of
Crev (s ) =
1 + 5.73s
16.45s 2 (1 + s tsc )
(4.54)
(1 + s tn )(1 + s tv )
s 2 ti2 (1 + s tsc )
(4.55)
99
(a)
revised PI
yr( )
conventional PID revised PID
t ss = 8.23
y o( )
t ss = 22.2
t ss = 8.13
= t/ T p1
(b)
S ( ju )
 T ( ju )
revised PID
conventional PID
revised PI
u=
T p1
results in
Ccon (s ) =
(1 + 2.2s )(1 + s )
2.42s 2 (1 + s tsc )
(4.56)
1 + s x + s 2 y
s 2 ti2 (1 + s tsc )
(4.57)
100
(a)
revised PID
conventional PID
revised PI
 Fol ( ju )
33
= 47.2
m
= 34.5
(u )
u=
Tp1
(b)
revised PI
t ss = 20.22
revised PID
t ss = 6.38
u( )
t ss = 6.92
conventional PID
= t/ Tp1
1 + 3.63s + 3.32s2
3.29s 2 (1 + s tsc )
1 + s (1.819 + 0.13i) 1 + s (1.819 0.13i)
3.29s 2 (1 + s tsc )
(4.58)
From (4.58), it is apparent that the revised PID tuning method has led to a controller
with conjugate complex zeros. In Fig. 4.6, it is shown that there is little difference both
in the step and frequency response of the closedloop control system. Specifically,
after comparing the step response and output disturbance rejection of the control
101
loop, it is clear that Fig. 4.6 settling time of disturbance rejection is 8.13 in case of
the revised tuning compared to 8.23 in case of the conventional tuning.
From the frequency response and regarding both PID tuning methods, see Fig. 4.6
of T , S, robustness of the control loop is practically the same. Note at that point that
the conventional tuning method fails to tune a PI control action, see Sect. 3.2.3, in
contrast with the proposed method. From Fig. 4.6b, it is clear that the peak value of
the PI control action u( ) is significantly lower than the one provided by the PID
control action. This advantage can be critical in a realworld application, since high
peak command signal values might not be available by the constraints of the hardware
of the actuator unit.
From the frequency response and the phase diagram, see Fig. 4.6b it is apparent
that the phase margin of the Fol (s) is (u) = 47.2 , whereas in the case of the
revised PID tuning the phase margin is (u) = 34.5 in the case of PID control. The
level of the phase margin is ((u) < 45 ) is justified also by the overshoot of the step
response of the closedloop control system, see Fig. 4.5 which is higher the 50 %.
1
(1 + s )(1 + s )(1 + 0.01s )(1 + 0.001s )(1 + 0.0001s )
(4.59)
is considered. After the application of the revised PI control law, it is found that
Crev (s ) =
1 + 7.82s
30.56s 2 (1 + s tsc )
(4.60)
In similar fashion, the conventional and the revised PID control action are given by
(Fig. 4.7)
Ccon (s ) =
(1 + 4.44s )(1 + s )
9.87s 2 (1 + s tsc )
(4.61)
and
Crev (s ) =
1 + 5.1s + 6.1s2
2
6.84s (1 + s tsc )
(4.62)
respectively. From the step response of the closedloop control system, see Fig. 4.8
it is found that the revised PID control action leads to faster disturbance rejection
compared to the conventional tuning, since the settling time tss in the first case is
tss = 11.1 compared to tss = 18 in the second case.
102
Fig. 4.7 Control of a process
with two dominant time
constants defined by (4.59).
Comparison between the
conventional and the revised
PID tuning method. (Black)
revised PI tuning, (black
dotted) conventional PID
tuning, (gray) revised PID
tuning. a Step response of the
control loop. b Frequency
response of sensitivity S and
complementary sensitivity T
(a)
revised PI
yr ( )
revised PID
conventional PID
yo ( )
t ss = 11.1
t ss = 18
t ss = 30.2
= t/ Tp1
(b)
S( ju )
 T ( ju )
revised PID
conventional PID
revised PI
u=
Tp1
In Fig. 4.7b, it is shown that the revised PID control action has improved the
robustness of the closedloop control system, since the magnitude of complementary
sensitivity T (ju) remains equal to one in a wider range compared to the conventional
tuning. The same result holds for sensitivity S, since the amplitude of S(ju) remains
equal to zero in a wider range in the case of the revised control action.
The phase margin introduced to the closedloop control system via the conventional tuning is equal to (u) = 36.3 , whereas in the case of the revised tuning the
phase margin is equal to (u) = 45.7 , see Fig. 4.8a. From Fig. 4.8b, it is apparent that the revised tuning requires less effort on the command signal side since
the settling time is tss = 8.6 compared to tss = 15.1 which is required by the
conventional tuning.
103
(a)
F ol (ju )
revised PID
revised PI
(u )
conventional
PID
m = 45.7
m = 36.3
m = 32.9
u = T p1
(b)
revised PID
t ss = 8.66
revised PI
t ss = 28.08
u ( )
conventional PID
t ss = 15.1
= t/ T p1
(4.63)
which introduces two zeros on the right half plane. In this case, the calculated PI
control action via the revised method is given by
Crev (s ) =
1 + 29.9s
451.2s 2 (1 + s tsc )
(4.64)
104
(a)
revised PI
revised PID
t ss
t ss = 105
= 71.6
conventional PID
= t/ T p1
(b)
 T ( ju )
S ( ju )
revised PID
revised PI
conventional PID
u = T p1
In the case of PID control, the conventional and the revised controllers are given by
Ccon (s ) =
(1 + 16.4s )(1 + s )
134.48s 2 (1 + s tsc )
(4.65)
and
Crev (s ) =
1 + 24.4s + 65.7s2
2
236.7s (1 + s tsc )
(4.66)
respectively.7 From Fig. 4.9a, it is apparent that the conventional PID tuning method
fails to tune a control loop with acceptable performance. By contrast, the revised
7
Let it be noted that the conventional tuning has never been tested to nonminimum phase processes
within the academic literature.
105
(a)
 Fol ( ju )
conventional PID
revised PID
revisedPI
(u )
m = 4.76
m = 28.2
m = 30.7
u = T p1
(b)
t ss = 95.7
t ss = 147
revised PI conventional
PID
t ss = 64.6
revised PID
u( )
= t/ T p1
method succeeds in tuning the PI, PID control action achieving fast disturbance
suppression. The oscillatory behavior of the control loop involving the conventional
tuning is also observed in the frequency domain where the magnitude of T (ju)
exhibits a high peak, ten times greater than the unity.
From Fig. 4.10b, it is clear that the oscillatory behavior in the control loop with the
conventional tuning is the result of the unacceptable command signal which results
from the poor tuning of the controller.
106
(1 + s )5
e5s .
(4.67)
Regarding the PI control action, the revised method results in the controller
Crev s =
1 + 34.8s
606.5s 2 (1 + s tsc )
(4.68)
whereas the corresponding PID controller via the conventional and the revised methods are given by
Ccon (s ) =
(1 + 16.4s )(1 + s )
134.48s 2 (1 + s tsc )
(4.69)
and
Crev (s ) =
1 + 27.22s + 82.3s2
2
288.22s (1 + s tsc )
(4.70)
respectively. Note that in this case, controller (4.69) fails to tune a stable control
loop. On the contrary, the proposed method leads to a satisfactory step response
and disturbance rejection, see Fig. 4.11a. Note that, the PID controller exhibits
an increased robustness regarding disturbances, see Fig. 4.11b. The introduction
of the D term decreases dramatically the settling time of disturbance rejection,
from tss = 123 to tss = 81.9 . This also reflected by the step response of
the command signal u( ) where the settling time of the PID control action is
tss = 71 compared to tss = 117 in the case of PI control action, see
Fig. 4.12b.
(1 + 1.5s )
.
(1 + s )(1 + 0.9638s )(1 + 0.4061s )(1 + 0.2392s )(1 + 0.1751s )
(4.71)
107
(a)
ovs = 60.9%
ovs = 59.8%
revised PI
y r ( )
y o ( )
revised PID
td =
t ss = 81.9
t ss = 123
= t/ T p1
(b)
 T ( ju )
S ( ju )
revised PID
revised PI
u = T p1
The aforementioned feature is reflected also in the step response of the closedloop control system since the overshoot introduced in the case of the revised control action is almost equal to 50 % both for the PI and the PID controller, see
Fig. 4.13a.
After applying PI control action, the revise controller is defined by
Crev (s ) =
1 + 4.32s
10.49s 2 (1 + s tsc )
(4.72)
108
(a)
 Fol ( ju )
revised PID
revised PI
(u )
m = 29.2
m = 29.8
u = T p1
(b)
revised PI
t ss = 117
u ( )
revised PID
t ss = 71
= t / T p1
Fig. 4.12 Control of a process with long time delay defined by (4.67). Comparison between the
conventional and the revised PID tuning method. (Black) revised PI tuning, (black dotted) conventional PID tuning, (gray) revised PID tuning. a Frequency response, phase diagram of the open loop
transfer function Fol (s). b Step response of the command signal u( ) in the presence of a change
on the reference signal r(s)
whereas the corresponding PID controller for the conventional and the revised tuning
are given by
Ccon (s ) =
(1 + 7.53s )(1 + s )
28.4s 2 (1 + s tsc )
(4.73)
109
(a)
revised PI
conventional PID
y r ( )
revised PID
y o ( )
t ss = 10.2
t ss = 25
t ss = 22.5
= t/ T p1
(b)
S ( ju )
 T ( ju )
conventional PID
revised PID
revised PI
u = T p1
and
Crev (s ) =
1 + 3.73s + 4.73s2
3.34s 2 (1 + s tsc )
(4.74)
respectively. In Fig. 4.13a, it is shown that the revised PID controller leads to a much
faster output disturbance rejection, since the settling time in the case of conventional
tuning is tss = 25 , whereas in the case of revised tuning is tss = 10.2 . However, the
disadvantage of the revised PID control action is apparent from the phase diagram
of the openloop transfer function Fol (s), since the phase margin in the case of
conventional tuning is (u) = 58.5 > 45 in contrast with the revised method which
is equal to (u) = 25.4 > 45 , see Fig. 4.14.
110
revised PID
conventional PID
revised PI
m = 58.5
m = 25.4
m = 37.3
u = T p1
Fig. 4.14 Control of a process with large zeros defined by (4.71). Comparison between the conventional and the revised PID tuning method. (Black) revised PI tuning, (black dotted) conventional
PID tuning, (gray) revised PID tuning. Frequency response, phase diagram of the openloop transfer
function Fol (s)
Ideal source
Line impedance
Transformer model
I feed
I ar
RM
CDC
LM
V 50Hz
V net
DC LinkIload
V ar
AC/DC
V DC
Fig. 4.15 Grid connected active rectifier on system level. The interfaced signals are: V net (V)
and net (Hz) stand for the threephase voltage measured at PCC; V ar (V) I ar (A) stand for
the threephase voltage and current at the grid converter respectively. VDC (V) and Ifeed (A) is
the DC link voltage and the feeding current from grid converter and IC (A), Iload (A) is the
DC link capacitor current and the load current respectively. System parameters that should be
known or estimated for controlling purpose are CDC (C) from the DC link capacitor, L (H)
(leakage inductance of the transformer), R ( ) (leakage resistance of the transformer), LM (H)
(magnetizing inductance of the transformer), RM ( ) (magnetizing resistance of the transformer), Lnet (H) (equivalent line inductance of the network), Rnet ( ) (equivalent line resistance
of the network)
111
purpose is to maintain the DC link voltage that supplies power for a drive typically,
or another network. The network may be modeled as a voltage source along with its
grid impedance that reflects its strength. The grid transformer is modeled through its
magnetizing and leakage impedance.
Id act (s)
1
1
=
=
Id ref (s)
(1 + sTp1 )(1 + sTp )
(1 + sTp1 )(1 + s Tp1 )
(4.75)
for which can be chosen sufficiently small compared to the current control loops
time constant, modeling any parasitic time constants of the inner current control loop
itself. The transfer function regarding the DC link voltage control loop (outer control
V act (s)
or
loop, see Fig. 4.16) is equal to Tv (s) = VDC
DC (s)
ref
Cv (s)TI (s)G(s)
Tv (s) =
=
1 + Cv (s)TI (s)G(s)
1
sCDC
1
1 + Cv (s)TI (s)
sCDC
Cv (s)TI (s)
(4.76)
where Cv is the DC link voltage controller and G(s) is the integrating process for
the voltage control loop, which in this case is the capacitor bank path within the DC
link, see Fig. 4.16. Substituting (4.75) into the voltage control loop transfer function,
VDCact (s)
V DCref(s)
eVDC(s)
CV(s)
Idref(s)
+
Iqref(s)
Cex(s)
voltage controller
CId(s)
u q(s)
+
Vqnet(s)
current controller
command signal
CIq(s)
current controller
L
u d(s)
++
current controller
command signal
current controller
eiq(s)
Iqact(s)
Ipff(s)
id
e (s)
Idact(s)
1/R
L
s+1
R
Park
transformation
current measurement
current measurement
M AR
kpesTd
inverter
Vqctrl(s)
V2dctrl + V2qctrl
Vdctrl
atan
Vqctrl
Vdctrl (s)
construction of
modulation index
Park
transformation
Vdnet (s)
1
sCDC
Iload(s) = di(s)
Iqact(s)
[Tdq+1 ]
[Tdq+1 ]
Idact(s)
integrating process
for the outer voltage control lo op
VDCact (s)
Fig. 4.16 Cascaded control loop for AC/DC grid converters. The interfaced signals are Mar : modulation index of grid converter AC voltage, Vdctrl (V): voltage
control value (activepart), Vqctrl (V): voltage control value (reactivepart), Vdnet (V): grid voltage measurement (activepart), Vqnet (V): grid voltage measurement
(reactivepart), ID (A): grid current measurement (activepart), IQ (A): grid current measurement (reactivepart), Ipff (A): current feedforward from load drive,
Idref (A): grid current reference (activepart), Iqref (A): grid current reference (reactivepart), VDCref (V): DC link voltage reference
VDCref(s)
external filter
2DoF Controller
112
4 TypeII Control Loops
113
(a)
do ( )
conventional
V DC act ( )
proposed
= t/ T p1
(b)
step response
u ( )
proposed
conventional
voltage controllers
command signal
= t/ T p1
results after in
Tv (s) =
Cv (s)
sCDC (1 + sTp1 )(1 + s Tp1 ) + Cv (s)
(4.77)
where Cv (s) is designed every time according to the conventional and the revised
symmetrical optimum criterion presented in Sects. 4.2, 4.2.3 and 4.3.
Since the control loop is of typeII, an external filter is added on the reference
signal, VDCref for dealing with the high overshoot in VDCact in case of step changes
in VDCref . Note that VDCref is set on the electric drive, in practice it does change that
often. The choice of the filter time constant of Cex (s) is chosen such that it cancels
the zero of the calculated voltage controller as it is extensively discussed in [11, 13].
In Fig. 4.17, the step response of VDCact is presented. Disturbance rejection applied
at = 30 shows significant improvement compared to the conventional method,
decrease of settling time tss from 45.5 32 .
114
4.6 Summary
An explicit PID tuning solution for controlling integrating processes has been presented. The proposed method lies in the principle of the symmetrical optimum criterion and can be applied to any linear single input single output process regardless of
its complexity. The control laws proof does not involve any model reduction techniques which often lead to poor tuning as it happens in the case of the conventional
PID tuning procedure.
For justifying the tuning performance, the proposed control law is compared with
the current state of the art relevant to the PID tuning via the symmetrical optimum
criterion. This comparison focuses on the performance of the required control action,
in terms of reference tracking and disturbance rejection. Since the proposed method
concentrates on the PID controller which is often used in many industry applications,
the control of the actual DC link voltage on an AC/DC converter arrangement was
chosen as an example from the field of electric motor drives so that the feasibility
in terms of the methods implementation is also justified. The presented comparison
study reveals a satisfactory and promising improvement in terms of reference tracking
and disturbance rejection.
References
1. strm KJ, Hang CC, Lim BC (1994) A new Smith predictor for controlling a process with
an integrator and long dead time. IEEE Trans Autom Control 39(2):343345
2. Bahrani B, Kenzelmann S, Rufer A (2011) MultivariablePIbased current control of voltage source converters with superior axis decoupling capability. IEEE Trans Ind Electron
58(7):30163026
3. Habetler TG (1993) A space vectorbased rectifier regulator for AC/DC/AC converters. IEEE
Trans Power Electron 8(1):3036
4. Kessler C (1955) UG ber die Vorausberechnung optimal abgestimmter regelkreise teil III. Die
optimale einstellung des reglers nach dem betragsoptimum. Regelungstechnik 3:4049
5. Kessler C (1958) Das symmetrische optimum, Regelungstechnik pp 395400 and 432426
6. Li S, Pingxi Y, Lifei L, Lin C, Liuxin B, Gang C, Chao Z (2012) Research on gridconnected
operation of novel variable speed constant frequency (VSCF) shaft generator system on modern
ship. In: 15th international conference on electrical machines and systems. ICEMS, IEEE,
Sapporo, pp 15
7. Matauek MR, Micic AD (1999) On the modified Smith predictor for controlling a process
with an integrator and long deadtime. IEEE Trans Autom Control 44(8):16031606
8. Mirecki A, Roboam X, Richardeau F (2007) Architecture complexity and energy efficiency of
small wind turbines. IEEE Trans Ind Electron 54(1):660670
9. Nishikata S, Tatsuta F (2010) A new interconnecting method for wind turbine/generators in a
wind farm and basic performances of the integrated system. IEEE Trans Ind Electron 57(2):468
475
10. Oldenbourg RC, Sartorius H (1954) A uniform approach to the optimum adjustment of control
loops. Trans ASME 76:12651279
11. Papadopoulos KG, Margaris NI (2012) Extending the symmetrical optimum criterion to the
design of pid typep control loops. J Process Control 12(1):1125
References
115
12. Papadopoulos KG, Siemaszko D, Margaris NI (2012) Optimal automatic tuning of PID controllers applied to grid converters. In: Electrical systems for aircraft, railway and ship propulsion
(ESARS). IEEE, Bologna, Italy, pp 16
13. Papadopoulos KG, Papastefanaki EN, Margaris NI (2013) Explicit analytical PID tuning rules
for the design of typeIII control loops. IEEE Trans Ind Electron 60(10):46504664
14. Parlmor ZJ (1996) Time delay compensationSmith predictor and its modifications. In: Levine
WS (ed) Control handbook, Boca Raton, FL: CRC Press, vol 53, pp 224237
15. Richard JP (2003) Timedelay systems: an overview of some recent advances and open problems. Automatica 39(10):16671694
16. Sartorius H (1945) Die zweckmssige festlegung der frei whlbaren regelungskonstanten.
Master thesis, Technische Hochscule, Stuttgart, Germany
17. Shafiei Z, Shenton AT (1994) Tuning of PIDtype controllers for stable and unstable systems
with time delay. Automatica 30(10):16091615
18. Smith OJM (1959) Closed control of loops with deadtime. Chem Eng Sci 53:217219
19. Umland WJ, Safiuddin M (1990) Magnitude and symmetric optimum criterion for the design
of linear control systems: what is it and how does it compare with the others? IEEE Trans Ind
Appl 26(3):489497
20. Wang QC, Hang CC, Yang PX (2001) Singleloop controller design via IMC principles. Automatica 37(12):20412048
21. Watanabe K, Ito M (1981) A process model control for linear systems with delay. IEEE Trans
Autom Control 26(6):12611268
22. Zhang W, Xu X, Sun Y (1999) Quantitative performance design for integrating processes with
time delay. Automatica 35(4):719723
Chapter 5
Abstract In this chapter, the problem of designing PID typeIII control loops is
investigated. On a theoretical basis and if frequency domain modeling is followed,
typeIII control loops are characterized by the presence of three pure integrators in
the open loop transfer function, see Sect. 2.1. Therefore, such a control scheme has
the advantage of tracking fast reference signals since it exhibits zero steady state
position, velocity and acceleration error, see Sect. 2.1. This advantage is considered
critical in many industry applications, i.e. control of electrical motor drives, control
of power converters, since it allows the output variable, i.e., DClink voltage or speed,
to track perfectly step, ramp and parabolic reference signals. In a similar fashion,
with Chaps. 3 and 4, the proposed PID control law (1) consists of analytical expressions that involve all modeled process parameters (2) can be straightforward applied
to any process regardless of its complexity since for its development a generalized
transfer function process model is employed consisting of npoles, mzeros plus
unknown time delayd (3) allows for accurate investigation of the performance of
the control action to exogenous and internal disturbances in the control loop, investigation of different operating points. For justifying the potential of the proposed
control law, several examples of process models met in many industry applications
are investigated.
5.1 Introduction
From a conceptual point of view, the advantage of typeIII control loops compared to
typeI or typeII systems is obvious, since the former are able to track a step, ramp,
and parabolic reference input by achieving zero steady state position, velocity, and
acceleration error, respectively. Therefore, such control loops are capable of tracking
very fast reference signals.
A first attempt of designing typeIII control loops for singleinput singleoutput
processes has been proposed in [3, 5, 7, 8] and is presented in Sect. 5.2.1 for the
sake of completeness of the proposed theory. In this case, the design of the control
loop is developed in the frequency domain, and the principle of the Symmetrical
Optimum criterion is once more adopted [2, 4]. The proposed PID typeIII control
law is based on polezero cancellation as the conventional Symmetrical Optimum
Springer International Publishing Switzerland 2015
K.G. Papadopoulos, PID Controller Tuning Using the Magnitude Optimum Criterion,
DOI 10.1007/9783319072630_5
117
118
implies, and therefore an accurate estimation of the dominant time constants of the
process is required. Further to this constraint, the PID controller zeros are restricted to
be tuned only with real values, and not with conjugate complex if needed. Moreover,
the process model used to develop the aforementioned control law is simple (second
order process model), and therefore other dynamics of the process are neglected.
All aforementioned constraints, regarding the conventional PID typeIII control
law proposed in [5] can be summarized as follows:
1. the PID controller parameters are tuned as a function of the processs dominant
time constants. Unmodeled dynamics of the plant are approximated by a first
order lag time constant,
2. the principle of polezero cancellation is followed,
3. the PID controller zeros are allowed to be tuned only with real values,
4. a simple second order model is employed for the development of the proposed
PID control law regarding typeIII control loops.
From the above, it is apparent that when the complexity of the process increases,
the conventional PID typeIII control law presented in Sect. 5.2.1, and according
to [5, 8] fails sometimes to tune a stable control loop as it is shown in the sequel.
One way to improve the control law presented in Sect. 5.2.1 is to introduce a more
complex process model, and explicitly tune the PID parameters without following
model reduction techniques, see Sect. 5.2.2 and Appendix B.3.
To cope with this model reduction approximation issue, a first attempt of designing
typeIII control loops without using a simple process model has been reported in [8].
In this work, for modelling the process, a transfer function consisting of n poles
and unknown time delay d has been employed whereas any zeros of the process
are not taken into account. The potential of this PID control law is tested on a
nonminimum phase process and a process with dominant time constants achieving
promising results.
For that reason and motivated by the promising results in [8], scope of this chapter
is to tune analytically a PID type controller, regardless of the process complexity
(n poles, m zeros plus time delay d), so that the final closed loop control system
exhibits zero steady state, position velocity and acceleration error. At this point,
the assumptions presented in [5, Sect. 5.2.1] are disregarded and for developing the
proposed theory
1. the PID controller parameters are tuned explicitly as a function of all n poles, m
zeros plus time delay d,
2. the principle of polezero cancellation is not followed,
3. a more flexible form is introduced and the PID controller zeros are allowed to be
tuned both with real values and conjugate complex values if needed, see [9],
4. no model reduction techniques are going to be followed, see Sect. 5.2.2.
In this case, zeros of the PID controller are not forced to be compensated by the
plants dominant time constants since a more flexible form of the PID controller is
introduced. This form tunes the zeros of the controller as a function of all modeled
process parameters allowing its values to become conjugate complex if needed. To
5.1 Introduction
119
this end, control engineers are able to design PID typeIII control loops regardless of
the process complexity and analyze accurately the control loops performance before
proceeding on a real time implementation.
For developing the proposed control law, once more the concept of Symmetrical
Optimum criterion is employed [2, 4, 9]. Thus, for extracting the explicit solution
for the proposed PID control law, the principle of the Magnitude Optimum criterion
presented in Appendix A.1 is utilized once more.
For justifying the proposed control law, several process models are employed for
testing the control loops response to step, ramp, and parabolic reference signals. The
proposed control law is compared with the conventional PID tuning via the Symmetrical Optimum criterion of Sect. 5.2.1. The proposed method achieves satisfactory
performance in terms of reference tracking (zero steady state position, velocity, and
acceleration error) compared to the conventional tuning, where its resulting response
is oscillatory and most of the times unstable, Sect. 5.2.3. The robustness of the proposed control law to model uncertainties is also discussed, see Sect. 5.2.3.4. As a
result, control engineers are given the ability apart from designing a typeIII control
loop, to test on a simulation basis the performance of the proposed control law before
integrating it on a real time application.
