See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/4141781

PID tuning rules for second order systems

Conference Paper · August 2004

DOI: 10.1109/ASCC.2004.184806 · Source: IEEE Xplore

CITATIONS READS

19 16,148

2 authors:

Jing-Chung Shen Huann-Keng Chiang

National Formosa University National Yunlin University of Science and Technology
46 PUBLICATIONS 841 CITATIONS 124 PUBLICATIONS 1,076 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Jing-Chung Shen on 26 June 2014.

The user has requested enhancement of the downloaded file.

 PID Tuning Rules for Second Order Systems
Jing-Chung Shen Huann-Keng Chiang*

Department of Automation Engineering

National Huwei Institute of Technology
Huwei, Yunlin, Taiwan
Email: [email protected]

*Department of Electrical Engineering

National Yunlin University of Science and Technology
Toulou, Yunlin, Taiwan

ABSTRACT model of systems and the design method of PID controller

This paper presents PID tuning rules for second are described. Section 3 presents the way to derive the
systems. These tuning rules are derived by optimizing the tuning rules. Simulation and experimental results are given
integrated absolute errors of set point and load disturbance in Section 4. Finally, conclusions are drawn in Section 5.
responses under robustness and bandwidth constrains. For
deriving the tuning formulas, PID controllers for
2. MODEL AND DESIGN METHOD
normalized systems were designed. The relationship
Consider the following system
between the controller parameters, the parameters that
characterize the system dynamics and the normalized gain K sw n2
Gs = (1)
crossover frequency are determined and the tuning s2 + 2zw n s + w n2
formulas are then derived. Simulation examples and
experimental results are provided to demonstrate the where K s is the static gain, w n is the undamped natural
effectiveness of these tuning rules. frequency and z is the damping ratio. For over-damped
systems ( z > 1), the models in (1) can be rewritten as
Key Words: bandwidth, PID controllers, tuning rules, Ks
second order system. Gs = (2)
(1 + sT )(1 + sT1 )
1. INTRODUCTION where
PID controller is the most common control algorithm 1/T = w n (z - z 2 -1) , 1/T1 = w n (z + z 2 -1) . (3)
and is widely used. There are a lot of tuning rules for PID
controllers [1-19]. Most of the tuning rules are derived for When T1 < 10T (z > 1.74) , the models in (2) can be
process control and are derived under idealize assumptions, well approximated by
such as infinite bandwidth. In fact, in most applications, the Ks
measurement noise, the range of manipulated variable and Gso = . (4)
(1+ sT)
the sample rate of the system limit the closed-loop
bandwidth. In this paper, the systems with z £ 2 are considered.
In this paper, PID tuning rules for second order Suppose that the following PID controller is
systems are derived. For deriving the tuning rules, PID employed to control the systems:
controllers for some normalized second order systems are Ê 1 de ˆ
designed. These PID controllers are designed by optimizing u = K ÁÁ (by r - y ) + Ú edt + Td ˜˜ ,
the integrated absolute errors (IAE) of set point and load Ë Ti dt ¯
disturbance responses under constraints on robustness and e = yr - y . (5)
crossover frequency (the frequency where the loop gain
equals one). Note that the closed-loop bandwidth can be where u , y r , y , K , b , Ti , Td and e are the
approximated by the crossover frequency [20]. Therefore, controller output, set point, system output, controller gain,
the PID controllers are designed under bandwidth set-point weighting, integral time, derivative time, and
constrain. When the PID controllers for normalized systems error, respectively. Notice that the controller parameters
are designed, the curve fitting technique is used to derive K , b , Ti and Td must be positive constants.
simple formulas that describe the relationships among
normalized PID controller parameters, parameters that Denote the loop transfer function of the closed-loop
characterize the system dynamics and the normalized system as Gl (s ) and define M s as
crossover frequency. Once these formulas are derived, they
1
can be used to tune the PID controllers. Simulation M s = max .
examples and experimental results show that these w 1 + Gl ( jw )
formulas give satisfactory results.
This paper is organized as follows: In Section 2, the M s is the inverse of the shortest distance from the Nyquist
