REVISION OF PAST CONTROLLER DESIGN METHODS BY EXAMPLES

THE SECOND METHOD OF ZIEGLER NICHOLS:

Example 1: Given the characteristic equation 𝑄(𝑠) = 𝑠 3 + 3𝐾𝑠 2 + (𝐾 + 2)𝑠 + 4 = 0, determine the
critical gain 𝐾𝑐 of the system.

The critical gain is determine using the Routh-Hurtwiz method.

The Routh-Hurwits table for the system is given as:

𝑠3 1 K+2
𝑠2 3K 4
𝑠1 3𝐾2 +6𝐾−4
2 marks
3𝐾
0
𝑠 4 1 mark

2 marks for the rest of the Table (5 marks total for Table)

For a stable system: 1>0 . (1)

3K>0 or K>0 ….(2) 1 mark

3𝐾 2 + 6𝐾 − 4 > 0 → 3(𝐾 2 + 2𝐾 − 1.3333) > 0 … (3) 3 mark

Hence: (𝐾 + 2.5275)(𝐾 − 0.5275) > 0 implies that:

(i) (𝐾 + 2.5275)(𝐾 − 0.5275) > 0 → (𝐾 + 2.5275) > 0 𝑎𝑛𝑑 (𝐾 − 0.5275) > 0

Hence: K > -2.5275 and K > 0.5275; yielding K > 0.5275 …. (4) 3 mark

(ii) (𝐾 + 2.5275)(𝐾 − 0.5275) > 0 → (𝐾 + 2.5275) < 0 𝑎𝑛𝑑 (K-0.5275) < 0

Hence: 𝐾 < − 2.5275 < 0 𝑎𝑛𝑑 K < 0.5275 < 0: yielding K < -2.5275 … (5) 3 mark

Since from (2) K>0 and from (4), K> 0.5275

Hence, for stability, K > 0.5275 2 mark

The critical values of K are: -2.5275 and 0.5275 8 marks

Example 2: Ziegler Nichols Method
Example 3 (Root-Locus Plotting): Using the graph paper provided, draw a detailed root-locus plot of
5𝐾
the system 𝐺(𝑠) = (𝑠+2)(𝑠+6)(𝑠+8)
 Example 4: Controller Design Using the Root-Locus Approach: An open-loop transfer function is
𝐾
given by 𝐺1 𝐻1 (𝑠) = (𝑠+2)(𝑠+8)(𝑠+10):

Using the graph paper provided, draw a detailed roots locus plot of the system, showing all branches,
break-away point(s), the centroid, all asymptotes and the points where the branches cross the
imaginary axis [28]

Root Locus
25

20 𝑗10.8
15
Centroid
10
6 marks for
5
plot showing
Imaginary Axis

all key details

0

-5

-10

-15

-20

-25
-35 -30 -25 -20 -15 -10 -5 0 5 10
Real Axis

A detailed roots locus plot.

Question 2.1
The given system has the poles: -2,-8,-10; m=3 . No poles: n=0; m-n=3
−2−8−10 (2𝑘+1)180
Centroid= = −6.33 √, asymptote angles: ; 600 , 1800 , 3000 √√√
3 3
1 1 1
For the breakaway point 𝜎: 𝜎+2
+ 𝜎+8 + 𝜎+10 = 0 √. Hence,

(𝜎 + 8)(𝜎 + 10) + (𝜎 + 2)(𝜎 + 10) + (𝜎 + 2)(𝜎 + 8) = 0. This gives

3𝜎 2 + 40𝜎 + 116 = 0; √∴ 𝜎 = −9.07, −4.26. √√ True value is 𝜎 = −4.26 √

For the cross of the imaginary axis, The closed loop transfer function is
𝐾 𝐾
𝑇(𝑠) = (𝑠+2)(𝑠+8)(𝑠+10)+𝐾 = 𝑠3 +20𝑠2 +116𝑠+160+𝐾. √ The characteristic equation is

𝑄(𝑠) = 𝑠 3 + 20𝑠 2 + 116𝑠 + 160 + 𝐾 = 0 √

The Routh Table:

𝑠3 1 116
For the onset of oscillations, 1220-160-K=0, K=2160 √
𝑠2 20 For the frequency of oscillations, 20𝑠 2 + 160 +
160+K
𝑠1 2320−160−𝐾
√√ 2160 = 0 → 𝑠 2 + 116 = (𝑗𝜔)2 + 116 = 0 √
0
20
𝑠0 160+K √√ Therefore, 𝜔 = √116 = 10.77𝑟𝑎𝑑/𝑠𝑒𝑐 √
The branches
2.2 Given that the system operates with a damping factor cross the
of 0.5, locate imaginary
on the axis plot
root locus at the
±𝑗10.77 𝑟𝑎𝑑/𝑠𝑒𝑐
point of operation of the system, and determine the corresponding value of K. [10] √√
 Root Locus The damping factor of 0.5 corresponds to an angle
0.5 of 𝜃 = 𝑐𝑜𝑠 −1 (0.5) = 600 . √√ Show the 60 deg line on
15
the plot. √ This angle intercepts the point on the
System: G1 plot 𝑠 ∗ = −2.88 + 𝑗5.04; √
Gain: 320
10 Pole: -2.88 + 5.04i
Damping: 0.497 For the roots locus gain at the point𝑠 ∗ = −2.88 +
Imaginary Axis

