Vous êtes sur la page 1sur 12

Article

Application of stability region centroids


in robust PI stabilization of a class of
second-order systems

Transactions of the Institute of


Measurement and Control
34(4) 487498
The Author(s) 2011
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0142331211400117
tim.sagepub.com

Mohammad Amin Rahimian and Mohammad Saleh Tavazoei

Abstract
In this paper we offer a tuning method for the design of stabilizing PI controllers that utilizes the stability region centroid in the controller parameter
space. To this end, analytical formulas are derived to describe the stability boundaries of a class of relative-degree-one linear time invariant secondorder systems, the stability region of which has a closed convex shape. The so-called centroid stable point is then calculated analytically and the resultant
set of algebraic formulas are utilized to tune the controller parameters. The freedom to choose the surface density function in the calculation of
centroid stable point provides the designer with the possibility to incorporate optimal or robustness requirements in the controller design process.
The proposed method uses the stability regions in the controller parameter space to ensure closed-loop stability, and, while offering robust stability
properties, it does not rely on predetermined information with regard to the nature or range of parameter variations and coefficient uncertainty
bounds. Being situated away from the boundaries of the stability region in the controller parameter space, controllers designed based on the centroid
method are both robust and non-fragile.

Keywords
Numerator dynamics, optimal tuning, PI controller, robust tuning, second-order system, stability region, stabilizing controllers, surface density

Introduction
Favoured for their simple structure, eectiveness and robustness, PID controllers remain of paramount importance in the
control of industrial processes (Astrom and Hagglund, 1995,
2005). Moreover, good man-machine interfaces for the specication of controller structure and parameters, as well as
ecient tuning tools, are necessary requirements for widespread adoption of any industrial controller; and methods
that can expedite, ameliorate and simplify the tuning process
are highly sought after (Astrom and Hagglund, 2001).
Among various design criteria to be satised by a controller, closed-loop stability is the most basic, and the stability of
nal closed-loop systems for both fractional-order and integer-order PID controllers have been the subject of many previous studies (Nusret Tan et al., 2006; Hamamci, 2007; Fang
et al., 2009). The set of controller parameters yielding a stable
closed-loop system is known as the stability region, and a lot
can be learned about PID control through analysis of stability
regions. The problem of nding all stabilizing PID controllers
for a given plant, which leads to the stability domains in the
controllers parameter space, has traditionally been dealt with
using Nyquist plot, stability boundary locus, characteristic
equation and frequency-based methods. Applications of the
HermiteBiehler theorem (Caponetto et al., 2010), as well as
elaborate polynomial calculations (Hohenbichler and
Ackermann,
2003a;
Soylemez
and
Baki,
2003;
Hohenbichler, 2009), have been the driving force behind
some of the recent results on this topic. The author in
Hamamci (2008) has introduced a method for investigating

the stabilization of fractional-order PI controllers that is based


on plotting the global stability region in the (kp, ki) plane. A
similar method is adopted by the authors in Hamamci and
Koksal (2010) to derive the stability regions for fractionalorder PD controllers in the (kp, kd) plane. The authors in
Rahimian and Tavazoei (2010a,b) have provided an extension
of the method used in Hamamci (2008) to plot the stability
regions for the integer-order approximations of PIl and PDm
controllers. The same methods form the basis of the design
scheme used in Rahimian et al. (2010) for the stabilization of
two-mass drive systems with elastic coupling.
The main theme of this paper is the calculation of the centroid point for the stable region in the controller parameter
space. This so-called centroid stable point serves as the design
choice, according to which controller parameters are set. The
centroid stable point has the advantage that it lies at the farthest possible distance from the boundaries of the stability region
and will therefore ensure the robust stability of the resulting
closed-loop control system. Accordingly, while oering robust
stability properties, the proposed method does not rely on any
information pertaining to the nature or range of parameter
variations. Furthermore, the freedom to choose the surface
Electrical Engineering Department, Sharif University of Technology, Iran
Corresponding author:
Mohammad Saleh Tavazoei, Electrical Engineering Department, Sharif
University of Technology, Tehran, Iran
Email: tavazoei@sharif.edu

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

488

Transactions of the Institute of Measurement and Control 34(4)

density in the calculation of centroid stable point provides a


more exible tuning strategy, and therefore an easier way to
achieve control requirements such as optimality or robustness.
The remainder of this paper is organized as follows. The
method of Hohenbichler and Ackermann (2003b), Hamamci
(2008) and Hamamci and Koksal (2010) is used in the next
section to derive an analytical description for the boundaries
of the stability region in a closed-loop system comprised of a
PI controller and a strictly proper second-order plant transfer
function. Next, in the section titled Centroid stable point
and the proposed tuning formulas, a set of conditions are
set forth under which the stability region described in its preceding section has a closed convex shape, for which the centroid exists and is meaningful stability-wise. Following the
specication of constraints on the plant parameters, a constant value for the surface density function is assumed and the
centroid of the stability region is analytically calculated. The
section titled Choice of surface density for robust and optimal
tuning elaborates on the eects of the choice of surface density
on the calculation of the centroid stable point, and illustrates
how various surface density functions can be harnessed to provide the designed control system with robustness and optimality
properties. To this end, a specic example, where surface density function is adjusted to procure optimal disturbance rejection properties, is discussed analytically, and algebraic formulas
for the corresponding centroid stable point are proered. Using
constant and tailored surface density functions, two choices of
centroid stable points are computed for an example system in
sections Centroid stable point and the proposed tuning formulas and Choice of surface density for robust and optimal
tuning respectively, and the corresponding closed-loop systems are simulated and compared in the section titled
Simulation results and discussion. A second example in the
latter section compares the performance of the proposed tuning
rules with classical ZieglerNichols rules. Lastly, some concluding remarks are provided in the nal section concerning the
scope and possible extensions of the proposed approach.

