Professional Documents
Culture Documents
ARROW@DIT
Conference papers School of Electrical Engineering Systems
2000-01-01
Recommended Citation
O'Dwyer, Aidan : A summary of PI and PID controller tuning rules for processes with time delay. Part 2: PID controller tuning rules.
Proceedings of PID ’00: IFAC Workshop on Digital Control, pp. 242-247, Terrassa, Spain, April 4-7, 2000.
This Conference Paper is brought to you for free and open access by the
School of Electrical Engineering Systems at ARROW@DIT. It has been
accepted for inclusion in Conference papers by an authorized administrator
of ARROW@DIT. For more information, please contact
yvonne.desmond@dit.ie, arrow.admin@dit.ie.
Aidan O’Dwyer
School of Control Systems and Electrical Engineering, Dublin Institute of Technology, Kevin
St., Dublin 8, Ireland.
Abstract: The ability of proportional integral (PI) and proportional integral derivative (PID)
controllers to compensate many practical industrial processes has led to their wide acceptance
in industrial applications. The requirement to choose either two or three controller parameters
is perhaps most easily done using tuning rules. A summary of tuning rules for the PID control
of single input, single output (SISO) processes with time delay is provided in this paper.
Copyright ©2000 IFAC
1. INTRODUCTION Tuning rules for these and other such PID controller
structures are explicitly indicated; in all cases,
This paper summarises some of the most directly numerical data is quoted to a maximum of two places
applicable tuning rules for PID controllers that have of decimals. Most authors recommend application of
been developed to compensate SISO processes with the tuning rules for a range of model time delay to
time delay, modeled in either first order lag plus time constant ( τ m Tm ) between 0.1 and 1.0; this
delay (FOLPD) form or integral plus delay (IPD) data, together with other relevant comments, is
form. It is a companion paper to that of O’Dwyer provided by O’Dwyer (2000b). Results from the
(2000a) and the two papers have similar structure. A analytical calculation of robustness criteria associated
comprehensive summary of PID controller tuning with a number of tuning rules, for a range of τ m Tm
rules for processes with time delay is available from
values, are presented in Section 4. A list of symbols