1
sTm (1 + sTp1 )(1 + sTp )
(5.1)
is adopted, where Tp1 stands for the dominant time constant of the process and
Tm , Tp stand for the integrators time constant and the unmodeled plant dynamics,
respectively [1]. Supposing that the dominant time constant Tp1 is evaluated, the
proposed IPID controller is defined by
120
r (s)
+
+
di ( s )
controller
e(s)
C (s)
u (s)
+
do (s)
y f (s)
S
kp
G (s)
+
+
y (s)
kh
+
n o (s)
Fig. 5.1 Block diagram of the closedloop control system. G(s) is the plant transfer function, C(s)
is the controller transfer function, r (s) is the reference signal, do (s) and di (s) are the output and
input disturbance signals, respectively, and n r (s), n o (s) are the noise signals at the reference input
and process output, respectively. kp stands for the plants dc gain and kh is the feedback path
C(s) =
(5.2)
where Tc1 , Tc2 are known and sufficiently small time constants compared to
Tp1 , arising from the controllers implementation. By setting Tx = Tp1 (polezero
cancellation) and assuming that
Tc = Tc1 + Tc2 , and Tc1 Tc2 0
(5.3)
T (s)
s 2 kp Tn Tv + skp (Tn + Tv ) + kp
Ti Tm T s 4 + Ti Tm s 3 + s 2 kp kh Tn Tv + skp kh (Tn + Tv ) + kp kh
(5.4)
(5.5)
(Ti Tm T)2 8 + Ti Tm Ti Tm 2kp kh Tn Tv T 6
D() = +kp kh 2T i Tm T
2 (Tn + Tv ) Ti Tm + kp kh Ti2 Tm2 4
+(kp kh )2 Tn2 + Tv2 2 + (kp kh )2
(5.6)
121
2kp kh Tn Tv T
.
Tm
(5.7)
In a similar fashion, setting the term of 4 equal to zero and along with the aid of
(5.7) results in
4T2 4 Tn + Tv T + Tn Tv = 0.
(5.8)
4(n 1)
T .
n4
(5.9)
Summarizing the relations (5.7) and (5.9), the aforementioned PID control law
defined in (5.2) results in
Tp1
Tx
nT
Tv
4(n 1)
T
Tn =
n4
Ti
n(n 1)T3
8kp kh
kh
(n 4)Tm
1
(5.10)
Proper selection of parameter n, (n > 4 must hold by) leads to a feasible IPID
control law. Substituting Eq. (5.10) into the closed loop transfer function results in
T (s) =
.
(5.11)
.
(5.12)
Note that the control loop defined in (5.12) is of typeIII, since the terms of s j , j =
0, 1, 2, are equal, a0 = b0 , a1 = b1 , a2 = b2 , see Sect. 2.5. The respective step and
frequency responses of (5.12) for two different values of parameter n, are presented
122
(a)
n = 7.46
y r ( )
n = 4.1
y o ( )
n = 7.46
= t/ T
(b)
 T ( ju )
S ( ju )
n = 4.1
u n = 0.28
n = 7.46
u n = 0.85
u = T
in Fig. 5.2. In addition, in Fig. 5.3 the open loop frequency response is shown. Its
transfer function is given by
Fol (s) =
(5.13)
From Fig. 5.2b it is concluded that the magnitude of the complementary sensitivity
T ( ju) is practically independent of the parameter n. Sensitivity S( ju) becomes
maximum if n = 4.1 and minimum, if n = 7.46. If n = 7.46 then Tn = Tv holds by.
For every other value of parameter n, the shape of the open loop frequency response
is preserved exactly as presented in Fig. 5.2b.
The same conclusion holds also for the overshoot of the step response of the typeIII control loop which remains almost equal to 50 % regardless of the parameter n.
123
 Fol ( ju )
n = 7.46
n = 4.1
u c = 0.5 u c = 1
(u )
n = 4.1
n = 7.46
m = 35
u = T
s m m + s m1 m1 + + s 2 2 + s1 + 1 sTd
e
s n1 an1 + + s 3 a3 + s 2 a2 + sa1 + 1
(5.14)
is adopted where n 1 > m. Since the target of the design is a typeIII control
loop, according to the analysis presented in Sect. 2.5, three integrators in Fol (s) =
kh kp G(s)C(s) must exist. Therefore, the proposed IIPID controller is given by
C(s) =
1 + s X + s2Y
.
s 3 Ti3 (1 + sTpn )
(5.15)
Parameter Tpn stands for the parasitic controller time constant as mentioned in
Sect. 5.2.1. In contrast with Sect. 5.2.1, the flexible form of the numerator Nc (s) =
1 + s X + s 2 Y defined in (5.15) allows its parameters X, Y to become complex
conjugates if possible, see [9].
Purpose of this section is to determine explicitly controllers parameters, as a
function of all plant parameters, without following the principle of polezero cancellation and ignoring other possible fundamental dynamics of the process. In that case,
124
j
2
j=0 (s j )(1 + s X + s Y ) sTd
e
n
s 3 Ti3 i=0
(s i pi )
(5.16)
n
j
(s i pi ) = (1 + sTpn ) n1
where i=0
j=0 (s a j ). According to Fig. 5.1, the closed
loop transfer function is given by
T (s) =
kp C(s)G(s)
1 + kp kh C(s)G(s)
(5.17)
n
3
s 3 Ti
kp (1 + s X + s 2 Y )
j=1
(s j
pj
)esTd
m
j=0 (s
+ kp kh (1 + s X
j)
+ s2Y )
m
j=0 (s
j)
(5.18)
where p0 = 1, 0 = 1. For the need of the analysis, a general purpose time constant
c1 is considered. Therefore all time constants involved within the control loop are
normalized by setting s = sc1 . This results in the following substitutions
x=
d=
X
,
c1
y=
Y
Ti
, ti = ,
2
c1
c1
pj
Td
i
, r j = i , j = 1, . . . , n, z i = i
c1
c1
c1
(5.19)
(5.20)
es d =
7
1
k=0
k!
k
s dk .
(5.21)
T (s ) =
(5.22)
125
r1 + d
r2 + r1 d + d 2
q1
2!
q2
1 2
1 3
r3 + r2 d + d r1 + d
q3 =
2!
3!
q4
1 2
1 3
1 4
r4 + r3 d + d r2 + d r1 + d
q5
2!
3!
4!
1 2
1 3
1 4
1 5
r5 + r4 d + d r3 + d r2 + d r1 + d
2!
3!
4!
5!
(5.24)
(5.25)
or finally
D1 (s ) =
k
ti3 q j s ( j+3)
(5.26)
j=0
where
k
1 i
qk =
d
r(ki)
i!
(5.27)
i=0
126
kh N (s ) = kh kp
p
(r )
yz (r 2) + x z (r 1) + z (r ) .
s
(5.28)
r =0
Finally,
D(s ) = D1 (s ) + kh N s
=
p
k
(r )
yz (r 2) + x z (r 1) + z (r ) .
(ti3 q j )s ( j+3) + kh kp s
(5.29)
r =0
j=0
3
j=0 (ti q j )s
p
r =0 s
( j+3)
kp
T (s ) = k
(r ) (yz
+ kh kp
+ x z (r 1) + z (r ) )
.
(r ) yz
s
(r 2) + x z (r 1) + z (r )
r =0
(r 2)
p
(5.30)
(5.31)
8
Cjx j = 0
(5.32)
j=0
y=
ti3
1 2 1 2
x + (z 1 2z 2 )
2
2
kh kp y 2 (z 12 2z 2 )2 + z 22 2z 1 z 3 + 2z 4
=
.
2
x + z 1 q1
(5.33)
(5.34)
127
0.254
s (1 + s )5
(5.35)
(5.36)
(5.37)
is selected.
Note that x, y and tn , tv are the solutions coming from the conventional and
the revised control law, respectively, so that the comparison between the two tuning
128
(a)
y ( )
step response
revised symmetrical
optimum
di ( ) = r ( )
conventional symmetrical d ( ) = 0.25r ( )
o
optimum
= t/ ( T p1 )
(b)
u ( )
step response
revised symmetrical
optimum
do ( ) = 0.25r ( )
conventional symmetrical
optimum
di ( ) = r ( )
= t/ ( T p1 )
CPIDSO (s ) =
(1 + 29.03s + 421.5s 2 )
.
s 2 942.8(1 + 0.1s )
(5.38)
(5.39)
129
(a)
ram presponse
y ( )
conventional symmetrical
optimum
revised symmetrical
optimum
r ( )=
= t/ T p1
(b)
parabolic response
y ( )
r ( )= 2
conventional symmetrical
optimum
revised symmetrical
optimum
= t/ T p1
Fig. 5.5a, b present the ramp and parabolic response of the final closed loop control
system when the PID controller is tuned via (5.38) and (5.39), respectively. From
Fig. 5.5a it is apparent that the revised tuning reaches steady state at = 64 in contrast
with the conventional tuning where its response remains practically unstable.
5.2.3.2 Process with Time Delay Equal to Its Dominant Time Constant
In this example the process to be controlled is defined by
G(s ) =
0.254
s (1 + s )(1 + 0.5s )(1 + 0.2s )2 (1 + 0.1s )
es
(5.40)
130
which exhibits a time delay constant Td equal with the dominant time constant Tp1 ,
d = TTpd = 1. The resulting PID controller parameters according to Sects. 4.2.3 and
1
5.2.2 are equal to
(1 + s tn )(1 + s tv )(1 + s tx )
s 2 ti (1 + s tsc1 )(1 + s tsc2 )
(1 + 7.46s )(1 + 2.98s )(1 + s )
=
s 2 4.52(1 + 0.1s )(1 + 0.1s )
CPIDSO (s ) =
(5.41)
and
CPID (s ) =
=
(1 + s x + s 2 y)
s 2 ti (1 + s tsc1 )
(1 + 13.63s + 92.9s 2 )
(5.42)
s 2 97.6(1 + 0.1s )
respectively. In Fig. 5.6a, b the response of the control loop for y( ), u( ) to a step
reference input r (s ) = s1 , a step output and input disturbance do (s ) = 0.25
s , di (s ) =
r (s ) is presented both for the conventional and the revised control law, respectively.
The PID controller via the conventional PID tuning (Symmetrical Optimum criterion) leads to unacceptable response in terms of overshoot, input and output disturbance rejection, see Fig. 5.6a, b. Let it be noted that in both cases, conventional and
revised PID tuning, the reference r (s ) is filtered by an external controller Cex (s )
defined by (5.36) and (5.37), respectively. Ramp response of the revised tuning,
settles faster than the conventional tuning, Fig. 5.7a.
(5.43)
The resulting PID control law according to the conventional and the revised tuning
are given by
CPIDSO (s ) =
=
(1 + s tn )(1 + s tv )(1 + s tx )
s 2 ti (1 + s tsc1 )(1 + s tsc2 )
(1 + 14.55s )(1 + 7.47s )(1 + s )
1 + s x + s 2 y
CPID (s ) = 2
s ti (1 + s tsc1 )
(5.44)
131
(a)
y ( )
step response
revised symmetrical
optimum
di ( )= r ( )
conventional symmetrical
optimum
do ( )= 0.25r ( )
= t/ T p1
(b)
u ( )
step response
revised symmetrical
optimum
do ( )= 0.25r ( )
conventional symmetrical
optimum
di ( )= r ( )
= t/ T p1
=
1 + 22.02s + 242.8s 2
s 2 2577.41(1 + 0.1s )
.
(5.45)
Once more the conventional PID tuning fails to tune a stable typeIII control loop,
Fig. 5.8a, b. Ramp response of the conventional tuning reaches the steady state operation much faster than the conventional control loop, Fig. 5.9.
132
Fig. 5.7 Comparison
between the conventional and
the revised PID control law.
Control of a process with time
delay equal to its dominant
time constant. a Ramp
response of the output y( ) of
the control loop. b Parabolic
response of the output y( ) of
the control loop
(a)
ramp response
conventional symmetrical
optimum
y ( )
revised symmetrical
optimum
= t/ T p1
(b)
parabolic response
y ( )
r ( )= 2
conventional symmetrical
optimum
revised symmetrical
optimum
= t/ T p1
G(s ) =
(1 + s )5
e4s .
(5.46)
1+a
(1 + s )5
e4(1+b)s
(5.47)
133
(a)
y ( )
step response
revised symmetrical
optimum
di ( )= r ( )
d ( )= 0.25r ( )
conventional symmetrical o
optimum
= t/ T p1
(b)
step response
revised symmetrical
optimum
u ( )
do ( )= 0.25r ( )
conventional symmetrical
optimum
di ( )= r ( )
= t/ T p1
Cex (s ) =
1
1 + xs + y 2 s
(5.48)
134
Fig. 5.9 Comparison
between the conventional and
the revised PID control law.
Control of a nonminimum
phase process. a Ramp
response of the output y( ) of
the control loop. b Parabolic
response of the output y( ) of
the control loop
(a)
ramp response
conventional symmetrical
optimum
y ( )
revised symmetrical
optimum
r ( )=
= t/ T p1
(b)
parabolic response
y ( )
conventional symmetrical
optimum
r ( )= 2
revised symmetrical
optimum
= t/ T p1
type p control loop there should exist an PI p D, or PI( p1) D, or PID and so on, if
the process is of type0 or typeI or type( p 1), respectively. A typeII control
loop achieves zero steady state position and velocity error, a typeIII control loop
achieves zero steady state position, velocity and acceleration error, and therefore a
type p control loop is expected to track both faster reference signals and eliminate
higher order errors at steady state.
For deriving the proposed control law, a transfer function containing dominant
time constants and the plants unmodeled dynamics has been considered in the frequency domain. The final control law consists of analytical expressions that involve
both dominant dynamics and model uncertainty of the plant. For justifying the potential of the proposed theory, simulation results for representative processes met in
many real world applications are presented.
135
(a)
b = 0.25
do ( )
di ( )
y ( )
optimal tuning
approximate tuning
(b)
= t/ T
a = 0.25
di ( )
do ( )
optimal tuning
y ( )
approximate tuning
= t/ T
Tm s q
1
n s
,
(1
+
T
m j s) k=1 (1 + Tsk s)
j=1
n m
(5.49)
136
(1 + Tsk s) = 1 + Ts s
(5.50)
k=1
where
Ts =
ns
Tsk
(5.51)
k=1
stands for the process small unmodeled time constants. Since the target of the design
is the final closedloop control system to be of type p, according to the analysis
presented in Sect. 5.2.1, the proposed PID type controller is given by
p1
n m
r =1 1 + Tnr s
j=1 1 + Tm j s
C(s) =
.
(5.52)
c
1 + Tcz s
Ti s pq nz=1
Thus, according the design of typeII (4.2.3), typeIII control loops (5.2.1), the PID
type controller has to contain n m zeros equal to the Tm j dominant time constants
( j = 1, 2, . . . , n m ) so that exact polezero cancellation is achieved.
Moreover, in order the denominator of the final closedloop transfer function T (s)
is a full polynomial in terms of the s j coefficients, it is easily proved after manipulating algebraically T (s), that p 1 zeros must exist. Furthermore, the controller must
introduce p q integrators, so that the final closedloop is of type p. Finally, in order
the controller transfer function is strictly causal, denominators order must be greater
or equal to p 1 + n m . The unmodeled controllers dynamics are represented by
nc
1 + Tcz s = 1 + Tc s
(5.53)
z=1
where
Tc =
nc
Tcz .
(5.54)
z=1
p1
r =1 1 + Tnr s
s
c
Fol (s) = kp kh G(s)C(s) = kp kh
(5.55)
Ti Tm s p nk=1
(1 + Tsk s) nz=1
(1 + Tcz s)
or by substituting (5.50) and (5.52)(5.54) results in
p1
Fol (s) = kp kh
r =1 (1 + Tnr s)
Ti Tm s p (1 + T s)
(5.56)
137
Ti Tm T
s p+1
p1
r =1 (1 + Tnr s)
p1
Ti Tm s p + kp kh r =1 (1 +
kp
Tnr s)
(5.57)
b p1 s p1 + b p2 s p2 + + b3 s 3 + b2 s 2 + b1 s + b0
a p+1 s p+1 + a p s p + a p1 s p1 + + a3 s 3 + a2 s 2 + a1 s + a0
(5.58)
where
b p1 =
p1
p1
Tp j = Tp1 Tp2 Tp p1 , b3 = kp
b2 = kp
Tni Tn j Tnk ,
(5.59)
i= j=k=1
j=1
p1
Tni Tn j , b1 = kp
i= j=1
p1
Tni , b0 = kp ,
(5.60)
i=1
and
a p+1 = Ti Tm T , a p = Ti Tm ,
a3 = k p k h
p1
Tni Tn j Tnk , a2 = kp kh
i= j=k=1
a1 = k p k h
p1
(5.61)
Tni Tn j ,
(5.62)
i= j=1
p1
Tni , a0 = kp kh .
(5.63)
i=1
(5.64)
Ti = 2kp kh
(5.65)
r =1
This can be proved as follows. For a process of one dominant time constant defined
by (4.1) where (q = 1), then in order the final control loop is of typeII p = 2, the
PID type controller (according to the Symmetrical Optimum criterion) is given by
138
C(s) =
(5.66)
for which
Tn2 = Tp1
(5.67)
and (1 + sTp )(1 + sTc ) 1 + sT have been set. In that, the openloop transfer
function is given by
Fol (s) = kp kh
(1 + Tn1 s)
Ti Tm s 2 (1 + T s)
(5.68)
kp (1 + Tn1 s)
.
Ti Tm s 2 (1 + T s) + kp kh (1 + Tn1 s)
(5.69)
According to the analysis presented in Sect. 5.2.1, the integrators time constant is
calculated if
a22 = 2a1 a3
(5.70)
is set, as another means of optimizing the magnitude of (5.69) [11, 12]. The resulting
integrators time constant proves to be equal to
Ti = 2kp kh
T
Tn
Tm 1
(5.71)
T2
and if Tn1 = 4T is chosen, then Ti = 8kp kh Tm , see Sect. 5.2.1.
According to Sect. 5.2.1, for a process of one dominant time constant defined
again by (5.1) where (q = 1) then in order the final control loop is of typeIII p = 3,
the PID type controller is given by
C(s) =
(5.72)
(5.73)
and (1 + sTp )(1 + sTc ) 1+sT the openloop transfer function Fol (s) becomes
Fol (s) = kp kh
(5.74)
139
kp 1 + Tn1 s 1 + Tn2 s
T (s) =
Ti Tm s 3 (1 + T s) + kp kh 1 + Tn1 s 1 + sTn2
(5.75)
According to (5.70) and since n = 2, the integrators time constant is calculated via
a32 = 2a2 a4 .
(5.76)
Finally, after some algebraic manipulation it was shown that the integrators time
constant is equal to
Ti = 2kp kh
T
Tn Tn .
Tm 1 2
(5.77)
In a similar fashion, for a process of one dominant time constant defined by (5.1)
and if n = k 1, in order the final control loop is of type p, the PID type controller
is given by
C(s) =
(5.78)
According to the analysis presented previously, it can be claimed regarding the integrators time constant Tik1 , that
T
Tn j .
Tm
k1
Tik1 = 2kp kh
(5.79)
j=1
k1
T
= 2kp kh
Tn j Tnk = Tik Tnk .
Tm
k
Ti = 2kp kh
(5.80)
j=1
According to the design of type p control loops, the PID type controller is given by
C(s) =
(5.81)
140
for which Tnk+1 = Tp1 is set, assuming design via polezero cancellation. Since again
(1 + sTp )(1 + sTc ) 1 + sT , the open and closedloop transfer functions are
given by
k
Fol (s) = kp kh
T (s) =
kp
Ti Tm
s k+1 (1 +
j=1 (1 + Tn j s)
,
Ti Tm s k+1 (1 + T s)
k
j=1 (1 +
Tn j s)
k
T s) + kp kh
j=1 (1 +
(5.82)
Tn j s)
(5.83)
or
kp rk s k + rk1 s k1 + + r2 s 2 + r1 s + 1
T (s) =
rk s k + rk1 s k1 +
Ti Tm T s k+2 + Ti Tm s k+1 + kp kh
+r2 s 2 + r1 s + 1
(5.84)
(5.85)
(Ti Tm )2 = 2kp kh Ti Tm T rk
(5.86)
Ti Tm = 2kp kh T rk .
(5.87)
or
or
k1
T
= 2kp kh
Tn j Tnk = Tik Tnk
Tm
k
Ti = 2kp kh
(5.88)
j=1
which is equal to (5.80). In that case, if (5.88) is substituted into (5.83), results in
T (s) =
2kp kh T2
p1
r=1
Tnr
s p+1
p1
(1 + Tnr s)
.
p1
p1
+ 2kp kh T r=1 Tnr s p + kp kh r=1 (1 + Tnr s)
kp
r=1
(5.89)
For determining now parameters Tnr , it is shown that in order the magnitude of (5.89)
satisfies condition T ( j) 1, controller time constants Tnr must satisfy condition
4T2
p3
Tni 4T
i=1
p2
Tni +
i=1
141
p1
Tni = 0.
(5.90)
i=1
This is justified as follows. In typeII control loops for determining parameter Tn1
we make use of a12 2a2 a0 = 0 [see (A.11)]. This results in
kp2 Tn21 = 2kp (2kp Tn1 T )
(5.91)
Tn1 4T = 0.
(5.92)
or finally
In a similar fashion, in typeIII control loops for determining parameters Tn1 , Tn2
we make use of a22 2a3 a1 + 2a4 a0 = 0, see (A.11). This results in
4T2 Tn1 Tn2 4T Tn1 Tn2 (Tn1 + Tn2 ) + Tn21 Tn22 = 0
(5.93)
(5.94)
or finally,
According to the above, and based on (5.79) if the closedloop control system is of
type p, then for determining parameters Ti and Tn j , ( j = 1, 2, . . . , k), the following
optimization conditions are claimed to be,
2
= 2ak2 ak 2ak3 ak+1 ,
ak1
ak2
= 2ak+1 ak1
(5.95)
(5.96)
kp (rk1 s k1 + rk2 s k2 + + r2 s 2 + r1 s + 1)
.
rk1 s k1 + rk2 s k2
k+1
k
Ti Tm T s
+ Ti Tm s + kp kh
+ + r2 s 2 + r1 s + 1
(5.97)
2
2rk2 kp 2kp T rk1 + 2kprk3 2kp T rk1 T = 0,
kp2 rk1
(5.98)
which yields
rk1 4T rk2 + 4T2 rk3 = 0.
(5.99)
142
(5.100)
(5.101)
which yields
4T2 rk2 4T rk1 + rk = 0
(5.102)
or finally
4T2
p2
Tni 4T
p1
i=1
Tni +
i=1
p
Tni = 0.
(5.103)
i=1
p3
Tni 4T Tn p
p2
i=1
Tni + Tn p
i=1
p1
Tni = 0
(5.104)
i=1
or finally
4T2
p3
i=1
Tni 4T
p2
i=1
Tni +
p1
Tni Tn p = 0
(5.105)
i=1
(5.106)
143
4 Tn1 + Tn2 + Tn3 T2 4 Tn1 Tn2 + Tn2 Tn3 + Tn1 Tn3 T + Tn1 Tn2 Tn3 = 0.
(5.107)
TypeIII control loops:
4T2 4 Tn1 + Tn2 T + Tn1 Tn2 = 0.
(5.108)
(5.109)
Note that (5.108) and (5.109) are equal to (4.16), respectively. In similar fashion with
typeIII control loops and for the sake of simplicity of the analysis, if we choose
Tn1 = Tn2 = = Tn p1 = nT
(5.110)
the respective open Fol (s) and closedloop T (s) transfer functions are given by
Fol (s)
(1 + nT s) p1
p
2n p1 T s p (1 + T s)
(5.111)
and
T (s)
(1 + nT s) p1
p+1
2n p1 T s p+1
+ 2n p1 T s p + (1 + nT s) p1
(5.112)
The optimal value of parameter n depends on the type of the control loop we want
to design. If we substitute (5.110) into (5.107)(5.109), we have consequently,
TypeV control loops:
n 2 n 2 16n + 24 T4 = 0
n opt = 14.32.
(5.113)
n opt = 10.89.
(5.114)
n opt = 7.46.
(5.115)
With respect to the above, for the design of a typeIV control loop, a PID type
controller of three zeros in its transfer function is required. Therefore, if we chose
144
(5.116)
4n(n 2)
4n (n 2)
T =
T
n 2 8n + 4
(n 0.536)(n 7.464)
(5.117)
Based on the above, the corresponding Fol (s) and T (s) transfer functions are given
by
3
4n (n 2)T3 s 3 + n 2 (n 2 12)T2 s 2
+2n 2 (n 6)T s + (n 0.536)(n 7.464)
,
(5.118)
Fol (s) =
8n 3 (n 2)T5 s 5 + 8n 3 (n 2)T4 s 4
T (s) =
a5
s5
b3 s 3 + b2 s 2 + b1 s + b0
+ a4 s 4 + a3 s 3 + a2 s 2 + a1 s + a0
(5.119)
where
(n 2 12) 2
T
n2
(5.120)
(n 6)
(n 0.536) (n 7.464)
T , b0 =
n2
n2
(5.121)
b3 = 4n 3 (n 2) T3 , b2 = n 2
b1 = 2n 2
and
a5 = 8n 3 (n 2)T5 , a4 = 8n 3 T4
a3 = 4n 3 T3 , a2 = n 2
a1 = 2n 2
(n 2 12) 2
T
n2
(n 0.536) (n 7.464)
(n 6)
T , a0 =
.
n2
n2
(5.122)
(5.123)
(5.124)
145
(a)
n = 10.89
y r ( )
n = 7.5
n = 7.5
y o ( )
n = 10.89
= t/ T
(b)
S ( ju )
 T ( ju )
n = 10.89
n = 7.5
u = T
There, it is shown how the controllers zeros are affected in case of variations in design
parameter n. Similar results are also observed in the time domain, Fig. 5.11a.
The step response of the typeIV closedloop control system exhibits an overshoot
of 50 %, which is justified by the phase margin ( = 32 < 45 ) of the openloop Fol (s) frequency response, Fig. 5.12. For decreasing the overshoot of the final
closedloop control system, the two degrees of freedom controller structure is again
be exploited. If n = 10.89, then the closedloop transfer function in terms of time
constants form is given by
T (s) =
where
N1 (s)
,
D1 (s)D2 (s)D3 (s)
(5.125)
146
 Fol ( j )
n = 10.89
n = 10.89
n = 7.5
1
10.89
m = 32
u = T
Fig. 5.12 Openloop frequency response of a typeIV control loop for various values of parameter n
(5.126)
(5.127)
D3 (s) = (14.75)
(5.128)
T2 s 2
+ 1.9(14.75)T s + 1.