Stability regions for the PI controller


Consider the basic closed-loop control system depicted in
Figure 1, where y is the output and r is the reference input.
The inputoutput relation for the closed-loop system of
Figure 1 in the Laplace domain is given by
Ys
CsGs

,
Rs 1 CsGs

with G(s) and C(s) indicating the plant and compensator


transfer functions, respectively; the characteristic equation

C(s)

G(s)

Figure 1 A basic closed-loop control system, using unity negative


feedback.

for the closed-loop system will be derived from setting the


denominator of (1) to zero, and is as follows:
1 CsGs 0:

The numerator of (2) is called the characteristic polynomial


and is denoted by P(s). This paper focuses on the problem of
controlling a general second-order plant given by
Gs

as b ,
s2 cs d

with a PI controller given by


Cs kp

ki
:
s

Numerous real-world processes can be shown to observe the


transfer function in (3), and second-order processes with
numerator dynamics are the focus of Seborg et al. (2004:
Section 6.1). There, it is explained that second-order systems
with right-half-plane (RHP) zeros exhibit the inverseresponse phenomenon, where the direction of an initial
response to a step input contradicts that of its nal steady
state. The authors in Seborg et al. (2004) then describe two
practical scenarios, involving the liquid levels in a distillation
column and the temperature of a chemical reactor with
exothermic reactions. In both cases, competing dynamic
eects that operate on two dierent time scales lead to the
inverse-response behaviour. Accordingly, second-order systems with numerator dynamics that exhibit inverse-response
or overshoot behaviour can occur whenever two physical
eects act on the process output variable in dierent ways
and with dierent time scales.
Another classical example of a second-order process with
numerator dynamics is a mass-spring-dashpot (MSD) system
in series, where a mass, m, is attached to a spring with constant k, which is attached to a dashpot with damping coecient h, which is attached to a wall. Accordingly, the transfer
function from the force F, acting on the mass m, to its speed n,
is given by
ns
hs k
:

Fs mhs2 mks kh

Moreover, several of the plants that are investigated in the


literature are formulated in form of (3). These include the
transfer function from motor torque to load torque in a
two-mass drive system with elastic coupling in Rahimian
et al. (2010), and the transfer function from steering
angle to tilt angle for the unrideable bicycle in Klein
(1989). Other examples from robotics include the free load
linear actuator model for the biomimetic robot in Robinson
et al. (1999) and the transfer function of the end eector
motion/command position for the single-link manipulator
model in Xu and Paul (1988). The transfer function
from the elevator deection (de) to the pitch rate (q) in
ight control systems (Oosterom and Babuska, 2006;
El-Mahallawy et al., 2011) also follows the same dynamics
as described by (3).

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

Rahimian and Tavazoei

489

In addition to the aforementioned examples, in which real


world plants are modelled by the same type of transfer functions as (3), there are several cases where approximation,
order-reduction or identication methods lead to a formulation of the process dynamics that is in accordance with (3).
The transfer function for every First-Order Plus Dead-Time
(FOPDT) system, in which the time delay is replaced by a
rst-order Pade approximation, takes the form of a secondorder system with a non-minimum phase zero in the numerator, as does the transfer function for every Second-Order Plus
Dead-Time (SOPDT) system, in which the exponential term is
approximated by the rst two terms in its Taylor series. The
author in Luus (1999) introduces a method to optimally
choose the coecients for a second-order reduced model,
the transfer function for which is given by (3). The optimizations are performed in the frequency domain, using a multipass optimization technique, and the resulting second-order
reduced models can closely follow the Nyquist plots of the
original fth- and eighth-order systems. Last, but not least,
plant models of the type in (3) can be the result of an identication procedure. The authors in Wang and Chen (2009) use
subspace system identication methods to determine the
transfer function matrices for the Multi-Input Multi-Output
dynamics of a proton exchange membrane fuel cell (PEMFC)
system. The resulting transfer functions, after being converted
into continuous-time by zero-order-hold, are all in the
form of (3).
According to (2), the characteristic polynomial can be calculated as
Ps s3 c akp s2 d aki bkp s bki :

The Hurwitz stability criterion for the characteristic polynomial in (6) is satised if and only if every root of P(s) lies in
the left half plane (LHP). In other words, the imaginary axis
and the origin are the only places where the stability shift of
the system will occur (Fang et al., 2009). However, the positions of roots of P(s) will change continuously as long as its
coecients are continuous functions of the plant or controller
parameters and those parameters are changed continuously.
Thus, a stable polynomial, P(s), whose roots all lie in the
LHP, becomes unstable if and only if at least one root crosses
the imaginary axis. Accordingly, the set of coecients that
would lead to a stable closed-loop system can be denoted as
stability domains in the parameter space of the characteristic
polynomial P(s). These stability domains are determined by
the following three boundaries, which describe the roots
crossing from the LHP to the RHP and vice versa
(Hohenbichler and Ackermann, 2003b; Hamamci, 2008;
Hamamci and Koksal, 2010):