the author (O’Dwyer, 2000b).
and abbreviations used in the paper is provided in the
appendix.
The ideal continuous time domain PID controller for
a SISO process is expressed in the Laplace domain as
follows:
1 K m e − sτ
m
Rule Kc Ti Td Rule Kc Ti Td
0.96 0.75 0.93
Astrom 0.94Tm Zhuang . ⎛ Tm ⎞
153 Tm ⎛ Tm ⎞ ⎛τ ⎞
⎜ ⎟ ⎜ ⎟ 0.41Tm ⎜ m ⎟
and Kmτ m 2τ m 0.5τ m and Km ⎝ τm ⎠ 0.97 ⎝ τ m ⎠ ⎝ Tm ⎠
Hagglund Atherton 2 0.71 0.60 0.85
(1995) (1993) – . ⎛ Tm ⎞
159 Tm ⎛ Tm ⎞ ⎛τ ⎞
⎜ ⎟ ⎜ ⎟ 0.41Tm ⎜ m ⎟
min. Km ⎝ τm ⎠ 0.96 ⎝ τ m ⎠ ⎝ Tm ⎠
Chien et
al. (1952) 0.95Tm 2.38τ m 0.42τ m ISTES
–regulator Kmτm Servo
– 0% o.s. Rovira et . ⎛ Tm ⎞
109
0.87
Tm ⎛τ ⎞
0.91
min. ITAE N = 10
Rule Kc Ti Td Rule Kc Ti Td
Robust Direct synthesis
Fruehauf 5Tm Tsang and 0.81Tm Tm 0.5τ m ,
et al. 9τ m K m 5τ m ≤ 0.5τ m Rad K mτ m N=5
(1993) (1995)
Tm Tsang et aTm Tm 0.25τ m ,
2τ m K m Tm ≤ 0.5τ m
al. (1993) Kmτ m N = 2.5
Ultimate cycle
a ξ a ξ a ξ
Zhuang 0.05Tu
1.68 0.0 0.86 0.4 0.54 0.8
and 0.51K u (3.30K K + 1) 013
. Tu
m u 1.38 0.1 0.75 0.5 0.50 0.9
Atherton 1.16 0.2 0.67 0.6 0.46 1.0
servo
(1993) – 4 0.99 0.3 0.60 0.7
min. Robust
ISTSE K c ( 3) Ti ( 3) 014
. Tu
1 ⎛ Tm ⎞ Tm 0.5τ m ,
regulator ⎜ ⎟
Chien K m ⎝ λ + 0.5τ m ⎠ N = 10
Classical controller – (1988) λ = [τ m , Tm ]
⎛ 1 ⎞ ⎛ 1 + Td s ⎞
G c (s) = K c ⎜ 1 + ⎟⎜ ⎟ 1 ⎛ 0.5τ m ⎞ 0.5τ m Tm ,
⎝ Ti s ⎠ ⎝ 1 + Td s N ⎠ ⎜ ⎟
K m ⎝ λ + 0.5τ m ⎠ N = 10
Process reaction λ = [τ m , Tm ]
Hang et al. 0.83Tm . τm
15 0.25τ m , Ultimate cycle
(1993) Kmτm N = 10 Shinskey 0.95Tm Kmτm . τm
143 0.52 τ m
(1988) 0.95Tm Kmτm . τm
117 0.48τ m
Witt and aTm τm τm min. IAE 114
, . Tm K mτ m . τm
103 0.40τ m
Waggoner Kmτ m N = [10,20] – regulator
(1990) 139
. Tm Kmτ m 0.77 τ m 0.35τ m
a = [0.6,1] – varying
Regulator τ m Tm
Kaya and 0.98 ⎛ Tm ⎞
0.76
Tm ⎛ Tm ⎞
1.05
⎛τ ⎞
0.90 Industrial controller –
⎜ ⎟ ⎜ ⎟ 0.60Tm ⎜ m ⎟
Scheib Km ⎝ τm ⎠ 0.91 ⎝ τ m ⎠ ⎝ Tm ⎠ ⎛ 1 ⎞⎛ 1 + Td s ⎞
(1988) – U(s) = K c ⎜ 1 + ⎟ ⎜ R (s) − Y(s)⎟
N = 10 ⎝ Ti ⎠⎝
s 1 + T d s N ⎠
min. IAE
Kaya and 0.90 0.88 Regulator
. ⎛ Tm ⎞ ⎛τ ⎞
0.95
112 Tm ⎛ Tm ⎞
Scheib ⎜ ⎟ ⎜ ⎟ 0.55Tm ⎜ m ⎟ Kaya and 0.91 ⎛ Tm ⎞
0.79
Tm ⎛ Tm ⎞
1.00
⎛τ ⎞
0.78
Km ⎝ τm ⎠ 0.80 ⎝ τ m ⎠ ⎝ Tm ⎠ ⎜ ⎟ ⎜ ⎟ 0.54Tm ⎜ m ⎟
(1988) – Scheib Km ⎝ τm ⎠ . ⎝ τm ⎠
101 ⎝ Tm ⎠
N = 10 (1988) –
min. IAE N = 10
Kaya and 1.06 1.04 min. IAE
0.78 ⎛ Tm ⎞ ⎛τ ⎞
0.71
Tm ⎛ Tm ⎞
