Ch.04 Time Response

25/02/2023
System Dynamics and Control 1 Time Response System Dynamics and Control 2 Time Response
Learning Outcome
After completing this chapter, the student will be able to
• use poles and zeros of transfer functions to determine the time
response of a control system
• describe quantitatively the transient response of 1st-order systems
• write the general response of 2nd-order systems given the pole
location
• find the damping ratio and natural frequency of a 2nd-order system
• find the settling time, peak time, percent overshoot, and rise
time for an underdamped 2nd-order system
• approximate higher-order systems and systems with zeros as
1st/2nd-order systems
• describe the effects of nonlinearities on the system time response
• find the time response from the state-space representation
HCM City Univ. of Technology, Faculty of Mechanical Engineering Nguyen Tan Tien HCM City Univ. of Technology, Faculty of Mechanical Engineering Nguyen Tan Tien
§1.Introduction §2.Poles, Zeros, and System Response

- This chapter - The output response of a system is the sum of two responses
• analyze of system transient response • the forced response
• demonstrate applications of the system representation by • the natural response
evaluating the transient response from the system model - Output response of a system depends on the positions of the
• evaluate the response of a subsystem prior to inserting it into poles and zeroes and their relationships
the closed-loop system → study
• describe a valuable analysis and design tool, poles and zeros • the concept of poles and zeroes
• analyze the models to find the step response of 1st- and 2nd- • fundamental to the analysis and design of control systems
order systems • simplifies the evaluation of a system’s response
§2.Poles, Zeros, and System Response §2.Poles, Zeros, and System Response
- Transfer function (TF) - Example of poles and zeros of a first order system
𝑈(𝑠) 𝑌(𝑠) 𝑌(𝑠) 𝑗𝜔
𝐺 𝑠 = 𝐺(𝑠)
input 𝑈(𝑠) output 1 𝑠 − plane
𝑅 𝑠 = 𝑠+2 𝐶(𝑠)
𝑠 𝜎
𝑠+5 −5 −2
𝑚 𝑚−1
𝑌 𝑠 𝑏0 𝑠 + 𝑏1 𝑠 + ⋯ + 𝑏𝑚−1 𝑠 + 𝑏𝑚 system showing input and output pole-zero plot of the system
𝐺 𝑠 = =
𝑈 𝑠 𝑎0 𝑠 𝑛 + 𝑎1 𝑠 𝑛−1 + ⋯ + 𝑎𝑛−1 𝑠 + 𝑎𝑛 The unit step response of the system
• The poles of a TF are roots of 1 𝑠+2 𝑠 +2 𝐴 𝐵 2/5 3/5
𝐶 𝑠 = × = = + = +
𝑎0 𝑠 𝑛 + 𝑎1 𝑠 𝑛−1 + ⋯ + 𝑎𝑛−1 𝑠 + 𝑎𝑛 = 0 𝑠 𝑠+5 𝑠 𝑠+5 𝑠 𝑠+5 𝑠 𝑠+5
• The zeros of a TF are roots of 𝑠+2 2 𝑠+2 3
𝐴= ቤ = ,𝐵 = ቤ =
𝑏0 𝑠 𝑚 + 𝑏1 𝑠 𝑚−1 + ⋯ + 𝑏𝑚−1 𝑠 + 𝑏𝑚 = 0 𝑠 + 5 𝑠→0 5 𝑠 𝑠→−5 5
2 3
→ 𝑐 𝑡 = + 𝑒 −5𝑡
5 5
1
25/02/2023
§2.Poles, Zeros, and System Response §2.Poles, Zeros, and System Response
input pole system zero system pole - Ex.4.1 Evaluating Response using Poles
1 𝑗𝜔 𝑗𝜔 1 𝑗𝜔
𝑠 𝑠 − plane
𝑠+2
𝑠 − plane 𝑠+5 𝑠 − plane Write the output, 𝑐 𝑡 , in general terms, for the given system
𝜎 𝜎 𝜎 1
𝑅 𝑠 = 𝑠+3 𝐶(𝑠)
−2 −5 𝑠
(𝑠 + 2)(𝑠 + 4)(𝑠 + 5)
output transform 𝐶(𝑠) =

2/5
+
3/5 Specify the forced and natural parts of the solution
𝑠 𝑠+5
Solution
2 3 −5𝑡
output time response 𝑐 𝑡 = +
5 5
𝑒
• Each system pole generates an exponential as part of the
forced response natural response natural response
Remark Evolution of a system response
• The input’s pole generates the forced response
• A pole of the input function generates the form of the forced response
𝐾1 𝐾2 𝐾3 𝐾4
• A pole of the TF generates the form of the natural response, Thus 𝐶 𝑠 = + + +
𝑠 𝑠+2 𝑠+4 𝑠+5
an exponential response of 𝑒−𝑎𝑡 forced response natural response
• The zeros and poles generate the amplitudes for both the →𝑐 𝑡 = 𝐾1 + 𝐾2 𝑒 −2𝑡 + 𝐾3 𝑒 −4𝑡 + 𝐾4 𝑒 −5𝑡
forced and natural responses
§3.First Order Systems §3.First Order Systems 𝑐(𝑡)

initial slope =
1
time constant
=𝑎
1.0
- A first-order system without zeros can be described by the TF 1.Time Constant 𝑇𝑐 0.9
𝑗𝜔 1 0.8
𝐺(𝑠) - Time constant 𝑇𝑐 = 0.7
𝑅 𝑠 𝑎 𝐶(𝑠)
𝑠 − plane 𝑎 0.6
63% of final value at
𝜎 𝑡 = one time constant
𝑠+𝑎 0.5
−𝑎 0.4
first order system pole plot
0.3
- If the input is step 0.2
0.1
1 𝑎 𝑎 1 1 0.0 𝑡
𝐶 𝑠 =𝑅 𝑠 𝐺 𝑠 = × = = − 1 2 3 4
𝑠 𝑠+𝑎 𝑠 𝑠+𝑎 𝑠 𝑠+𝑎 𝑒 −𝑎𝑡
ቚ =𝑒 −1
= 0.37 𝑎 𝑎 𝑎 𝑎
1 𝑇𝑟
- Taking the inverse transform, the step response is given by 𝑡= 𝑇𝑠
𝑎
1st-order system response to a unit step
𝑐 𝑡 = 𝑐𝑓 𝑡 + 𝑐𝑛 𝑡 = 1 − 𝑒 −𝑎𝑡 𝑐 𝑡 ቚ 1 = 1 − 𝑒 −𝑎𝑡 ቚ 1 = 0.63
𝑡= 𝑡=
• input pole, 0, generates the forced response 𝑐𝑓 𝑡 = 1 𝑎 𝑎
→The time constant is the time it takes for the step response to
• system pole, −𝑎, generates the natural response 𝑐𝑛 𝑡 = −𝑒−𝑎𝑡
rise to 63% of its final value
1 𝑎 𝑎 1 1
𝐶 𝑠 =𝑅 𝑠 𝐺 𝑠 = × = = − → 𝑐 𝑡 = 𝑐𝑓 𝑡 + 𝑐𝑛 𝑡 = 1 − 𝑒 −𝑎𝑡
𝑠 𝑠+𝑎 𝑠 𝑠+𝑎 𝑠 𝑠+𝑎
§3.First Order Systems 𝑐(𝑡)