(b) Complex Root Boundary (CRB): A pair of complex roots


will cross over the imaginary axis at s jv, and their corresponding locations in the parameter space are obtained by
substituting s jv in P(s) and setting its real and imaginary
parts to zero.
(c) Innite Root Boundary (IRB): The last possibility is for a
root to cross the imaginary axis at innity (|s| ! ), and it
can be determined by setting the coecient for the term
with the largest power of s in (6) to zero. Here, since the
term with the largest power corresponds to s3 and its coecient cannot be zero, the IRB does not exist.
Once the coecients of (6) are described in terms of controller parameters, the three boundaries given by Denition 1
will translate into stability boundaries in the parameter space
of the PI controller, i.e. {kp, ki}. These stability boundaries
separate dierent regions in the parameter space, and so, to
determine the stability of a given region, it suces to check
the stability of one test point within that region (e.g. by the
Nyquist criterion or through investigation of the roots of the
corresponding characteristic polynomial).
Next, substituting s jv in (6) and equating the real and
imaginary parts to zero results in the following two equations,
respectively:
bki c akp v2 ,

v2 d bkp aki :

Eliminating the parameter v between the two equations in (8)


and (9) yields
ki

cd cb ad kp abkp 2
:
b  ca  a2 kp

In light of the previous discussion, the stability boundaries for


the control system described by Figure 1 and equations (3)
and (4) are given by equations (7) and (10), which specify the
RRB and CRB, respectively.
Lastly, in order to determine the stability regions in the
controller parameter space, we should test the stability of
single test points in each of the regions generated from the
intersection of the boundaries given by (7) and (10).
Example 1. An unstable non-minimum-phase second-order
relative-degree-one linear time invariant (LTI) system.
In order to demonstrate the aforementioned procedure,
the following unstable non-minimum-phase second-order
relative-degree-one LTI system, which is taken from Roup
and Bernstein (2003: Example 1), is considered:

Denition 1. The Three Root Boundaries:


G1 s
(a) Real Root Boundary (RRB): A real root crosses over the
imaginary axis at the origin (s 0), and for P(s) given by
(6) it can be determined as
bki 0 ! ki 0,
i.e. by setting the constant term to zero.

10

s5
:
s2 15s  5

11

The sample plant in (11) can be derived from (3) by setting


a 1, b 5, c 15 and d 5, and it is used throughout the
rest of the paper to demonstrate the applicability of the proposed method. Having an unstable pole at p 0.3262 and a
non-minimum phase zero at z 5, the system in (11) is a

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

490

Transactions of the Institute of Measurement and Control 34(4)

typical hard-to-control second-order plant (Astrom, 2000).


The presence of the RHP zero imposes a fundamental limitation on control, and high controller gains will induce closedloop instability (Skogestad and Postlethwaite, 1996). The stability region for (11) can be computed using the boundaries in
(7) and (10), and is depicted in Figure 2. The closed and
convex shape of the stability region in Figure 2, which occupies a limited portion of the plane, is a further indication that
the underlying closed-loop system is not well-behaved and is
hard to control.
In the next section we nd the centroid of the stability
region depicted in Figure 2 analytically, and derive a set of
algebraic formulas that can be used as a suitable choice for
the unknown controller parameters, kp and ki. Such a choice
has the advantage that it lies at the farthest possible distance
from the boundaries of the stability region and would therefore ensure the robust stability of the resulting closed-loop
control system.

Centroid stable point and the proposed


tuning formulas
This section focuses on the calculation of stability region centroids. This is motivated by the fact that the geometric centre
of an objects shape can be a best option if the aim is to avoid
its boundaries and exterior as much as possible. This is true
provided that the underlying object has a closed convex
shape. Accordingly, prior to the calculation of centroid
stable points, a set of constraints on the plant parameters a,
b, c and d are needed to ensure that the stability regions have
indeed a closed convex shape, as was the case with Figure 2,
for which the stability region was an upward semi-parabola,
capped at the top by the kp axis. These conditions ensure that

calculation of the centroid for the stable region is justied


from the perspective of closed-loop system stability. Such
systems, for which the stability region is bounded, would in
eect be hard to control, and examples often include plant
transfer functions that are unstable or non-minimum phase.
Peering into (10), it will dawn on us that the CRB crosses
the kp axis at the following two points:
kp r1  c ,
a

12

kp r2  d ,
b

13

and it has a vertical asymptote at


kp a b 2ac :
a

Depending on the relative locations of the two roots in


(12) and (13) and the asymptote in (14), several cases may
arise, each of which is realized under a particular set of conditions on the plant parameters. Among the many possibilities, those specied in Tables 1 to 4 culminate in stability
regions with a closed convex shape. The conditions in
Table 1 and Table 2 yield a downward semi-parabola that
is capped at the bottom by the kp axis, while those in
Table 3 and Table 4 yield an upward semi-parabola that is
capped at the top by the kp axis. Table 1 and Table 3 correspond to the cases where ab < 0, while Table 2 and Table 4 list
the conditions for the cases where ab > 0. The plant numerator dynamics in the former case include a non-minimum
phase zero, leading to closed-loop systems that are particularly hard to control. Each table lists the additional

100
CRB
RRB

50
0
50

ki

100
150

Stability region

200
250
300
350
400
40

30

20

14

10
kp

10

20

Figure 2 The stability region is plotted for the control system of Figure 1, where G(s) and C(s) are given by (11) and (4), respectively.