curve of the loop transfer function to the critical point –1 B = B / w n . If we can find the relations between z , B
and is a measure of stability robustness. Typical value of
M s is in the range from 1.4 to 2.0 and the standard value and the normalized control parameters, and represent K,
is 2.0 [10]. Ti , Td and b as functions of z and B , these functions
Let es denote the error caused by a unit step set-point can be used to tune the PID controllers systems that
modeled by Gs .
change and ed denote the error caused by a unit step For deriving the tuning rules, the PID controllers for
disturbance at the system input, respectively. Define the the normalized systems Gs with z =0.1, 0.2, … 2.0 and
performance index as
• • normalized upper bound B = 1, 2, … 10 were designed.
J = Ú es (t ) dt + Ú ed (t ) dt . Then the normalized controller parameters were plotted as
0 0
In this study, the PID controller parameters were functions of z and B . We then utilized curve fitting
chosen such that the performance index J is minimized technique to find the relations between the normalized
under the following constraints: controller parameters, z and B .
Fig. 1 shows the designed results of normalized
K > 0 , 1 ≥ b > 0 , Ti > 0 , Td > 0 ,
systems. It was tried to express the normalized controller
M s £ m and w g £ B gain as
where constant m > 1 represents the minimal requirement K = f (z ,B)
and analogous expressions for other parameters. By the
of stability robustness (In this study, we let m = 2.0 .), w g
data in Fig. 1, it can be found that the normalized gain K
denotes the crossover frequency and B > 0 represents the increases rapidly as B increases and the variation of Ti ,
upper bound of crossover frequency. Clearly, this is a
Td and b are large for B £ 2 . This makes it difficult to do
constrained nonlinear optimization problem. In this study,
the genetic algorithm described in [17] is used to solve this curve fitting. In order to obtain better fitting, the data is
optimization problem. separated into two groups ( B £ 2 and 10 ≥ B > 2 ) for
curve fitting. After some trials, we found that for
3. THE TUNING RULES 10 ≥ B > 2 the function of K , Ti , Td and b could be
The PID controller design method described in last well approximated by function of the form
section is a time consuming procedure. If we can find f (z ,B) = a0 + a1z + a2z 2 + B(a3 + a4z + a5z 2 )
simple formulas that describe the relations among the
parameters of PID controller, the parameters that +B 2 (a6 + a7z + a8z 2 ) + B 3 (a9 + a10z + a11z 2 )
characterize the system dynamics and the upper bound of (7)
crossover frequency, the user can obtain proper PID Table 1 shows the coefficients a0 , a1 , … a11 of the
controller parameters easily and need not to run the entire functions of the form as in (7) that were least squares fitted
design procedure. In this section, the way to derive the to the data in Fig. 1.
tuning rules will be described.
The closed-loop transfer function of the system Gs
controlled by the PID controller described in (5) is
1 K sw n2
K(b + + Td s)( 2 )
Ti s s + 2zw n s + w n2
Gy r w =
1 K sw n2
1+ K(1+ + Td s)( 2 )
Ti s s + 2zw n s + w n2
1
K (b + + Td s )Gs ( s )
Ti s
= (6)
1
1 + K (1 + + Td s )Gs ( s )
Ti s
where K = KK s , Ti = Tiw n , Td = Td w n are the
normalized controller parameters, s = s / w n and Fig. 1. Design results for normalized control system
1 Gs . ( B = 1: o, B = 2 : x, B = 3 : +, B = 4 : *,
Gs = B = 5 : , B = 6 : ‡, B = 7 : —, B = 8 : D, B = 9 :
(s + 2zs + 1)
2
<, B = 10 : >)
is the normalized system model. Systems with the same
K , Ti , Td , b and z will have similar responses, both in For deriving the tuning rules for B £ 2 , more
time and frequency domain. The only difference is in the controllers were designed. Fig. 2 depicts the designed
scale of time and frequency axis (scaled by w n ). For results. After some trial, we found that it was difficult to fit
convenience, the upper bound B is normalized as the data in Fig. 2 reasonably well to a function like (7),
especially for Td . Therefore, the curves in Fig. 2 were 6.23 and the closed-loop bandwidth is about 8.9.
Fig. 3 depicts the control result of these two
treated separately for each value of B and express the
controllers. Clearly, the performance of the proposed method
normalized controller parameters as functions of z . Table
is better than Haeri’s method.
2 lists the results of fitting.