(1 + 2.3T s) (14.75)2 T2 s 2 + 1.9(14.75)T s + 1
(1 + 10.89T s)3 (1 + T s)
(5.129)
(5.130)
In all three examples, it is assumed that the sum T of all time constants of the
process is accurately measured. Time constant
147
ovs 56.5%
ovs 14.75%
with Cex (s )
y r ( )
y o ( )
= t/ T
Fig. 5.13 The effect of the two degrees of freedom controller structure to the step response of the
typeIV closedloop control system
Fig. 5.14 Variations of
parameters Tn1 , Tn2 , Tn3
according to variations of
parameter n
Tn 1
T
Tn 2
T
Tn 3
T
T =
k
Tp j + Tc
(5.131a)
j=1
(5.131b)
includes both plants and controllers unmodeled dynamics. Since typeIII control
loops are designed
Tn1 = Tp1 ,
4 (n 1)
T ,
Tn2 =
n4
Tn3 = nT .
(5.132a)
(5.132b)
(5.132c)
148
Parameter n has been chosen equal to n = 7.46. The integrators time constant
is calculated through
T
Tnr = 2kp kh Tn2 Tn3 T .
Ti = 2kp kh
Tm
p1
(5.133)
r =1
2
(1 + s ) (1 + 0.84s )(1 + 0.78s )(1 + 0.57s )(1 + 0.28s )
(5.134)
is considered. From Fig. 5.15a it is apparent that the typeIII closedloop control system exhibits an undesired overshoot of 87.4 % which is decreased by the filtering the
reference with an external controller Cex1 (s). Settling time remains almost unaltered,
tss = 143 . Note that disturbance rejection has remained the same since the external
controller Cex1 (s) acts only only at the reference signal outside of the control loop.
For manipulating the overshoot of the output, if
Cex2 (s) =
1
(tn2 tn3 )s 2 + (tn2 + tn3 )s + 1
(5.135)
reference filter is to be used, then the overshoot is decreased to 6.2 %. Since the closedloop control system is of typeIII, the output of the process can track perfectly both
ramp and parabolic reference signals, Fig. 5.16. External filter of the form
Cex (s) =
(0.45tn2 tn3
)s 2
1
+ (tn2 + 0.45tn3 )s + 1
(5.136)
2
es
(1 + s )(1 + 0.99s )(1 + 0.57s )(1 + 0.28s )(1 + 0.1s )
(5.137)
is assumed in this example. Note that the proposed control law does not take into
account the effect of the time delay and therefore in this example the robustness
149
(a)
typeIII control loop
ovs = 87.4%
without Cex (s)
ovs = 6.2%
di ( ) = 0.1r ( )
with Cex2 (s)
do ( ) = 0.1r ( )
= t/ T p1
(b)
command signal u ( )
u ( )
di ( ) = 0.1r ( )
do ( ) = 0.1r ( )
= t/ T p1
of the method to model uncertainties is also tested. If no external filter is used for
reference tracking, the control loop exhibits an overshoot of 100.4 %, Fig. 5.17a.
The use of both Cex1 (s), Cex2 (s) eliminates the overshoot to 9.4 and 0 %, respectively, Fig. 5.17a. Disturbance rejection remains unaltered. Cex2 (s) is of the same
form as in the previous example. Note that control signal u( ) is improved in case
the reference signal is filtered, Figs. 5.15b and 5.17b. External filter of the form
1
Cex1 (s) =
2
0.45tn2 tn3 s + tn2 + 0.45tn3 s + 1
is used for decreasing the overshoot of the output.
(5.138)
150
(a)
typeIII control loop
y ( )
r ( )=
ramp response
= t/ T p1
(b)
y ( )
parabolic response
r ( )= 2
= t/ T p1
1.34(1 0.771s )
(1 + s )(1 + 0.33s )(1 + 0.12s )(1 + 0.056s )(1 + 0.038s )
(5.139)
is adopted for testing the robustness of the proposed control law. The step response
of (5.139) is presented in Fig. 5.18. In addition, in Fig. 5.19a, b the step response of
the output y( ) and the control signal u( ) are presented, respectively. If no external
filter is used, the overshoot of the step response is 59.9 %. Since this is undesirable, if
r (s) is filtered by Cex1 (s), Cex2 (s) then the overshoot is reduced to 0 % in both cases.
Output and input disturbance rejection remain unaltered since the external filter does
not participate into
151
(a)
typeIII control loop
ovs = 100.4%
without Cex (s)
di ( )= 0.1r ( )
y ( )
do ( )= 0.1r ( )
= t/ T p1
(b)
command signal u ( )
u ( )
with Cex 1 (s)
u ( )
do ( )= 0.1r ( )
= t/ T p1
Si (s) =
y (s)
di (s)
(5.140)
So (s) =
y (s)
di (s)
(5.141)
and
Cex1 (s) =
2
0.45tn2 tn3 s + tn2 + 0.45tn3 s + 1
is used for decreasing the overshoot of the output.
(5.142)
152
= t/ T p1
(a)
ovs = 59.9%
y ( )
= t/ T p1
(b)
command signal u ( )
without Cex (s)
u ( )
with Cex 2 (s)
with Cex 1 (s)
di ( )= 0.1r ( )
u ( )
do ( )= 0.1r ( )
= t/ T p1
153
n!
.
s n+1
(5.143)
3!
s 3+1
4!
s 4+1
(5.144a)
(5.144b)
respectively. According to the proposed theory for a typeIV, typeV control loop the
proposed PID type controllers are given by
C(s) =
1 + Tn1 s 1 + Tn2 s 1 + Tn3 s 1 + Tn4 s
C(s) =
,
Ti s 4 (1 + Tc1 s)(1 + Tc2 s)
(5.145)
(5.146)
respectively. For determining parameters Tn1 , Tn2 , Tn3 , Tn4 , Ti in (5.145) according
to the proposed theory, we set Tn4 = Tp1 and Tn1 = Tn2 = nT according to (5.110).
For that reason, (5.107) becomes
4(2nT + Tn3 ) 4(n 2 T + 2nTn3 ) + n 2 Tn3 = 0
(5.147)
or finally
Tn3 =
4n(n 2)
T .
(n 2 8n + 4)
(5.148)
(5.149)
In a similar fashion, for the (5.146) PID type controller and since the control loop
is of typeV, we set Tn5 = Tp1 and Tn1 = Tn2 = Tn3 = nT . Accordingly, (5.106)
becomes
154
4(3n 2 T2 + 3nT Tn4 )T 4(n 3 T3 + 3n 2 T2 Tn4 ) + n 3 T2 Tn4 = 0
(5.150)
4n 2 (n 3)
4n (n 3)
T = 2
T .
2
n 12n + 12
n n 12n + 12
(5.151)
(5.152)
The process in this example is defined by (5.154). The respective response to r (t) =
t 3 and r (t) = t 4 reference signals for the typeIV and the typeV control loop are
presented in Fig. 5.20b.
1.23
(1 + s )(1 + 0.872s )(1 + 0.367s )(1 + 0.287s )(1 + 0.11s )
. (5.153)
From Fig. 5.21a, b it is apparent that if an error of 30 % when measuring Tp1 occurs,
a small change is observed in the overshoot of the closed loop control system. In
addition, both input and output disturbance rejection remain almost unaltered.
5.3.3.2 Comparison Between a TypeI and a TypeIII Control Loop
For showing the advantages of designing a higher order faster control loop, the
following process
G(s ) =
1.23
(5.154)
(1 + s )(1 + 0.992s )(1 + 0.692s )(1 + 0.139s )(1 + 0.107s )
is adopted. For this process, a typeI, typeIII closed control loop is designed. For
designing the PID typeI control loop the conventional Magnitude Optimum criterion
155
(a)
y ( )
parabolic response
y ( )
r ( )= 3
= t/ T p1
(b)
y ( )
y ( )
r ( )= 4
= t/ T p1
Fig. 5.20 Response of a typeIV and a typeV control loop. Parameter n has been chosen equal
to n = 14.32 according to (5.114). a Response of the typeIV control loop to reference signal
r ( ) = 3 , parameter n has been chosen equal to n = 10.89 according to (5.114). b Response of
the typeV control loop to reference signal r ( ) = 4
(see 3.2.4 and 5.2.1) is employed. Note that for determining controllers zeros, exact
pole zero cancellation has to take place (see 3.2.4 and 5.2.1), [10]. From Fig. 5.22 it
is apparent that the typeI control loop fails to track both the ramp and the parabolic
reference signal exhibiting nonzero steady state velocity and acceleration error.
156
(a)
a = 0.3
a=0
do ( )= 0.1r ( )
y ( )
a = 0.3
di ( )= 0.1r ( )
= t/ T p1
(b)
without Cex
command signal u ( )
di ( )= 0.1r ( )
u ( )
do ( )= 0.1r ( )
= t/ T p1
1
(1 + s )(1 + as )(1 + a 2 s )(1 + a 3 s )(1 + a 4 s )
(5.155)
is adopted. As proved in Sects. 4.2.3 and 5.2.1 the proposed control law depends on
polezero cancellation and time constant T which models the process unmodeled
dynamics (poles of the process far from the origin), see (4.5) and (5.3) where T =
Tc + Tp and Tp is the process parasitic time constant and Tc Tp . In Fig. 5.23
the process is modeled by a = 0.15 containing a relatively large dominant time
157
(a)
y ( )
typeII control loop
r ( ) =
= t/ T p1
(b)
y ( )
= t/ T p1
constant and in the next case a = 0.6 the parasitic time constant of the process is
comparable to its dominant time constant.
Since
Tp
=
a j,
Tp1
4
(5.156)
j=1
it is apparent that when a = 0.15 then Tp = Tp1 4j=1 a j = 0.1764Tp1 and
when a = 0.6 then Tp = Tp1 4j=1 a j = 1.3056Tp1 . The conclusion according
to Fig. 5.23 is that the less accurate the model of the process in terms of zeros,
time delay, poles compared to the dominant time constant (T Tp j ), the poorer the
performance becomes, (see settling time of the output and input disturbance rejection
Fig. 5.23).
158
Fig. 5.23 Step response of
the PID typeIII control loop
when a = 0.15 and a = 0.6
for a process defined by
(5.155)
= t/ T p1
5.4 Summary
Explicit PID tuning rules have been presented towards the design of typeIII control
loops and regardless of the process complexity in Sect. 5.2.2. The proposed control
law is considered feasible for many real world applications since it is of PID type. For
the definition of the optimal control law, the powerful principle of the Symmetrical
Optimum criterion was adopted. The advantage of typeIII control loops compared
to typeI, typeII (control of integrating processes) is obvious since the higher the
type of the control loop, the faster reference signals can be tracked by the output of
the process.
This advantage has been justified through simulation examples for the control of
a variety of process models as show in Sect. 5.2.3. It was shown that the conventional PID tuning (typeII control loops, current state of the art) via the Symmetrical
Optimum criterion fails to track parabolic reference signals. Even in cases when the
conventional tuning is used for the design of a typeIII control loop, the performance
is still suboptimal especially in cases when the process complexity is increased.
In contrast to this, the proposed PID control law tracks with zero steady state
position, velocity and acceleration error step, ramp, and parabolic reference signals regardless of the plant complexity. The robustness of the proposed control
law was also tested to parameters variations showing finally promising results, see
Sect. 5.2.3.4. To this end, control engineers are capable of designing typeIII control
loops, firstly on a simulation level before going finally on a real time implementation.
Moreover, the Symmetrical Optimum criterion has been extended for the design of
type p control loop in Sect. 5.3. Based on the design of typeIII control loops (design
with polezero cancellation), the proposed control law was extended for tuning PID
type p control loops so that tracking of faster reference signals is achieved.
5.4 Summary
159
The development of the proposed control is carried out in the frequency domain
where the transfer function of the process involves the dominant time constants and
the plants unmodeled dynamics, see Sect. 5.3.1. Once more, the proposed theory
has been evaluated for the control of representative plants met in many industry
applications, see Sect. 5.3.3. The robustness of the proposed control law achieves
promising results (see Sect. 5.3.3.3) also for the control of processes with parameters
the control law disregards, such as nonminimum phase processes and processes with
time delay.
References
1. strm KJ, Hgglund T (1995) PID controllers: theory, design and tuning, 2nd edn. Instrument
Society of America, Research Triangle Park
2. Kessler C (1958) Das symmetrische optimum. Regelungstechnik, pp 395400 and 432426
3. Margaris NI (2003) Lectures in applied automatic control (in Greek), 1st edn. Tziolas
4. Oldenbourg RC, Sartorius H (1954) A uniform approach to the optimum adjustment of control
loops. Trans ASME 76:12651279
5. Papadopoulos KG, Margaris NI (2012) Extending the symmetrical optimum criterion to the
design of PID typep control loops. J Process Control 12(1):1125
6. Papadopoulos KG, Mermikli K, Margaris NI (2011a) Optimal tuning of PID controllers for
integrating processes via the symmetrical optimum criterion. In: 19th mediterranean conference
on control & automation (MED), IEEE, Corfu, Greece, pp 12891294
7. Papadopoulos KG, Papastefanaki EN, Margaris NI (2011b) Optimal tuning of PID controllers
for typeIII control loops. In: 19th mediterranean conference on control & automation (MED),
IEEE, Corfu, Greece, pp 12951300
8. Papadopoulos KG, Papastefanaki EN, Margaris NI (2012a) Automatic tuning of PID typeIII
control loops via the symmetrical optimum criterion. In: International conference on industrial
technology, (ICIT), IEEE, Athens, Greece, pp 881886
9. Papadopoulos KG, Tselepis ND, Margaris NI (2012b) Revisiting the magnitude optimum
criterion for robust tuning of PID typeI control loops. J Process Control 22(6):10631078
10. Papadopoulos KG, Papastefanaki EN, Margaris NI (2013) Explicit analytical PID tuning rules
for the design of typeIII control loops. IEEE Trans Ind Electron 60(10):46504664
11. Poulin E, Pomerleau A (1999) PI settings for integrating processes based on ultimate cycle
information. IEEE Trans Control Syst Technol 7(4):509511
12. Preitl S, Precup RE (1999) An extension of tuning relation after symmetrical optimum method
for PI and PID controllers. Automatica 35(10):17311736
Chapter 6
Abstract In this chapter, analytical tuning rules for digital PID typeI, typeII,
typeIII control loops are presented. Controller parameters are determined explicitly
as a function of the process parameters and the sampling time Ts of the controller. For
developing the proposed theory in typeI, typeII control loops, a generalized singleinput singleoutput stable process model is used consisting of npoles, mzeros plus
unknown time delayd. As far as typeIII control loops is concerned the principle
of polezero cancellation according to the method proposed in Sect. 5.2.1, see [3], is
followed. The derivation of the proposed PID control law lies in the principle of the
Magnitude Optimum criterion and the optimization conditions proved in Appendix
A.1 are used for extracting the explicit solution. For all control loop types, a performance comparison is presented in terms of simulation examples. The comparison
focuses on the effect of the sampling time Ts to the control loops response both in
the time and frequency domain.
1 + s X + s2Y
sTi
1 esTs
sTs
.
(6.2)
Note in this case that C (s) stands for the digital representation of the PID control
law and CZOH (s) stands for the zero order hold unit. Ts stands for the sampling period
of the controller.
Springer International Publishing Switzerland 2015
K.G. Papadopoulos, PID Controller Tuning Using the Magnitude Optimum Criterion,
DOI 10.1007/9783319072630_6
161
162
r ( s)
+

di ( s )
controller
Ts
C ( s)
e( s)
+
CZOH ( s) u ( s)+
do ( s)
kp
G ( s)
kh
++
++
y ( s)
n o ( s)
Fig. 6.1 Block diagram of the closedloop control system. G(s) is the plant transfer function, C(s)
is the controller transfer function, r (s) is the reference signal, y(s) is the output of the control loop,
yf (s) is the output signal after kh , do (s) and di (s) are the output and input disturbance signals respectively and n r (s), n o (s) are the noise signals at the reference input and process output respectively.
kp stands for the plants dc gain and kh is the feedback path
Normalizing all time constants in the frequency domain with the sampling period
Ts and substituting with s = sTs results in
s m zm + + s 4 z4 + s 3 z3 + s 2 z2 + s z1 + 1
es d
G(s ) = kp
n
n1
5
4
3
s rn + s rn1 + + s r5 +s r4 + s r3
+s 2 r2 + s r1 + 1
(6.3)
C(s ) = C (s )CZOH (s ) =
1 + s x + s2 y
s ti
1 es
s
(6.4)
for the controller, respectively. Let it be noted that for normalizing both (6.1), (6.2)
p
the substitutions x = TXs , y = TY2 , ti = TTsi , d = TTds , r j = T ij , j = 1, . . . , n, and
s
es 1
z1
=
s =
.
z
es
(6.5)
Since z = es , the digital PID type controller takes now the form
C(s ) = C (s )CZOH (s )
1 (1 + x + y)e2s (x + 2y)es + y
ti
es (es 1)
(6.6)
163
For simplifying the calculation of the final closed loop transfer function, the substitutions
x = 2 y x 2 and y = x y + 1
(6.7)
1 (1 es )x + (e2s 1) y + 1
.
ti
es (es 1)
(6.8)
In addition, the respective open Fol (s ) and closed loop T (s ) transfer functions
become
Fol (s ) = kh C(s )G(s )
(6.9)
or
Fol (s ) = kh
kp
ti
s m zm + + s 3 z3
s
2s
(1 e )x + (e 1) y + 1
+s 2 z 2 + s z 1 + 1
n
s rn + + s 3r3 + s 2 r2 s (d+1) s
e
(e 1)
+s r1 + 1
(6.10)
and
T (s ) =
N (s )
N (s )
C(s )G(s )
=
=
1 + kh C(s )G(s )
D(s )
D1 (s ) + kh N (s )
(6.11)
or
s m zm + + s 3 z3
2s 1) y +1
(1 es )x+(e
2
+s z 2 + s z 1 + 1
.
(d+1) s
s n rn + + s 3 r3
s
ti
(e 1)
e
+s 2 r2 + s r1 + 1
(1 es )x
s m zm + + s 3 z3
2s
+kh kp
+(e 1) y
2
+s z 2 + s z 1 + 1
+1
kp
T (s ) =
(6.12)
Substituting the time delay constant by the all pole series approximation
es = 1 + s +
1 2
1
1
1
1
s + s 3 + s 4 + s 5 + s 6 +
2!
3!
4!
5!
6!
(6.13)
164
yields as proved in the Appendix C.1, that the corresponding polynomials for both
the numerator N (s ) and denominator D(s ) of the closed loop transfer function are
given by
N (s ) = + kp (z 6 + y6 y x6 x)s
+ kp (z 5 + y5 y x5 x)s
+ kp (z 4 + y4 y x4 x)s
5
+ kp (z 3 + y3 y x3 x)s
+ kp (z 2 + y2 y x2 x)s
+ kp (z 1 + 2 y x)s
+ kp
(6.14)
and
6
D(s ) = D1 (s ) + kh N (s ) = + ti q6 + kh kp (z 6 + y6 y x6 x)
s
4
5
5
+ ti q5 s + kh kp (z 5 + y5 y x5 x)
s + ti q4 + kh kp (z 4 + y4 y x4 x)
s
3
2
+ ti q3 + kh kp (z 3 + y3 y x3 x)
s + ti q2 + kh kp (z 2 + y2 y x2 x)
s
+ ti + kh kp (z 1 + 2 y x)
s + kh kp .
(6.15)
As it is proved in the Appendix C.1, the final PID control law is defined by
kh = 1
(6.16)
1
ti = 2kh kp r1 + d z 1 x
2
(6.17)
x a1 y = b1 and x a2 y = b2
(6.18)
where
a1 =
b1 =
a2 =
2(q22 q3 ) (q2 y2 y3 )
(q22 q3 ) (q2 x2 x3 )
2
q2 z 4 q3 z 3 + q4 z 2
q3 2q2 q4 + 2q5 (q2 z 1 ) +
q5 z 1 z 5 + q6
b2 =
.
(q3 x3 )q3 (q4 x4 )q2 (q2 x2 )q4
+ (q5 x5 )
(6.19)
(6.20)
(6.21)
(6.22)
165
a1 b2 a2 b1
,
a1 a2
y =
b2 b1
.
a1 a2
(6.23)
From the definition of the integrators time constant (6.17) or (C.48) it is critical to
point out that
Ti
1
= 2kh kp r1 + d z 1 x
Ts
2
(6.24)
(6.25)
i=1
In other words as it was proved in (3.72) and (C.36), integrators time constant is
equal
1
Tidig = Tian 2kh kp Ts ,
2
(6.26)
where Tidig and Tian are the optimal values for the integrators time constant regarding
the analog and digital design, respectively.
166
In the sequel, two curves are plotted within each figure. Given the transfer function
of the plant G(s), the response of the output y( ) and the command signal u( ) are
investigated. To do this, both control loops are normalized with the sampling time
Ts , the digital controller is implemented according to the relation s = sTs .
0.5
(1 + 2s ) (1 + 1.56s ) (1 + 1.34s ) (1 + 0.67s ) (1 + 0.6s )
(6.27)
where ratio Tps1 = 2. From Fig. 6.2b it is clear that output disturbance rejection
regarding analog control action outperforms in terms of settling time the response
coming from the digital control action. This conclusion holds also for the step
response and input disturbance rejection, see Fig 6.2a.
In the case of output disturbance rejection, settling time is tss = 12 and tss =
27.5 regarding analog and digital control respectively.
The response of the output y( ) is also reflected by the command signal u( )
response, see Fig. 6.3a, b. There, it is shown that the fast input disturbance suppression
observed in Fig. 6.2b is achieved by the strong and fast command effort observed in
Fig. 6.3a. The same conclusion holds also for the output disturbance rejection applied
at = 76, see Figs. 6.2a and 6.3b.
0.5
(1 + 10s ) (1 + 7.79s ) (1 + 6.73s ) (1 + 3.39s ) (1 + 2.97s )
(6.28)
for which Tps1 = 20 holds by. The sampling time of the controller has been decreased
to twenty times less the dominant time constant of the process. In this case, it is
apparent from the response in the time domain, see Fig. 6.4 the reference tracking,
input and output disturbance rejection exhibit almost the same behavior, see Fig. 6.4a.
This result is also reflected by the response of the command signal at the presence
of an input disturbance di ( ) = 0.25r ( ).
However, the response in the frequency domain both for T ( ju), S( ju), see
Fig. 6.5 shows that, the region for which T ( ju) 1 becomes shorter in case when
T
the sampling time of the controller is chosen such that Tps1 = 20. This result is against
the target of the Magnitude Optimum criterion the goal of which it to try to maintain
T ( ju) 1 in the widest possible frequency range.
167
(a)
T p1
Ts
= 2
di ( ) = 0.25r ( )
t ss = 12 t ss = 27.5
analog control action
y ( )
PID control
= t/ Ts
(b)
do ( ) = 0.75r ( )
PID control
output disturbance rejection
y ( )
= t/ Ts
Fig. 6.2 Comparison of the analog and digital control action for the control of the process defined
by (6.27). a Response of the output y( ) in the presence of input di ( ) = 0.25r ( ) and output
do ( ) = 0.75r ( ) disturbance. b Output disturbance rejection
On the other hand, the frequency range for which complementary sensitivity is
T
S( ju) 0, is shorter in the case where Tps1 = 2 compared to the region in the case
T
where Tps1 = 20. As mentioned in Sect. 2.6 this behavior is not desired, since such a
control loop becomes more sensitive to possible disturbances in the low frequency
range region.
For that reason, control engineers have to find a compromise as far as the choice
of the sampling time Ts of the controller is concerned. This is also the goal of this
chapter. The introduction of the sampling time Ts in the control action along with
168
(a)
PID control
T p1
Ts
= 2
u ( )
analog control action
(b)
= t/ Ts
T p1
Ts
= 2
PID control
= t/ Ts
Fig. 6.3 Command signal response in the presence of input and output disturbance. Comparison of
the analog and digital control action for the control of the process defined by (6.27). a Response of
the command signal u( ) in the presence of input disturbance. b Response of the command signal
u( ) in the presence of output disturbance
the explicit definition of the PID controller parameters allows for such accurate
investigation before the final integration of the control law within a realtime system.
169
(a)
do ( ) = 0.75r ( )
di ( ) = 0.25r ( )
y ( )
PID control
= t/ Ts
(b)
di ( ) = 0.25r ( )
= t/ Ts
Fig. 6.4 Comparison of the analog and digital control action for the control of the process defined
by (6.28). a Response of the output y( ) in the presence of input di ( ) = 0.25r ( ) and output
do ( ) = 0.75r ( ) disturbance. b Response of the command signal u( ) at the presence of an output
disturbance
G s =
T
0.5
(1 + 2s ) (1 + 1.56s ) (1 + 1.34s ) (1 + 0.67s ) (1 + 0.6s )
(6.29)
for which Tps1 = 2. From Fig. 6.6 it is apparent that in case when the the actuators
gain changes by +20 % the overshoot of the step response increases from 6 % to
21 %.
In the opposite case, when the error is = 20 % the step response of the control
loop exhibits an overshoot of 0 % while the rise time has been increased. Let it be
170
 S ( ju )
Tp
 T ( ju )
 S ( ju )
T p1
Ts
T p1
Ts
1
Ts = 2
 T ( ju )
= 20
= 2
 T ( ju )
u = Ts
Fig. 6.5 Frequency response of the digital control loop when for the controlled of the same process
T
T
the sampling time of the digital PID controller is chosen such that Tps1 = 2 and Tps1 = 20. Decrease
of the sampling time Ts of the controller decreases the bandwidth of T ( ju) for which T ( ju) 1
= 0.2
kp = k p ( 1 + )
= 0
y ( )
= 0.2
PID control
= T p 1 / Ts
Fig. 6.6 Effect on the step response of the closed loop control system due to changes on the plants
dc gain kp
noted that such an investigation is critical in the field of electric motor drives, since
the gain kp in the frequency domain stands for the modulation policy followed in
vectorcontrolled electrical drives.