initial slope =
1
time constant
constant
=𝑎 §3.First Order Systems 𝑐(𝑡)
initial slope =
1
time constant
constant
=𝑎
𝑑 −𝑎𝑡 1.0
2.Rise Time 𝑇𝑟 1.0
𝑒 ቤ = −𝑎𝑒 −𝑎𝑡 ቚ = −𝑎 0.9 0.9
𝑑𝑡 𝑡=0
𝑡=0 0.8
- The time for the waveform to go 0.8
0.7 0.7
63% of final value at 63% of final value at
𝑎: the initial rate of change 0.6
𝑡 = one time constant from 0.1 to 0.9 of its final value 0.6 𝑡 = one time constant
0.5 0.5
of the exponential at 𝑡 = 0 0.4 - Rise time is found by solving 0.4
0.3 0.3
- Transient response specifications 0.2 𝑐 𝑡 = 1 − 𝑒 −𝑎𝑡 0.2
0.1 0.1
𝑇𝑐 : time constant 0.0 𝑡 𝑐 𝑡1 = 0.1 0.0 𝑡
1 2 3 4 1 2 3 4
ln0.9
𝑇𝑟 : rise time 𝑎 𝑎 𝑎 𝑎 1 − 𝑒 −𝑎𝑡1 = 0.1 → 𝑡1 = − 𝑎 𝑎 𝑎 𝑎
𝑇𝑟
𝑇𝑠 𝑎 𝑇 𝑟
𝑇𝑠
𝑇𝑠 : settling time 1st-order system response to a unit step 𝑐 𝑡2 = 0.9 1st-order system response to a unit step
- 𝑎 relates to the speed at which the system responds to a step ln0.1
input: the farther the pole from the imaginary axis, the faster the 1 − 𝑒 −𝑎𝑡2 = 0.9 → 𝑡2 = −
𝑎
transient response ln0.1 ln0.9 0.11 2.31
→ 𝑡2 − 𝑡1 = − + =− +
- 𝑎 has the units 1/𝑠, or frequency → the exponential frequency 𝑎 𝑎 𝑎 𝑎
2.2
→ 𝑇𝑟 =
𝑎
2
25/02/2023
§3.First Order Systems 𝑐(𝑡)

initial slope =
1
time constant
constant
=𝑎 §3.First Order Systems
1.0
3.Settling Time 𝑇𝑠 0.9
4.First-Order Transfer Functions via Testing
0.8
- The time for the response to 0.7
- Consider a simple first-order system with the step input
63% of final value at
reach, and stay within, 2% of its 0.6
𝑡 = one time constant 1 𝐺(𝑠)
0.5 𝑅 𝑠 = 𝐶(𝑠)
final value 0.4 𝑠 𝐾
0.3
- Settling time is found by solving 0.2
𝑠+𝑎
𝑐 𝑡 = 1 − 𝑒 −𝑎𝑡 0.1
0.0 𝑡 1 𝐾 𝐾/𝑎 𝐾/𝑎
𝑐 𝑇𝑠 = 0.98 1 2 3 4 𝐶 𝑠 =
× = −
𝑎 𝑎 𝑎 𝑎 𝑠 𝑠+𝑎 𝑠 𝑠+𝑎
1 − 𝑒 −𝑎𝑇𝑠 = 0.98 𝑇𝑟
𝑇𝑠 If we can identify 𝐾 and 𝑎 from laboratory testing, we can
ln0.02 1st-order system response to a unit step
→ 𝑇𝑠 = − obtain the TF of the system
𝑎
4
→ 𝑇𝑠 =
𝑎
§3.First Order Systems §3.First Order Systems

5
- Example using the plot of 𝐺 𝑠 = Skill-Assessment Ex.4.2
𝑠+7
From the plot 𝑐(𝑡)
Problem Find the time constant, 𝑇𝑐 , rising time, 𝑇𝑟 , and settling
• Steady state value 𝑐𝑠𝑠 = 0.72 0.8
0.72
time, 𝑇𝑠 , of the system with the following TF
50
• Time constant 0.6
𝐺 𝑠 =
amplitude
0.45 𝑠 + 50
0.63 × 𝑐𝑠𝑠 = 0.45 → 𝑇𝑐 = 0.14𝑠 0.4
Solution From the TF
Therefore 0.2
𝑎 = 50
1 1 0.0 𝑡
𝑇𝑐 = → 𝑎 = = 7.14 0.10.140.2 0.3 0.4 0.5 0.6 1 1
𝑎 0.14 time (sec) Time constant 𝑇𝑐 = = = 0.02𝑠
𝐾 𝑎 50
= 𝑐𝑠𝑠 → 𝐾 = 𝑎𝑐𝑠𝑠 = 7.14 × 0.72 = 5.14 2.2 2.2
𝑎 Rising time 𝑇𝑟 = = = 0.044𝑠
5.14 𝑎 50
→ The TF of the plot 𝐺 𝑠 = 4 4
𝑠 + 7.14 Settling time 𝑇𝑠 = = = 0.08𝑠
Note: 1st order system has no overshoot and nonzero initial slope 𝑎 50
§4.Second Order Systems: Introduction §4.Second Order Systems: Introduction