Fig. 3. Set point and load disturbance responses of

Fig. 2. PID controller design results for normalized G2 ( s ) controlled by a PID controller tuned by the
systems Gs for B = 1(o), B = 1.25 (x), proposed method and Haeri’s method.
B = 1.5 (+), B = 1.75 (*) and B = 2 ( ).
Example 2: Consider a system
1
G2 (s) =
4. EXAMPLES (1+ s)(1+ 0.2s)
In order to demonstrate the performance and use of the
A step test obtained K s = 1. w n and z were determined by
tuning rules, these tuning rules were applied to a few
systems. Comparisons will be made with Astrom and relay feedback test as 2.16 and 1.318 respectively. For
Hagglund’s method [10, 18], Haeri’s method [19] and comparison, the tuning rules proposed by Astrom and
Shen’s [17] method. Hagglund [10], and Shen [17] also used to tune the PID
In the examples, the model of system was determined controller. For their method, the approximated model
by step and relay feedback test [10]. Step test can determine e -0.105 s /(1 + 1.11s ) was used. Using this model, Astrom
the static gain K s of Gs , while a relay feedback test (with and Hagglund’s rule gives K = 40.52 , Ti = 0.29 ,
an integrator inserted between the relay and the system under
test) can determine the ultimate gain K u and ultimate Td = 0.076 , and b = 0.23 . The crossover frequency of
frequency w u of G p /s . By the definition of ultimate gain the closed-loop system with this controller is 16.88 and the
bandwidth is 22. Shen’s method gives K = 79.55 ,
and frequency, w n and z can be obtained as
KK Ti = 0.51 , Td = 0.08 , and b = 0.75 . The crossover
w n = w u, z = s u . (8) frequency of the closed-loop system with this controller is 33
2w u
and the bandwidth is 38. Select B = 7 , the proposed
Example 1: Consider a under damped system method gives K = 42.73 , Ti = 0.45 , Td = 0.061 , and
G1 (s) =
3 b = 0.84 . The bandwidth of the closed-loop system with
s + s+ 3
2 this controller is similar to that tuned by Astrom and
A step test obtained K s = 1. w n and z were determined by Hagglung’s method.
relay feedback test as 1.73 and 0.288 respectively. Applying Fig. 4 shows the set point and load disturbance
the exact parameters to Haeri’s method [19] gives responses of G2 (s) controlled by the controllers derived
K = 8.383 , Ti = 8.45 , Td = 0.929 and b = 1 . The above. Shen’s method provides best performance. But the
bandwidth of the closed-loop system is much beyond the
crossover frequency of G1 (s) controlled with this PID bandwidth of G2 (s) . With similar bandwidth, the proposed
controller is 23.48 and the closed-loop bandwidth is about 23 method provides better performance than Astrom and
(about thirteen times wider than that of G1 (s) ). Let Hagglung’s method.
B = 3.5 and apply the parameters of approximated model
to the proposed tuning rules, the following controller
parameters can be obtained: K = 8.55 , Ti = 0.67 ,
Td = 0.226 , and b = 0.76 . The gain of this controller is
similar to that tuned by Haeri’s method. The crossover
frequency of G1 (s) controlled with this PID controller is
 Fig. 4. Set point and load disturbance responses Fig. 5. Set point and load disturbance responses of the
of G2 (s) controlled by a PID controller tuned system in example 3 controlled by the PID controller
by the proposed method, Astrom and tuned by the proposed method.
Hagglund’s method, and Shen’s method.