Scheib ⎜ ⎟ ⎜ ⎟ 0.57Tm ⎜ m ⎟ Kaya and . ⎛ Tm ⎞
111
0.90
Tm ⎛ Tm ⎞
0.88
⎛τ ⎞
0.91
Km ⎝ τm ⎠ . ⎝ τm ⎠
114 ⎝ Tm ⎠ ⎜ ⎟ ⎜ ⎟ 0.57Tm ⎜ m ⎟
(1988) – Scheib Km ⎝ τm ⎠ 0.93 ⎝ τ m ⎠ ⎝ Tm ⎠
N = 10 (1988) –
min. ITAE N = 10
Servo min. ISE
0.89 0.99 0.97
Kaya and 1.04 1.08 Kaya and 0.71 ⎛ Tm ⎞ Tm ⎛ Tm ⎞ ⎛τ ⎞
0.65 ⎛ Tm ⎞ Tm ⎛τ ⎞ ⎜ ⎟ ⎜ ⎟ 0.60Tm ⎜ m ⎟
Scheib ⎜ ⎟ τm 0.51Tm ⎜ m ⎟ Scheib Km ⎝ τm ⎠ . ⎝ τm ⎠
103 ⎝ Tm ⎠
Km ⎝ τm ⎠ 0.99 + 010
. ⎝ Tm ⎠
(1988) – Tm (1988) –
N = 10 N = 10
min. IAE min. ITAE
Kaya and 1.03 0.86 Servo
0.72 ⎛ Tm ⎞ Tm ⎛τ ⎞
Scheib ⎜ ⎟ τm 0.55Tm ⎜ m ⎟ Kaya and 0.82 ⎛ Tm ⎞
1.00
Tm ⎛τ ⎞
0.97
Km ⎝ τm ⎠ . − 018
113 . ⎝ Tm ⎠ ⎜ ⎟ 0.44Tm ⎜ m ⎟
Scheib τ
(1988) – Tm Km ⎝ τm ⎠ . − 0.22 m
109 ⎝ Tm ⎠
N = 10 (1988) – Tm
min. ISE N = 10
Kaya and 0.80 1.01 min. IAE
. ⎛ Tm ⎞
113 Tm ⎛τ ⎞
Scheib ⎜ ⎟ τm 0.43Tm ⎜ m ⎟ Kaya and . ⎛ Tm ⎞
114
0.94
Tm ⎛τ ⎞
0.78
Km ⎝ τm ⎠ . + 0.03
100 ⎝ Tm ⎠ ⎜ ⎟ 0.35Tm ⎜ m ⎟
Scheib τ
(1988) – Tm Km ⎝ τm ⎠ 0.99 − 0.35 m ⎝ Tm ⎠
(1988) – Tm
N = 10
4 min. ISE
0.76 1.11
Kaya and 0.83 ⎛ Tm ⎞ Tm ⎛τ ⎞
4.43K m K u − 0.97 ⎜ ⎟ τm 0.44Tm ⎜ m ⎟
Kc ( 3)
= Ku , Scheib Km ⎝ τm ⎠ . + 0.01
100 ⎝ Tm ⎠
. K m K u + 173
512 . (1988) – Tm
N = 10
. K m K u − 0.61
175 min. ITAE
Ti ( 3) = Tu
3.78K m K u + 139
.
Preprints of Proceedings of PID ’00: IFAC Workshop on Digital Control, Terrassa, Spain,
April 4-7, 2000, pp. 242-247.
K m e− sτ m
Rule Kc Ti Td
1
Ideal controller G c (s) = K c (1 + + Td s)
Ts
i
Process reaction
Ford 1.48
(1953) Kmτ m 2τ m 0.37 τ m
Astrom 0.94 - =
and K mτ m 2τ m 0.5τ m +=
Hagglund o=
(1995) Ratio of τ m to Tm
Direct synthesis Figure 1: Gain margin
0.96
Cluett and Kmτ m 3.04 τ m 0.39 τ m - = Ziegler-Nichols (1942)
Wang + = Astrom-Hagglund (1995)
0.62
(1997) – o = Cohen-Coon (1953)
designed Kmτ m 5.26τ m 0.26τ m * = Chien et al. (1952) – reg – 20% o.s.
closed 0.47
loop time K mτ m 7.23τ m 0.21τ m
constant
equals τ m 0.38
Kmτ m 9.19 τ m . τm
017
to 6τ m ,
respectively 0.31
K mτ m . τm
1116 . τm
015
0.27
Kmτ m . τm
1314 . τm
013
Classical controller –
⎛ 1 ⎞ ⎛ 1 + Td s ⎞
G c (s) = K c ⎜ 1 + ⎟⎜ ⎟
⎝ Ti s ⎠ ⎝ 1 + Td s N ⎠
Regulator Ratio of τ m to Tm
Shinskey 0.93 . τm
157 0.56τ m Figure 2: Phase margin
(1996) – Kmτ m
min. IAE
0.93 . τm
160 0.58τ m , Figure 3: Gain margin
Shinskey Kmτ m N = 10
(1994) –
min. IAE
0.93 1.48τ m 0.63τ m ,
K mτ m N = 20
4. SIMULATION RESULTS
5. CONCLUSIONS