Overshoot (%): 16.6

Frequency (rad/sec): 5.81 1
5 𝑗5.04 , 𝐾 = ⌈𝐺 ∗ .
1 1 (𝑠 )⌉
𝐻

0
∴ 𝐾 = |(𝑠 ∗ + 2)(𝑠 ∗ + 8)(𝑠 ∗ + 10)|√√
𝐾 = |𝑠 ∗ + 2||𝑠 ∗ + 8||𝑠 ∗ + 10|
-5
0.5 𝐾 = |−2.88 + 𝑗5.04 + 2||−2.88 + j5.04 +
-10 -5 0 5
Real Axis
8||−2.88 + j5.04 + 10| √√

𝐾 = 5.1162𝑥7.1844𝑥8.723 = 320.645√√
The damping factor line and associated gain

Design a lag controller for the system so that the positional error is reduced by a factor of 30. Assume
that the controller pole is located at s=-0.001 [15]
𝑠+𝑧 𝑧
Let the transfer function of the lag controller be 𝐺𝑐 (𝑠) = 𝑠+𝑝𝑐 ; lim 𝐺𝑐 = 𝑝𝑐 √
𝑐 𝑠→0 𝑐

320
Now for 𝐺1 𝐻1 (𝑠) = (𝑠+2)(𝑠+8)(𝑠+10), the positional error constant for the given system is

320 320
𝐾𝑝 = lim (𝑠+2)(𝑠+8)(𝑠+10) = 2𝑥8𝑥10 = 2 √√
𝑠→0

1 1
The positional error is given by 𝑒𝑝 = 1+𝐾 = 1+2 = 0.333 √√
𝑃

0.333
For the reduction of the error by a factor of 30, the new error is 𝑒𝑛𝑒𝑤 = 30
= 0.0111 √√
1 1
Correspondingly 𝑒𝑛𝑒𝑤 = = 0.011√→ 𝐾𝑃𝑛𝑒𝑤 = − 1 = 89.909 √√
1+𝐾𝑃𝑛𝑒𝑤 0.011

𝑧𝑐 𝑧𝑐 89.909
Also, 𝐾𝑛𝑒𝑤 = 89.909 = 2 → = = 44.95 √√
𝑧𝑝 𝑧𝑝 2

∴ 𝑧𝑐 = 44.95𝑧𝑝 = 44.95𝑥(−0.001)= − 0.04495 √

𝑠+𝑧 𝑠+0.04495
The lag controller is given by 𝐺𝑐 (𝑠) = 𝑠+𝑝𝑐 = 𝑠+0.001
√√√
𝑐
Example 5: Root Locus Based Controller Design
𝐾
The third-order control system 𝐺7 (𝑠) = is known to operate with a damping factor of
𝑠(𝑠+4)(𝑠+5)
𝜉 = 0.5:

4.1 Using the graph paper provided, make a detailed root-locus plot of 𝐺7 (𝑠) . [16]

4.2 Using the root-locus plot, evaluate the system gain, K, at the damping factor of 𝜉 = 0.5. [7]

4.3 Compute the steady-state error of the system. [3]

4.4 If the given system is to be controlled by a proportional controller of gain 𝑘𝑝 , determine the
controller gain required to reduce the steady-state error of the system by a factor of 8. [4]

𝐾 𝐾
Given that 𝐺7 (𝑠) = = :
𝑠(𝑠+4)(𝑠+5) 𝑠3 +9𝑠2 +20𝑠

Poles: 0, -4, -5; number of poles m=3 (1 marks) No finite zeros. N=0, n-m=3-0-3 (1 mark)

0−4−5
Centroid= = −3 1 mark
3

(2𝑘+1)𝜋
Asymptote angles 𝜃 = = 60(2𝑘 + 1), 𝑘 = 0,1,2. Hence, 𝜃 = 600 , 1800 , 3000 1 marks
3
𝐾
Now, closed-loop transfer function 𝑇7 (𝑠) = 𝑠3 +9𝑠2 +20𝑠+𝐾; 𝑄(𝑠) = 𝑠 3 + 9𝑠 2 + 20𝑠 + 𝐾 = 0

Form the Routh-Hurwitz table: 𝑠3 1 20 From the table, the gain for the
𝑠2 9 K crossing of the imaginary axis is:
𝑠1 (180-K)/9
3 marks 180 − 𝐾
𝑠0 K 𝐾𝑐𝑟 = = 0 → 𝐾𝑐𝑟 = 180
9
2 marks
Root Locus
5 0.5