Experimental results: In order to show the 5. CONCLUSIONS

applicability of the proposed tuning rules, experiments were In this paper, PID tuning rules for second order
carried out. In the experiments, the PID controller was systems are proposed. These tuning rules take the bandwidth
implemented with back-calculation based anti-windup [10] limitation into consideration. Therefore, the user can tune the
and a low-pass filter with time constant Td /10 was PID controller according to the bandwidth limitation of the
connected to the derivative part. The input of this low-pass system. Simulation examples and experiment are provided to
filter was connected to system output directly. That is, the demonstrate the performance and the use of the proposed
derivation acted on filtered system output directly. Moreover, tuning rules.
the sampling frequency was chosen to be 200Hz. In the ACKNOWLEDGEMENTS
following, the results on a motor speed control systems are Part of this work was supported by the National
presented. Science Council, Republic of China (Taiwan) under grant
NSC91-2213-E-150-002.
Example 3: Consider a disk that was driven by DC
motor with a flexible shaft. In the experiment, the speed was REFERENCES
estimated by an encoder. This resulted in a measure noise [1] Ziegler, J. G. and N. B., Nichols, “Optimal settings for
about ±0.4 rad/sec. automatic controllers”, Trans. ASME, Vol. 64, pp. 759-
The K s of this system was obtained by a step test as 768 (1942).
[2] Cohen, G. H. and G. A., Coon, “Theoretical
350. While w n and z were estimated as 19.64 and 0.454 consideration of retarded control”, Trans. ASME, Vol.
respectively. 75, pp. 827-834 (1953).
For speed control, a PID controller was tuned. We [3] Smith, C. A. and C. B., Corripio, Principals and
hope the fluctuation of controller output keeps inside ±0.02 Practices of Automatic Control, Wiley, New York,
( ±0.2 volt) when the system is in steady state. For the (1985).
implemented PID controller, the high frequency gain from [4] Shinkey, F. G., Process Control System Application,
system output to controller output is (1+ 10)K . Therefore, Design and Tuning, 3rd ed., McGraw-Hill, New York,
K should satisfy the following inequality: (1988).
0.4(1+ 10)K £ 0.02 fi K £ 0.0045 [5] Morari, M. and E., Zafiriou, Robust Process Control,
Selecting B = 1, the controller parameters can be obtained Prentice-Hall, Englewood Cliffs, NJ, (1989).
as K = 0.0024 , Ti = 0.055 , Td = 0.052 , and b = 0.95 . [6] Chien, I. L. and P. S., Fruehauf, “Consider IMC tuning to
Fig. 5 shows the controlled result of this controller. As the improve controller performance”, Chem. Eng. Prog. Vol.
results shown, the performance is good. 86, pp. 33-41 (1990).
[7] Hang, C. C., K. J., Astrom, and W. K., Ho, “Refinements
of the Ziegler-Nichols tuning formula”, IEE Proc.-D:
Control Theory Appl., Vol. 138, pp. 111-118 (1991).
[8] Astrom, K. J., T., Hagglund, C. C. Hang, and W. K., Ho,
“Automatic tuning and adaptation for PID controllers-A
survey”, IFAC J. of Control Eng. Practice, Vol. 1, pp.
699-714 (1993).
[9] Zhaung, M. and D. P., Atherton, “Automatic tuning of
optimal PID controllers”, IEE Proc.-D: Control Theory
Appl., Vol. 140, pp. 216-224 (1993).
[10] Astrom, K. J. and T., Hagglund, PID Controllers:
Theory, Design and Tuning, ISA, Research Triangle
 Par, NC, (1995). Control, Vol. 2, pp. 31-41 (2000).
[11] Ho, W. K., C. C., Hang, and L. S., Cao, “Tuning of PID [16] Shen, J. C., “Fuzzy neural networks for tuning PID
controllers based on gain and phase margin controller for plants with under-damped responses”,
specification”, Automatica, Vol. 31, pp. 497-502 IEEE Trans. On Fuzzy Systems, Vol. 9, pp. 333-342
(1995). (2001).
[12] Abbas, A., “A new set of controller tuning relations”, [17] Shen, J. C., “New tuning method for PID controller”,
ISA Trans. Vol. 36, pp. 183-187 (1997). ISA Trans., Vol. 41, pp. 473-484 (2002).
[13] Ho., W. K., C. C., Hang, and J., Zhou, “Self-tuning PID [18] Hagglund, T., and K. J., Astrom, “Revisiting the
control of a plant with under-damped response with Ziegler-Nichols tuning rules for PI control”, Asian
specifications on gain and phase margins”, IEEE Journal of Control, Vol. 4, 364-380 (2002).
Trans. Control Syst. Technol., Vol. 5, pp. 446-452 [19] Haeri, M., “Tuning rules for PID controller using a
(1997). DMC strategy”, Asian Journal of Control, Vol. 4, pp.
[14] Wang, Q. G., T. H., Lee, H. W., Fung, Q., Bi, and Y., 410-417 (2002).
Zhang, “PID tuning for improved performance”, IEEE [20] Franklin, G. F., J. D., Powell, and A., Emami-Naeini,
Trans. Control Syst. Technol., Vol. 7, pp. 457-465 Feedback Control of Dynamic Systems, Addison-
(1999). Wesley, (1986).
[15] Shen, J. C., “New tuning method for PID control of a
plant with under-damped response”, Asian Journal of