171
s m m + + s 4 4 + s 3 3 + s 2 2 + s1 + 1
s n pn + s n1 pn1 + + s 5 p5 + s 4 p4 + s 3 p3
+s 2 p2 + s p1 + 1
esTd
(6.30)
is introduced where n > m. Note that since the control loop is of typeII, two pure
integrators must be included in the open loop Fol (s) transfer function. As a result,
the proposed PID type controller is given by
1 + s X + s2Y
s 2 Ti
1 esTs
s
(6.31)
for which the second integrator is introduced by the control action since it is of IPID.
Let it be noted that the same analysis holds for the control of integrating processes
since one integrator is introduced by the process itself and one more by the controller
from the PID control action. For calculating the closed loop transfer function T (s),
both the controller and the process are normalized with the sampling period Ts of
the zero order hold. In that after substituting with s = sTs , relations (6.30), (6.31)
become
s m zm + + s 4 z4 + s 3 z3 + s 2 z2 + s z1 + 1
es d
G(s ) = kp
(6.32)
n
n1
5
4
3
s rn + s
rn1 + + s r5 + s r4 + s r3
+s 2 r2 + s r1 + 1
and
C(s ) = C (s )CZOH (s ) = Ts
for which x =
i
Tsi
X
Ts ,
y =
Y
,t
Ts2 i
Ti
Ts ,
1 + s x + s2 y
s 2 ti2
d =
Td
Ts
1 es
s
and r j =
pj
Ts
(6.33)
, j = 1, . . . , n,
es 1
z1
=
z
es
1
Ts z
Ts es
=
=
2
s2
(z 1)2
(es 1)
s =
(6.34a)
(6.34b)
172
(6.35)
or
Ts
C(s ) = 2
ti
(x + y)e2s (x + 2y Ts )es + y
(es 1)
(6.36)
or finally
x
y
+
2
T
Ts
Ts
C(s ) = s2
ti
e
2s
x
y
y
+ 2 1 es +
Ts
Ts
Ts
.
2
(es 1)
(6.37)
x
y
+2 1
Ts
Ts
(6.38)
x
y
+
Ts
Ts
(6.39)
and
y =
takes place. This results in
x
= 2 y x 1
Ts
(6.40)
y
= x y + 1.
Ts
(6.41)
and
(1 es )x + (e2s 1) y + 1
(es 1)
(6.42)
With respect to the above, the corresponding open Fol (s ) and closed loop T (s )
transfer functions become
Fol (s ) = kh C(s )G(s )
(6.43)
173
or
kh C(s )G(s ) = kh
Ts2 kp
ti2
s m zm + + s 3 z3 + s 2 z2 + s z1 + 1
(1 es )x + (e2s 1) y + 1
2 (6.44)
s n r n + + s 3 r 3 + s 2 r 2 + s r 1 + 1 es d es 1
and
N (s )
N (s )
C(s )G(s )
=
=
1 + kh C(s )G(s )
D(s )
D1 (s ) + kh N (s )
m
kp (s z m + + s 3 z 3 + s 2 z 2 +
s z 1 + 1)
(1 es )x + (e2s 1) y + 1
=
.
ti2 (s n rn + + s 3r3 + s 2 r2 + s r1 + 1)es d (es 1)2
kh k (s m z m + s 3 z 3 + s 2 z 2 + s z 1 + 1)
+
(1 es )x + (e2s 1) y + 1
T (s ) =
(6.45)
Finally, the corresponding polynomials N (s ), D(s ) for both the numerator and
denominator of the closed loop transfer function are given by
N (s ) =
m
j
kp y j y x j x + z j s
(6.46)
j=0
where y1 = 2, x1 = 1, z 0 = 1 and
n
i
D s =
qi ti2 + kh kp yi y xi x + z i s .
(6.47)
i=0
From the application of the optimization conditions presented in A.1, the final PID
control action as proved in Appendix C.2 is defined by
1
2kh kp (2q3 y2 )
1 2kh kp (x2 q3 )
ti2
x =
D
E
.
0
E)]
0
0
2 [(2D+E)Z +D(AD+B
y
2
(2D+E)
(z 2 + q3 z 1 q4 )
D(2B Z +C D)+Z 2
2
(2D+E)
(6.48)
174
1
1 + s xex + s 2 yex
(6.49)
is added in series after the reference signal r ( ), where x, y are the zeros of the
corresponding PID controller. Specifically, once the x, y controller parameters are
determined by the explicit solution from (4.44), (4.45) for the analog and (6.48) for
the digital control law respectively, the external filter is tuned then according to these
values, with xex = xan , yex = yan and xex = xdig , yex = ydig .
0.8147
(1 + s )(1 + 0.99s )(1 + 0.69s )(1 + 0.13s )(1 + 0.1s )
e0.6s . (6.50)
From (6.50) it is apparent that Ts = T p1 . From Fig. 6.7a it is apparent that the digital
control action leads to an unsatisfactory step response of the control loop. The digital
control loop exhibits an overshoot of around 16 % compared to the analog control
loop which is around 0.5 %.
Of course this response can be finely tuned by properly choosing the parameters
of the external filter Cex (s) in (6.49). From Fig. 6.7b it is clear that spends less effort
in terms of overshoot compared to the analog control action at the presence of input
1
Since the control loop in both cases analog and digital control law is of typeII a high overshoot
at the output y( ) is expected at step changes on the reference signal r ( ).
175
(a)
PID control
digital control action
di ( ) = 0.25r ( )
y ( )
analog control action
do ( ) = 0.75r ( )
(b)
= t/ Ts
PID control
digital control action
do ( ) = 0.25r ( )
di ( ) = 0.75r ( )
u ( )
= t/ Ts
Fig. 6.7 Response of the output y( ) and the controllers command signal u( ) for the control
loop with the plant defined by (6.50). a Response of the output y( ) in the presence of input
di ( ) = 0.25r ( ) and output do ( ) = 0.75r ( ) disturbance. b Response of the command signal
u( ) in the presence of input di ( ) = 0.25r ( ) and output do ( ) = 0.75r ( ) disturbance
and output disturbance. In Fig. 6.8a, the settling time of the analog control loops
response is faster compared to the digital control loops response, which is also
reflected by the effort spent from the digital controller.
6.2.1.2 Sampling Time 10 Less Than the Plants Dominant Time Constant
In this case, the sampling time of the controller has been decreased to 10 less the
dominant time constant of the process. The plants transfer function is given by
176
(a)
PID control
do ( ) = 0.75r ( )
y ( )
(b)
= t/ Ts
command signal
PID control
= t/ Ts
Fig. 6.8 Output disturbance rejection and command signal response at the presence of output
disturbance. a Output disturbance rejection. b Command signal response at the presence of output
disturbance do ( ). Control effort in the case of digital control action is less aggressive compared to
the analog control action
G(s ) =
e9.4s
(6.51)
for which two zeros also exist. The response of both the control loops output
y( ) and the controllers command signal is presented in Fig. 6.9. From there it
is clear that the digital controller spends less effort, see Fig. 6.9b for achieving
almost the same output response in terms of settling time of disturbance rejection, see
Fig. 6.9a.
177
(a)
di ( ) = 0.25r ( )
PID control
(b)
= t/ Ts
command signal
u ( )
digital control action
PID control
= t/ Ts
Fig. 6.9 Response of the output y( ) and the controllers command signal u( ) for the control loop
with the plant defined by (6.51). a Response of output y( ) in the presence of output disturbance.
b Command signal response at the presence of output disturbance do ( )
0.96
(1 + 1s )(1 + 0.91s )(1 + 0.72s )(1 + 0.7s )(1 + 0.03s )
(6.52)
for which the nominal dc gain of the process is kp = 0.96. Initially, the digital controller is tuned according to (6.48). Let it be noted that the plants dc gain is involved
178
(a)
do ( ) = 0.75r ( )
PID control
= 0
= 0.2
y ( )
(b)
= t/ Ts
command signal
digital control action
= 0
= 0.2
u ( )
= 0.2
PID control
= t/ Ts
Fig. 6.10 Variation of the plants dc gain kp = kp (1 + ), = 20 %. a Effect of the plants dc kp
gain variation to the quality of response of y( ). b Effect of the plants dc kp gain variation to the
quality of response of u( )
only within the integrators closed form expression, see (C.100) after parameters
x, y or x,
y are optimally determined.
To this end, the change on the plants dc gain affects only the tuning of the
integrators time constant. In this example, the first tuning of the digital PID controller
is done based on (6.48) and the nominal measured gain kp = 0.96 whereas in the
other case, the controller stays tuned with its initial nominal value and the plants dc
gain changes by 20 %, kp = kp (1 + ).
In Fig. 6.10a, b the response of the output y( ) and the command signal u( )
is presented. From there it is apparent that variations of the plants dc gain up to
= 20 % cause a change in the settling time of output disturbance suppression by
179
30.09 %. Initially, settling time is tss = (156 127) = 29 whereas in the case
where = 20 %, settling time is tss = (171 127) = 44 .
1
.
sTm (1 + sT p1 )(1 + sTp )
(6.53)
(1 esTs )
sTs
(6.54)
where Ts stands for the controllers sampling period. Again all time constants in the
control loop are normalized in the frequency domain with the sampling period Ts
and the substitution s = sTs takes place. In that, (6.53) and (6.54) become
G(s ) =
1
st
(1 + s t
(6.55)
p1 )(1 + s tp )
and
C(s ) = C (s )CZOH (s ) = Ts
for which ti =
Tm
Ts , t p1
(1 + s tn )(1 + s tv )(1 + s tx )
s 2 ti (1 + s tc1 )(1 + s tc2 )
(1 es )
s
(6.56)
Tc
Tn
Tx
Tv
2
= Ts , tn = Ts , tv = Ts , tx = Ts , tm =
Tc
Ti
1
Ts , tc1 = Ts , tc2
T
tp = Tsp has been set.
= Tps1 ,
In similar fashion with the analog design procedure in Section B.3, the open loop
transfer function Fol (s ) is given by
Fol (s ) = kp kh C(s )G(s )
1 + s tn 1 + s tv 1 + s t x
(1 es )
= Ts
s
s 2 ti 1 + s tc1 1 + s tc2
kp kh
.
s tm (1 + s t p1 )(1 + s tp )
(6.57)
180
For moving from the L{.} to the Z{.} domain, the substitutions below are considered
es
z
1
=
,
=
s
z 1
es 1
1
Ts z
Ts es
=
=
.
2
s2
(z 1)2
(es 1)
(6.58)
(6.59)
es
1 + s tn 1 + s tv 1 + s t x
ti 1 + s tc1 1 + s tc2
(es 1)
(6.60)
(6.61)
Fol (s ) =
kp kh es (1 + s tn )(1 + s tv )
(6.62)
and after setting kp = kp Ts2 . In similar fashion with the analog design it is set
tc1 tc2 0 and tc = tc1 + tc2 .
This results in (1 + s tp )(1 + s tc1 )(1 + s tc2 ) = (1 + s tp )(1 + s tc ).
Moreover if tc tp 0 and t = tc + tp then (6.62) becomes equal to
Fol (s ) =
kp kh es (1 + s tn )(1 + s tv )
s tm ti (1 + s t )(es 1)
(6.63)
kp es (1 + s tn )(1 + s tv )
T (s ) =
kp C(s )G(s )
=
1 + kp kh C(s )G(s )
s tm ti (1 + s t )(es 1)
kp es (1 + s tn )(1 + s tv )
1 + kh
2
s tm ti (1 + s t )(es 1)
kp es (1 + s tn )(1 + s tv )
s tm ti (1 + s t )(es 1) + kh kp es (1 + s tn )(1 + s tv )
(6.64)
181
Since (6.64) is in the form of (A.1), the optimization conditions (A.9)(A.12) can be
applied for determining the optimal digital PID control law.
In Appendix C.3, it is proved that parameters kh , tx , tn , tv , ti are determined
finally by
t p1
kh
tn =
2[nt2 (4 n) 2(2B 1)]
tv
nt
n v
(6.65)
tm
where variables B, are also process dependent parameters.
Cexan =
(6.66)
Cexdig
(6.67)
is utilized.2 The presence of the external filter is necessary for higher than typeI
control loops to avoid high overshoot on the output y( ) when step changes on r ( )
occur, see Sects. 4.2.3, 5.2.3 also [2, 3], where the 2DoF (two Degree of Freedom
controller) is described.
Parameter is chosen so that the overshoot of y( ) satisfies a certain value (depending on the
application) when step changes on the reference signal r ( ) occur.
2
182
0.1
0.8s (1 + 2s )(1 + 1.6s )
(6.68)
is introduced in this example. The calculated digital and analog controllers according
to the theory presented in Sects. 5.2.1, 6.3 are defined by
Can (s ) =
(1 + s 12.68)(1 + s 12.69)(1 + s 2)
(1 + s tn )(1 + s tv )(1 + s tx )
=
s 2 ti (1 + s tc1 )(1 + s tc2 )
s 2 171(1 + s 0.1)(1 + s 0.1)
(6.69)
Cdig (s ) = Ts
(1 + s tn )(1 + s tv )(1 + s tx )
(6.70)
From Fig. 6.11a, it is apparent that when the ratio Tps1 = 2, the response of the
digital control action is oscillatory compared to the analog control action and exhibits
an undesired overshoot of 22 %. The same unsatisfactory behavior is observed also
as far as output disturbance rejection is concerned, see Fig. 6.11a.
This is the result of the oscillating command signal which comes out of the digital
controller, see Fig. 6.12a. In the frequency domain, the response of sensitivity S( ju)
and complementary sensitivity T ( ju) is shown in Fig. 6.12b where the two systems
exhibit almost the same behavior.
In Fig. 6.13a the sampling time of the control loop has been increased, see ratio
T p1
Ts = 500. In this case, the output disturbance rejection has been significantly
improved, see Fig. 6.13b but on the contrary, the region where the magnitude of
complementary sensitivity T ( ju) 1 has been also reduced.
This contradicts of course with the principle of the Magnitude Optimum criterion,
for which the closed loop control system is designed such that T ( j) 1 in the
widest possible frequency range.
183
(a)
ovs 22%
step response
y ( )
analog control action
digital control action
T p1
Ts
= 2
= t/ Ts
(b)
y ( )
T p1
Ts
= 2
= t/ Ts
Fig. 6.11 Control of an integrating process defined by (6.68). a Step response of the analog and
digital control action. b The control loops output y( ) in the presence of input r ( ) and do ( )=
0.25r ( ) output disturbance at = 500
G 2 (s ) =
0.1
s (1 + 5s )(1 + 4.5s )
e2.5s
(6.71)
and the calculated analog and digital PID control actions are given by
Can (s ) =
=
(1 + s tn )(1 + s tv )(1 + s tx )
s 2 ti (1 + s tc1 )(1 + s tc2 )
(6.72)
184
(a)
T p1
Ts
= 2
u ( )
= t/ Ts
(b)
frequency response
Tdig ( ju )
S an ( ju )
T p1
Ts
S dig ( ju )
= 2
Tan ( ju )
u = Ts
Fig. 6.12 Control of an integrating process defined by (6.68). Analog and digital control loop:
T
ratio Tps1 = 2. An output disturbance do ( ) = 0.25r ( ) is applied at = 500 at the presence
of r ( ). a Step response of the command signal. b Frequency response of sensitivity S( ju) and
complementary sensitivity T ( ju)
and
Cdig (s ) = Ts
(1 + s tn )(1 + s tv )(1 + s tx )
(6.73)
185
(a)
output disturbance rejection
T p1
Ts
= 500
= t/ Ts
(b)
frequency response
 Tan ( ju )
 Tdig ( ju )
 S an ( ju )
 S dig ( ju )
T p1
Ts
= 500
u = Ts
T
Fig. 6.13 Control of an integrating process defined by (6.68). Ratio Tps1 = 500. a An output
disturbance do = 0.25r ( ) is applied at = 0. b Increase of the sampling time T (s) has improved
disturbance rejection in the time domain but reduced the region for which T ( ju) 1 is satisfied
T
Note that in this example the ratio Tps1 has been chosen equal to Tps1 = 5. Within
the digital control action Fig. 6.14b, control effort has a more oscillatory behavior
than the analog control action. In Sect. 6.3.2 the choice of the sampling time is
discussed so that such behavior is avoided.
6.3.1.3 A Nonminimum Phase Process
In this example, the nonminimum phase process described by
G 3 (s ) =
0.1(1 10s )
10s (1 + 50s )(1 + 40s )
(6.74)
186
(a)
step response
do ( ) = 0.25r ( )
y ( )
di ( ) = 0.25r ( )
analog control action
T p1
Ts
= 5
= t/ Ts
(b)
step response
digital control action
di ( ) = 0.25r ( )
do ( ) = 0.25r ( )
= 5
= t/ Ts
Fig. 6.14 Control of an integrating process with time delay half of the dominant time constant
defined by (6.71). Response of the output y( ) and the command signal u( ) in the presence of
input do ( ) = 0.25r ( ) and output disturbance di ( ) = 0.25r ( ). a Step response of the analog
and digital control action of y( ). b Step response of the analog and digital control action of u( )
is controlled both by the analog and digital PID control action which are described
by
Can (s ) =
=
(1 + s tn )(1 + s tv )(1 + s tx )
s 2 ti (1 + s tc1 )(1 + s tc2 )
(6.75)
187
(a)
step response
do ( ) = 0.25r ( )
y ( )
di ( ) = 0.25r ( )
T p1
Ts
= 50
= t/ Ts
(b)
step response
digital control action
di ( ) = 0.25r ( )
= 50
= t/ Ts
Fig. 6.15 Control of an integrating nonminimum phase process defined by (6.74). Response of the
output y( ) and the command signal u( ) in the presence of input do ( ) = 0.25r ( ) and output
disturbance di ( ) = 0.25r ( ). a Step response of the analog and digital control action of y( ). b
Step response of the analog and digital control action of u( )
and
Cdig (s ) = Ts
=
(1 + s tn )(1 + s tv )(1 + s tx )
(6.76)
188
the control effort u( ) is oscillating compared to the analog command signal of the
PID controller.
0.1
.
2s (1 + 10s )(1 + 9.5s )
(6.77)
1
st
(1 + s t
p1 )(1 + s tp )
(6.78)
but the integrators time constant ti in (6.54) stays still tuned according to (6.54)
which is equal to
ti =
(6.79)
In Fig. 6.16 changes to are forced, which are equal to = 0.2. From the step
response Fig. 6.16a and the output disturbance rejection Fig. 6.16b it is apparent that
a non significant change is caused in the settling time and the overshoot of the output
y( ) of the control loop.
G 1 (s ) =
for which tm =
Tm
Ts , t p1
=
T
kp
s tm (1 + t p1 s )(1 + tp s )
T p1
Ts , tp
Tp
Ts
(6.80)
Three different ratios of Tps1 are investigated regarding the performance of the digital
control action compared to the analog control law, both in the time and frequency
domain.
189
(a)
step response
T p1
Ts
= 10
= 0.2
y ( )
= 0
= 0.2
= t/ Ts
(b)
= 0.2
T p1
Ts
= 10
= 0
y ( )
= 0.2
output disturbance rejection
= t/ Ts
Fig. 6.16 Robustness of the proposed digital control law to model uncertainties. A change in the dc
gain of the process kp is provoked of the form k per = kp (1 + ) while the integrators time constant
of the controller stays tuned with its initial nominal value kp . a Step response of the closed loop
control system. b Output disturbance rejection
0.1
.
0.4s (1 + 2s )(1 + 1.8s )
(6.81)
190
(a)
ovs 18%
y ( )
analog control action
digital control action
T p1
Ts
= 2
step response
= t/ Ts
(b)
T p1
Ts
y ( )
= 2
= t/ Ts
Fig. 6.17 Step and ramp response of the digital and analog control loop when
T p1
Ts
= 2 for the
process defined by (6.81). a Step response of the digital and analog control loop when
b Ramp response of the digital and analog control loop when
T p1
Ts
T p1
Ts
= 2.
=2
In Fig. 6.17 the step (Fig. 6.17a), ramp (Fig. 6.17b) response is presented along with
the frequency response and output disturbance rejection. For avoiding the great overshoot at the output of the control loop, a firstorder reference filter has been added
of the form (6.66) and (6.67) for the analog and digital controller.
Parameter has been chosen equal to = 0.75 and parameters tvan = tv in
(6.66), tvdig = tv in (6.67) are coming from the optimal control law (5.10) and (6.65),
respectively.
From Fig. 6.18b it becomes apparent that the frequency response of the closed loop
control system is almost the same both for the analog and the digital implementation.
191
(a)
analog control action
y ( )
= 2
= t/ Ts
(b)
 S an ( ju )
 Tdig ( ju )
 Tan ( ju )
 S dig ( ju )
frequency response
T p1
Ts
= 2
u = Ts
Fig. 6.18 Output disturbance rejection and frequency response of sensitivity S and complementary
sensitivity T for the analog and digital control action when the plant is defined by (6.81). a Output
disturbance rejection for the analog and digital control action. b Frequency response of sensitivity
S and complementary sensitivity T for the analog and digital control action
On the contrary, since Tps1 = 2 output disturbance rejection of the digital control
action is poor compared to the analog control loop, see Fig. 6.18a.
0.1
2s (1 + 10s )(1 + 9.5s )
(6.82)
192
(a)
T p1
Ts
step response
= 10
= t/ Ts
(b)
T p1
Ts
ramp response
= 10
y ( )
r( ) =
analog control action
= t/ Ts
Fig. 6.19 Step and ramp response of the analog and digital control loop for the plant defined by
T
(6.82). Ratio Tps1 = 10. a Step response of the analog and digital control loop. b Ramp response of
the analog and digital control loop
T
from which it is apparent that Tps1 = 10. The time domain performance of the
digital controller has been significantly improved, see Figs. 6.19a, b and 6.20a. On
the contrary, the magnitude of T ( ju) is equal to 0.707 at u = 0.09, see Fig. 6.20b
whereas in Fig. 6.18b this takes place at u = 0.47.
6.3.2.3 Sampling Time 100 Less Than the Plants Dominant Time Constant
The results from the previous example are also confirmed in the following case for
which the process defined by
193
(a)
digital control action
y ( )
T p1
Ts
= 10
= t/ Ts
(b)
 Tdig ( ju )
frequency response
 S an ( ju )
T p1
Ts
 Tan ( ju )
= 10
 S dig ( ju )
u = Ts
Fig. 6.20 Output disturbance rejection and frequency response of sensitivity S and complementary
sensitivity T for the analog and digital control action when the plant is defined by (6.82). Bandwidth
of T has been decreased compared to Fig. 6.18 but time domain performance has been significantly
T
improved compared to Fig. 6.18. Ratio Tps1 = 10. a Output disturbance rejection for the analog and
digital control action. b Frequency response of sensitivity S and complementary sensitivity T for
the analog and digital control action
G 3 (s ) =
0.1
s (1 + 2.5s )(1 + 2.25s )
(6.83)
has been considered. The region of u for which T ( ju) 1 has been reduced even
more while the performance in the time domain delivers similar satisfactory results
T
with the previous example at which Tps1 = 10. The frequency of T ( ju) at which
T ( ju) 0.707 is equal to u = 0.01.
T
From the above analysis, it is apparent that in the case where the Tps1 decreases
(i.e.,
T p1
Ts
= 10, 100), the frequency region where the magnitude of T ( ju) remains
194
(a)
T p1
Ts
step response
= 100
( )
= t/ Ts
(b)
ramp response
y ( )
= 100
r( ) =
analog control action
= t/ Ts
Fig. 6.21 Step and ramp response of the analog and digital control loop for the plant defined by
T
(6.83). Ratio Tps1 = 100. a Step response of the analog and digital control loop. b Ramp response
of the analog and digital control loop
equal to one, T ( ju) 1 becomes smaller compared to magnitude where the ratio
T
T
of the control loop Tps1 is equal to Tps1 = 2. This feature contradicts with the principle
of the Magnitude Optimum criterion, see Section A.1 which requires that T ( ju)
must be maintained equal to the unity in the widest possible frequency range.
As a result, the digital control design has to satisfy all requirements both in the
time and frequency domain. Therefore it has to satisfy an acceptable behavior in
the time domain (step response) and comply with the principle T ( ju) 1 in the
widest possible frequency range. Therefore, sampling time Ts has to be chosen such
that the command signal is noise free and the magnitude of the closed loop transfer
function is T ( ju) 1 in the widest possible frequency range. The latter feature is
195
(a)
T p1
Ts
= 100
digital control action
y ( )
= t/ Ts
(b)
 Tdig ( ju )
frequency response
T p1
Ts
 S an ( ju )
= 100
 Tan ( ju )
 S dig ( ju )
u = Ts
Fig. 6.22 Output disturbance rejection and frequency response of sensitivity S and complementary
sensitivity T for the analog and digital control action when the plant is defined by (6.83). Bandwidth
of T has been decreased compared to Figs. 6.18 and 6.20 but time domain performance has been
T
significantly improved compared to Fig. 6.17. Ratio Tps1 = 100. a Output disturbance rejection for
the analog and digital control action. b Frequency response of sensitivity S and complementary
sensitivity T for the analog and digital control action
1
y(s)
=
do (s)
1 + kp kh C(s)G(s)
(6.84)
to be equal to zero in the widest possible frequency range starting from the low
frequency region (Fig. 6.21).
196
6.4 Summary
Analytical expressions for the digital PID controller tuning have been presented
regarding the control of typeI, typeII, typeIII control loops. The explicit control
law takes into account all modeled process parameters (model of n poles, m zeros
plus unknown time delay d) plus the controllers sampling time Ts . Basis of the
proposed theory is the Magnitude Optimum criterion and the proof of each one of
the control actions for typeI, typeII, typeIII control loops is presented in Appendix
C. One big advantage of the proposed theory is the introduction of the sampling time
Ts within the explicit closed form expressions regarding the determination of the PID
parameters. This idea gives the benefit to control engineers to investigate the effect
of the sampling time Ts to the control loops performance both in the time and the
frequency domain (Fig. 6.22).