- The general 2nd-order system - Second-order systems, pole plots, and step responses
𝑅 𝑠 =
1 𝐺(𝑠)
𝐶(𝑠) • Overdamped
𝑠 𝑏 𝑐(𝑡)
𝑗𝜔
𝑠 2 + 𝑎𝑠 + 𝑏 1 𝑠 − plane
𝐺(𝑠) 1.0
𝑅 𝑠 = 𝐶(𝑠)
general 𝑠 9
𝜎 0.5
𝑠 2 + 9𝑠 + 9 −7.854 −1.146
- Depending on component values, the second order system overdamped pole plot 0.0 𝑡
0 1 2 3 4 5
exhibits a wide range of responses such as
• first-order system 9
𝐶 𝑠 =
• damped oscillation 𝑠 𝑠 + 1.146 𝑠 + 7.854
• pure oscillation → 𝑐 𝑡 = 1 + 0.171𝑒 −7.854𝑡 − 0.171𝑒 −1.146𝑡
Poles: two real at −𝜎1 and −𝜎2
Natural response: 𝑐𝑛 𝑡 = 𝐾1 𝑒 −𝜎1 𝑡 + 𝐾2 𝑒 −𝜎2 𝑡
[r,p,k]=residue([9],[1,9,9,0]), 𝐶 𝑠 = 1/𝑠 → 𝑐 𝑡 = 1, 𝐶 𝑠 = 1/ 𝑠 + 𝑎 → 𝑐 𝑡 = 𝑒 −𝑎𝑡

3
25/02/2023
• Underdamped 𝑐(𝑡) 𝑐(𝑡)

𝑗𝜔 𝑗𝜔
1 𝑠 − plane 1 𝑠 − plane
𝐺(𝑠) 1.0 𝐺(𝑠) 1.0
𝑅 𝑠 = 𝐶(𝑠) +𝑗 8 𝑅 𝑠 = 𝐶(𝑠) +𝑗 8
𝑠 9 𝑠 9
𝜎 0.5 𝜎 0.5
𝑠 2 + 2𝑠 + 9 −1 𝑠 2 + 2𝑠 + 9 −1
−𝑗 8 −𝑗 8
underdamped 0.0 𝑡 underdamped 0.0 𝑡
0 1 2 3 4 5 0 1 2 3 4 5
𝑐(𝑡)
9 exponential decay generated by
𝐶 𝑠 = 2.0 𝐴𝑒 −𝜎 𝑑 𝑡 real part of complex pole pair
𝑠 𝑠 2 + 2𝑠 + 9
1.5
9
= 1.0
𝑠 𝑠 − −1 − 𝑗 8 𝑠 − −1 + 𝑗 8
sinusoidal oscilation generated by
0.5
→ 𝑐 𝑡 = 1 − 1.06𝑒 −𝑡 cos 8𝑡 − 19.470 imaginary part of complex pole pair
0.0 𝑡
Poles: two complex at −𝜎𝑑 ± 𝑗𝜔𝑑 0 1 2 3 4 5 6 7
Natural response: 𝑐𝑛 𝑡 = 𝐴𝑒 −𝜎𝑑 𝑡 cos 𝜔𝑑 𝑡 − 𝜙

• Undamped 𝑐(𝑡) • Critically damped

𝑐(𝑡)
𝑗𝜔 2.0 𝑗𝜔
1 𝑠 − plane 1 𝑠 − plane
𝐺(𝑠) 𝐺(𝑠) 1.0
𝑅 𝑠 = 𝐶(𝑠) +𝑗3 𝑅 𝑠 = 𝐶(𝑠)
𝑠 9 𝑠 9
𝜎 1.0 𝜎 0.5
𝑠2 + 9 𝑠 2 + 6𝑠 + 9 −3
−𝑗3
undamped 0.0 𝑡 critically damped 0.0 𝑡
0 1 2 3 4 5 0 1 2 3 4 5
9 1 𝑠 9
𝐶 𝑠 = = − 2 𝐶 𝑠 =
𝑠 𝑠2 + 9 𝑠 𝑠 + 32 𝑠 𝑠 2 + 6𝑠 + 9
→ 𝑐 𝑡 = 1 − cos3𝑡 1 1 3
= − −
Poles: two imaginary at ±𝑗𝜔1 𝑠 𝑠+3 𝑠+3 2
→ 𝑐 𝑡 = 1 − 3𝑡𝑒 −3𝑡 − 𝑒 −3𝑡
Natural response: 𝑐𝑛 𝑡 = 𝐴cos 𝜔1 𝑡 − 𝜙
Poles: two real at −𝜎1
Natural response: 𝑐𝑛 𝑡 = 𝐾1 𝑒 −𝜎1 𝑡 + 𝐾2 𝑡𝑒 −𝜎1 𝑡
𝐶 𝑠 = 𝑠/ 𝑠 2 + 𝑎 2 → 𝑐 𝑡 = cos𝑎𝑡 𝐶 𝑠 = 𝑎/ 𝑠 + 𝑎 2 → 𝑐 𝑡 = 𝑎𝑡𝑒 −𝑎𝑡

• Step responses for second-order system damping cases Skill-Assessment Ex.4.3
𝑐(𝑡) Problem For each of the following transfer functions, write, by
undamped
2.0 inspection, the general form of the step response
1.8
400 900
1.6
a. 𝐺 𝑠 = 2 b. 𝐺 𝑠 = 2
1.4
underdamped
𝑠 + 12𝑠 + 400 𝑠 + 90𝑠 + 900
1.2 critically 225 625
1.0 damped c. 𝐺 𝑠 = 2 d. 𝐺 𝑠 = 2
0.8 𝑠 + 30𝑠 + 225 𝑠 + 625
0.6
overdamped
0.4
0.2
0.0 𝑡
0 1 2 3 4 5
4
25/02/2023