Table 1. Tuning formula for systems that can be modeled by Gs with 2 < B £ 10 .

0 < z £ 2 , 2 < B £ 10

KK s Tiw n Td w n b
a0
1.8476 1.0743 1.4274 0.8712
a1 -6.7604 0.7686 -1.8460 -0.1955
a2 2.8846 -0.0150 0.5692 0.1043
a3 -0.8778 0.0512 -0.5047 -0.1514
a4 5.7533 -0.3071 0.7723 0.2142
a5 -1.9453 -0.0036 -0.251 -0.0828
0.6445 -0.01 0.0703 0.0339
a6 -0.7925 0.0329 -0.1107 -0.0454
a7 0.4080 0.0045 0.0365 0.0168
a8 0.0071 0.0002 -0.0033 -0.0019
a9 0.0414 -0.0008 0.0052 0.0027
-0.0248 -0.0005 -0.0017 -0.0010
a10
a11
 Table 2. Tuning formula systems that can be modeled by Gs with 1 £ B £ 2 .

B KK s Tiw n Td w n b
2.1266z - 4.6156z
2
1 1.7034z + 0.0713 -0.2382z 2 + 1.1225z -0.0553z + 1.023
+2.5748 for 0 < z < 1.2
+0.6064 0.0104z 2 - 0.0372z
+0.0376 for 1.2 £ z £ 2
0.8093z 2 - 2.1177z
1.25 2.149z + 0.2730 -0.129z 2 + 0.7975z -0.089z + 0.9811
+1.4476 for 0 < z < 1.3
+0.8269 0.0232z 2 - 0.0868z
+0.0877 for 1.3 £ z £ 2
0.4542z 2 -1.3187z
1.5 2.4511z + 0.7056 0.0349z 2 + 0.2337z -0.0509z + 0.8618
+1.0058 for 0 < z < 1.4
+1.1508 -0.0004z 2 + 0.0005z
+0.0052 for 1.4 £ z £ 2
0.3205z 2 - 0.9502z
1.75 2.7891z + 1.264 0.0135z 2 + 0.2029z 0.0256z + 0.7451
+0.7874 for 0 < z < 1.5
+1.2133 0.3177z 2 -1.2488z
+1.2260 for 1.5 £ z £ 2
0.2586z 2 - 0.8177z
2 3.1821z + 1.8282 0.0178z 2 + 0.2139z 0.0592z + 0.7083
+0.7209 for 0 < z < 1.6
+1.1847 -0.39z 2 + 1.2784z
-0.9926 for 1.6 £ z £ 2

View publication stats