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

Rahimian and Tavazoei

491

constraints that should be satised for various combinations


of signs of the parameters c and d.
The analytical formulas derived for the CRB in the previous
section provide us with the possibility of selecting a point inside
the stable region that is at the farthest possible distance from
every point on the boundary. The coordinates (xc, yc) for the
centroid of a graph in the (x, y) plane are given by
xc
yc

RR
RR

sx, yx dxdy
,
M

15

sx, yy dydx
,
M

16

where I1, I2 and Mc are given by


I1

I2 

Z Z

and s(x, y) denotes the surface density function (Thomas and


Finney, 1999). The double integrals in (15), (16) and (17) are
calculated over a stable region such as the one in Figure 2.
Under the assumption that no information is available as to
the range or type of variations against which the closed-loop
system should exhibit robustness, the surface density s(x, y) is
set to be a constant number and is thus eliminated from the
numerators and denominators of (15) and (16). Using (7) and
(10) for the boundaries of the stability region, and assuming
that one of the condition sets in Tables 1 to 4 are satised, the
centroid stable point for the closed-loop system of Figure 1,
with the plant and controller given by (3) and (4) respectively,
can be calculated analytically. The results are the following
algebraic tuning rules:
kp c

I1
,
Mc

18

ki c

I2
,
Mc

19

S2
S P
1  Q,
2a3 b a4 b

21
22

and P, Q, R and Sn are dened as

17

sx, y dydx,

20

S3
PS2 P2 3b2 RS1 R b2

Q,
6a5 b 2a6 b
2a7 b
a3
Mc 

where M is given by
M

S3
PS2
RS
b Q,

6 1 ac 
a2
3a4 b2 2a5 b2
a

P a2 d b2 ,

23

 
R ,
Q bR
ln
a5
b2

24

R P  abc,

25

Sn an dn  bn cn , n 2 N:

26

The algebraic tuning formulas introduced in (18) and (19) can


be applied to Example 1 and the corresponding centroid
stable point can be computed as
kp c1 8:8929,
ki c1 8:9291:

27

Figure 3 denotes the centroid stable point along with the


stability region for the system in (11).
Here it should be noted that the proposed choice of the
stability region centroid as the tuning rule for setting controller parameters is inherently conservative. The centroid stable
point tuning rules in (18) and (19) are, in fact, a sucient but
not necessary condition for closed-loop system stability, and
using them as the basis for design is justied only when stabilization is the rst and foremost criterion that is expected to
be satised by the control system. Such conservatism in

Table 1 Constraints on the plant parameters for the case where a < 0
and b > 0

Table 3 Constraints on the plant parameters for the case where a > 0
and b < 0

Signs of parameters c and d

Additional conditions

Signs of parameters c and d

Additional conditions

c < 0,
c > 0,
c < 0,
c > 0,

 ac \  db \
 ac \  db \
 db \ bac
a2
None

c < 0,
c > 0,
c < 0,
c > 0,

bac
a2
bac
a2
bac
a2

d>0
d<0
d<0
d>0

bac
a2
bac
a2

or  db \  ac
or  db \  ac

d>0
d<0
d<0
d>0

\  db \  ac or  ac \  db
\  db \  ac or  ac \  db
\  db
None

Table 2 Constraints on the plant parameters for the case where a < 0
and b < 0

Table 4 Constraints on the plant parameters for the case where a > 0
and b > 0

Signs of parameters c and d

Additional conditions

Signs of parameters c and d

Additional conditions

c < 0,
c > 0,
c > 0,
c < 0,

bac
a2
bac
a2
bac
a2

c > 0,
c < 0,
c > 0,
c < 0,

 ac \  db \
 ac \  db \
 db \ bac
a2
None

d<0
d>0
d<0
d>0

\  db \  ac or
\  db \  ac or
\  db
None

d
b
d
b

\
\

c
a
c
a

d>0
d<0
d<0
d>0

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

bac
a2
bac
a2

or  db \  ac
or  db \  ac

492

Transactions of the Institute of Measurement and Control 34(4)

60
50
40
CRB
Centroid stable point
RRB

30

ki

20
10
0
10
20
30
40

30

20

10
kp

10

20

Figure 3 The centroid stable point is depicted for the control system given by Figure 1 and equations (11) and (4).

design is best justied in the case of highly uncertain systems,


where plant parameters can only be determined with a limited
precision and are always subject to variations. Lack of information with regard to the nature and range of such variations
is a second factor which favours the use of the stability region
centroid as the tuning point. Another class of systems for
which the tuning rules of (18) and (19) are particularly
useful is the class of highly unstable systems with narrow
stability regions, where it becomes critically important to
avoid the boundaries of the stability region. Being situated
away from the boundaries of the stability region in the controller parameter space, controllers designed based on the
centroid method are both robust and non-fragile.
Accordingly, such controllers can tolerate certain variations
in the parameters of both the controller and the plant (Keel
and Bhattacharyya, 1997; Makila et al., 1998). The parameter
uncertainties can be due to either inherent physical properties
or modelling diculties, and having control systems which
remain operational in the face of parameter variations is
desirable in both cases.
In the next section, the eects of the choice of surface
density function on the calculation of the centroid stable
point in (15) to (17) is investigated. It is further demonstrated
through an example that the choice of surface density function can, in fact, be leveraged to produce desirable optimality
or robustness properties.