400 225 400
Solution a. 𝐺 𝑠 = 2 c. 𝐺 𝑠 = 2 =
𝑠 + 12𝑠 + 400 𝑠 + 30𝑠 + 225 𝑠 + 15 2
400 The system has two repeated real poles at −15
=
𝑠 − −6 − 𝑗19.08 𝑠 − −6 + 𝑗19.08 The general form of the step response
The system has two complex poles at −6 ± 𝑗19.08 𝑐 𝑡 = 𝐴 + 𝐵𝑒 −15𝑡 + 𝐶𝑡𝑒 −15𝑡
The general form of the step response 625 625
𝑐 𝑡 = 𝐴 + 𝐵𝑒 −6𝑡 cos 19.08𝑡 + 𝜙 d. 𝐺 𝑠 = 2 =
𝑠 + 625 𝑠 2 + 252
900 900 The system has two imaginary poles at ±𝑗25
b. 𝐺 𝑠 = 2 =
𝑠 + 90𝑠 + 900 𝑠 + 78.54 𝑠 + 11.46 The general form of the step response
The system has two real poles at −78.54 and −11.46 𝑐 𝑡 = 𝐴 + 𝐵cos 25𝑡 + 𝜙
The general form of the step response
𝑐 𝑡 = 𝐴 + 𝐵𝑒 −78.54𝑡 + 𝐶𝑒 −11.46𝑡
𝐺 𝑠 has poles at −𝜎𝑑 ± 𝑗𝜔𝑑 → natural response 𝑐𝑛 𝑡 = 𝐴𝑒 −𝜎𝑑 𝑡 cos 𝜔𝑑 𝑡 − 𝜙 𝐺 𝑠 has poles at −𝜎1 , −𝜎1 → natural response 𝑐𝑛 𝑡 = 𝐾1 𝑒 −𝜎1 𝑡 + 𝐾2 𝑡𝑒 −𝜎1𝑡
𝐺 𝑠 has poles at −𝜎1 , −𝜎2 → natural response 𝑐𝑛 𝑡 = 𝐾1 𝑒 −𝜎1𝑡 + 𝐾2 𝑒 −𝜎2𝑡 𝐺 𝑠 has poles at ±𝑗𝜔1 → natural response 𝑐𝑛 𝑡 = 𝐴cos 𝜔1 𝑡 − 𝜙
§5.The General Second Order Systems §5.The General Second Order Systems
1.Natural frequency 𝜔𝑛 - Ex.4.5 Finding ζ and 𝜔𝑛 For a Second-Order System
- The natural frequency of a second-order system is the Given the transfer function of Eq. (4.23), find ζ and 𝜔𝑛
frequency of oscillation of the system without damping 36
𝐺 𝑠 = 2 (4.23)
2.Damping ratio ζ 𝑠 + 4.2𝑠 + 36
- The natural frequency of a second-order system is the Solution
frequency of oscillation of the system without damping From observation
exponential decay frequency 1 natural period • The natural frequency of the given system
ζ= =
natural frequency 2𝜋 exponential time constant 𝜔𝑛 = 𝑏 = 36 = 6
- Consider the general second order system time constant • The damping ratio of the given system
𝑏 𝜔𝑛2 𝑎 4.2
𝐺 𝑠 = 2 ≡ ζ= = = 0.35
𝑠 + 𝑎𝑠 + 𝑏 𝑠 2 + 2ζ𝜔𝑛 𝑠 + 𝜔𝑛2 2 𝑏 2×6
→ 𝜔𝑛 = 𝑏 (4.18)
𝑎
ζ= (4.20)
2 𝑏
§5.The General Second Order Systems §5.The General Second Order Systems
- The poles of the transfer function - Ex.4.4 Characterizing Response from the Value of ζ
𝜔𝑛2 Find the value of ζ and report the kind of response expected
𝐺 𝑠 = 2 → 𝑝1,2 = −ζ𝜔𝑛 ± 𝜔𝑛 ζ 2 − 1
𝑠 + 2ζ𝜔𝑛 𝑠 + 𝜔𝑛2 Solution
→ Second-order response as a function of damping ratio ζ
𝑅 𝑠 12 𝐶(𝑠) 𝑎 8
ζ=0 𝑐(𝑡) 0<ζ<1 𝑐(𝑡) 𝑠 2 + 8𝑠 + 12
ζ= = = 1.16
𝑗𝜔 𝑠 − plane 𝑗𝜔 𝑠 − plane 2 𝑏 2 12
+𝑗𝜔𝑛 +𝑗𝜔𝑛 1 − ζ 2
→ overdamped
𝜎 𝜎
−ζ𝜔𝑛
𝑅 𝑠 16 𝐶(𝑠) 𝑎 8
−𝑗𝜔𝑛 −𝑗𝜔𝑛 1 − ζ 2
𝑠 2 + 8𝑠 + 16
ζ= = =1
undamped 𝑡 underdamped 𝑡 2 𝑏 2 16
ζ=1 𝑐(𝑡) ζ>1 𝑐(𝑡)
𝑗𝜔
𝑠 − plane
𝑗𝜔
𝑠 − plane → critically damped
𝑅 𝑠 20 𝐶(𝑠) 𝑎 8
𝜎 𝜎
𝑠 2 + 8𝑠 + 20
ζ= = = 0.89
−ζ𝜔𝑛
2 𝑏 2 20
critically damped 𝑡 overdamped 𝑡 → underdamped
5
25/02/2023
§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

- The step response for the general second-order system - Second-order underdamped response specifications
𝐺(𝑠)
𝜔𝑛2 𝑅 𝑠 =
1
𝜔𝑛2 𝐶(𝑠)
𝑐(𝑡) • Rise time, 𝑇𝑟
𝑠
𝐶 𝑠 = 𝑐𝑚𝑎𝑥
𝑠 𝑠 2 + 2ζ𝜔𝑛 𝑠 + 𝜔𝑛2 𝑠 + 2ζ𝜔𝑛 𝑠 + 𝜔𝑛 2 2
1.02𝑐𝑓𝑖𝑛𝑎𝑙 ±2%𝑐𝑓𝑖𝑛𝑎𝑙 𝑇𝑟 = 𝑡ቚ − 𝑡ቚ
𝑐𝑓𝑖𝑛𝑎𝑙 0.9𝑐𝑓𝑖𝑛𝑎𝑙 0.1𝑐𝑓𝑖𝑛𝑎𝑙
1 −ζ𝜔𝑛 𝑡 0.98𝑐𝑓𝑖𝑛𝑎𝑙
𝑐 𝑡 =1− 𝑒 cos 𝜔𝑛 1 − ζ 2 𝑡 − 𝜙 (4.28) 0.90𝑐𝑓𝑖𝑛𝑎𝑙 • Peak time, 𝑇𝑝
1 − ζ2
𝑇𝑝 = 𝑡ቚ
𝜙 = tan−1 ζ/ 1 − ζ 2 𝑐𝑚𝑎𝑥
𝑐(𝜔𝑛 𝑡)
ζ = 0.1 0.10𝑐𝑓𝑖𝑛𝑎𝑙 • Percent overshoot, %𝑂𝑆
1.8
1.6
0.2
0.3 𝑡 𝑐𝑚𝑎𝑥 − 𝑐𝑓𝑖𝑛𝑎𝑙
0.4
0.5
𝑇𝑟 𝑇𝑝 𝑇𝑠 %𝑂𝑆 = × 100
1.4
1.2
0.6
0.7
0.8
𝑐𝑓𝑖𝑛𝑎𝑙
1.0
0.8
• Settling time, 𝑇𝑠
0.6
0.4
The time required for the transient’s damped oscillations to
0.2 reach and stay within 2% of the steady-state value
0.0 𝜔𝑛 𝑡
0 5 10 15 20
Second-order underdamped responses for damping ratio values