4 System: G
Gain: 33.3
Pole: -1.11 + 1.92i
3 Damping: 0.502
Overshoot (%): 16.2
Frequency (rad/sec): 2.21
Imaginary Axis

-1 7 marks for plot

-2 0.5
-6 -5 -4 -3 -2 -1 0 1 2
Real Axis
4.2 To evaluate the system gain at the given damping factor, 𝑠 ∗ = −1.11 + 𝑗1.92: 1 mark

𝐾 = |𝑠 ∗ ||4 + 𝑠 ∗ ||5 + 𝑠 ∗ | = |−1.11 + 𝑗1.92||2.89 + 𝑗1.92||3.89 + 𝑗1.92| = 33.38

3 marks +3 marks
33.3
4.3 The velocity error constant 𝐾𝑣 = = 1.665. 1 mark
4𝑥5
1 1
The steady state velocity error 𝑒𝑉 = 𝐾 = 1.665 = 0.600 2 mark
𝑣

0.6006
4.4 To reduce the error by a factor of 8, 𝑒𝑉𝑛𝑒𝑤 = 8
= 0.075075 1 mark
1 1
The new error constant 𝐾𝑣𝑛𝑒𝑤 = = = 13.32 1 mark
𝑒𝑉𝑛𝑒𝑤 0.075075

𝐾𝑉 𝑛𝑒𝑤 13.32
Now, 𝐾𝑉 𝑛𝑒𝑤 = 𝑘𝑝 𝐾𝑣 → 𝑘𝑝 = 𝐾𝑣
= 1.665 = 8 2 marks
Example 6: Nichols chart Plotting
90
Given the third-order transfer function 𝐺(𝑠) = :
(𝑠+2.5)(𝑠+4)(𝑠+5)

the expression for 𝜑(𝑗𝜔) , the phase response of the system. [4]

Write the expression for |𝐺(𝑗𝜔|, the magnitude-frequency response of the system. [4]

Complete Table 1. [6]

Using frequency data presented in Table 1, plot the Nichol’s chart for the system. [6]

Evaluate the phase and gain margins of the system using the Nichol’s chart. [6]

𝝎 (𝒓𝒂𝒅𝒔−𝟏 ) 0.1 0.5 1.0 4 8 15 20

𝝋(𝒋𝝎) (deg) -24.1 -141.7 -227.2 -237.5
|𝑮(𝒋𝝎| 1.8 1.59 0.13
M=20log|𝑮(𝒋𝝎|dB 5.09 4.02

Table 1: Frequency response table for a given third-order system.

Example 7: Nichols Chart
6
You are required to estimate the percentage overshoot of the open-loop control system 𝐺𝐻(𝑠) =
𝑠(𝑠+1)(𝑠+3)
based on the system indices obtainable from the Nichols chart of GH(s):

1.1 Complete Table 1 with the following the frequency response information of GH(s). [19]

Frequency (𝜔 𝑟𝑎𝑑/𝑠) 0.1 0.5 0.8 1.2 2 6 10 Marks

Phase 𝜑(𝑗𝜔) degrees -97.6 -126 -143.6 -162 -187.1 -234 -247.6 4
Magnitude (M) 19.89 3.53 1.88 0.99 0.37 0.02 0.006 4
20 log 𝑀 25.97 10.95 5.51 -0.082 -8.59 -32.2 -44.8 4
Closed-loop 0.1 1.3 3.5 8 -5 -32.2 -44.2 7
Magnitude in dB

Table 1: The frequency response table for GH(s)

1.2 Make the Nichols chart on the chart provided and use the chart to evaluate the following indices: [16]

The resonant peak 𝑀𝑝 (𝑜𝑟 𝑀𝑟 )=8dB (3 marks)

The phase margin 𝜑𝑀 =24 deg (3 marks)

The gain margin 𝐺𝑀 =6dB (3 marks) . Remaining seven marks on the drawn Nichol’s chart.

1.3 Estimate the percentage overshot to a step input. [12]

Now 𝑀𝑟 = 20𝑙𝑜𝑔𝑀 = 8 → log 𝑀 = 208 . ∴ 𝑀 = 100.4 = 2.5119 √

1 1
Also, 𝑀 = = 2.5119 → = 0.19905 = 𝜉√1 − 𝜉 2 √√
2𝜉√1−𝜉 2 2𝑥2.5119

Then (1 − 𝜉 2 )𝜉 2 = 𝜉 2 − 𝜉 4 = 0.199052 = 0.03962 → 𝜉 4 − 𝜉 2 + 0.03962 = 0, 𝜉 = ±0.9791, ±0.2033

√√ √√ √√
1
Since the resonance peak does not exist when 𝜉 > , 𝜉 = 0.2033 √
√2

− 𝜋𝜉 0.2033
√1−𝜉2 −
The percentage overshoot %𝑂𝑆 = 100𝑒 (2) = 100𝑒 √1−0.20332 = 52.08% (1)
7 marks for chart