Choice of surface density for robust and


optimal tuning
In the previous section it was pointed out that in the absence
of any extra information the surface density function in (15),
(16) and (17) is considered to be a constant number, which is

then cancelled out in the calculations. The uniform density,


in eect, corresponds to the case where robust stability is to
be guaranteed and no extra information is available with
regard to variations of plant parameters and coecient
uncertainty bounds. This, however, need not necessarily be
the case. For instance, if we know how the shape of the
stability region is aected by variations in the plant parameters, then we might be able to place the centroid so that it
is properly distanced from sensitive locations. Accordingly,
the undesirable locations are discriminated against by being
assigned lower weights when designating the surface density
function across the stability region. The following paragraphs elaborate on the choice of surface density function
and how it can be modied to satisfy various design
specications.
By and large, any tuning formula can be thought of as a
centroid stable point which is derived through (15), (16) and
(17) with an appropriate choice of surface density function s.
In general, the surface density function will be adjusted so
that less weight is given to the undesirable regions and the
centroid stable point is drawn toward more favourable
regions because of their higher weights. To clarify this
point, an example is given in which the surface density function is designed to provide the tuned control system with
optimal disturbance rejection characteristics.
Tuning methods based on an optimization approach have
received
considerable
attention
in
the
literature
(Panagopoulos et al., 2002). The problem of designing an
optimal PI controller, C*(s), given by (4), can be formulated
as determining the optimal parameters kp  and ki  so that a
set of constraints are satised and a cost function J(kp, ki) is
minimized. In general, the parameter optimization design
process consists of searching the space of variable controller
parameters as a function of some performance index J to

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

Rahimian and Tavazoei

493

determine where the performance index is minimized (Ogata,


2009).
The occurrence of eective constraints which necessitate the
optimal solution to reside at the boundaries of the admissible
parameter domain would mean that there is no guarantee for
the necessary control requirements to remain satised under
various parameter variations and coecient uncertainties.
Using the centroid approach, however, the centroid point is
guaranteed to maintain a safe distance from admissible
region boundaries, while the optimality properties are more
or less preserved through the appropriate choice of surface
density function s(kp, ki). Such a choice of s(kp, ki) would
have to be a positive decreasing function of the cost J so that
the regions with higher values of J are discriminated against.
The positive s would ensure that the centroid is within the
admissible region, provided that it has a closed convex shape.
In eect, it suces for s to not change sign across the region for
which the centroid is being calculated. This is due to the fact
that the actual positive or negative sign of a uni-sign function s
is cancelled out from the numerator and denominator of both
(15) and (16), and hence it does not aect the nal result.
The freedom in choosing the function s can be a valuable
asset in the management of the important trade-o between
performance and robustness in control systems (Kristiansson
and Lennartson, 2002). On the one hand, aggressive choices
of s, which decrease drastically with increasing J, would culminate in centroids that are closer to the truly optimal solutions. Such aggressive choices, on the other hand, may
compromise the closed-loop system robustness by being situated too close to the admissible region boundaries. In other
words, with the choice of s, the designer can decide to move
away from a conservative design and closer to an optimal
one, or vice versa.
Examples of decreasing functions which may be used to
describe the surface density function in terms of a positive
cost function J include

where e(t) is the error to a step input function. If the closedloop system output is denoted by y(t), then e(t) is given by
et 1  yt, 8t>0:

31

Here, the cost function to be minimized is chosen as follows:


J jIEj:

32

The integral performance index in (30) can be used to evaluate the performance of a designed control system or, as here,
for optimal tuning of xed structure controllers. In the latter
case, the parameters of the control system are optimized by
minimizing the integral performance index given by (32). In
Astrom and Hagglund (1995) it is shown that for a stable
closed-loop system, if the error is initially zero and a unit
step disturbance is applied at the plant input, then
IE 1 :
ki

33

Accordingly, the closed-loop system disturbance rejection can


be optimized by minimizing the IE criterion, and it is a natural choice for the control of quality variables for processes
where the product is sent to a mixing tank (Astrom and
Hagglund, 1995). Several PID tuning rules based on minimization of the IE criterion have been presented, and some
incorporate gain and phase margin specications to ensure
stability and robustness of the controlled systems (Astrom
et al., 1998). The authors in Hwang and Hsiao (2002), for
instance, present an eective solution to the non-convex optimization problem, which arises in the design of stabilizing PI
and PID controllers based on the minimization of the integral
of the error caused by a step disturbance and subject to constraints on maximum sensitivity Ms.
Using the results in (29), (32) and (33), we can write

s 1N , N 1, 2, 3, 4, . . . ,
J

28

where lower values of N correspond to more conservative


choices, and higher values of N are chosen when aggressively
optimal solutions that are near the truly optimal one are
preferred. Depending on the relative importance of robustness or optimality properties, the power of J in the denominator of (28) may be decreased or increased. In the following
paragraphs an optimal tuning scenario is considered, in which
s is chosen as
s 1 :
J

34

Next, supposing that one of the condition sets in Tables 1 to 4


is satised, the surface density function given in (34) can be
used in the context of (15) to (17) to calculate the centroid
stable point for the stability boundaries given by (7) and (10).
The resultant centroid stable point is given by
kp 