1.Evaluation of 𝑇𝑝 2.Evaluation of %𝑂𝑆
𝑐𝑚𝑎𝑥 − 𝑐𝑓𝑖𝑛𝑎𝑙
𝑇𝑝 = 𝑡ቚ = 𝑡ቚ %𝑂𝑆 = × 100
𝑐𝑚𝑎𝑥 ሶ
𝑐=0 𝑐𝑓𝑖𝑛𝑎𝑙
𝜔𝑛
(4.28) → 𝑐ሶ 𝑡 = 𝑒 −ζ𝜔𝑛 𝑡 sin𝜔𝑛 1 − ζ 2 𝑡 (4.28) → 𝑐𝑚𝑎𝑥 = 𝑐 𝑇𝑝
1 − ζ2
−ζ𝜋/ 1−ζ 2 ζ
𝑐ሶ 𝑡 = 0 → 𝜔𝑛 1 − ζ 2 𝑡 = 𝑛𝜋 = 1−𝑒 cos𝜋 + sin𝜋
𝑛𝜋 1−ζ2
→𝑡=
𝜔𝑛 1 − ζ 2 −ζ𝜋/ 1−ζ 2
= 1+𝑒
𝜋
→ 𝑇𝑝 = (4.34) For the unit step 𝑐𝑓𝑖𝑛𝑎𝑙 = 1
𝜔𝑛 1 − ζ 2 ζ𝜋
−
→ %𝑂𝑆 = 𝑒 1−ζ × 100
2
(4.38)
𝑒 −ζ𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 1 − ζ 2 𝑡 − 𝜙 𝑒 −ζ𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 1 − ζ 2 𝑡 − 𝜙

1 1
𝑐 𝑡 =1− (4.28) 𝑐 𝑡 =1− (4.28)
1−ζ 2 1−ζ 2

ζ can be found from given %𝑂𝑆 3.Evaluation of 𝑇𝑠
ln %𝑂𝑆/100 - To find the settling time → to find the time it takes for the
(4.38) → ζ = − (4.39)
𝜋 2 + ln2 %𝑂𝑆/100 amplitude of the decaying sinusoid in (4.28) to reach 0.02
1
Percent overshoot versus damping ratio 𝑒 −ζ𝜔𝑛 𝑡 = 0.02
1 − ζ2
−ln 0.02 1 − ζ 2
→ 𝑇𝑠 =
ζ𝜔𝑛
- An approximation for the settling time that will be used for all
values of ζ
4
𝑇𝑠 = (4.42)
ζ𝜔𝑛
𝑒 −ζ𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 1 − ζ 2 𝑡 − 𝜙
1
𝑐 𝑡 =1− (4.28)
−ζ𝜋/ 1−ζ 2 1−ζ2
%𝑂𝑆 = 𝑒 × 100 (4.38)
6
25/02/2023

4.Evaluation of 𝑇𝑟 - Ex.4.5 Finding 𝑇𝑝 , %𝑂𝑆, 𝑇𝑠 , and 𝑇𝑟 from a TF
When a precise analytical relationship between rise time and Finding 𝑇𝑝 , %𝑂𝑆, 𝑇𝑠 , and 𝑇𝑟 from given transfer function
damping ratio, ζ, cannot be found → using numerical method 100
Solution 𝐺 𝑠 = 2
- Ex. Normalized rise time versus damping ratio for a second- 𝑠 + 15𝑠 + 100
order underdamped response From observation
𝜔𝑛 = 𝑏 = 100 = 10𝑟𝑎𝑑/𝑠
𝑎 15
ζ= = = 0.75
2 𝑏 2 100
𝜋 𝜋
𝑇𝑝 = = = 0.475𝑠
𝜔𝑛 1 − ζ 2 10 1 − 0.752
−ζ𝜋/ 1−ζ 2 1−0.752
𝑂𝑆 = 𝑒 × 100 = 𝑒−0.75𝜋/ × 100 = 2.84%
4 4
𝜔𝑛 𝑇𝑟 ≈ 1.76ζ 3 − 0.417ζ 2 + 1.039ζ + 1 𝑇𝑠 = = = 0.533𝑠
with maximum error less than 0.5% for 0 < ζ < 0.9
ζ𝜔𝑛 0.75 × 10

- Ex.4.5 Finding 𝑇𝑝 , %𝑂𝑆, 𝑇𝑠 , and 𝑇𝑟 from a TF - The relation between peak time, percent overshoot, and settling
Finding 𝑇𝑝 , %𝑂𝑆, 𝑇𝑠 , and 𝑇𝑟 from given transfer function time to the location of the poles
100 • Pole plot for an underdamped second-order system
Solution 𝐺 𝑠 = 2
𝑠 + 15𝑠 + 100 𝑗𝜔 𝜔𝑑 ≡ 𝜔𝑛 1 − ζ 2
From observation 𝜔𝑛
𝜎𝑑 ≡ ζ𝜔𝑛
+𝑗𝜔𝑛 1 − ζ 2 = +𝑗𝜔𝑑
𝜔𝑛 = 10𝑟𝑎𝑑/𝑠 cos ζ = 𝜃 −1
𝜎 cos𝜃 = ζ
−ζ𝜔 = −𝜎
ζ = 0.75
𝑛 𝑑
𝜔𝑑 : damped frequency of oscillation