For that reason, during the comparison between the analog and the digital control loop design, all time constants within the control loop have been normalized
according to the relation s = sTs . One interesting result observed in typeI control
loops is the trade of the control engineer is faced with, regarding the choice of the
sampling time against the control loops performance. Specifically, it was shown
T
that the higher the ratio Tps1 is, the more the analog response y( ) is identical
to the digital as far as the time domain is concerned. However, the decrease of the
sampling time versus the dominant time constant affects the bandwidth of T ( j)
in the frequency domain.
For two different sampling times Ts1 and Ts2 for which Ts1 < Ts2 , it was shown
that the frequency range BW for which T ( ju) 1, is decreased in the case where
the controller has been designed with sampling time Ts = Ts1 compared to the
controller designed with sampling time Ts2 . This is a feature against the principle
of the Magnitude Optimum criterion, for which the controller is designed such that
T ( ju) 1 in the widest possible frequency range. Therefore, control engineers
have to find a compromise between the desired bandwidth of T and the desired
response in the time domain so that these two basic requirements of the design are
satisfied.
References
1. Papadopoulos KG, Margaris NI (2012) Extending the symmetrical optimum criterion to the
design of PID typep control loops. J Process Control 12(1):1125
2. Papadopoulos KG, Papastefanaki EN, Margaris NI (2013) Explicit analytical PID tuning rules
for the design of typeIII control loops. IEEE Trans Ind Electron 60(10):46504664
3. Papadopoulos KG, Tselepis ND, Margaris NI (2013) Type III control loopsdigital PID controller
design. J Process Control 23(10):14011414
Part III
Chapter 7
199
200
201
of electric motor drives where demanding requirements are often met regarding the
speed, current and flux PI controllers (SFOC, RFOC).1
Last but not least, such a method should consider an adaptive behavior of the
controller in case the process model changes rather frequently. In other words, the
controller should have the benefit of retuning its parameters in cases when variations
of the plant parameters occur.
In order to develop such a tuning technique able to satisfy all the aforementioned
requirements, the advantages introduced by the Magnitude Optimum criterion are
exploited throughout this chapter, see [20, 28]. The Magnitude Optimum criterion,
introduced by Sartorius and Oldenbourg is based on the idea of designing a controller which renders the magnitude of the closedloop frequency response as close
as possible to unity in the widest possible frequency range [20]. The conventional
tuning of the PID controller based on this principle has been thoroughly discussed
in Chap. 3 (see Sect. 3.2) where a revised PID control law has also been presented,
see Sect. 3.3 and [20].
One important feature of both the conventional and the revised PID control law
presented in Sect. 3.2 [20] is the preservation of the shape of the step2 and frequency
response of the final closedloop control system regardless of the process complexity.
The preservation of the shape means that the output of the control loop exhibits
a specific overshoot (4.4 %), settling and rise time in the time domain, whereas the
amplitude of the closedloop transfer function remains as close as possible to unity
in the widest possible frequency range.
The second important feature of both the conventional and the revised PID control
law is the coupling analytical relation between the PID control parameters when the
principle of the Magnitude Optimum criterion is followed. This coupling relation
gives the flexibility to express all three control parameters based on one, and therefore
by tuning only one parameter (zero of the PID controller) all two other parameters
of the PID controller are tuned automatically. To this end, target of the proposed
method is to tune automatically only one parameter of the PID controller (all others
are tuned automatically) by achieving the prescribed aforementioned performance
of the step response in terms of overshoot (4.4 %), settling and rise time.
For the sake of a clear presentation of this chapter, in Sect. 7.2, the direct tuning
of the conventional PID tuning is presented. There it is shown how the step and
frequency response are preserved when the plant is controlled under I, PI, and PID
control via the Magnitude Optimum criterion. In Sect. 7.2.5, the proposed method
is presented. In Sects. 7.3 and 7.4.3, evaluation results demonstrate the potential of
the proposed automatic tuning method where a comparison between the explicit
solution presented in Sect. 3.3 and the solution provided by automatically tuned PID
controller takes place.
1
2
202
r (s)
+
+
e(s)
controller
C (s)
di ( s )
u (s)
+
do (s)
y f (s)
S
kp
kh
G (s)
+
+
y (s)
+
+
n o (s)
Fig. 7.1 Block diagram of the closedloop control system. G(s) is the plant transfer function, C(s) is
the controller transfer function, r (s) is the reference signal, y(s) is the output of the control loop, yf (s)
is the output signal after kh , do (s) and di (s) are the output and input disturbance signals, respectively,
and nr (s), no (s) are the noise signals at the reference input and process output, respectively. kp stands
for the plants dc gain and kh is the feedback path
1
,
1 + sTp
(7.2)
where Tp = ni=1 Tpi is the equivalent sum time constant of the plant. When the
information about the plant is limited, the control that can consciously be applied
is limited to integral control, so that the system exhibits at least zero steady state
position error.
203
1
,
sTiI (1 + sTc )
(7.3)
kp C(s)G(s)
1 + kp kh C(s)G(s)
(7.4)
(s) =
T
(7.5)
for which Tc Tp 03 and T = Tc + Tp has been considered. Note that Tc
stands for the controllers unmodeled dynamics arising from its implementation.
According to the conventional design via the Magnitude Optimum principle see
Sect. 3.2.1, the integration time constant TiI along with the feedback path kh prove
to be equal to
kh = 1 and TiI = 2kp kh T .
(7.6)
Condition kh = 1 implies that the closedloop system has zero steady state position
error. Substituting (7.6) into (7.5), leads to
(s) =
T
1
.
2T2 s 2 + 2sT + 1
(7.7)
1
.
2s 2 + 2s + 1
(7.8)
At this point, it is necessary to declare that by using only the integration time constant
TiI and if kh = 1, which results in the above closedloop dynamic behavior, the sum
time constant of the closedloop system T can be estimated by the relation
3
The controllers unmodeled dynamics Tc are negligible compared to the plants unmodeled
dynamics Tp , Tc Tp .
204
Test =
TiI
Ti
= I.
2kp kh
2kp
(7.9)
n+1
s
2
Tn
2
n
3
j=1 Tp j + + s
+ 2s + 2s + 1
2
T2
T
T
p
p
i= j=1 i
j
n
(7.10)
as it was proved Sect. 3.2.2. There, it was shown that depending on the ratio =
the step and frequency response exhibits certain performance characterized by
Tp1
T
Mean rise time tr = 4.40T (4.7T for 0.9 and 4.1T for = 0.3).
Mean settling time tss = 7.86T (8.40T for 0.9 and 7.32T for = 0.3).
Mean overshoot 4.47 % (4.32 % for 0.9 and 4.62 % for = 0.3).
Gain margin m = 205 db.
Phase margin m = 65.27 .
1
,
(1 + sTp1 )(1 + sT1p )
(7.11)
n
where T1p = i=2
Tpi is the parasitic time constant of the plant. Since the plant
has a dominant time constant, PI control of the form
C(s) =
1 + sTn
,
sTiPI (1 + sTc )
(7.12)
2T21 s 2
1
.
+ 2T1 s + 1
(7.13)
205
1
,
2s 2 + 2s + 1
(7.14)
for which the optimal PI control action has been proved in Sect. 3.2.3 to be given by
kh = 1,
Tn = Tp1 ,
TiPI = 2kp kh T1
(7.15)
(7.16)
(7.17)
Comparing (7.14) with (7.8), it is concluded that with the application of PI control via
the conventional design of the Magnitude Optimum criterion, a closedloop system
with time and frequency response of the same shape results.
However, the response of (7.14) is faster, because the timescale is smaller (T1 <
T ). In other words, the compensation of the dominant time constant Tp1 has left the
shape (performance features) of the system time and frequency responses unaltered
and produced only a change both in the time and frequency scale, respectively. In
addition, through the new integration time constant TiPI , with which a step response
with mean overshoot 4.47 % is achieved, the parasitic time constant of the closedloop system can be estimated through the relation
T1est =
TiPI
Ti
= PI .
2kp kh
2kp
(7.18)
1
,
(1 + sTp1 )(1 + sTp2 )(1 + sT2p )
(7.19)
n
Tpi stands for the parasitic time constant of the plant.
where again T2p = i=3
Since the plant has two dominant time constants, PID control defined by
C(s) =
(7.20)
is imposed to (7.19). In similar fashion, in Sect. 3.2.3, it was proved that the end
closedloop transfer function is given by
206
(s) =
T
2T22 s 2
1
.
+ 2T2 s + 1
(7.21)
1
2s 2
+ 2s + 1
(7.22)
(7.23)
Tn = Tp1 ,
Tv = Tp2 ,
(7.24)
(7.25)
(7.26)
Comparing (7.22) with (7.14) and (7.8), it becomes evident that with the application
of PID control, we end up again, in a closedloop system with time and frequency
responses of the same shape (performance features), but with even smaller timescale
(T2 < T1 < T ) and consequently even faster.
Moreover, with the integration time constant TiPID , with which we achieve a step
response with 4.47 % mean overshoot, we can estimate the new parasitic time
constant of the closedloop system using the relation
T2est =
TiPID
Ti
= PID .
2kp kh
2kp
(7.27)
s0
(7.28)
and if the process G(s) is stable. In vector controlled induction motor drives, kp
stands for the pulse width modulator gain which is a priori known for the whole
operating range of the motor. Moreover, an estimation of the sum time constant Tp
of the plant can be derived from the step response according to
207
(a)
k pest
t ss
td
t
(b)
ovs = 8%
ovs = 7%
ovs = 4.4%
= t/ T p1
Fig. 7.2 Typical step response after an openloop experiment of the process and screen shots of
the automatic tuning procedure. a Typical step response of the process. b A series of small step
variations of the reference input with alternating sign are imposed for tuning the PID controllers
parameters
Tpest
tss
,
4
(7.29)
208
r (s)
di ( s )
controller
++
+ +
Cx ( s )
do (s)
kp
G (s)
y (s)
y f (s)
kh
+
+
n o (s)
ovs act
PI
 max/ min
ovsre f
Fig. 7.3 Block diagram of the closedloop control system and the tuning loop. kp is the plants dc
gain and kh stands for the feedback path. C x stands for the automatically tuned controller. ovsact is
the measured overshoot of y(s) and ovsref is set equal to 4.47 %
During these variations, the overshoot (undershoot) is being measured and is compared with the reference overshoot (undershoot). According to the preceding analysis
in Sect. 3.2, the absolute value of the reference overshoot is 0.0447. The error is fed
into a PI controller, which tunes the controller C x (s) in succession, so that the overshoot (undershoot) of the closedloop step response to be 4.47 %. According to the
analysis presented in Sects. 3.2 and 3.2.4, the controller C x (s) is being given the form
C x (s) =
(2kp kh Tx
(7.30)
where Tx , Tnx and Tvx are time constants that must be determined automatically.
Step 2: Determination of the time constant Tx . In (7.30) Tnx = Tvx = 0 is set.
In succession, a series of step variations on the reference input is imposed and time
constant Tx is tuned such, so that the overshoot (undershoot) is 4.47 %.
According to Sect. 3.2.1, this occurs when Tx T . Tuning of Tx , (or Tix ) is
described in Fig. 7.4
Step 3: Determination of the time constant Tnx . With the value of Tx given from
Step 2, Tvx = 0 is set in (7.30).
C x (s) =
1 + sTnx
.
(2kp kh Tx 2kp kh Tnx )s(1 + sTc )
(7.31)
A series of step variations of the reference input is again imposed and Tnx is tuned,
so that the overshoot (undershoot) becomes again 4.47 %. As shown in Fig. 7.4a,
this occurs when Tnx Tp1 , Sect. 3.2.3, PI control. If the parasitic time constant
209
(a)
Tn x > T p 1
Tn x < T p 1
= t/ T 1
(b)
Tv x > T p 2
Tv x < T p 2
= t / T 2
Fig. 7.4 Tuning of the PID controller. According to step 2: let at step (k) a series of step pulses is
applied at the reference input. If ovsact < ovsref then at (k +1) step Tix (k +1) < Tix (k). The amount
of this change is based on the parameters of the PI controller (gray box) which is tuned heuristically.
The PI controller takes the error between ovsact , ovsref at step k and returns the Tix (k + 1) for the
next step. Note that at step 2, Tnx = Tvx = 0. At step 3, let at step (k), a series of step pulses is
applied at the reference input. If ovsact < ovsref then at (k + 1) step Tnx (k + 1) > Tnx (k). Since Tix
is tuned automatically (see (7.30)) while all other parameters remain constant, ovsact is controlled
only by tuning Tnx . Since Tnx is the zero of the openloop transfer function, if Tnx (k) > Tnx (k + 1)
then ovsact (k + 1) < ovsact (k). The same tuning procedure stands for Tvx . a Tuning of parameter
Tnx . b Tuning of parameter Tvx
T1 = Tx Tnx is relatively large, the procedure can be continued by attempting
step 4. If the parasitic time constant is sufficiently small, PI control is retained.
Step 4: Determination of the time constant Tvx . Given the values of Tx and Tnx ,
Tvx is tuned in such a way, so that the overshoot is again 4.47 %, by imposing again
a series of step variations on the reference input. As shown in Fig. 7.4b, this occurs
210
when Tvx Tp2 . If the parasitic time constant T2 = Tx Tnx Tvx still remains
relatively large, a fact that shows that other relatively large time constants may exist,
the tuning procedure can be continued incorporating in cascade with the controller
C x (s) the necessary number of highpass stages of the form
Ch (s) =
1 + sTa
,
1 + sTb
(7.32)
where Ta > Tb . For one additional highpass stage, the controller C x (s) must take
the form
(1 + sTnx )(1 + sTvx )(1 + sTa )
2kp s(Tx Tnx Tvx Ta )(1 + sTc )(1 + sTb )
(1 + sTnx )(1 + sTvx )(1 + sTa )
C x (s) =
(7.33)
where Tbc = Tc + Tb is considered as the new parasitic time constant of the
controller.
However, this can only occur if the noise level, that accompanies the controlled
physical quantities, allows it. If this is not possible, but the design of a faster closedloop system is required, then different control techniques should be followed, as
cascade control, for example [14]. Obviously, the controllers tuning of the inner
loops can be achieved using the same procedure.
and
Tx Tnx
2
(7.34)
211
(a)
y r ( )
(b)
= t / T p1
y r ( )
= t/ T p1
At this point, it should be noted that, as long as the controller parameters are
determined, every repetition of the procedure is faster. In Fig. 7.5, it is shown the
application of the suggested tuning procedure when the starting conditions are quite
different from the nominal ones.
212
1.31
(1 + s )(1 + 0.2s )(1 + 0.1s )(1 + 0.05s )(1 + 0.02s )
(7.35)
1 + s tn
s ti (1 + s tsc )
(7.36)
is given finally by
CPIaut (s ) =
1 + s 0.99
s 1.243(1 + s t
sc )
(7.37)
However, the optimal PI control action calculated analytically according to Sect. 3.3
is defined by
CPIopt (s ) =
1 + s 1.02
.
s 1.17(1 + s tsc )
(7.38)
From (7.36) and (7.38) it is apparent that the parameters calculated from both methods
are practically the same. This is also justified by the step response of the closedloop control system in Fig. 7.6a, b where reference tracking and output disturbance
rejection is depicted.
In a similar fashion, by tuning automatically the PID controller of the form
C(s ) =
(1 + s tn )(1 + s tv )
s ti (1 + s tsc )
(7.39)
In this case, only an openloop experiment is required to the process for initializing the algorithm
and no other information.
5 In this case, the transfer function is assumed accurately modeled.
213
(a)
ovs = 5.48%
ovs = 4.51%
y ( )
optimal tuning
PI control
= t/ T p1
(b)
PI control
automatic tuning
t ss = 3.12
y ( )
t ss = 3.34
= t/ T p1
we ended up in
CPIDaut (s ) =
(1 + s 0.99)(1 + s 0.209)
s 0.692(1 + s tsc )
(7.40)
1 + s 1.252 + s 2 0.25
.
s 0.57(1 + s tsc )
Let it be noted that the zeros of (7.41) are real positive values since
(7.41)
214
(a)
ovs = 6.76%
ovs = 4.45%
y ( )
optimal tuning
PID control
= t/ T p1
(b)
PID control
automatic tuning
t ss = 1.46
y ( )
t ss = 1.86
= t/ T p1
Fig. 7.7 PID control of a plant with one dominant time constant defined by (7.35). a Step response
of the closed loop control system. b Output disturbance rejection
CPIDopt (s ) =
(1 + s 0.99)(1 + s 0.25)
s 0.57(1 + s tsc )
(7.42)
After comparing (7.40) with (7.42) it is apparent that both controllers tuning results
in almost the same step and frequency response of the final closed control system,
see also Fig. 7.7a, b.
215
0.84
.
(1 + s )(1 + 0.9s )(1 + 0.1s )(1 + 0.05s )(1 + 0.02s )
(7.43)
1 + s tn
s ti (1 + s tsc )
is finally given by
CPIaut (s ) =
(7.44)
1 + s 1.179
s 1.7(1 + s t
sc )
(7.45)
1 + s 1.3
.
s 1.46(1 + s tsc )
(7.46)
(1 + s tn )(1 + s tv )
s ti (1 + s tsc )
(7.47)
results in
CPIDaut (s ) =
(1 + s 1.179)(1 + s 0.718)
,
s 0.492(1 + s tsc )
(7.48)
1 + s 1.91 + s 2 0.91
,
s 0.437(1 + s tsc )
(7.49)
the zeros of which are conjugate complex since (7.49) can be rewritten in the form
of
CPIDopt (s ) =
(7.50)
216
(a)
ovs = 6.01% ovs = 4.44%
y ( )
optimal tuning
PI control
= t/ T p1
(b)
PI control
automatic tuning
t ss = 6.44
y ( )
t ss = 7.5
= t/ T p1
1.81
e1.048s . (7.51)
(1 + s )(1 + 0.69s )(1 + 0.3s )(1 + 0.13s )(1 + 0.1s )
1 + s tn
s ti (1 + s tsc )
(7.52)
217
(a)
ovs = 5.46%
ovs = 4.52%
y ( )
optimal tuning
PID control
= t/ T p1
(b)
PID control
automatic tuning
y ( )
t ss = 1.83
t ss = 2.2
= t/ T p1
Fig. 7.9 PID control of a plant with two dominant time constants defined by (7.43). a Step response
of the closed loop control system. b Output disturbance rejection
resulted in
CPIaut (s ) =
1 + s 1.066
,
s 8.3(1 + s tsc )
(7.53)
1 + s 1.44
.
s 6.98(1 + s tsc )
(7.54)
218
(1 + s tn )(1 + s tv )
s ti (1 + s tsc )
(7.55)
(1 + s 1.066)(1 + s 0.25)
s 7.38(1 + s tsc )
(7.56)
we ended up in
CPIDaut (s ) =
1 + s 2.049 + s 2 1.16
,
s 4.78(1 + s tsc )
(7.57)
the zeros of which are conjugate complex since (7.57) can be rewritten as follows
CPIDopt (s ) =
(7.58)
From (7.56) and (7.58), it is apparent significant difference in the value of the integrators time constant and the zeros of the controller. This difference is depicted also
in Figs. 7.10, and 7.11 regarding the step response of the closedloop control system
and disturbance rejection.
Specifically, the settling time tss of disturbance rejection in the case of PI control is
equal to tss = 14.7 and tss = 11.3 when the controller is tuned automatically and
optimally, respectively. This difference becomes bigger in the case of PID control
where the corresponding settling time is equal to tss = 13.1 and tss = 7.05 when
the controller is tuned automatically and optimally.
7.3.4 Plant with Dominant Time Constants, Zeros, and Time Delay
In this example, the process defined by
G(s ) = 1.31
es
(7.59)
1 + s tn
s ti (1 + s tsc )
(7.60)
219
(a)
ovs = 5.99%
ovs = 4.47%
y ( )
t d = 1.08
= t/ T p1
(b)
PI control
optimal tuning
automatic tuning
y ( )
t ss = 14.7
t ss = 11.3
= t/ T p1
Fig. 7.10 PI control of a plant with dominant time constants and time delay defined by (7.51).
a Step response of the closed loop control system. b Output disturbance rejection
CPIaut (s ) =
1 + s 1.06
(7.61)
1 + s 1.67
.
s 5.84(1 + s tsc )
(7.62)
s 7.423(1 + s t
sc )
220
(a)
ovs = 4.49%
ovs = 5.67%
y ( )
t d = 1.08
= t/ T p1
(b)
PID control
automatic tuning
y ( )
t ss = 7.05
t ss = 13.1
= t/ T p1
Fig. 7.11 PID control of a plant with dominant time constants and time delay defined by (7.51).
a Step response of the closed loop control system. b Output disturbance rejection
C(s ) =
(1 + s tn )(1 + s tv )
s ti (1 + s tsc )
(7.63)
(1 + s 1.06)(1 + s 0.31)
,
s 6.6(1 + s tsc )
(7.64)
results in
CPIDaut (s ) =
221
CPIDopt (s ) =
1 + s 2.39 + s 2 1.61
.
s 3.94(1 + s tsc )
(7.65)
Note again that the zeros of (7.65) are conjugate complex since its numerator can be
rewritten in the form of
CPIDopt (s ) =
(7.66)
Note also in this case, the difference in the performance of the final closedloop
control system regarding reference tracking and output disturbance rejection when
the PID controller is tuned both automatically and analytically, see Figs. 7.12 and
7.13.
es .
(7.67)
(7.68)
1 + 1.44s
,
s 11.18(1 + s tsc )
(7.69)
C(s ) =
resulted in
CPIaut (s ) =
1 + 2.2s
.
9.3s (1 + s tsc )
(7.70)
(1 + s tn )(1 + s tv )
s ti (1 + s tsc )
(7.71)
222
(a)
ovs = 6.38%
ovs = 4.49%
y ( )
t d = 1
PI control
= t/ T p1
(b)
PI control
optimal tuning
automatic tuning
t ss = 18.5
t ss = 13.2
= t/ T p1
Fig. 7.12 PI control of a plant with dominant time constants, zeros, and time delay defined by
(7.59). a Step response of the closed loop control system. b Output disturbance rejection
CPIDaut (s ) =
1 + 1.91s + 0.68s 2
,
s 9.95(1 + s tsc )
(7.72)
1 + 3.12s + 2.81s 2
s 6.91(1 + s tsc )
(7.73)
223
(a)
ovs = 4.5%
ovs = 6.64%
y ( )
t d = 1
PID control
= t/ T p1
(b)
PID control
optimal tuning
automatic tuning
y ( )
t ss = 16.6
t ss = 8.41
= t/ T p1
Fig. 7.13 PID control of a plant with dominant time constants, zeros, and time delay defined by
(7.59). a Step response of the closed loop control system. b Output disturbance rejection
Note that in this case, zeros of (7.73) are real values, since the numerator can be
rewritten in the form of
CPIDaut (s ) =
(7.74)
In this case, it is apparent that the closedloop control system with the automatically
tuned PI, PID controller exhibits poor performance compared to the optimal PI, PID
tuning via the explicit control law, see Figs. 7.14 and 7.15. Specifically, the settling
224
(a)
ovs = 4.47%
ovs = 5.88%
y ( )
optimal tuning
t d = 1
PI control
= t/ T p1
(b)
PI control
automatic tuning
y ( )
t ss = 19.4
t ss = 26.2
= t/ T p1
Fig. 7.14 PI control of a nonminimum phase plant defined by (7.67). a Step response of the closed
loop control system. b Output disturbance rejection
time of output disturbance rejection in the case of PID control is tss = 23.2 for the
automatically tuned controller and tss = 12.9 for the explicitly tuned controller.
225
(a)
ovs = 5.83%
ovs = 4.49%
y ( )
optimal tuning
PID control
(b)
= t/ T p1
PID control
automatic tuning
t ss = 12.9
t ss = 23.2
= t/ T p1
Fig. 7.15 PID control of a nonminimum phase plant defined by (7.67). a Step response of the
closed loop control system. b Output disturbance rejection
of the step response of the control loop is achieved as described in Sect. 7.2. This
feature leads effortlessly to the automatic tuning of the PID type controller which is
finally presented in Sect. 7.4.2.
226
G(s) =
1
, <1
(1 + 2 T s + T 2 s 2 )(1 + sTp )
(7.75)
T (s) =
where T
results in
T (s ) =
kp (1 + s x + s 2 y)
.
2 ti s 4 + ti 2 +
s 3 + 2 ti + ti + kh kp y s 2
2
+ ti + k h k p x s + k h k p
(7.78)
ti = 2kp kh (1 + 2 ) ,
(7.79)
T
and ti = TTi . Integral control law (7.79) is proved as
where = TT = Tc +T
p
follows. From (7.78), if x = y = 0 then
T (s ) =
kp
.
3
2 ti s + ti 2 + 2 s + ti (1 + 2 ) s 2 + ti s + kh kp
4
(7.80)
According to A.1 and (7.80), where the principle of the Magnitude Optimum criterion
is presented, it is apparent that kh kp = kp or finally
kh = 1.
(7.81)
The application of (A.10) into (7.80) results in a12 = 2a2 a0 since the terms b1 , b2 of
(7.80) are b1 = b2 = 0. Therefore it is apparent that ti2 = 2kp kh ti (1 + 2 ) or
ti = 2kp kh (1 + 2 ) .