I1 
,
I2

35

ki 

I2 
,
I2

36

29

Once the cost function J in (29) is described in terms of controller parameters kp and ki, the function s s*, given by
(29), can be used in the context of (15) to (17) to calculate
the corresponding centroid stable point. Integrated Error (IE)
is an integral performance index, which is dened (Astrom
and Hagglund, 1995) as
Z
IE
et dt,
30
0

s kp , ki 1 jki j:
J

where I2 is given by (21). I1  and I2  in (35) and (36) are


given by




I 1  f 1 kp r 2  f 1 kp r 1 ,

37





I 2  f 2 kp r 2  f 2 kp r 1 ,

38

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

494

Transactions of the Institute of Measurement and Control 34(4)

where kp r1 and kp r2 are given by (12) and (13), and the functions f1() and f2() are dened in the Appendix.
The algebraic tuning formulas introduced in (35) to (38)
can be applied to Example 1, and the corresponding optimal
centroid stable point can be computed as
kp 1 9:2808,

39

ki 1 12:8080:

Figure 4 denotes the optimal centroid stable point given by


(39), along with the centroid stable point in (27) computed for
the system in Example 1.
In the next section the two choices of controllers given by
the two centroid points computed in this and the previous
sections for the sample plant (11) are simulated, and the performance of the corresponding closed-loop systems are compared. A second example compares the performance of two
PI controllers: one is designed using the proposed centroid
scheme, the other is tuned using the ZieglerNichols frequency domain method.

Simulation results and discussion


The output responses of the closed-loop system of Figure 1
with the sample system in (11) and the two controllers given
by (27) and (39) are plotted in Figure 5. The results show that,
being situated at a farther distance from the stability region
boundaries, the controller in (27) yields a less oscillatory
response to the unit step input applied at t 0. The controller
in (39) is, however, more eective in the rejection of the unit
step disturbance which is applied to the plant input at t 5.
The Integrated Error (IE) is a common performance index,

which is dened as in (30). The value of IE for a unit step


disturbance has been calculated in Astrom and Hagglund
(1995) and is given by (33). Accordingly, the corresponding
values for the controllers in (27) and (39) are 0.1120 and
0.0781. This conrms that the integral of the error caused
by the unit disturbance at t 5 has a lower absolute value for
the controller in (39). This is to be expected, since the controller in (39) has been designed for an optimal performance
with regard to the IE performance index, and for its disturbance rejection.
Next, in order to compare the robust stability of the two
systems, uncertainties are introduced in plant parameters a
and c and their eects are investigated. First, a is increased
from its nominal value of a 1 until the corresponding system
is no longer stable. The closed-loop system with the controller
in (27) destabilizes at a 1.5, whereas the one with the suboptimal controller given by (39) becomes unstable at a 1.34.
This conrms the superior robustness of the centroid stable
point as compared to its optimal version. Repeating the
experiment by decreasing c from its nominal value of c 15
results in a similar observation. The closed-loop systems with
the controllers in (27) and (39) destabilize at c 10.37 and
c 11.51 respectively. It is worth highlighting that the true
optimal solution, which minimizes the cost function in (32)
subject to the stability condition, corresponds to the base of
the downward semi-parabola in Figure 4, and thus lies on the
boundary of the stability region.
Last, but not least, it is worth mentioning that the presence
of a non-minimum phase zero at z 5 for the plant in (11)
imposes inherent restrictions on the disturbance rejection
properties of the closed-loop system (Astrom, 2000). Due to
the so-called push-pull eect, low-frequency non-minimum

Centroid stable points

60
50

s = constant
s = 1 = |k i |
J

40
30

ki

20
10
0
10
20
30
40

30

20

10
kp

10

Figure 4 The two centroid stable points, derived using a constant and an optimal s, are depicted above.

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

20

Rahimian and Tavazoei

495

phase zeros inhibit superior disturbance rejection at low frequencies (Freudenberg and Looze, 1998). Accordingly,
increasing the parameter a from its nominal value of a 1
causes the zero at z 5a to move closer to the origin, thus
further degrading the disturbance rejection at low frequencies. The next example compares the performance of the

proposed tuning rules with classical ZieglerNichols rules


for an inverse response process.
Example 2. PI-control of an inverse response process.
Here, Skogestad and Postlethwaite (1996: Example 2.3) is
considered and the performance of the proposed centroid

Unit step and disturbance responses


2.5

2
s = constant

1.5

s = 1 = |k i|
J

y(t)

0.5

0.5

10

t
Figure 5 The responses for the closed-loop system of Figure 1, with (11) as the plant and the two PI controllers specified by (27) and (39).

Unit step responses


1.8
Ziegler and Nichols
Centroid stable point

1.6
1.4
1.2

y(t)

1
0.8
0.6
0.4
0.2
0
0.2

20

40

60
t (s)

80

100

120

Figure 6 The responses for the closed-loop system of Figure 1, with (40) as the plant and the two PI controllers specified by (41) and (42).