2
−𝑗𝜔 1 − ζ = −𝑗𝜔
𝑇𝑝 = 0.475𝑠 𝑛 𝑑
𝜎𝑑 : exponential damping frequency

𝑂𝑆 = 2.84% - The peak time and settling time in terms of the pole location
𝑇𝑠 = 0.533𝑠 𝜋 𝜋
𝑇𝑝 = = (4.44)
𝜔𝑛 𝑇𝑟 ≈ 1.76ζ 3 − 0.417ζ 2 + 1.039ζ + 1 𝜔𝑛 1 − ζ 2 𝜔𝑑
= 1.76 × 0.75 3 − 0.417 × 0.75 2 + 1.039 × 0.75 + 1 4 4
= 2.2870 𝑇𝑠 = = (4.45)
ζ𝜔𝑛 𝜎𝑑
→ 𝑇𝑟 = 0.2287𝑠

- Lines of constant peak time, 𝑇𝑝 , settling time, 𝑇𝑠 , and percent - Step responses of the second-order underdamped systems as
overshoot, %𝑂𝑆 (𝑇𝑠2 < 𝑇𝑠1 , 𝑇𝑝2 < 𝑇𝑝1 , %𝑂𝑆1 < %𝑂𝑆2 ) poles move with constant real part
%𝑂𝑆2 𝑐(𝑡)
𝑗𝜔 𝑗𝜔
same envelope 𝐴𝑒 −𝜎 𝑑 𝑡
𝑠 − plane
𝑇𝑝 2 1
constant setting time 𝑇𝑠 𝐺 𝑠 = 2
%𝑂𝑆1 constant overshoot %𝑂𝑆
𝑠 + 2ζ𝜔𝑛 𝑠 + 𝜔𝑛2 +𝜔𝑑
𝑇𝑝 1 𝑠 − plane
constant peak time 𝑇𝑝
1
= 𝜎
𝑠 − −𝜎𝑑 − 𝑗𝜔𝑑 𝑠 − −𝜎𝑑 + 𝑗𝜔𝑑 −𝜎𝑑
𝜎 −𝜔𝑑
𝑇𝑠2 𝑇𝑠1
pole motion 𝑡
t=0:0.01:5;
sigmad=-1; omegad=[3 5 7];
for j=1:length(omega)
N=[1]; D=poly([sigmad-omegad(j)*i, sigmad+omegad(j)*i]); T=tf(N,D);
step(T/dcgain(T),t); hold on;
axis([0,5,0,2]);
end
7
25/02/2023

- Step responses of the second-order underdamped systems as - Step responses of the second-order underdamped systems as
poles move with constant imaginary part poles move with constant damping ratio part
𝑐(𝑡) 𝑐(𝑡) ζ𝜋
−
𝑗𝜔 𝑗𝜔 1−ζ 2
same damped frequency same over shoot 𝑒
𝑠 − plane 𝑠 − plane
1 1
𝐺 𝑠 = 𝐺 𝑠 =
𝑠 2 + 2ζ𝜔𝑛 𝑠 + 𝜔𝑛2 +𝜔𝑑 𝑠 2 + 2ζ𝜔𝑛 𝑠 + 𝜔𝑛2 +𝜔𝑑
1 1 cos −1 ζ = 𝜃
= 𝜎 = 𝜎
𝑠 − −𝜎𝑑 − 𝑗𝜔𝑑 𝑠 − −𝜎𝑑 + 𝑗𝜔𝑑 −𝜎𝑑 𝑠 − −𝜎𝑑 − 𝑗𝜔𝑑 𝑠 − −𝜎𝑑 + 𝑗𝜔𝑑 −𝜎𝑑
−𝜔𝑑
−𝜔𝑑
pole motion pole motion

𝑡 𝑡
t=0:0.01:5; t=0:0.01:5;
sigmad=-[0.25 0.75 1.25]; omegad=3; sigmad=-[1 2 3]; omegad=2*sigmad;
for j=1:length(sigmad) for j=1:length(sigmad)
N=[1]; D=poly([sigmad(j)-omegad*i, sigmad(j)+omegad*i]); T=tf(N,D); N=[1]; D=poly([sigmad(j)-omegad(j)*i, sigmad(j)+omegad(j)*i]); T=tf(N,D);
step(T/dcgain(T),t); hold on; step(T/dcgain(T),t); hold on;
axis([0,5,0,2]); axis([0,5,0,2]); zeta=D(2)/(2*sqrt(D(3))); OS=exp(-(zeta*pi)/sqrt(1-zeta^2))*100,
end end

- Ex.4.6 Finding 𝑇𝑝 , %𝑂𝑆, and 𝑇𝑠 from Pole Location - Ex.4.7 Transient Response Through Component Design
Given the pole plot, find ζ; 𝜔𝑛 ; 𝑇𝑝 ; %𝑂𝑆, and 𝑇𝑠 𝑇(𝑡)𝜃(𝑡) Find 𝐽 and 𝐷 to yield 20% overshoot
𝑗𝜔 𝐽
Solution 𝐷
and a settling time of 2𝑠 for a step input
𝑗7 = 𝑗𝜔 𝐾 = 5𝑁𝑚/𝑟𝑎𝑑
𝑑
The damping ratio of torque 𝑇 𝑡

𝜔 𝑛
Solution 1
−1
cos ζ = 𝜃 ζ = cos𝜃 = cos tan−1 7/3 = 0.394
𝜎 𝐽
−𝜎𝑑 = −3
The natural frequency The rotational mechanical system 𝐺 𝑠 =
𝐷 𝐾
𝑠2 + 𝑠 +
−𝑗7 = −𝑗𝜔𝑑 𝜔𝑛 = 72 + 32 = 7.616𝑟𝑎𝑑/𝑠 𝐽 𝐽
−ln %𝑂𝑆/100
The peak time Requirement %𝑂𝑆 = 20 → ζ = = 0.456
𝑇𝑝 = 𝜋/𝜔𝑑 = 𝜋/7 = 0.449𝑠 𝜋 2 + ln2 %𝑂𝑆/100
𝑇𝑠 = 4/ζ𝜔𝑛 = 2 → 𝜔𝑛 = 2/0.456 = 4.386
The percent overshot
−ζ𝜋/ 1−ζ 2 Observation 𝜔𝑛 = 𝐾/𝐽 = 4.386
%𝑂𝑆 = 𝑒 × 100 = 26%
The approximate settling time 2ζ𝜔𝑛 = 𝐷/𝐽 = 2 × 0.456 × 4.386 = 38.474
𝑇𝑠 = 4/𝜎𝑑 = 4/3 = 1.333𝑠 Therefore 𝐷 = 1.04𝑁𝑚𝑠/𝑟𝑎𝑑, 𝐽 = 0.26𝑘𝑔𝑚2

