# ELEC ENG 4CL4

:

Control System Design
Notes for Lecture #14
Wednesday, February 4, 2004
Dr. Ian C. Bruce
Room: CRL-229
Phone ext.: 26984
Email: ibruce@mail.ece.mcmaster.ca

Chapter 6

Tuning of PID Controllers
Because of their widespread use in practice, we
present below several methods for tuning PID
controllers. Actually these methods are quite old and
date back to the 1950’s. Nonetheless, they remain in
In particular, we will study.

Ziegler-Nichols Oscillation Method
Ziegler-Nichols Reaction Curve Method
Cohen-Coon Reaction Curve Method

Increase the gain until the loop starts oscillating. Prentice Hall 2000 (1) Ziegler-Nichols (Z-N) Oscillation Method This procedure is only valid for open loop stable plants and it is carried out through the following steps ◆ ◆ Set the true plant under proportional control. Salgado. with a very small gain. Note that linear oscillation is required and that it should be detected at the controller output.©Goodwin. Chapter 6 Graebe. .

but the version described here is. there is some controversy regarding the PID parameterization for which the Z-N method was developed. applicable to the parameterization of standard form PID. Salgado. . to the best knowledge of the authors. Adjust the controller parameters according to Table 6.1 (next slide). Pc. Prentice Hall 2000 Record the controller critical gain Kp = Kc and the oscillation period of the controller output.©Goodwin. Chapter 6 ◆ ◆ Graebe.

5Pc Td Pc 8 . Chapter 6 Table 6.50Kc PI 0. Prentice Hall 2000 Ziegler-Nichols tuning using the oscillation method Tr P Kp 0.1: Graebe.60Kc Pc 1.©Goodwin. Salgado.45Kc PID 0.2 0.

©Goodwin. Salgado. The next plot shows the resultant closed loop step ∆ responses as a function of the ratio x = γτ . γ0 > 0 γ 0s + 1 then one can obtain the PID settings via ZieglerNichols tuning for different values of τ and νγ0. ν0 . Chapter 6 Graebe. Prentice Hall 2000 General System If we consider a general plant of the form: K0e− sτ G0(s) = .

Salgado. Ziegler−Nichols (oscillation method) for different values of the ratio x= τ/νγ o x=0.0 0.5 x=0.3: PI Z-N tuned (oscillation method) control ∆τ loop for different values of the ratio x = νγ00 .1 Plant response 1. Prentice Hall 2000 Figure 6. Chapter 6 Graebe.5 0 0 1 2 3 4 5 Time [t/τ] 6 7 8 9 10 .©Goodwin.5 1 x=2.

Chapter 6 Graebe.©Goodwin. Obtain a graph of the response to a unit step input reference and to a unit step input disturbance. Prentice Hall 2000 Numerical Example Consider a plant with a model given by Go (s) = 1 (s + 1)3 Find the parameters of a PID controller using the Z-N oscillation method. Salgado. .

Salgado. Td = 0.45 The closed loop response to a unit step in the reference at t = 0 and a unit step disturbance at t = 10 are shown in the next figure. Tr = 0.1.6 ∗ Kc = 4. Chapter 6 Graebe.©Goodwin. Prentice Hall 2000 Solution Applying the procedure we find: Kc = 8 and ωc = √3. . Hence.8.125 ∗ Pc ≈ 0.81. from Table 6.5 ∗ Pc ≈ 1. we have Kp = 0.

Chapter 6 Graebe.5 0 0 2 4 6 8 10 Time [s] 12 14 16 18 20 .©Goodwin.5 1 0. Prentice Hall 2000 Figure 6. Salgado.4: Response to step reference and step input disturbance PID control tuned with Z−N (oscillation method) Plant output 1.

in the notation used in (6. we have Ks = 4. but with the Z-N settings applied to the series structure. To obtain an appreciation of these differences we evaluate the PID control loop for the same plant in Example 6.2. Salgado.5).1 .1.©Goodwin.81 Ds = 0.45 γs = 0. Prentice Hall 2000 Different PID Structures? A key issue when applying PID tuning rules (such as Ziegler-Nichols settings) is that of which PID structure these settings are applied to.e. i.8 Is = 1. Chapter 6 Graebe.

5 0 0 2 4 6 8 10 Time [s] 12 14 16 18 20 .5: PID Z-N settings applied to series structure (thick line) and conventional structure (thin line) Z−N tuning (oscillation method) with different PID structures 2 Plant output 1.©Goodwin. Salgado.5 1 0. Prentice Hall 2000 Figure 6. Chapter 6 Graebe.

Chapter 6 Graebe. to which form of PID the tuning rules are applied. . Prentice Hall 2000 Observation In the above example. it has not made much difference.©Goodwin. the reader is warned that this can make a difference in general. However. Salgado.

apply a step change to the plant input. 2. Cont/. using the following procedure: 1. Graebe. from u0 to u∞ (this should be in the range of 10 to 20% of full scale). Say that the plant output settles at y(t) = y0 for a constant plant input u(t) = u0. With the plant in open loop. At an initial time. Salgado.. . t0..Chapter 6 ©Goodwin. take the plant manually to a normal operating point. Prentice Hall 2000 (2) Reaction Curve Based Methods A linearized quantitative version of a simple plant can be obtained with an open loop experiment.

6.t. Record the plant output until it settles to the new operating point. Salgado. This curve is known as the process reaction curve. Assume you obtain the curve shown on the next slide.s. Compute the parameter model as follows Ko = y∞ − yo . Chapter 6 Graebe. stands for maximum slope tangent. 4. m. u∞ − uo τ o = t1 − to . νo = t2 − t1 . In Figure 6. Prentice Hall 2000 3.©Goodwin.

©Goodwin.2. Salgado.s. y∞ yo to t1 t2 Time (sec.t. Chapter 6 Graebe. Prentice Hall 2000 Figure 6.6: Plant step response The suggested parameters are shown in Table 6. m.) .

2: Graebe.©Goodwin. Prentice Hall 2000 Ziegler-Nichols tuning using the reaction curve P PI PID Kp νo Ko τo 0.2νo Ko τo Tr Td 3τo 2τo 0.9νo Ko τo 1.5τo . Salgado. Chapter 6 Table 6.

Salgado. Prentice Hall 2000 General System Revisited Consider again the general plant: K0e− sτ G0( s) = γ 0s + 1 The next slide shows the closed loop responses resulting from Ziegler-Nichols Reaction Curve ∆ tuning for different values of x = γτ . ν0 .©Goodwin. Chapter 6 Graebe.

5 1 x=2. Salgado. Prentice Hall 2000 Figure 6.©Goodwin.0 0. Chapter 6 Graebe.5 x=0.7: PI Z-N tuned (reaction curve method) control loop Ziegler−Nichols (reaction curve) for different values of the ratio x= τ/νγ o 2 Plant response x=0.1 1.5 0 0 5 10 Time [t/τ] 15 .

. Prentice Hall 2000 Observation We see from the previous slide that the ZieglerNichols reaction curve tuning method is very sensitive to the ratio of delay to time constant. Chapter 6 Graebe. Salgado.©Goodwin.