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

496

Transactions of the Institute of Measurement and Control 34(4)

stable point controller is compared with a PI controller that is


designed according to the classical tuning rules of Ziegler and
Nichols. The plant model (time in seconds) is given by

32s 1
:
G2 s
5s 110s 1

kp ZN 1:1400,

41

ki ZN 0:0898:

Table 5 Comparison of the performance measures for centroid stable


point and frequency domain ZieglerNichols tuning rules
Centroid

ZN

Phase margin
Gain margin
Phase crossover frequency
Gain crossover frequency
Rise time (0  90%)
Settling time (65%)
Overshoot
Decay ratio
Maximum sensitivity (Ms)

19.30198
1.6298
0.3375 rad/s
0.2370 rad/s
7.9916 s
65.2532 s
63.0855%
0.3602
3.9373

22.9355
1.7250
0.353 rad/s
0.2363 rad/s
8.0669 s
52.7526 s
53.5915%
0.3046
3.4579

kp c2 1:1563,
ki c2 0:0730:

40

The ZieglerNichols PI controller parameters, which are


derived from the ultimate gain and ultimate period according
to the frequency response method (Astrom and Hagglund,
1995), are set as follows (Skogestad and Postlethwaite, 1996):

Performance measure

The algebraic tuning formulas introduced in (18) and (19) can


be applied to the plant in (40) and the corresponding centroid
stable point can be computed as
42

The output responses and control signals for the closed-loop


system of Figure 1 with the sample system in (40) and the two
controllers given by (41) and (42) are plotted in Figures 6 and
7 respectively. It is important for the control signals not to
violate the actuator and plant dynamics due to the presence of
nonlinearities such as saturation.
The presence of integral action in form of a PI controller
removes the steady-state oset in the step response of the
closed-loop system. Table 5 compares various performance
measures for the two control systems and their corresponding
step responses. The results conrm that the ZieglerNichols
tuning rules are somewhat aggressive, culminating in a closedloop system with smaller stability margins and a more oscillatory response (Skogestad and Postlethwaite, 1996). On the
other hand, the conservative nature of the centroid tuning
rules leads to a closed-loop system with larger stability margins and a superior transient response. The conservative rules,
proposed in (18) and (19), are particularly useful for the
tuning of systems with narrow stability regions, such as the
one in Example 2, for which it is critically important to avoid
the boundaries.

Conclusion
This paper oered a set of algebraic tuning formulas that
would ensure robust stability of the closed-loop system without

Control signals for a unit step reference input


1.6
Ziegler and Nichols
Centroid stable point

1.4
1.2
1

u (t)

0.8
0.6
0.4
0.2
0
0.2
0.4

20

40

60
t (s)

80

100

120

Figure 7 The control signals for the closed-loop system of Figure 1, with (40) as the plant and the two PI controllers specified by (41) and (42).

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

Rahimian and Tavazoei

497

exploiting any prior knowledge as to the nature or range of


parameter variations in the plant or controller. The controller
parameters are chosen so that the designed system will lie at
the farthest possible distance from every point on the boundary of the stability region in the controller parameter space.
Analytical formulas for the stability regions of a general
second-order plant were derived and the desired controller
parameters were selected as the centroid of the stability
regions. Moreover, a set of conditions were introduced
which will ensure that the stability regions have a closed
convex shape, and the calculation of centroids is therefore
meaningful stability-wise. Unlike classical robust stabilization
techniques, the stability region centroid approach proposed in
this paper does not require the coecient uncertainty bounds
to be known or satisfy any inequality constraints. The conservative nature of the proposed tuning rules can prove particularly useful in the control of closed-loop systems with
narrow stability regions or highly uncertain parameters. The
devised scheme, which is developed for PI controllers and a
class of second-order relative-degree-one LTI plants, can be
further extended to PD and PID controllers as well as to the
general case of controllers for which the stability region has a
closed convex shape. In the case of PID and other three-parameter controllers, however, the design parameter space
involves three unknowns (kp, ki, kd) and the resulting calculations, which involve the centre of mass for a three-dimensional gure, are analytically cumbersome.

Acknowledgment
The authors would like to thank Samira Rahimian for her help with
calculation of centroids and derivation of conditions in Tables 1 to 4.

Funding
This research received no specic grant from any funding agency in
the public, commercial or not-for-prot sectors.

Conflict of interest
None declared.

References
Astrom K (2000) Limitations on control system performance.
European Journal of Control 6: 119.
Astrom KJ and Hagglund T (1995) PID Controllers: Theory, Design
and Tuning, 2nd edn. Durham, NC, USA: ISA.
Astrom KJ and Hagglund T (2001) The future of PID control.
Control Engineering Practice 9: 11631175.
Astrom KJ and Hagglund T (2005) Advanced PID Control. Durham,
NC, USA: ISA.
Astrom KJ, Panagopoulos H and Hagglund T (1998) Design of PI
controllers based on non-convex optimization. Automatica 34:
585601.
Caponetto R, Dongola G, Fortuna L and Gallo A (2010) New results
on the synthesis of FO-PID controllers. Communications in
Nonlinear Science and Numerical Simulation 15: 9971007.
El-Mahallawy A, Yousef H, El-Singaby M, Madkour A and Youssef
A (2011) Robust flight control system design using H
loop-shaping and recessive trait crossover genetic algorithm.
Expert Systems with Applications 38: 169174.

