Asian Journal of Control, Vol. 17, No. 3, pp. 1–11, May 2015
Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/asjc.931
A NEW TUNING METHOD FOR STABILIZATION TIME DELAY SYSTEMS
USING PI
D
CONTROLLERS
Sami Hafsi, Kaouther Laabidi, and Rihem Farkh
ABSTRACT
This paper presents a new design procedure to tune the fractional order PI
D
controller that stabilizes a frst order
plant with time delay. The procedure is based on a suitable version of the Hermite–Biehler Theorem and the Pontryagin
Theorem. A Theorem and a Lemma are developed to compute the global stability region of the PI
D
controller in the
(k
p
, k
i
, k
d
) space. Hence, this Theorem and Lemma allow us to develop an algorithm for solving the PI
D
stabilization
problem of the closed loop plant. The proposed approach has been verifed by numerical simulation that confrms the
effectiveness of the procedure.
Key Words: Fractional PID controller, delay system, stability, tuning method.
I. INTRODUCTION
Fractional order dynamic systems and controls
have gained increasing attention in research communities
during the last three decades, mainly due to their demon-
strated applications in numerous, seemingly diverse and
widespread, felds of science and engineering including
control theory, mechanics, physics, chemistry and signals
processing [1–7,9–12,14]. The mathematical modeling
and simulation of systems and processes, based on the
description of their physical laws, leads to differential
equations of a fractional order. In fact, it is natural
that many authors have tried to solve the fractional
differential equations and propose different defnitions
that involve defnite fractional integrals and fractional
derivatives. The earliest systematic works on fractional
calculus were made by researchers such as Leibniz,
Liouville and Riemann [6]. In [14] Oustaloup demon-
strated the superior performance of the CRONE
(Commande Robuste d’Ordre Non Entier) method over
the proportional–integral–derivative (PID) controller
for the control of dynamic systems.
The fractional order proportional integrative
derivative is a matter of almost exclusive interest not
only because of the number of tunable parameters in
a controller, but also its design fexibility and perfor-
mance benefts from its inherent structure [3]. Podlubny
proposed a generalization of the classic PI and PID
Manuscript received March 17, 2013; revised September 20, 2013; accepted
December 21, 2013.
Sami Hafsi (corresponding author, e-mail: samihafsi@ymail.com), Kaouther
Laabidi, and Rihem Farkh are with the Université de Tunis El Manar, Ecole
Nationale d’Ingénieurs de Tunis, LR-11-ES20 Laboratoire d’Analyse, Concep-
tion et Commande des Systèmes, 1002, Tunis, Tunisie.
controllers defned as PI
and PI
D
where the order
integrator and the order differentiator assume real
non-integer values. Also he proved that this type of
fractional controller can achieve better control of
dynamic systems [12]. Using PI
D
allows one to choose,
in addition to the parameters of the classical PID (k
p
, k
i
and k
d
), the orders of integration and derivation . It
is well known that the classical PID allows a phase shift
only of ±
2
, whereas the PI
D
controller may provide
more intermediate information, since the phase shifts
are between -
2
and
2
[3].
Numerous methods and approaches using some
extensions of classical control theory have been proposed
to defne new effective tuning techniques for fractional
controllers. It has been shown in these methods that
the fractional controllers provide a better response than
the integer order controllers when used both for the
control of integer-order systems and fractional-order sys-
tems [1–14]. In [11] Maione and Lino introduce a new
design approach for PI
controllers inspired by the clas-
sical symmetrical optimum method. More recently in [3],
the authors present some effective synthesis methods for
fractional-order PI
, PD
and PI
D
to fulfl different
design specifcations for some closed-loop systems. In [6]
Hamamci use the real root boundary (RRB), the infnite
root boundary (IRB) and the complex root boundary
(CRB) to determine the set of global stability regions
of PI
and PI
D
for different values of the fractional
orders and , but this technique, which is based on
analytical expressions derived from the stability domain
boundaries, is biased in the case where + is equal
to 2.
In general, real systems often involve time delay,
so the dynamic behavior of many industrial plants can
© 2014 Chinese Automatic Control Society and Wiley Publishing Asia Pty Ltd