(7.82)
227
(a)
y ( )
= 2.5, = 0.2
= 2, = 0.2
= t / T
(b)
= 0.1, = 0
ovs = 4.4%
y ( )
= t/ T
1
2 2 (1 + 2 )s 4
+ 2 (1 + 2 )( + 2 )s 3
2 2
+2(1 + 2 ) s + 2(1 + 2 )s + 1
(7.83)
From Fig. 7.16a it is apparent that the final closedloop control system is not stable
, . However, (7.83) becomes stable if is forced 0, Fig. 7.16b. In this
case (7.83) becomes equal to
228
= 0, = 0
y r ( )
y o ( )
= t/ T
Fig. 7.17 I controlstep response of the closedloop control system for a secondorder process
with conjugate complex poles and if 0. The final transfer function of the control loop is
defined by (7.84)
T (s) =
1
2s 2 + 2s + 1
(7.84)
which is equivalent to (3.10), (3.25) and (3.41) presented in Sect. 3.2. Therefore,
according to Sect. 3.2, the step response of the closedloop control system exhibits
overshoot 4.4 %, see Fig. 7.17. From this point and based on the analysis in Sects. 7.2
and 7.4.1 and the determination of the integrators time constant, see (7.82), a method
for the automatic tuning of the PID controllers parameters is proposed in the sequel.
1 + s X x + s 2 Yx
sTix (1 + sTc )
(7.85)
T
.
Tc + Tp
229
(7.86)
1
.
sTi1 (1 + sTx ) (1 + sTc )
(7.87)
T
Tx + Tp + Tc
T
1
Tx
(7.88)
where Tx = Tx + Tp + Tc is the equivalent sum time constant of the closed loop.
Again, as mentioned in Sect. 7.4.1, it is assumed in our analysis that Tc Tp 0,
Tc Tp Tx 0 and Tx (Tc + Tp ) 0. In that case and according to (7.77) (X =
Y = 0), the respective closedloop transfer function is equal to
kp C x (s)G(s)
Fol (s)
=
1 + kh Fol (s)
1 + kp kh C x (s)G(s)
1
1
kp
sTi1 (1 + sTx ) (1 + sTc ) (1 + 2 T s + T 2 s 2 )(1 + sTp )
=
1
1
1 + kh kp
sTi1 (1 + sTx ) (1 + sTc ) (1 + 2 T s + T 2 s 2 )(1 + sTp )
(7.89)
T (s) =
or finally
T (s) =
kp
+ Ti1 T T + 2 Tx s 3 + Tx + 2 T Ti1
+ Ti1 s + kp kh
T 2 Ti1 Tx s 4
(7.90)
(7.91)
Since Tx is known,7 Ti1 is tuned such, so that the overshoot of the closedloop control
system becomes equal to 4.4 %. The tuning of Ti1 is made as follows.
Step 1: Determination of the gain kp . Initially, the gain kp is determined from the
step response of the plant at steady state, Fig. 7.20. Therefore,
6
7
Tx is a design parameter.
This time constant was chosen sufficiently large, so that 0.
230
r (s)
di ( s )
controller
++
+ +
Cx ( s )
do (s)
kp
G (s)
y (s)
y f (s)
kh
n o (s)
ovs act
PI
max/ min
ovsre f
Fig. 7.18 Block diagram of the closedloop control system and the tuning loop in the frequency
domain. kp is the plants dc gain and kh stands for the feedback path. C x stands for the automatically
tuned controller. ovsact is the measured overshoot of y(t) and ovsref is set equal to 4.4 %
s0
(7.92)
Fol (s) =
kp
1
sTi1 (1 + sTx ) (1 + sTc ) (1 + 2 T s + T 2 s 2 )(1 + sTp )
(7.93)
it is easily seen that the ovsact at the next step increases and the rise time decreases,
if the change at the Ti1 is done such that Tix (k + 1) < Tix (k). The amount of this
change is based on the parameters of the PI controller (gray box), the tuning of
The amplitude of these pulses is small enough, so that the output of the control loop y(t) does not
diverge far from its operating point.
231
(a)
ovs = 4.32%
y ( )
= t/ T x
(b)
ovs = 8%
ovs = 7%
ovs = 4.4%
= t/ T p1
Fig. 7.19 Determination and automatic tuning of the Ti1 time constant during Ilag control action.
a Tuning of the integrators time constant Ti1 so that the overall parasitic time constant T of the
closed loop is determined. b series of small step variations of the reference input with alternating
sign are imposed for tuning the Ilag controller and the PID controllers parameters
which is heuristic and trivial,9 [23]. The PI controller is fed with the error between
ovsact , ovsref at step k and returns the Tix (k + 1) for the next step.
Scope of this tuning is the determination of the overall parasitic time constant
T = Tp + Tc of the closed loop. When the overshoot of the closed loop becomes
9
The PI controller can be avoided and a simple bangbang control with a hysteresis band in the
output overshoot reference can be introduced.
232
ovs%
kp
t ss
y (t )
t
Fig. 7.20 Typical step response of the approximate secondorder process with conjugate complex
poles
(7.94)
Note that after that step Ti1 is known. Thus, for determining Tx through (7.94)
a measurement of kp , via an openloop experiment to the process, Fig. 7.20 is
required.
From Fig. 7.20 it is apparent that
kp = yrss = yr (),
1 2 .
M =e
(7.95)
(7.96)
An accurate estimation of the overshoot Fig. 7.20 is related to the damping ratio
through
n 2 M
,
(7.97)
est
2 +
n 2 M
where
max yr (t)
M=
1.
kp
Moreover, an accurate estimation of Test can be obtained through
(7.98)
tsst
4
4
tss
= T, Test est est .
n
233
(7.99)
Ti1
Tx 2est Test .
2kp kh
(7.100)
Note that kh = 1, kp is measured from (7.95), Tx is known and est , Test are measured
from (7.97) and (7.99) respectively. As a result, C x (s) in (7.87) is finally replaced
by the PIDtype controller
C z (s) =
2 s2
1 + 2est Test s + Test
.
sTi2 (1 + sTc )
(7.101)
kp
kp
2
.
sTi2 (1 + sT ) + kh kp
s Ti2 T + sTi2 + kh kp
(7.103)
Therefore, Ti2 is tuned exactly as Ti1 so that the overshoot of the closedloop control
system becomes equal to 4.4 %. In this case, Ti2 is then equal to
Ti2 2kh kp T .
(7.104)
=
2
2
2
2kp T T s + 2kp T s + kh kp
2T s + 2T s + 1
(s)
T
s 2 Ti2 T
(7.105)
234
G(s) =
(7.106)
G(s) =
(7.107)
The step and frequency response of the process is shown in Fig. 7.24a, b respectively.
In Fig. 7.25a, b the tuning of the Ilag controller (Ti1 ) and the PID controller (Ti2 ) is
presented.
235
(a)
step response open loop experiment
ovs%
kp
y (t )
t ss
(b)
frequency response open loop experiment
 G ( j )
Fig. 7.21 Responses in the time and frequency domain after an openloop experiment of the process
G defined by (7.106). a Step response of the process G(s). b Frequency response of the process
G(s)
It is critical to mention that poor initialization of the Ilag controller, Fig. 7.25a
can lead to a high overshoot at the output of the control loop. For that reason, initial
values both when tuning Ti1 and Ti2 have to lead to at least 0 % overshoot of the
control loop. In this case and according to the Ilag controller tuning, Tx is initialized
with Tx = Test which is measured from the openloop experiment of the process.
236
(a)
ovs%
y ( )
(b)
ovs%
y (t )
7.5 Summary
In this chapter, an automatic tuning algorithm for the PID controllers parameters
has been presented. The method requires only measurements from an openloop
experiment of the process, which serves for initializing the proposed algorithm.
The method assumes access to the output of the process and not to the states as it
7.5 Summary
237
(a)
PID control
di ( t )
y (t )
do (t )
Ti2
y (t )
(b)
PID control
Ti2
do (t )
di ( t )
Ilag control
Ti1
u (t )
di ( t )
u (t )
t
Fig. 7.23 Step response of the control loop. Output do ( ) = ( ) and input di ( ) = 0.5r ( )
disturbance is applied at t = 3 and t = 6 respectively, where r (s) = 1s . a Response of the output
y(t) in the presence of output do (t) = r (t) and input di (t) = 0.5r (t) disturbance. Ilag control and
PID control. b Response of the command signal u(t) in the presence of output do (t) = r (t) and
input di (t) = 0.5r (t) disturbance
238
(a)
ovs%
kp
y (t )
t ss
step response
open loop experiment
t
(b)
frequency response open loop experiment
 G ( j )
Fig. 7.24 Responses in the time and frequency domain after an openloop experiment of the process
G defined by (7.107). a Step response of the process G(s). b Frequency response of the process
G(s)
of the proposed method was evaluated via simulation examples. An extensive simulation test batch was presented in Sect. 7.3 comparing the control action resulting
from the proposed method, (very little knowledge of the process) see Sect. 7.2.5,
with the control action resulting from the explicit solution (exact knowledge of the
process model) presented in Sect. 3.3.
7.5 Summary
239
(a)
ovs%
y (t )
(b)
ovs%
y (t )
The proposed method was also extended to processes with conjugate complex
poles. Such processes are often met in many industry applications, i.e., field of
electric motor drives where the problem there is known as design of active damping
regulators. The method requires an openloop experiment of the process, so that
basic information (overshoot and the time constant of the process) is measured,
which serves for initializing the proposed algorithm. The method assumes access
only to the output of the process and not to the states. The proposed method was
tested at processes with damping ratio very close to zero achieving promising results.
240
(a)
y (t )
PID control
Ti2
do (t )
Ti1
Ilag control
di ( t )
(b)
u (t )
do (t )
PID control
Ti2
do (t )
di ( t )
Ti1
u (t )
Ilag control
t
Fig. 7.26 Step response of the control loop. Output do ( ) = ( ) and input di ( ) = 0.5r ( )
disturbance is applied at t = 3 and t = 6 respectively, where r (s) = 1s . a Response of the output
y(t) in the presence of output do (t) = r (t) and input di (t) = 0.5r (t) disturbance. Ilag control and
PID control. b Response of the command signal u(t) in the presence of output do (t) = r (t) and
input di (t) = 0.5r (t) disturbance
References
1. Ang KH, Chong G, Li Y (2005) PID control system analysis, design, and technology. IEEE
Trans Control Syst Technol 13(4):559576
2. str KJ, Hgglund T (1995) PID controllers: theory, design and tuning, 2nd edn. Instrument
Society of America, North Carolina
3. strm KJ, Wittenmark B (1973) On self tuning regulators. Automatica 9(2):185199
4. strm KJ, Hgglund T, Hang CC, Ho WK (1993) Automatic tuning and adaptation for PID
controllersa survey. Control Eng Pract 1(4):699714
5. Bakhshai AR, Joos G, Jain PK, Hua J (2000) Incorporating the overmodulation range in
space vector pattern generators using a classification algorithm. IEEE Trans Power Electron
15(1):8394
References
241
6. Bi Q, Cai WJ, Wang QC, Hang CC, Lee EL, Sun Y, Liu KD, Zhang Y, Zou B (2000) Advanced
controller autotuning and its application in HVAC systems. Control Eng Pract 8(6):633644
7. Cameron F, Seborg DE (1982) A self tuning controller with a PID structure. Int J Control
38(2):401417
8. Chan FY, Moallem M, Wang W (2007) Design and implementation of modular FPGAbased
PID controllers. IEEE Trans Ind Electron 54(4):18981906
9. Chen CL (1989) A simple method for online identification and controller tuning. AIChE J
35(12):20372039
10. Cox CS, Daniel PR, Lowdon A (1997) Quicktune: a reliable automatic strategy for determining
PI and PPI controller parameters using FOPDT model. Control Eng Pract 5(10):14631472
11. Dannehl J, Fuchs FW, Hansen S, Thogersen PB (2009) Investigation of active damping
approaches for PIbased current control of gridconnected pulse width modulation converters with LCL filters. IEEE Trans Ind Appl 46(4):15091517
12. Dannehl J, Liserre M, Fuchs FW (2011) Filterbased active damping of voltage source converters with LCL filter. IEEE Trans Ind Electron 58(8):36233633
13. Friman M (1997) Automatic retuning of PI controllers in oscillating control loops. Ind Eng
Chem Res 36(10):42554263
14. Frhr F, Orttenburger F (1982) Introduction to electronic control engineering. Siemens, Berlin
15. Ho WK, Hang CC, Zhou J (1997) Selftuning PID control of a plant with underdamped
response with specifications on gain and phase margins. IEEE Trans Control Syst Technol
5(4):446452
16. Kerkman DRJ, Leggate Seibel BJ (1996) Operation of PWM voltage sourceinverters in the
overmodulation region. IEEE Trans Ind Electron 43(1):132141
17. Lee DC, Lee GM (1998) A novel overmodulation technique for spacevector PWM inverters.
IEEE Trans Power Electron 13(6):11441151
18. Liu F, Wu B, Zargari NR, Pande M (2011) An active damping method using inductorcurrent
feedback control for highpower PWM currentsource rectifier. IEEE Trans Power Electron
26(9):25802587
19. O Dwyer A (2003) Handbook of PI and PID controller tuning rules, 1st edn. Imperial College
Press, London
20. Oldenbourg RC, Sartorius H (1954) A uniform approach to the optimum adjustment of control
loops. Trans ASME 76:12651279
21. Papadopoulos KG, Margaris NI (2012) Extending the symmetrical optimum criterion to the
design of PID typep control loops. J Process Control 12(1):1125
22. Papadopoulos KG, Margaris NI (2013) Optimal automatic tuning of active damping PID regulators. J Process Control 23(6):905915
23. Papadopoulos KG, Tselepis ND, Margaris NI (2012a) On the automatic tuning of PID type
controllers via the magnitude optimum criterion. In: International conference on industrial
technology (ICIT), IEEE, Athens, Greece, pp 869874
24. Papadopoulos KG, Tselepis ND, Margaris NI (2012b) Revisiting the magnitude optimum
criterion for robust tuning of PID typeI control loops. J Process Control 22(6):10631078
25. Papadopoulos KG, Papastefanaki EN, Margaris NI (2013) Explicit analytical PID tuning rules
for the design of typeIII control loops. IEEE Trans Ind Electron 60(10):46504664
26. Rahimi AR, Syberg BM, Emadi A (2009) Active damping in DC/DC power electronic converters: a novel method to overcome the problems of constant power loads. IEEE Trans Ind
Electron 56(5):14281439
27. Saeedifard M, Bakhshai A (2007) Neurocomputing vector classification SVM schemes to
integrate the overmodulation region in neutral point clamped (NPC) converters. IEEE Trans
Power Electron 22(3):9951004
28. Umland WJ, Safiuddin M (1990) Magnitude and symmetric optimum criterion for the design
of linear control systems: what is it and how does it compare with the others? IEEE Trans Ind
Appl 26(3):489497
Chapter 8
Abstract In this chapter, a summary of the books contribution to the current state
of the art is presented. The summary concentrates on the contribution of the book
regarding both the direct and the automatic tuning procedure for the PID controller
via the Magnitude Optimum criterion. Open issues regarding both tuning approaches
are presented in Sect. 8.2.
243
244
the design of typeII control loops. For tuning the PID controller, the line of polezero
cancellation is followed in both tuning methods. This line is considered in this book
as the conventional way of tuning, which finally proves to lead at suboptimal or
even sometimes unstable control loops.
To cope with this issues, and regarding the design of typeI control loops, the so
called revised theory is proposed. The revised theory does not consider polezero
cancellation between the plants poles and the controllers zeros. On the contrary, it
determines analytically the PID controllers parameters as a function of all plants
time constants (poles, zeros, delay). In other words, the proposed revised method
tunes the controllers gains with all the available information coming from the plant.
This was not the case regarding the conventional tuning.
The same line is also followed for the design of typeII control loops where
the tuning of the PID controller via the Symmetrical Optimum criterion follows the
line of polezero cancellation between the processs dominant time constant and
the controllers zero, see Sect. 4.2.3. In this case, the conventional tuning proves to
lead to unstable closed loops, especially in cases where the plant contains dominant
time constants, right half plane zeros or long time delays.
Summarizing the aforementioned state of the art, one can argue that the conventional tuning
requires that the plants poles are canceled by the controllers zeros,
restricts the PID controllers zeros to be tuned with real values because of the
polezero cancellation method,
has been tested to simple process models,
leads to unstable control loops or control loops with unacceptable performance
when the complexity of the process is increased,
no tuning rules or guidelines are presented relevant to the choice of the sampling
time Ts , in cases where the controller is implemented digitally.
In contrast to these open points, the revised theory comes to fill this gap by optimally
tuning the controller since it
does not require polezero cancellation between the processs poles and the
controllers zeros. Therefore, it determines the PID controllers parameters as a
function of all time constants coming from the process. This explicit solution is
defined by closed form expressions, all of which are proved in the appendix.
allows the controllers zeros to be tuned with conjugate complex values if needed.
outperforms the conventional tuning regarding the closed loops system response,
both in time and frequency domain.
is able to handle the design of higher order control loops (typeII, typeIII, typeIV,
typeV) and even when the complexity of the process is increased.
introduces the sampling time Ts of the controller within the closed form
expressions, which determine the controllers gains. This allows for accurate investigation of the effect of the sampling time on the control loops performance both
in the time and frequency domain.
245
In particular, in Part II of the book, the explicit PID tuning solution is proposed for
typeI (see Chap. 3), typeII (see Chap. 4), typeIII (see Chap. 5) control systems.
In Chap. 3, the current state of the art and the so called conventional PID tuning
procedure is presented in Sect. 3.2. In Sect. 3.2.5 all drawbacks of the conventional
tuning are summarized. The proposed revised tuning that follows is Sect. 3.3, the
proof of which is presented in Appendix B.1, is compared with the conventional
method in a series of simulation examples in Sect. 3.4. There it is shown that when
the controlled plant consists of one or two dominant time constants, then both methods
lead to the same performance. In any other case, the revised method outperforms the
conventional way of tuning both in the time and frequency domain.
Specifically, given a certain plant, the settling time of output disturbance rejection
in the control loop can be reduced up to 45 %, when this plant is controlled via
the revised method compared to the settling time coming from the conventional
control action. Moreover, the revised control loop is less sensitive to input and output
disturbances, since the range for which T ( j) 1 is greater compared to the range
coming from the conventional tuning.
Furthermore, the proposed revised control action is able to control plants with
large zeros. The crystal clear definition of the integrators time constant, see (4.42)
allows the control engineer to understand and decide when the D term has to be
added or omitted and when the PID controller has to be turned to PIDlag, in order
to cope with the existence of large zeros within the plants transfer function.
The same results are also observed regarding the control of integrating processes
which are discussed in Chap. 4. In Sect. 4.2, the conventional PID tuning method via
the Symmetrical Optimum criterion is presented. There it is shown that this kind of
tuning fails to tune a stable PI control action (see Sect. 4.2.2) and tunes only a PID
controller which is based on polezero cancellation, see Sect. 4.2.3. This restriction,
proves to be suboptimal in Sect. 4.4.1, and especially at cases of certain processes,
the control loop proves to be even unstable. In contrast with the conventional tuning,
the revised method is again proposed in Sect. 4.3 as in similar fashion with typeI
control loops. The proof of the control law is presented in Sect. B.2. Once more,
according to the revised method, all three PID parameters are determined in closed
form expressions as a function of all time constants coming from the process. To
this end, the proposed control laws proof does not involve any model reduction
techniques, see Appendix B.2.
In Sect. 4.4, the conventional tuning is again compared with the revised method.
This comparison focuses on the performance of the required control action, in terms
of reference tracking and disturbance rejection. It is interesting to mention that the
conventional tuning leads to unstable control loops in case where the complexity of
the process is increased, see examples in Sects. 4.4.34.4.5.
The introduction of the design of typeIII control loops is presented in Chap. 5
for first time within the literature, see also [2, 3]. Given the principle from polezero cancellation coming from the Magnitude and Symmetrical Optimum criteria, a
similar methodology is presented for the design of typeIII control loops, which is
finally extended to the design of typep control loops.
246
247
References
1. Jang JSR (1993) ANFIS: adaptivenetworkbased fuzzy inference systems. IEEE Trans Syst
Man Cybern B, Cybern 23(3):665685
2. Papadopoulos KG, Margaris NI (2012) Extending the symmetrical optimum criterion to the
design of PID typep control loops. J Process Control 12(1):1125
3. Papadopoulos KG, Papastefanaki EN, Margaris NI (2013) Explicit analytical PID tuning rules
for the design of typeIII control loops. IEEE Trans Ind Electron 60(10):46504664
4. Skogestad S, Postlethwaite I (2005) Multivariable feedback control: analysis and design. Wiley,
New York
5. Takagi T, Sugeno M (1985) Fuzzy identification of systems and its applications to modeling and
control. IEEE Trans Syst Man Cybern 15(1):113
6. Zadeh LA (1965) Fuzzy sets. Inf Control 8(3):338353
7. Zadeh LA (1973) Outline of a new approach to the analysis of complex systems and decision
processes. IEEE Trans Syst Man Cybern 3(1):2844
Appendix A
T (s) =
(A.1)
N ( j)2
D ( j)2
(A.2)
or
T ( j) =
(A.3)
Separating the real from the imaginary part in (A.3), polynomials N ( j) and D( j)
are rewritten as follows:
N ( j) + b8 8 b6 6 + b4 4 b2 2 + b0
+ j b7 7 + b5 5 b3 3 + b1
(A.4)
249
250
and
D( j) + a8 8 a6 6 + a4 4 a2 2 + a0
+ j a7 7 + a5 5 a3 3 + a1
(A.5)
or
D( j)2 + a82 16 + (a72 a8 a6 )14 + (a62 + 2a4 a8 2a5 a7 )12
+ (a52 + 2a3 a7 2a2 a8 2a4 a6 )10
+ (a42 + 2a0 a8 + 2a2 a6 2a1 a7 2a3 a5 )8
+ (a32 + 2a1 a5 2a6 a0 2a2 a4 )6 + (a22 + 2a0 a4 2a1 a3 )4
+ (a12 2a0 a2 )2 + a0 0
(A.6)
and
N ( j)2 + b82 16 + (b72 b8 b6 )14 + (b62 + 2b4 b8 2b5 b7 )12
+ (b52 + 2b3 b7 2b2 b8 2b4 b6 )10
+ (b42 + 2b0 b8 + 2b2 b6 2b1 b7 2b3 b5 )8
+ (b32 + 2b1 b5 2b6 b0 2b2 b4 )6
+ (b22 + 2b0 b4 2b1 b3 )4
+ (b12 2b0 b2 )2 + b0 0
(A.7)
N ( j)2
D ( j)
+ B4 8 + B3 6 + B2 4 + B1 2 + B0
+ A4 8 + A3 6 + A2 4 + A1 2 + A0
(A.8)
2a2 a0 =
b12
b22
b32
(A.9)
2b2 b0
(A.10)
2b3 b1 + 2b4 b0
(A.11)
(A.12)
251
a42 + 2a0 a8 + 2a6 a2 2a1 a7 2a3 a5 = b42 + 2b0 b8 + 2b6 b2 2b1 b7 2b3 b5
(A.13)
=
Equations (A.9)(A.13) are the basis for proving the optimal control law for typeI,
typeII, typeIII control loops, which is presented in Appendix.
Appendix B
Abstract In this chapter, the proof of the optimal PID control law for typeI, typeII,
typeIII control loops is presented. Basis of the design of the control law are the
optimization conditions (A.9)(A.13) of the magnitude T ( j) of the closed loop
transfer function presented in Sect. A.1.
s m m + s m1 m1 + + s 2 2 + s1 + 1 sTd
e
s n1 n1 + + s 3 3 + s 2 2 + s1 + 1
(B.1)
where n 1 > m. The proposed PIDtype controller is given by the flexible form
C(s) =
1 + s X + s2Y
sTi (1 + sTpn )
(B.2)
allowing its zeros to become conjugate complex. Time constant Tpn stands for the
unmodelled controller dynamics coming from the controllers implementation.
According to Fig. 3.1, the closed loop transfer function T (s) is given by
T (s) =
kp C(s)G p (s)
N (s)
N (s)
=
=
.
1 + kh kp C(s)G p (s)
D(s)
D1 (s) + kh N (s)
(B.3)
253
254
m
(s i i ),
(B.4)
i=0
n
(s j j )
(B.5)
j=0
m
(s i z i )
(B.6)
i=0
D1 (s ) = s ti es d
n
(s j r j )
(B.7)
j=0
respectively. The corresponding normalized terms involved in the control loop are
given by
x=
ri =
i
c1i
X
,
c1
y=
Y
Ti
Td
, ti = , d =
,
c1
c1
c12
, i = 1, 2, . . . , n, z j =
j
j
c1
, j = 1, 2, . . . , m.
The normalized time delay constant d is substituted with the all pole series approximation
es d =
n
1
k=0
k!
s k d k = 1 + s d +
1 2 2
1
s d + s 3 d 3
2!
3!
1 4 4
1
s d + s 5 d 5 +
4!
5!
(B.8)
(B.9)
i=1
where
k
1 i
d , k = 0, 1, 2, . . . n, r0 = 1
qk =
r(ki)
i!
i=0
(B.10)
255
or
q0 = 1
(B.11)
q1 = r 1 + d
(B.12)
q2 = r 2 + r 1 d +
q3 = r 3 + r 2 d +
q4 = r 4 + r 3 d +
q5 = r 5 + r 4 d +
1 2
d
2!
1 2
d r1 +
2!
1 2
d r2 +
2!
1 2
d r3 +
2!
(B.13)
1 3
d
3!
1 3
d r1 +
3!
1 3
d r2 +
3!
(B.14)
1 4
d
4!
1 4
1
d r1 + d 5
4!
5!