Fang J, Zheng D and Ren Z (2009) Computation of stabilizing PI


and PID controllers by using Kronecker summation method.
Energy Conversion and Management 50: 18211827.
Freudenberg J and Looze D (1998) Frequency Domain Properties of
Scalar and Multivariable Feedback Systems. New York: Springer
Verlag.
Hamamci SE (2007) An algorithm for stabilization of fractionalorder time delay systems using fractional-order PID controllers.
IEEE Transactions on Automatic Control 52: 19641969.
Hamamci SE (2008) Stabilization using fractional-order PI and PID
controllers. Nonlinear Dynamics 51: 329343.
Hamamci SE and Koksal M (2010) Calculation of all stabilizing
fractional-order PD controllers for integrating time delay systems.
Computers and Mathematics with Applications 59: 16211629.
Hohenbichler N (2009) All stabilizing PID controllers for time delay
systems. Automatica 45: 26782684.
Hohenbichler N and Ackermann J (2003a) Computing stable
regions in parameter spaces for a class of quasipolynomials.
In: Proceedings of the 4th IFAC Workshop on Time Delay
Systems.
Hohenbichler N and Ackermann J (2003b) Synthesis of robust PID
controllers for time delay systems. In: Proceedings of the European
Control Conference.
Hwang C and Hsiao CY (2002) Solution of a non-convex optimization arising in PI/PID control design. Automatica 38: 18951904.
Keel L and Bhattacharyya S (1997) Robust, fragile, or optimal? IEEE
Transactions on Automatic Control 42: 10981105.
Klein R (1989) Using bicycles to teach system dynamics. IEEE
Control Systems Magazine 9: 49.
Kristiansson B and Lennartson B (2002) Robust and optimal tuning
of PI and PID controllers. IEE Proceedings: Control Theory and
Applications 149: 1725.
Luus R (1999) Optimal reduction of linear systems. Journal of the
Franklin Institute 336: 523532.
Makila P, Keel L and Bhattacharyya S (1998) Comments on robust,
fragile, or optimal? [and reply]. IEEE Transactions on Automatic
Control 43: 12651268.
Nusret Tan CY, Kaya I and Atherton DP (2006) Computation of
stabilizing PI and PID controllers using the stability boundary
locus. Energy Conversion and Management 47: 30453058.
Ogata K (2009) Modern Control Engineering, 5th edn. Upper Saddle
River, NJ, USA: Prentice Hall.
Oosterom M and Babuska R (2006) Design of a gain-scheduling
mechanism for flight control laws by fuzzy clustering. Control
Engineering Practice 14: 769781.
Panagopoulos H, Astrom K and Hagglund T (2002) Design of PID
controllers based on constrained optimisation. IEE Proceedings:
Control Theory and Applications 149: 3240.
Rahimian MA and Tavazoei MS (2010a) Comparing the stability
regions for fractional-order PI controllers and their integerorder approximations. In: Proceedings of the 2010 IEEE
Conference on Decision and Control, pp. 720725.
Rahimian MA and Tavazoei MS (2010b) Stabilizing fractional-order
PI and PD controllers: An integer-order implemented system
approach. Proceedings of the Institution of Mechanical
Engineers, Part I: Journal of Systems and Control Engineering
224: 893903.
Rahimian MA, Tavazoei MS and Tahami F (2010) Fractional-order
PI speed control of a two-mass drive system with elastic coupling.
In: Proceedings of the 4th IFAC Workshop on Fractional
Differentiation and its Applications.
Robinson D, Pratt J, Paluska D and Pratt G (1999) Series elastic
actuator development for a biomimetic walking robot.
In: Proceedings of the IEEE/ASME International Conference on
Advanced Intelligent Mechatronics, pp. 561568.

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

498

Transactions of the Institute of Measurement and Control 34(4)

Roup A and Bernstein D (2003) Adaptive stabilization of secondorder non-minimum phase systems. In: Proceedings of the 2003
American Control Conference 1: pp. 225230.
Seborg DE, Edgar TF and Mellichamp DA (2004) Process Dynamics
and Control, 2nd edn. New York: Wiley.
Skogestad S and Postlethwaite I (1996) Multivariable Feedback
Control: Analysis and Design. New York: Wiley.
Soylemez NM, Munro T and Baki H (2003) Fast calculation of stabilizing PID controllers. Automatica 39: 121126.
Thomas GB and Finney RL (1999) Calculus and Analytic Geometry,
8th edn. Reading, MA, USA: Addison-Wesley Publishing
Company.
Wang FC and Chen HT (2009) Design and implementation of fixedorder robust controllers for a proton exchange membrane fuel cell
system. International Journal of Hydrogen Energy 34: 27052717.
Xu Y and Paul R (1988) On position compensation and force control
stability of a robot with a compliant wrist. In: Proceedings of the
IEEE International Conference on Robotics and Automation 2:
pp. 11731178.

Appendix
The centroid stable point with s*(kp, ki) |ki|
Functions f1() and f2(), which appear in the formulation of
the optimal centroid stable point in (37) and (38), are dened
as follows:

2 2
3
2 4
f1 x b x2 P x6 x 4
8a
4a
3a


bR2
ac  b hx  bPR kx,

gx

l x
a6
2a10
3 4
3
2
3
2 2
bR3
 b Px
11 2
 bP 7x
f2 x  b x3  P x
9
3a
12a
6a l x
3a5
2a


bP2 Rgx b3 R2
ac  b  b hx


gx

l x
a
a11
a11
2
2
2
 b11PR 2b PR
kx,
a l x
a7

where P and R are dened in (23) and (25), and l(x), g(x), h(x)
and k(x) are given by
l x a2 x  ca b,
gx lnl x,
2 

hx b 10R a2 c2 b2  2abc gx 0:5x2 a4  xa2 ac  b ,
a
b gx:
kx  x2 ac 
a
a4

Downloaded from tim.sagepub.com at INDIAN INST OF TECHNOLOGY on January 21, 2015

Vous aimerez peut-être aussi