Lecture09 - 5. Feedback Control Systems-2. Ziegler-Nichols Rules For Tuning PID Controllers
Lecture09 - 5. Feedback Control Systems-2. Ziegler-Nichols Rules For Tuning PID Controllers
[From Ogata's Modern
Control Engineering]
If the plant is not known or so complicated that its mathematical model cannot be easily obtained, then an
analytical or computational approach to the design of a PID controller is not possible. Hence, we need to
resort to experimental approaches to the tuning of PID controllers, leading to one of them, called Ziegler-
Nichols rules. In fact, the Ziegler-Nichols tuning rules ("as a heuristic method") give an educated guess
for the parameter values and provide a starting point for fine tuning, rather than giving the final settings
for 𝐾 , 𝑇 , 𝑇 . Ziegler-Nichols rules are the best in disturbance rejection, but are not necessarily optimal.
[Influences of individual control gains on control systems performance]
First Method: The response of the plant to a unit-step input is experimentally obtained. If the plant
involves neither integrator(s) nor dominant complex-conjugate poles, then such a unit-step response
curve may look S-shaped. Then, the transfer function 𝐶 𝑠 /𝑈 𝑠 may be approximated by a 1st-order
system with a transport lag.
⎯⎯⎯ 𝑒 ⎯⎯⎯⎯
where L implies a delay time and T denotes the time constant. Then, the parameters of PID controller can
be set according to Table 8-1. Note that the PID controller tuned by the 1st method of Ziegler-Nichols
rules gives
/
𝐷 𝑠 𝐾 1 ⎯⎯⎯ 𝑇 𝑠 1.2 ⎯ 1 ⎯⎯⎯ 0.5𝐿𝑠 0.6𝑇 ⎯⎯⎯⎯⎯⎯
The resulting PID controller has a pole at the origin and double zeros at 𝑠 1/𝐿.
Lecture09 Page 1
Example: Comparison of P‐, PD‐ and PID Controls (1a/4)
𝐷 𝑠 𝐾 1 ⎯⎯⎯ 𝑇 𝑠 where
𝐾 19, 𝑇 2, 𝑇 4/19 were
considered.
Unit‐Step Responses of the Closed‐Loop Control System without PID Controller (i.e., 𝐷 𝑠 1)
𝑇 8
⎯⎯ ⎯⎯⎯ 16
𝐿 0.5
• P-Controller: 𝑲𝒑 𝟏𝟔
𝑲𝒑
• PI-Controller: 𝑲𝒑 𝟏𝟒. 𝟒, 𝑻𝒊 𝟏. 𝟔𝟕 → 𝑲𝒊 ⎯⎯ 𝟖. 𝟔𝟐
𝑻𝒊
𝑲𝒑
• PID-Controller: 𝑲𝒑 𝟏𝟗. 𝟐, 𝑻𝒊 𝟏, 𝑻𝒅 𝟎. 𝟐𝟓 → 𝑲𝒊 ⎯⎯ 𝟏𝟗. 𝟐, 𝑲𝒅 𝑲𝒑 ⋅ 𝑻𝒅 𝟒. 𝟖
𝑻𝒊
Lecture09 Page 2
Example: Comparison of P‐, PD‐ and PID Controls (1b/4)
(1) P-Control
(2) PI-Control
(3) PID-control
Lecture09 Page 3
5. Feedback Control Systems - 2. Ziegler-Nichols Rules for Tuning PID Controllers (2/3)
Second Method: Considering the P-control action only (i.e., 𝐷 𝑠 𝐾 ), increase 𝐾 from zero to 𝐾
where the output first exhibits sustained oscillations. If the output does not exhibit sustained oscillations
for whatever value 𝐾 may take, then this method does not apply. The critical gain 𝐾 and the
corresponding period of the response 𝑃 are experimentally determined. Then, the parameters for PID
controller can be set according to the formula in Table 8-2.
Note that the PID controller tuned by the 2nd method of Ziegler-Nichols rules gives
1 1 𝑠 4/𝑃
𝐷 𝑠 𝐾 1 ⎯⎯⎯ 𝑇 𝑠 0.6𝐾 1 ⎯⎯⎯⎯⎯⎯ 0.125𝑃 𝑠 0.075 𝐾 𝑃 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
𝑇𝑠 0.5𝑃 𝑠 𝑠
The resulting PID controller has a pole at the origin and double zeros at 𝑠 4/𝑃 .
Note also that if there is a known mathematical model for a system (such as TF), then the root-locus
method can be used to find the critical gain 𝐾 and the frequency of the sustained oscillations 𝜔
2𝜋/𝑃 (at the intersection between the root-locus branch and 𝑗𝜔 axis). If the root-locus branches do not
cross the 𝑗𝜔 axis, this method does not apply.
Remarks: Neither of Ziegler-Nichols rules can be applied if the plant is unstable and without an
actuator – For example, 𝐺𝐻 ⎯⎯⎯⎯⎯⎯⎯⎯. However, if the plant includes an actuator, the 2nd method
can be tried – For example, 𝐺𝐻 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯.
Lecture09 Page 4
5. Feedback Control Systems - 2. Ziegler-Nichols Rules for Tuning PID Controllers (3/3)
Example: Open-loop transfer function is given as 𝐺𝐻 ⎯⎯⎯⎯⎯⎯⎯⎯⎯. Show that the 2nd method of
Ziegler-Nichols rules yields the PID controller as
1 1 𝑠 1.4235
𝐷 𝑠 𝐾 1 ⎯⎯⎯ 𝑇 𝑠 18 1 ⎯⎯⎯⎯⎯⎯ 0.35124𝑠 6.3223 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
𝑇𝑠 1.405𝑠 𝑠
⎯⎯
where 𝐾 30 and 𝜔 √5 so that 𝑃 ⎯⎯ 2.8099. Hence, 𝐾
⎯⎯ 0.6𝐾 18, 𝑇 0.5𝑃
√
1.405, 𝑇 0.125𝑃 0.35124.
Lecture09 Page 5