5.Second-Order Transfer Functions via Testing Skill-Assessment Ex.4.5
361
- Consider a simple second-order system with the step input Problem Find ζ, 𝜔𝑛 , 𝑇𝑠 , 𝑇𝑝 , 𝑇𝑟 , %𝑂𝑆 for 𝐺 𝑠 =
𝑠 2 + 16𝑠 + 361
1 𝐺(𝑠) Solution 𝜔𝑛 = 361 = 19𝑟𝑎𝑑/𝑠
𝑅 𝑠 = 𝜔𝑛2 𝐶(𝑠)
𝑠 16 16
𝑠2 + 2ζ𝜔𝑛 𝑠 + 𝜔𝑛2
2ζ𝜔𝑛 = 16 → ζ = = = 0.421
2𝜔𝑛 2 × 19
From testing data in laboratory 4 4
• The response curve for percent overshoot, %𝑂𝑆 , and 𝑇𝑠 = = = 0.5𝑠
ζ𝜔𝑛 8
settling time, 𝑇𝑠 𝜋 𝜋
𝑇𝑝 = = = 0.182𝑠
%𝑂𝑆, 𝑇𝑠 → ζ, 𝜔𝑛 𝜔𝑛 1 − ζ 2 19 1 − 0.4212
• The expected steady-state values, 𝑐𝑓𝑖𝑛𝑎𝑙 1.4998 1.4998
𝜔𝑛 𝑇𝑟 = 1.4998 → 𝑇𝑟 = = = 0.079𝑠
𝑐𝑓𝑖𝑛𝑎𝑙 →𝐾 𝜔𝑛 19
−ζ𝜋
- A problem at the end of the chapter illustrates the estimation −0.421𝜋
1−ζ2
of a second-order transfer function from the step response %𝑂𝑆 = 𝑒 × 100% = 𝑒 1−0.4212 × 100% = 23.3%
𝜔𝑛 𝑇𝑟 ≈ 1.76ζ 3 − 0.417ζ 2 + 1.039ζ + 1
8
25/02/2023
§6.Underdamped Second Order Systems §7.Systems Response with Additional Poles

Matlab - The formulas describing percent overshoot, settling time, and
numg=361; peak time were derived only for a system with two complex
deng=[1 16 361];
omegan=sqrt(deng(3)/deng(1))
poles and no zeros
zeta=(deng(2)/deng(1))/(2*omegan) → for a system with more than two poles or with zeros?
Ts=4/(zeta*omegan)
Tp=pi/(omegan*sqrt(1-zeta^2)) - Before apply the derived formulas, a system with more than two
pos=100*exp(-zeta*pi/sqrt(1-zeta^2)) poles or with zeros must be approximated as a 2nd-order
Tr=(1.768*zeta^3 -0.417*zeta^2+1.039*zeta+1)/omegan system that has just two complex dominant poles
Result
ζ = 0.421, 𝜔𝑛 = 19, 𝑇𝑠 = 0.5𝑠, 𝑇𝑝 = 0.182𝑠, 𝑇𝑟 = 0.079𝑠,
%𝑂𝑆 = 23.3%
§7.Systems Response with Additional Poles §7.Systems Response with Additional Poles
Case 0: Case I: Case II: Case III:
- The conditions to exist in order to approximate the behavior of Original 2nd-order system Nondominant pole is near Nondominant pole is far Nondominant pole is at
a three-pole system as that of a two-pole system dominant 2nd-order pair dominant 2nd-order pair infinity
1 1 1 1
Consider a three-pole system with
𝑠 2 + 2𝑠 + 4 2
𝑠 + 2𝑠 + 4 𝑠 + 2 𝑠 2 + 2𝑠 + 4 𝑠 + 5 𝑠 2 + 2𝑠 + 4 𝑠 + 500
• two complex poles, −ζ𝜔𝑛 ± 𝑗𝜔𝑛 1 − and ζ 2, case 0 𝑝1 𝑗𝜔 case I 𝑝1 𝑗𝜔 case II 𝑝1 𝑗𝜔 case III 𝑝1 𝑗𝜔
• a third pole on the real axis, −𝛼𝑟 𝑝3 𝑝3 𝑝3