(B.15)
(B.16)
s i kp (z (i) + z (i1) x + z (i2) y) ,
N (s ) =
(B.17)
i=0
or in an expanded form by
N (s ) = + s kp (yz 4 + x z 5 ) + s 5 kp (yz 3 + x z 4 + z 5 )
6
b6
b5
+ s 4 kp (yz 2 + x z 3 + z 4 ) + s 3 kp (yz 1 + x z 2 + z 3 )
.
b4
(B.18)
b3
+ s 2 kp (y + x z 1 + z 2 ) +s kp (x + z 1 ) + kp
b2
b1
b0
D(s ) =
k
s
ti q( j1) + kp kh z ( j) + z ( j1) x + z ( j2) y
(B.19)
j=0
or in an expanded form by
D(s ) = D1 (s ) + kh N (s )
4
= + s ti q3 + kh kp (z 2 y + z 3 x + z 4 )
a4
+ s ti q2 + kh kp (z 1 y + z 2 x + z 3 )
3
a3
+ s ti q1 + kh kp (y + z 1 x + z 2 ) + s ti + kh kp (x + z 1 ) + kp kh ,
2
a2
a1
a0
(B.20)
256
i k
+
z
x
+
z
y
z
s
(i)
(i1)
(i2)
p
i=0
= k
.
j
ti q( j1) + kp kh z ( j) + z ( j1) x + z ( j2) y
j=0 s
n
(B.21)
Optimization Condition: a0 = b0 .
From the application of (A.9)(B.21) it is obtained
kh = 1.
(B.22)
Condition (B.22) renders the zero order terms of the numerator and denominator
polynomial of the closed loop transfer function equal, which means that the closed
loop system has zero steady state position error (typeI control loops). Note that if
kh = 1 then N (s ) = + kp and D(s ) = + kp kh .
Optimization Condition: a12 2a2 a0 = b12 2b2 b0 .
The application of (A.10) to (B.21) results in
ti = 2kp kh (q1 z 1 x)
1
1 i
r(1i) d z 1 x
= 2kp kh
i!
(B.23a)
(B.23b)
i=0
or
ti = 2kp kh (r1 + d z 1 x)
1
Td
b1
X
= 2kp kh
+
1
c1
c1
c11
c1
(B.24a)
(B.24b)
n
i=1
T pi + Td
m
i=1
Tzi X ,
(B.25)
257
or
n
m
Ti = 2kp kh T pi + Td
Tzi X .
i=1
i=1
(B.26)
It is critical to point out that in comparison to the conventional definition of Ti , the new
definition of the integral gain contains all the dynamics involved in the closed loop.
Optimization Condition: a22 2a3 a1 + 2a4 a0 = b22 2b3 b1 + 2b4 b0 .
The application of (A.10) to (B.21) results in
x a12 y = b11
(B.27)
where
q1 z 1
,
(q1 z 1 )q1 (q2 z 2 )
(q 2 2q2 )(q1 z 1 ) + q1 z 2 q2 z 1 + q3 z 3
.
= 1
(q1 z 1 )q1 (q2 z 2 )
a12 =
b11
(B.28)
(B.29)
(B.30)
where
a22 =
q22
q1 z 2 q2 z 1 + q3 z 3
2q1 q3 q2 z 2 + q1 z 3 + q3 z 1 + q4 z 4
(B.31)
and
b22 =
Q0 Q1 + Q2
Q3
(B.32)
where
Q 0 = q22 2q1 q3 + 2q4
(B.33)
Q 1 = q1 z 1
Q 2 = q2 z 3 q3 z 2 q1 z 4 + q4 z 1 q5 + z 5
(B.34)
(B.35)
Q 3 = q22 2q1 q3 q2 z 2 + q1 z 3 + q3 z 1 + q4 z 4 .
(B.36)
258
or finally by
1 2kp kh 0
2kp kh (q1 z 1 )
ti
0 1 a12 x
b1
0 1
a22 y
b2
0 0
1
1
kh
(B.37)
1 2kp kh 0
ti
2kp kh (q1 z 1 )
x 0 1 a12
b1
=
.
y 0 1
a22
b2
0 0
1
1
kh
(B.38)
s m m + s m1 m1 + + s1 + 1 sTd
e
s(s n1 an1 + + s 3 a3 + sa1 + 1)
(B.39)
1 + s X + s2Y
sTi2 (1 + sTpn )
(B.40)
where parameter Tpn stands for the parasitic controllers time constant and is considered known from the controllers implementation. Purpose of the following analysis is to determine analytically controller parameters as a function of all modeled
time constants within the control loop, X = f 1 (i , a j , Td ), Y = f 2 (i , a j , Td ),
Ti = f 3 (i , a j , Td ). According to (B.39), (B.40) the product C(s)G(s) is defined by
(1 + s X + s 2 Y ) mj=0 (s j j )
C(s)G(s) =
n
s 2 Ti2 esTd i=0
(s i pi )
where
n
n1
(s i pi ) = (1 + sTpn ) (s j a j ).
i=0
j=0
(B.41)
(B.42)
259
Ffp (s)
kp C(s)G(s)
=
1 + Fol (s)
1 + kp kh C(s)G(s)
(B.43)
where Ffp (s), Fol (s) stand for the forward path and the open loop transfer function,
respectively. Along with the aid of (B.41) T (s) becomes equal to
T (s) =
s 2 Ti2 esTd
n
kp (1 + s X + s 2 Y )
i=0
(s i
m
j=0 (s
pi ) + kp kh (1 + s X
j)
+ s2Y )
m
j=0 (s
j)
(B.44)
In the sequel, a general purpose time constant c1 is considered for normalizing all
time constants within the control loop. Therefore, frequency is normalized by setting
s = sc1 and the following substitutions
X
,
c1
x=
ri =
pi
c1i
y=
Y
ti
Td
, ti = , d =
2
c1
c1
c1
, i = 1, . . . , n, z j =
, j = 1, . . . , m
c1
(B.45)
(B.46)
are considered.
The time delay constant is approximated by the series
es d =
1 k k
s d .
k!
(B.47)
k=0
(B.48)
N (s )
N (s )
=
,
D1 (s ) + kh N (s )
D(s )
(B.49)
where
2
N (s ) = kp (1 + s x + s y)
m
j=0
(s j z j )
(B.50)
260
and
D1 (s ) =
s 2 ti2
7
1
k=0
k!
k
(s )d
n
(s i ri ).
(B.51)
i=0
2
s ti2
3
+ s ti2 (r1
4
+ d) + s ti2
(B.52)
(B.53)
(B.54)
1 2
d ,
2!
1
q3 = r 3 + r 2 d + d 2 r 1 +
2!
1
q4 = r 4 + r 3 d + d 2 r 2 +
2!
q2 = r 2 + r 1 d +
(B.55)
1 3
d ,
3!
1 3
1
d r1 + d 4
3!
4!
(B.56)
(B.57)
results in
D1 (s ) = + s 8 ti2 q6 + s 7 ti2 q5 + s 6 ti2 q4
+ s 5 ti2 q3 + s 4 ti2 q2 + s 3 ti2 q1 + s 2 ti2 q0
(B.58)
p
r
(s )(yzr 2 + x zr 1 + zr )
(B.59)
r =0
b6
b5
261
+ s kp (yz 2 + x z 3 + z 4 ) +s kp (yz 1 + x z 2 + z 3 )
4
b4
b3
2
+ s kp (y + x z 1 + z 2 ) +s kp (x + z 1 ) + kp
b2
b1
(B.60)
b0
As a result, the final polynomial D(s ) of the closed loop transfer function is
defined by
D(s ) = D1 (s ) + kh N (s )
=
k
p
r
+ kh kp (s ) yzr 2 + x zr 1 + zr
(B.61)
r =0
j=0
or in an expanded form
D(s ) = + s
6
ti2 q5 + kh kp z 5 y +s ti2 q4 + kh kp (z 4 y + z 5 x)
a7
+ s ti2 q3 + kh kp z 3 y + z 4 x + z 5
a6
5
a5
+ s ti2 q2 + kh kp z 2 y + z 3 x + z 4 .
4
a4
3
2
+ s ti2 q1 + kh kp (z 1 y + z 2 x + z 3 ) +s ti2 + kh kp (y + z 1 x + z 2 )
a3
a2
+ s kp (x + z 1 ) + kh kp
a1
(B.62)
a0
According to (B.49), (B.59) and (B.60), the resulting closed loop transfer function
is given by
p
r
r =0 (s )(yz r 2 + x z r 1 + z r )
p
2
( j+2) + k k
r
p h
j=0 (ti q j )(s )
r =0 (s ) yz r 2 +
T (s ) = k
kp
x zr 1 + zr
.
(B.63)
Since (B.61) is now written in the same form of (A.1), for determining the optimal
control law we can make use of the optimization conditions proved in Sect. A.1.
Equations (A.9)(A.12) are used for the derivation of the optimal control law.
Therefore, the problem to be solved is formulated as follows: given known the
parameters of the process, calculate explicitly the PID control action x, y, ti .
262
Optimization Condition 1: a0 = b0 .
The application of (A.9) to (B.63) results in
kh = 1
(B.64)
which implies that the final closed loop control system exhibits zero steady state
position and velocity error. From (B.63) it is apparent that if kh = 1, then N (s ) =
+ s kp (x + z 1 ) + kp and D(s ) = + s kp kh (x + z 1 ) + kp kh , respectively.
According to the analysis presented Sect. 2.5, the closed loop system is of typeII.
Optimization Condition 2: a12 2a2 a0 = 0.
By making use of a12 2a2 a0 = b12 2b2 b0 we end up with ti = 0. For that reason,
we set a12 2a2 a0 = 0 as another means of optimizing the magnitude of (B.63).
This results in,
ti2 =
1
kp kh x 2 2y + z 12 2z 2 .
2
(B.65)
Let it be noted, that in cases where no zeros exist in the plant transfer function
z i = 0, i = 1, . . . , m, the integral gain is equal to
ti2 =
1
kp kh (x 2 2y).
2
(B.66)
(B.67)
(B.68)
(B.69)
263
2 q12 2q1 z 1 + 2z 2 y + q12 2q2 z 12 2z 2
+ 4 (q1 z 3 + q3 z 1 q4 z 4 q2 z 2 ) = 0.
(B.70)
From (B.68) to (B.70), we finally end up with the control law given by,
kh = 1,
ti2 =
1
kp kh (x 2 2y + z 12 2z 2 ),
2
(B.71)
(B.72)
x 2 2 [q1 (q1 z 1 ) q2 + z 2 ]
4x q13 3q12 z 1 + 2q1 z 12 + q1 z 2 + q2 z 1 q3 + z 3 2z 1 z 2
+
q12 2q1 z 1 + 2z 2 z 12 + 2z 2 + 4q2 4q1 z 1 + q12 2q2
z 12 2z 2 + 4 (q1 z 3 + q3 z 1 q4 z 4 q2 z 2 ) = 0,
1
1
y = x 2 + 2(q1 z 1 )x (z 12 + 2z 2 + 4q2 4q1 z 1 ).
2
2
(B.73)
(B.74)
T (s ) = k
3
j=0 (ti q j )s
p
yz (r 2) + x z (r 1) + z (r )
.
p
+ kh kp r =0 s (r ) yz (r 2) + x z (r 1) + z (r )
r =0 s
( j+3)
kp
(r )
(B.75)
Therefore, for determining the optimal control law according to the Magnitude Optimum criterion, optimization conditions (A.9)(A.12) can be applied in (B.75).
According to the proposed PID control action proposed in (5.15), given known
the plant transfer function (parameters d, r j , z i , j = 1, . . . n, i = 1, . . . m) in
(5.14) our goal is to determine explicitly parameters ti , x, y plus the feedback kh as
a function of the plants parameters. The proof takes place on the normalized closed
loop transfer function T (s ) for which s = sc1 has been set, where c1 is a general
purpose normalizing time constant. From the first optimization condition (A.9) it is
apparent that
264
Optimization Condition 1: a0 = b0
The application of (A.9) to (B.75) leads to
kh = 1
(B.76)
which implies that the final closed loop control system exhibits steady state position,
velocity, and acceleration error. From (B.75) it is apparent that if kh = 1, then
numerators polynomial
N (s ) = + s 2 kp (y + x z 1 + z 2 ) + s kp (x + z 1 ) + kp
b2
b1
(B.77)
b0
and denominators
D(s ) = + s 2 kp kh (y + x z 1 + z 2 ) + s kp kh (x + z 1 ) + kp kh
a2
a1
(B.78)
a0
are resulted. According to the definition regarding the type of the control loop, in 2.5
the closed loop control system is said to be of typeIII.
Optimization Condition 2: a12 2a2 a0 = 0.
By making use of a12 2a2 a0 = b12 2b2 b0 results in kp = 0 and x = z 1 which
does not lead to a feasible control law. For that reason, a12 2a2 a0 = 0 is set, as
another means of optimizing the magnitude of (A.1). This results in
y=
1 2 1 2
x + (z 1 2z 2 ).
2
2
(B.79)
y 2 z 12 2z 2 + z 22 2z 1 z 3 + 2z 4
1
kh kp
.
ti3 =
2
x + (z 1 q1 )
(B.80)
kh kp
8
2
x 4 + 2 z 12 2z 2 x 2 3 z 12 2z 2 + 4 z 22 2z 1 z 3 + 2z 4
x + (z 1 q1 )
(B.81)
265
+x 2 z 22 + 2z 4 2z 1 z 3 + z 32 + 2z 1 z 5 2z 6 2z 2 z 4 = 0. (B.83)
After considering the following substitutions,
2y (z 1 q1 )
A
B =
y 2 z 12 2z 2
C
y 4z 1 z 3 4z 4 2z 22
(B.84)
1 2
z 1 2z 2 x 4 + 2z 14 8z 12 z 2 + 8z 22 x 2
4
+ z 16 6z 14 z 2 + 12z 12 z 22 8z 23
C=
(B.86)
1
4z 1 z 3 4z 4 2z 22 x 2
2
+ 4z 13 z 3 4z 12 z 4 2z 12 z 22 8z 1 z 2 z 3 + 8z 2 z 4 + 4z 23
+ k p ti3
(B.85)
(B.87)
266
4
1
2 2z x 4 + z 1 + 2z 2 2z 2 z
2
x
z
2
2
1 2
2
4 1
2
=0
6
+ kp
z1
3 4
2
2 2
3
+ z 3 + 2z 1 z 5 + 4 2 z 1 z 2 + 2z 1 z 2 + 2z 1 z 3
2z 12 z 4 4z 1 z 2 z 3 + 2z 2 z 4
(B.88)
or finally
ti6
!
(q1 z 1 ) x 2 + 2 (q1 z 1 q2 z 2 ) x
+ (q1 z 1 ) z 12 2z 2 + 2 (q1 z 2 q2 z 1 + q3 z 3 )
!
2
3
1 2 z 12 2z 2 x 4 + 2 z 12 2z 2 x 2 + z 12 2z 2
+ kp
= 0.
4
4 z 12 2z 2 z 22 + 2z 4 2z 1 z 3 z 32 2z 2 z 4 + 2z 1 z 5
(B.89)
kp ti3
Finally, it is set
q1 z 1
Q1
Q2 =
q1 z 1 q2 z 2
Q3
q1 z 2 q2 z 1 + q3 z 3
(B.90)
z 1 2z 3 2z 5
Z1
Z 2 = z 1 z 2 z 3 z 4 2 z 2 2z 4
0
0
z3
Z3
0
2
0
(B.91)
and
Z 0 = Z 13 4Z 1 Z 2 + 4Z 3 ,
Z 4 = 3Z 12 4Z 2 .
(B.92)
(B.93)
Combining (B.89) with (B.81) the optimal control law is finally derived
kh = 1
y=
ti3 =
kh kp
2
x 2 + z 12 2z 2
2
2
y 2 z 12 2z 2 + z 22 2z 1 z 3 + 2z 4
x + z 1 q1
(B.94)
(B.95)
(B.96)
267
(B.97)
j=0
(B.98)
C1 = 8 (4Q 1 Z 0 2Q 1 Q 2 Z 4 + Q 0 Z 4 )
(B.99)
(B.100)
(B.101)
(B.104)
C7 = 8Q 1
(B.105)
C8 = 1
(B.106)
For determining parameter x the real maximum positive value of the eighth"order
#
polynomial solution of (B.97) is always adopted. Therefore, x = max x j ,
x j > 0, x .
Appendix C
Abstract In this chapter the proof of the optimal PID control law for typeI, typeII, typeIII control loops is presented. Basis of the design of the control law are
the optimization conditions (A.9)(A.13) of the magnitude T ( j) of the closed
loop transfer function presented in Sect. A.1. Controller parameters are determined
explicitly as a function of the process parameters and the sampling time of the
controller Ts . For developing the proposed theory a generalized singleinput singleoutput stable process model is employed consisting of npoles, mzeros plus unknown
time delayd.
s m m + + s 4 4 + s 3 3 + s 2 2 + s1 + 1
esTd , n > m
s n pn + s n1 pn1 + + s 5 p5 + s 4 p4 + s 3 p3
+s 2 p2 + s p1 + 1
(C.1)
1 + s X + s2Y
sTi
1 esTs
sTs
(C.2)
where the C (s) controller stands for the digital representation of the PID control
law. CZOH (s) stands for the zero order hold module and Ts stands for the controller
sampling period.
Springer International Publishing Switzerland 2015
K.G. Papadopoulos, PID Controller Tuning Using the Magnitude Optimum Criterion,
DOI 10.1007/9783319072630
269
270
The analysis proceeds by normalizing all time constants in the frequency domain
with the sampling period Ts of the zero order hold. In that,
s = sTs
(C.3)
is set and the resulting expressions (C.1) and (C.2) take the form
G(s ) = kp
snr
s m zm + + s 4 z4 + s 3 z3 + s 2 z2 + s z1 + 1
+ s n1r
n1
+ + s 5r
+s 4 r
+ s 3r
+s 2 r
+ s r
+1
es d
(C.4)
and
C(s ) = C (s )CZOH (s ) =
x=
rj =
X
,
Ts
y=
1 + s x + s2 y
s ti
1 es
s
Y
Ti
Td
, ti = , d =
,
2
Ts
Ts
Ts
pj
i
, j = 1, . . . n, z i = i , i = 1, . . . m.
Tsi
Ts
(C.5)
(C.6)
(C.7)
The transition from the L{.} to the Z{.} domain takes place by making the
transformation
s =
es 1
z1
=
.
z
es
(C.8)
1 (1 + x + y)e2s (x + 2y)es + y
=
.
ti
es (es 1)
(C.9)
By setting
x = x + 2y and y = 1 + x + y
(C.10)
x = 2 y x 2 and y = x y + 1.
(C.11)
results in
271
2s 1) y +1
1 (1es )x+(e
.
ti
es (es 1)
(C.12)
In addition, the respective open and closed loop transfer functions become
Fol (s ) = kh C(s )G(s )
(C.13)
or
kp
Fol (s ) = kh
ti
s m z m + + s 3 z 3 +s 2 z 2 + s z 1 + 1 (1 es )x + (e2s 1) y + 1
s n rn + + s 3 r3 + s 2 r2 + s r1 + 1 es (d+1) (es 1)
(C.14)
and
T (s ) =
N (s )
N (s )
C(s )G(s )
=
=
1 + kh C(s )G(s )
D(s )
D1 (s ) + kh N (s )
(C.15)
or
s m zm + + s 3 z3
kp
(1 es )x + (e2s 1) y + 1
2
+s z 2 + s z 1 + 1
.
T (s ) =
(d+1) s
s n rn + + s 3 r3
s
ti
(e 1)
e
+s 2 r2 + s r1 + 1
s m zm + + s 3 z3
s
2s
+kh k p
(1 e )x + (e 1) y +1
+s 2 z 2 + s z 1 + 1
(C.16)
Substituting the time delay constant by the all pole series approximation
es = 1 + s +
1 2
1
1
1
1
s + s 3 + s 4 + s 5 + s 6 +
2!
3!
4!
5!
6!
(C.17)
results in
1 2 2
1
1
d s + d 3 s 3 + d 4 s 4
2!
3!
4!
1
1
+ d 5 s 5 + d 6 s 6 +
5!
6!
es (1+d) = 1 + d s +
(C.18)
272
where
d = 1 + d.
(C.19)
Additionally,
es (1+d) (es 1) = d1 s + d2 s 2 + d3 s 3 + d4 s 4
+ d5 s 5 + d6 s 6 + d7 s 7 +
where
d1
d2
d3
d4 =
d5
1 +
..
5!
.
and
(C.20)
1
2! + d
1
1 2
1
+
d
+
d
3!
2!
2!
1
1 1 2
1 3
1
4! + 3! d + 2! 2! d + 3! d
1
1 1 2
1 1 3
1 4
d
+
d
+
d
+
d
4!
2! 3!
2! 3!
4!
..
.
(C.21)
q1
r1 + d2
q2
r2 + r1 d2 + d3
q3
q4 =
.
r3 + r2 d2 + r1 d3 + d4
q5
r4 + r3 d2 + r2 d3 + r1 d4 + d5
..
.
..
.
(C.22)
(C.23)
(1 es )x + (e2s 1) y + 1
= 1 + (2 y x)s
1
1
1
2
3
4
+ (8 y x)s
+ (16 y x)s
.
+ (4 y x)s
2!
3!
4!
1
1
5
6
+ (64 y x)s
+
+ (32 y x)s
5!
6!
(C.24)
273
+ (z 6 + y6 y x6 x)s
6
+ (z 5 + y5 y x5 x)s
5 + (z 4 + y4 y x4 x)s
4
N (s ) = kp
3 + (z 2 + y2 y x2 x)s
2
+ (z 3 + y3 y x3 x)s
+ (z 1 + 2 y x)s
+1
(C.25)
where
xk =
k
z ( j1)
j=1
1
[k ( j 1)]!
(C.26)
and
yk = 2
k j1
2
j=1
j!
z k j
(C.27)
or in an expanded form
1
+ z1
1
x1
2!
x2
1
1
+
z
+
z
2
x3
3!
2! 1
1
1
1
x4
,
4! + 3! z 1 + 2! z 2 + z 3
=
x5
1
1
1
1
+
z
+
z
+
z
+
z
1
2
3
4
5!
4!
3!
2!
x6 1
+ 1 z1 + 1 z2 + 1 z3 + 1 z4 + z5
4!
3!
2!
..
6! 5!
.
..
.
(C.28)
and
2z 1 +
4
2!
y2
2z 2 + 2!4 z 1 + 3!8
y3
2z 3 + 2!4 z 2 + 3!8 z 1 + 16
y4
4!
y5 =
4
8
16
2z 4 + 2! z 3 + 3! z 2 + 4! z 1 + 32
5!
y6
4
8
16
32
..
2z 5 + 2! z 4 + 3! z 3 + 4! z 2 + 5! z 1 +
.
..
.
64
6!
(C.29)
274
Finally, the corresponding polynomials for both the numerator and denominator of
the closed loop transfer function are given by
s
N (s ) = + kp (z 6 + y6 y x6 x)
b6
+ kp (z 5 + y5 y x5 x)
s + kp (z 4 + y4 y x4 x)
s
5
b5
b4
3
+ kp (z 3 + y3 y x3 x)
s + kp (z 2 + y2 y x2 x)
s
b3
b2
+ kp (z 1 + 2 y x)
s + kp
b1
(C.30)
b0
and
6
D(s ) = D1 (s ) + kh N (s ) = + ti q6 + kh kp (z 6 + y6 y x6 x)
s
5
5
+ ti q5 s + kh kp (z 5 + y5 y x5 x)
s
a6
a5
4
+ ti q4 + kh kp (z 4 + y4 y x4 x)
s
a4
3
2
+ ti q3 + kh kp (z 3 + y3 y x3 x)
s + ti q2 + kh kp (z 2 + y2 y x2 x)
s
a3
+ ti + kh kp (z 1 + 2 y x)
s + kh kp .
a1
a2
(C.31)
a0
For determining the optimal PID controllers parameters, Eqs. (A.9)(A.13) are
applied to (C.16). For that reason, from the application of
Optimization Condition: a0 = b0 .
To the closed loop transfer function (C.15) and since a0 = kp kh , b0 = kp and within
(C.30), (C.31) results in
kh = 1,
(C.32)
which implies that the final closed loop control system exhibits zero steady position
error if kh = 1.
275
(C.33a)
a1 = ti + kh kp (z 1 + 2 y x),
(C.33b)
a2 = ti q2 + kh kp (z 2 + y2 y x2 x)
(C.33c)
b0 = kp ,
(C.34a)
b1 = kp (z 1 + 2 y x),
(C.34b)
b2 = kp (z 2 + y2 y x2 x).
(C.34c)
and
(C.35)
(C.36)
The application of the third optimization condition (A.12) to the closed loop transfer
function and after taking into account that
a3 = ti q3 + kh kp (z 3 + y3 y x3 x),
(C.37)
a4 = ti q4 + kh kp (z 4 + y4 y x4 x),
(C.38)
b3 = kp (z 3 + y3 y x3 x),
b4 = kp (z 4 + y4 y x4 x)
(C.39)
(C.40)
and
results in
(C.41)
276
a6 = ti q6 + kh kp (z 6 + y6 y x6 x),
(C.42)
(C.43)
b5 = kp (z 5 + y5 y x5 x),
(C.44)
b6 = kp (z 6 + y6 y x6 x)
(C.45)
and
leads to
[(q3 x3 )q3 (q4 x4 )q2 (q2 x2 )q4 + q5 x5 ] x
2q32 4q2 q4 + q2 y4 q3 y3 y5 + 2q5 + q4 y2 y
= (q32 2q2 q4 + 2q5 )(q2 z 1 )
(C.46)
+(q2 z 4 q3 z 3 + q4 z 2 q5 z 1 z 5 + q6 )
To that end, the optimal PID controllers parameters are given by
kh = 1
(C.47)
1
ti = 2kh kp r1 + d z 1 x
2
(C.48)
x a1 y = b1 and x a2 y = b2
(C.49)
where
a1 =
b1 =
a2 =
2(q22 q3 ) (q2 y2 y3 )
(q22 q3 ) (q2 x2 x3 )
(C.50)
(C.51)
(C.52)
q32 2q2 q4 + 2q5 (q2 z 1 ) + q2 z 4 q3 z 3 + q4 z 2 q5 z 1 z 5 + q6
b2 =
(q3 x3 )q3 (q4 x4 )q2 (q2 x2 )q4 + (q5 x5 )
(C.53)
277
a1 b2 a2 b1
,
a1 a2
y =
b2 b1
.
a1 a2
(C.54)
From the definition of the integrators time constant (C.48), it is critical to point
out that
Ti
1
= 2kh kp r1 + d z 1 x
(C.55)
Ts
2
or according to (C.6) and (C.7)
1
Ti = 2kh kp Tsr1 + Ts d Ts z 1 Ts x Ts
2
1
= 2kh kp p1 + Td 1 Ts x Ts
2
n
m
1
= 2kh kp
(T pi ) + Td
(Tzi ) X Ts .
2
i=1
(C.56)
i=1
(C.57)
where Tidig and Tian the optimal values for the integrators time constant regarding
the analog and digital design, respectively.
s m m + + s 4 4 + s 3 3 + s 2 2 + s1 + 1
esTd
s n pn + s n1 pn1 + + s 5 p5 + s 4 p4 + s 3 p3 +s 2 p2 + s p1 + 1
(C.58)
1 + s X + s2Y
s 2 Ti
1 esTs
s
(C.59)
278
where controller C (s) stands for the digital representation of the analog PID control
law. CZOH (s) stands for the zero order hold module and Ts stands for the controller
sampling period. The analysis proceeds by normalizing all time constants in the
frequency domain with the sampling period Ts . In that, we make the substitution
s = sTs .
(C.60)
s m zm + + s 4 z4 + s 3 z3 + s 2 z2 + s z1 + 1
s n rn + s n1 rn1 + + s 5 r5 + s 4 r4 + s 3 r3 +s 2 r2 + s r1 + 1
es d
(C.61)
and
C(s ) = C (s )CZOH (s ) = Ts
x=
rj =
pj
Ts
X
,
Ts
y=
1 + s x + s2 y
s 2 ti2
1 es
s
Y
Ti
Td
, ti = , d =
,
Ts2
Ts
Ts
, j = 1, . . . , n, z i =
(C.62)
(C.63)
i
, i = 1, . . . , m.
Tsi
(C.64)
The transition from the L{.} to the Z{.} domain takes place by utilizing the relation
s =
1
s2
es 1
z1
=
z
es
Ts z
(z 1)2
Ts es
(C.65a)
(es 1)
(C.65b)
or
Ts
C(s ) = 2
ti
(x + y)e2s (x + 2y Ts )es + y
(es 1)
(C.66)
!
(C.67)
279
or finally
x
y
+
2
T
T
T
s
s
C(s ) = s2
ti
e
2s
x
y
y
s
+2 1 e +
Ts
Ts
Ts
.
2
(es 1)
(C.68)
x
y
+2 1
Ts
Ts
(C.69)
x
y
+ ,
Ts
Ts
(C.70)
and
y =
which finally results in
x
= 2 y x 1
Ts
(C.71)
y
= x y + 1.
Ts
(C.72)
and
(1 es )x + (e2s 1) y + 1
(es 1)
!
.
(C.73)
With respect to the above, the corresponding open and closed loop transfer functions
become
Fol (s ) = kh C(s )G(s )
(C.74)
or
Ts2 kp
ti2
m
s zm
kh C(s )G(s ) = kh
+ + s 3 z3 + s 2 z2 + s z1 + 1
(1 es )x + (e2s 1) y + 1
2
s n rn ++s 3 r3 +s 2 r2 +s r1 +1 es d es 1
(C.75)
280
and
T (s ) =
C(s )G(s )
N (s )
N (s )
=
=
1 + kh C(s )G(s )
D(s )
D1 (s ) + kh N (s )
k (s m z m
p
+ + s 3 z 3 + s 2 z 2 + s z 1 + 1)
(1 es )x + (e2s 1) y + 1
t 2 (s n rn + + s 3r3 + s 2 r2 + s r1 + 1)es d (es 1)2
i
kh k p (s m z m + s 3 z 3 + s 2 z 2 + s z 1 + 1)
(1 es )x + (e2s 1) y + 1
(C.76)
where
k p = Ts2 kp .
(C.77)
Substituting the time delay constant with the all pole series approximation
es 1 = s +
1 2
1
1
1
1
s + s 3 + s 4 + s 5 + s 6 +
2!
3!
4!
5!
6!
(C.78)
4 2
8
16
32
64
s + s 3 + s 4 + s 5 + s 6 +
2!
3!
4!
5!
6!
(C.79)
results in
e2s 1 = 2s +
or
(es 1)2 = (s +
1 2
1 3
1 4
1 5
1 6
s + s + s + s + s + )2
2!
3!
4!
5!
6!
(C.80)
and finally in
es d = 1 + ds +
1 2 2
1
1
1
d s + d 3 s 3 + d 4 s 4 + d 5 s 5
2!
3!
4!
5!
(C.81)
es d (es 1)2 = s 2 + d3 s 3 + d4 s 4 + d5 s 5 + d6 s 6 +
(C.82)
1
+ d 6 s 6 +
6!
Additionally, we have
281
where
d3
d4
=
d5
d6
1+d
0.5833 + 1d + 21 d 2
0.25 + 0.5833d + 21 d 2 + 16 d 3
0.0861 + 0.25d +
0.5833 2
2 d
+ 16 d 3 +
(C.83)
1 4
24 d
= ti2 ( + q7 s + q6 s + q5 s + q4 s + q3 s + s )
7
(C.84)
where
q3
q4
q5 =
q6
q7
r1 + d3
r2 + d3 r1 + d4
r3 + d3 r2 + d4 r1 + d5
r4 + d3 r3 + d4 r2 + d5 r1 + d6
(C.85)
r5 + d6r1 + d3 r4 + d4 r3 + d5r2 + d7
(1 es )x + (e2s 1) y + 1
1
4
1
2
3
y x s
= 1 + (2 y x)s
+ 2 y x s +
2
3
6
1
1
1
2
32
64
4
5
6
y x s +
y x s +
y x s + (C.86)
+
3
24
5!
5!
6!
6!
polynomial kh N (s ) takes the form
7
kh N (s ) = + kh k p y7 y x7 x + z 7 s
6
+ kh k p y6 y x6 x + z 6 s
5
+ kh k p y5 y x5 x + z 5 s
4
+ kh k p y4 y x4 x + z 4 s
3
+ kh k p y3 y x3 x + z 3 s
2
+ kh k p y2 y x2 x + z 2 s
+ kh k p 2 y x + z 1 s + kh k p
(C.87)
282
where
xk =
k
j=1
z ( j1)
1
[k ( j 1)]!
(C.88)
or in an expanded form
+ z1
2!
x1
1
1
+ z1 + z2
x2
3! 2
x3
1
1
1
+ z1 + z2 + z3
x4 =
4! 6
2
x5
1
1
1
1
+ z1 + z2 + z3 + z4
x6
5! 24
6
2
x7
1
1
1
1
1
+ z1 + z2 + z3 + z4 + z5
6! 5!
24
6
2
1
1
1
1
1
1
z1 + z2 + z3 + z4 + z5 + z6 +
6!
5!
24
6
2
7!
(C.89)
and
1
2(1 + z 1 )
2 z1 + z2 +
y1
y2
2
1
y3
z
2
z
+
z
+
+
2
3
1
3
3
y4 =
.
2
1
16
y5
2
z
z
z
+
z
+
+
+
3
4
2
1
3
3
5!
y6
2
1
16
32
y7
2 z5 + z4 + z3 + z2 + z1 +
3
3
5!
6!
2
1
16
32
64
2 z6 + z5 + z4 + z3 + z2 + z1 +
3
3
5!
6!
7!
(C.90)
Finally, the corresponding polynomials N (s ), D(s ) for both the numerator and
denominator of the closed loop transfer function are given by
N (s ) =
m
j
k p y j y x j x + z j s
j=0
(C.91)
283
where y1 = 2, x1 = 1, z 0 = 1 and
n
i
qi ti2 + kh k p yi y xi x + z i s
D s =
(C.92)
i=0
k p
k p
b7
y5 y x5 x + z 5 s
b5
y3 y x3 x + z 3 s
5
+ k p
3
+ k p
b3
2 y x + z 1 s + k p
+
b6
4
y4 y x4 x + z 4 s
b4
2
y2 y x2 x + z 2 s
b2
k p
b1
(C.93)
b0
and
D(s ) = D1 (s ) + kh N (s ) =
7
6
+ q7 ti2 + kh k p y7 y x7 x + z 7 s + q6 ti2 + kh k p y6 y x6 x + z 6 s
a7
a6
a5
a4
5
4
+ q5 ti2 + kh k p y5 y x5 x + z 5 s + q4 ti2 + kh k p y4 y x4 x + z 4 s
3
2
+ q3 ti2 + kh k p y3 y x3 x + z 3 s + ti2 + kh k p y2 y x2 x + z 2 s
a3
a2
+ kh k p 2 y x + z 1 s + kh k p
a1
(C.94)
a0
N (s )
D(s ) + kh N (s )
j
m
j=0 k p y j y x j x + z j s
= $
m
%
n 2
qi ti + kh k p yi y xi x + z i s i + kh
k p y j y x j x + z j s j
i=0
j=0
(C.95)
284
(C.96)
which implies that the final closed loop control system exhibits steady state position and velocity error. From, (C.91), it is apparent
that if kh = 1 then the
respective terms s 0 , s 1 , of N (s ) = + k p 2 y x + z 1 s + k p and D(s ) =
+ kh k p 2 y x + z 1 s + kh k p are equal.
Optimization Condition: a12 2a2 a0 = b12 2b2 b0 .
By making use of a12 2a2 a0 = b12 2b2 b0 results in ti = 0. For that reason,
a12 2a2 a0 = 0 is set, as another means of optimizing the magnitude of (A.1). This
results in,
ti2 =
1
kh k p (2 y x)
2 2(y2 2z 1 ) y + 2(x2 z 1 )x + z 12 2z 2 .
2
(C.97)
1
kh k p (2 y x)
2 2y2 y + 2x2 x .
2
(C.98)
(C.99)
q
y
2q
+
y
q
x
y
+
3 3
4
5
4 2
+ q5 x 4
ti2 = 2
(C.100)
q3 2q4
(q4 z 2 q5 z 1 q3 z 3 + q6 + z 4 )
2kh k p
For determining the optimal controller parameters relations (C.97), (C.99) and
(C.99), (C.100) are manipulated together. Therefore, from (C.97), (C.99) it is apparent that
2
2 y x 2 (4q3 y2 2z 1 ) y + 2 (2q3 x2 z 1 ) x
+ z 12 + 2z 2 4q3 z 1 + 4q4 = 0.
(C.101)
285
(C.102)
(C.103)
(C.104)
(C.105)
(C.106)
(C.107)
(C.108)
(C.109)
D x + E y = Z
(C.110)
Z E y
.
D
(C.111)
y +
D(2BZ + CD) + Z 2
(2D + E)2
= 0.
(C.112)
1 2kh k p (x2 q3 )
2kh k p (2q3 y2 )
ti2
0
D
E
x
+ BE)]
0
0
2 [(2D + E)Z +D(AD
y
2
(2D + E)
(z 2 + q3 z 1 q4 )
Z
=
D(2BZ + CD)+Z 2
2
2
(2D + E)
(C.113)
286
or
1
2kh k p (2q3 y2 )
1 2kh k p (x2 q3 )
ti2
x =
D
E
0
+ BE)]
0
0
2 [(2D + E)Z +D(AD
y
2
(2D+E)
(z 2 + q3 z 1 q4 )
Z
(C.114)
D(2BZ + CD)+Z 2
2
(2D + E)
x
Ts
= 2 y x 1 and
y
Ts
G(s) =
(C.115)
(1 esTs )
.
sTs
(C.116)
Note again that CZOH (s) stands for the zero order hold transfer function and Ts stands
for the controllers sampling period. All time constants in the control loop are normalized in the frequency domain with the sampling period Ts . Therefore, by substituting
s = sTs ,
(C.117)
1
s tm (1 + s t p1 )(1 + s t p )
(C.118)
and
C(s ) = C (s )CZOH (s ) = Ts
(1 + s tn )(1 + s tv )(1 + s tx )
s 2 ti (1 + s tc1 )(1 + s tc2 )
(1 es )
s
(C.119)
287
where
ti =
Tc1
Tc2
Ti
, tc1 =
, tc2 =
,
Ts
Ts
Ts
(C.120)
Tn
Tv
Tx
, tv =
, tx =
,
Ts
Ts
Ts
(C.121)
T p
Tp
Tm
, t p 1 = 1 , t P =
.
Ts
Ts
Ts
(C.122)
tn =
tm =
In similar fashion with the analog design procedure in section B, the open loop
transfer function Fol (s) is given by
Fol (s ) = kp kh C(s )G(s )
!
!
1 + s tn 1 + s tv 1 + s t x
(1 es )
= Ts
s
s 2 ti 1 + s tc1 1 + s tc2
kp kh
.
s tm (1 + s t p1 )(1 + s t p )
(C.123)
For moving from the L{.} to the Z{.} domain, the following substitutions are considered
es
z
1
= s
,
=
s
z 1
e 1
1
Ts z
Ts es
=
=
.
2
s2
(z 1)2
(es 1)
(C.124)
(C.125)
(1 es )es Ts es
!
1 + s tn 1 + s tv 1 + s t x
ti 1 + s tc1 1 + s tc2
(es 1)(es 1)
(C.126)
or finally
kp kh Ts2
Fol (s ) =
s tm (1 + s t p1 )(1 + s t p )
es
!
1 + s tn 1 + s tv 1 + s t x
ti 1 + s tc1 1 + s tc2
(es 1)
(C.127)
288
(C.128)
(C.129)
Fol (s ) =
kp kh es (1 + s tn )(1 + s tv )
(C.130)
In similar fashion with the analog design it is set tc1 tc2 0 and tc = tc1 + tc2 .
This results in (1 + s t p )(1 + s tc1 )(1 + s tc2 ) = (1 + s t p )(1 + s tc ).
Moreover if tc t p 0 and t = tc + t p then (C.130) becomes equal to
kp kh es (1 + s tn )(1 + s tv )
Fol (s ) =
s tm ti (1 + s t )(es 1)
(C.131)
kp es (1 + s tn )(1 + s tv )
T (s ) =
kp C(s )G(s )
=
1 + kp kh C(s )G(s )
s tm ti (1 + s t )(es 1)
kp es (1 + s tn )(1 + s tv )
1 + kh
2
s tm ti (1 + s t )(es 1)
kp es (1 + s tn )(1 + s tv )
s tm ti (1 + s t )(es 1) + kh kp es (1 + s tn )(1 + s tv )
. (C.132)
D1 (s ) = es 1 = s +
1 2
1
1
1
1
s + s 3 + s 4 + s 5 + s 6 +
2!
3!
4!
5!
6!
and
D2 (s ) = (es 1)2
1 2
1 3
1 4
1 5
1 6
= (s + s + s + s + s + s + )2
2!
3!
4!
5!
6!
(C.133)
= + Ds + Cs + Bs + As + s
6
where A =
1
1 1
5! + 2! 4! +
1
1
2! + 2! , B
1 1
1 1
3! 3! + 4! 2!
289
2
(C.134)
1
4!
and D =
k p e s 1 + s tn 1 + s tv
=
2
s tm ti (1 + s t ) es 1 + kp kh es (1 + s tn ) (1 + s tv )
.
s 2 kp tn tv + s kp (tn + tv ) + kp (D1 (s ) + 1)
= 2
(s ti tm t + s tm ti )D2 (s )
+ s 2 kp kh tn tv + s kp kh (tn + tv ) + kp kh (D1 (s ) + 1)
(C.135)
Since (C.135) is in the form of (A.1), the optimization conditions (A.9)(A.12) can
be applied for proving the optimal digital PID control law.
For the simplification of the proof of the optimal control, the following substitutions are made. Within the numerator of (C.135) it is set
T (s )
z 1 = kp + kp (tn + tv )
1
z 2 = kp + kp (tn + tv ) + kp tn tv
2!
1
1
z 3 = kp + kp (tn + tv ) + kp tn tv
3!
2!
1
1
1
z 4 = kp + kp (tn + tv ) + kp tn tv .
4!
3!
2!
(C.136)
(C.137)
(C.138)
(C.139)
(C.140)
(C.141)
r 3 = k h z 3 + tm ti
r4 = kh z 4 + Atm ti + tm ti t
(C.142)
(C.143)
r5 = kh z 5 + Btm ti + Atm ti t
r6 = kh z 6 + Ctm ti + Btm ti t
(C.144)
(C.145)
r7 = kh z 7 + Dtm ti + Ctm ti t .
(C.146)
Since (C.135) is in the form of (A.1) we are now ready to apply the optimization conditions (A.9)(A.12) for determining the proposed analytical control law regarding
parameters tn , tv , ti .
290
Optimization Condition: a0 = b0 .
By applying the first optimization condition to the closed loop transfer function
(C.135) results in
kh = 1
(C.147)
which implies that the final closed loop control system exhibits steady state position,
velocity error. From (C.135) it is apparent that if kh = 1, then
N (s ) = kp tn tv s 2 + s kp (tn + tv ) + kp
(C.148)
D(s ) = + kp kh tn tv s 2 + s kp kh (tn + tv ) + kp kh
(C.149)
and
respectively.
Optimization Condition: a12 2a2 a0 = b12 2b2 b0 .
The application of (A.10) to (C.135) results in
tn2 + tv2 = 0,
(C.150)
which is not accepted since for both tn , tv , conditions tn > 0, tv > 0 must
hold by.
Optimization Condition: a22 2a3 a1 + 2a4 a0 = b22 2b3 b1 + 2b4 b0 .
In similar fashion, the application of (A.10) to (C.135) does not result in an acceptable
relation and therefore the right part of (A.10) is set to zero. In that a22 2a3 a1 +
2a4 a0 = 0 is used.
(C.151)
(C.152)
or finally
ti =
(C.153)
291
(C.154)
At this point, the same line is adopted, as the one followed in Sect. 5.2.1, regarding
the determination of parameter tv . Therefore
tv = nt
(C.155)
(C.156)
where
= n(n 1)2 t4 [16n 8(2B 1)(n 3)].
(C.157)
(C.158)
is satisfied. To this end, n > 0 and n > 1. If n is chosen such that n > 1 then it is
easily shown that n > 1, 16n 8(2B 1)(n 3) > 0.
Index
A
Acceleration error, ix, 117119, 134, 155, 158,
264
Actuator, 1214, 101, 168
Amplitude, 23, 34, 79, 86, 195, 201
Angle, 23, 79, 200
Automatic tuning, 199, 212, 227, 228
conjugate complex, 227, 228
conjugate complex poles, 228
PID regulators, 202204, 206, 207, 209
B
Bounded
input, 16
output, 16
reference, 16
signal, 16
C
Capacitor, 76
Capacitor bank, 79
Closed loop, 21, 26
Closed loop transfer function, 43
Command signal, xx, 1416, 18, 87, 98, 101,
102, 105, 106, 166, 174, 176, 177,
182, 187, 195, 234
Comparator, 12, 13
Conjugate, xi, 123
complex, 95, 226
complex poles, xi, xvi, 224, 225, 228
complex zeros, 43, 226
Control law
optimal, 26
Control loop, 1416, 21
current, 78
speed, 8, 20
type, 19
typeI, 20, 26
typeII, 20, 26
typeIII, 26
typep, 21
Control system, 27, 32
Controller, 12, 13, 15
H , 23
stable, 16
two degrees of freedom 2DoF, 94
Conventional PID tuning, 36
typeI control loops, 33
Conventional tuning, 32, 37, 39
drawbacks, 41
Converter
grid side, 86
motor side, 86
Criterion
magnitude optimum, 32, 45
symmetrical optimum, 32
Critical frequency, 93
Cross coupling, 80
Current
DC link capacitor, 79
load, 79
Current controller
integrators time constant, 78
PI controllers zero, 78
Current feed forward, 80
Current reference, 79
D
Damping ratio , 42
293
294
DC link voltage control, 76
DC link voltage controller, 79
Denominator, 42, 90
Direct torque control, 8, 20
Disturbance, 21
output, 15
rejection, 18
Disturbance rejection, 33
Domain
frequency, ix, xi, 31, 33, 35, 48, 85, 87, 94,
105, 117, 134, 144, 159, 162, 179, 182,
194, 196, 200, 286
time, 92, 145, 192194, 196, 201
Dominant time constant, 90
Dynamic behavior, 35
E
Electric
motor, 14
motor drive, 8, 14, 20, 32
Energy conversion
shaft generator, 86
wind energy, 86
Equivalent sum time constant, 34
Error
control, 33
steady state acceleration, 20
steady state position, 20, 86
steady state velocity, 20, 86
three phase, 79
External
disturbance, 12
filter, 151
External controller, xv, 93, 130, 132, 146, 148,
181
F
Feedback, 13
control loop, 26
output, 21
path, 34
Filter
external, 113, 174, 181
reference, 181, 190
time constant, 113
Final value theorem, 19
First order process, 33, 79
Frequency range, 11, 23, 34
Frequency region
low, 35
Frequency response, 23, 25, 92
Frequency spectrum, 21
Index
G
Gain
proportional, 14
Grid
connected converter, 76, 85
current controller, 79
current measurement active part, 80
current measurement reactive part, 80
current reference active part, 80
current reference reactive part, 80
impedance, 76
transformer, 76
voltage measurement, 80
H
Half plane
left, 16, 42
right, 16, 42
Higher order terms, 42
I
Imaginary half plane, 42, 269
Imaginary part, 47, 249
Impedance
leakage, 76
magnetizing, 76
Inertia, 8, 20
Input, 11
disturbance, 15
Integral control action, 34
Integrating process, 85, 119, 127, 158, 171,
179, 183185
nonminimum phase, 85, 187
time delay, 186
Integrator, 8, 20
Internal model control, 85
Island network, 86
L
Linear, 14
Load, 12
current, 79, 81, 87, 110, 111
disturbance, 21, 5659, 61
drive, 112
electric, 86
step response, 81
torque, 88
M
Magnitude, 21, 27, 90
Index
Main diesel engine, 86
Margin
phase, 36, 92, 101, 102, 109, 123, 145, 204
Modulation
angle, 79, 200
index, 79
Motor, 12
N
Negative
feedback, 12
Network
frequency, 79
Noise, 12
rejection, 21
Normalized
closed loop transfer function, 166, 179
control loop transfer function, 196
plant transfer function, 188
Normalized time constant, 43
O
Openloop transfer function, 93
Optimal
disturbance rejection, 21
Optimization
conditions, xii, 11, 25, 27, 97, 126, 141,
161, 173, 181, 249, 253, 256, 261, 263,
269, 274, 284, 289
Order
controller, 23
Output, 11, 13
control loop, 15
disturbance, 15
sensitivity, 15, 68
tracking, 12
Overshoot, 36, 82, 9294, 101, 107, 122, 123,
127, 130, 145, 146, 148151, 154,
169, 174, 181, 182, 188, 190, 199,
228231, 233, 235, 239
P
Phase locked loop, 79
PI control, 37
PID control, 38, 41
Plant
five dominant time constants, 47, 71
input, 33
large zeros, 51
nonminimum phase, 51, 74
one dominant time constant, 46
295
output, 33
time delay, 49, 73
Point of common coupling, 79
Polynomial
denominator, 21, 35
numerator, 21
Power
converters, 20
Power converter, 80, 86
Power electronics, 14
Process
controlled, 11, 12
Q
Quadratic reference signal, xv
R
Real process, 33
Real zeros, 41
Realworld application, 15
Reference, 12
input, 14
ramp, 86
signal, 21
Reference frame
d q, 78
synchronous, 79
Reference tracking, 33
Resonance frequency, 25
Revised control law
analog design
typeI control loops, 45
typeII control loops, 98
typeIII control loops, 126
digital design
typeII control loops, 165, 174, 181
Revised PID tuning
typeI control loops, 42
typeII control loops, 94
typeIII control loops, 123
Robust performance, 18
Robustness, 11
feedback path, 62
plants DC gain, 64
plants dominant time constant, 64
S
Sampling
period, x, xi, 161, 179, 269, 270, 278, 286
296
time, xv, 161, 165167, 170, 174, 175, 181,
182, 185, 189, 191, 194, 196, 269
Second order system, 42
Sensitivity, 1820
command signal, 15, 68
complementary, 11, 18, 20
input, 15
output, 36
Setpoint response, 32
Shape preservation, 35
Signal
bounded, 16
command, 13, 14, 16
disturbance, 16, 21
error, 16
reference, 16
Smith predictor, 85
Speed
PI control, 8, 20
Stability, 15
control loop, 16
internal, 11, 16
matrix, 16
Stable
real poles, 32
Steady state
acceleration error, 8, 158
position error, 4, 7, 8, 34, 134, 158
velocity error, ix, 4, 7, 8, 134, 158
Step response, 35, 92
T
Time
rise, 36
Time constant, 37
dominant, 37, 39
integrator, 45
parasitic, 39
Time delay, 14
Time delay all pole approximation, 43
Transfer function
T (s), Si (s), Su (s), 16
open loop, 15, 20, 86
Transformer
leakage
inductance, 80
resistance, 80
magnetizing
Index
inductance, 80
resistance, 80
Tuning
adjustable, 85
explicit, 85
TypeII control loops, 81
TypeIV control loop, xi, 143145, 153, 154
U
Unity, 11, 45
frequency response, 32
Unmodeled dynamics, 34
controller, 43
Unstable
I control action, 89
PI control action, 90
Unstable control loop, 89, 90
V
Vector control, 8, 20
cascaded, 77
Voltage, 14
DC link, 20
source, 76
source inverters, 14
W
Wind energy conversion system, 86
Wind turbine, 86
Winding time constant (stator), 89
Z
Zero
controller, 119, 136, 144
error, 20
order hold, 161, 171
PID controller, 174
plant, 165
pole cancellation, 140, 179, 180
sensitivity, 21, 195
steady state acceleration error, 134
steady state position error, 117, 134
steady state velocity error, 134
typeIV control loop, 143