𝜎 𝜎 𝜎 𝜎
−ζ𝜔𝑛 −𝛼𝑟1 −ζ𝜔𝑛 −𝛼𝑟1 −ζ𝜔𝑛 ∞ −𝛼𝑟1 −ζ𝜔𝑛
The output transform
𝑠 − plane 𝑠 − plane 𝑠 − plane 𝑠 − plane
𝐴 𝐵 𝑠 + ζ𝜔𝑛 + 𝐶𝜔𝑑 𝐷 𝑝2 𝑝2 𝑝2 𝑝2
𝐶 𝑠 = + +
𝑠 𝑠 + ζ𝜔𝑛 2 + 𝜔𝑑2 𝑠 + 𝛼𝑟
In time domain
𝑐 𝑡 = 𝐴𝑢 𝑡 + 𝑒 −ζ𝜔𝑛 𝑡 𝐵cos𝜔𝑑 𝑡 + 𝐶sin𝜔𝑑 𝑡 + 𝐷𝑒 −𝛼𝑟 𝑡
den=1; num=[1 2 4];
TF0=tf(den,num); step(TF0/dcgain(TF0)); hold on
TF1=tf(den,conv(num,[1 1.5])); step(TF1/dcgain(TF1))
TF2=tf(den,conv(num,[1 5.0])); step(TF2/dcgain(TF2))
TF3=tf(den,conv(num,[1 500])); step(TF3/dcgain(TF3))
- Ex.4.8 Comparing Responses of Three-Pole Systems Analytical solutions
24.542
Find the step response of each of the TFs and compare them 𝑇1 𝑠 = 2
𝑠 + 4𝑠 + 24.542
24.542 10 3 24.542
𝑇1 𝑠 = 2 , 𝑇 𝑠 = 𝑇 𝑠 , 𝑇3 𝑠 = 𝑇 𝑠 =
𝑠 + 4𝑠 + 24.542 2 𝑠 + 10 1 𝑠+3 1 𝑠 − −2 − 𝑗4.532 𝑠 − −2 + 𝑗4.532
Solution → 𝑐1 𝑡 = 1 − 1.09𝑒 −2𝑡 cos 4.532𝑡 − 23.80
𝑐 (𝑡)
den=24.542; num=[1 4 24.542];
1
𝑐2 (𝑡) 245.42
𝑇2 𝑠 =
TF1=tf(den,num); step(TF1/dcgain(TF1)); hold on 𝑠 + 10 𝑠 2 + 4𝑠 + 24.542
TF2=tf(10*den,conv(num,[1 10])); 𝑐3 (𝑡)
step(TF2/dcgain(TF2))
245.42
=
TF3=tf(3*den,conv(num,[1 3])); [ 𝑠 − −10 ] 𝑠 − −2 − 𝑗4.532 𝑠 − −2 + 𝑗4.532
step(TF3/dcgain(TF3))
→ 𝑐2 𝑡 = 1 − 0.29𝑒 −10𝑡 − 1.189𝑒 −2𝑡 cos 4.532𝑡 − 53.340
𝑐1 𝑡 : the pure second-order system response 73.626
𝑐2 𝑡 : with the third pole at −10 and farthest from the dominant 𝑇3 𝑠 =
𝑠 + 3 𝑠 2 + 4𝑠 + 24.542
poles, is the better approximation of 𝑐1 𝑡 73.626
=
𝑐3 𝑡 : with the third pole at −3, close to the dominant poles, [𝑠 − −3 ] 𝑠 − −2 − 𝑗4.532 𝑠 − −2 + 𝑗4.532
yields the most error → 𝑐3 𝑡 = 1 − 1.14𝑒 −3𝑡 + 0.707𝑒 −2𝑡 cos 4.532𝑡 + 78.630
9
25/02/2023
Run ch4p2 in Appendix B System responses can alternately be obtained using
Learn how to use MATLAB to Simulink. Simulink is a software package that is
• generate a step response for a transfer function integrated with MATLAB to provide a graphical user
interface (GUI) for defining systems and generating
• plot the response directly or collect the points for responses
future use
The reader is encouraged to study Appendix C, which
• solve Ex.4.8 contains a tutorial on Simulink as well as some
examples. One of the illustrative examples, Ex.C.1,
solves Ex.4.8 using Simulink
Another method to obtain systems responses is Skill-Assessment Ex.4.6
through the use of MATLAB’s LTI Viewer. An advantage Problem Determine the validity of a second-order approximation
of the LTI Viewer is that it displays the values of settling for each of these two transfer functions
time, peak time, rise time, maximum response, and the 700
final value on the step response plot a. 𝐺1 𝑠 =
𝑠 + 15 𝑠 2 + 4𝑠 + 100
The reader is encouraged to study Appendix E, which 360
contains a tutorial on the LTI Viewer as well as some b. 𝐺2 𝑠 =
𝑠 + 4 𝑠 2 + 2𝑠 + 90
examples. Ex.E.1 solves Ex.4.8 using the LTI Viewer
Solution Matlab
700 a=15
a. 𝐺1 𝑠 = numga=100*a;
[𝑠 − −15 ] 𝑠 − −2 − 𝑗9.798 𝑠 − −2 + 𝑗9.798 denga=conv([1 a],[1 4 100]);
Ta=tf(numga,denga);
The dominant poles have a real part of −2 and the
numg=100;
higher-order pole is at −15, i.e. more than five-times deng=[1 4 100];
further T=tf (numg,deng);
step(Ta, '.',T,'- ')
→ The 2nd-order approximation is valid
700
b. 𝐺2 𝑠 = Result
[𝑠 − −4 ] 𝑠 − −1 − 𝑗9.434 𝑠 − −1 + 𝑗9.434
The dominant poles have a real part of −1 and the
higher-order pole is at −4, i.e. not more than five-
times further
→ The 2nd-order approximation is not valid
10
25/02/2023
§8.Systems Response with Zeros §8.Systems Response with Zeros

1 - This analysis can be reasoned via the partial-fraction expansion
- Given the 2nd-order system 𝐺 𝑠 =
𝑠 2 + 2𝑠 + 9 1
- Consecutively add zeros at −3, −5, and −10. The results, 𝑇 𝑠 = 𝑠+𝑎 ×
𝑠+𝑏 𝑠+𝑐
normalized to the steady-state value 𝑠+𝑎
=
• the closer the zero is to the 𝑠+𝑏 𝑠+𝑐
dominant poles, the greater its −𝑏 + 𝑎 −𝑐 + 𝑎
𝐴 𝐵
effect on the transient response = + = −𝑏 + 𝑐 + −𝑐 + 𝑏
• as the zero moves away from the 𝑠+𝑎 𝑠+𝑏 𝑠+𝑏 𝑠+𝑐
dominant poles, the response If the zero is far from the poles, then 𝑎 is large compared to 𝑏
approaches that of the two-pole and 𝑐, and
𝑎 𝑎
system 𝑎
den=1; num=[1 2 9]; 𝑇 𝑠 ≈= −𝑏 + 𝑐 + −𝑐 + 𝑏 =
TF0=tf(den,num); step(TF0/dcgain(TF0)); hold on 𝑠+𝑏 𝑠+𝑐 𝑠+𝑏 𝑠+𝑐
TF1=tf(conv(den,[1 03]),num); step(TF1/dcgain(TF1)) Hence, the zero looks like a simple gain factor and does not change
TF2=tf(conv(den,[1 05]),num); step(TF2/dcgain(TF2))
TF3=tf(conv(den,[1 10]),num); step(TF3/dcgain(TF3)) the relative amplitudes of the components of the response

- Another way to look at the effect of a zero, which is more - The response of a system with a zero
general, is as follows (Franklin, 1991) 𝑠 + 𝑎 𝐶 𝑠 = 𝑠𝐶 𝑠 + 𝑎𝐶 𝑠 (4.70)
Let 𝐶 𝑠 be the response of a system, 𝑇 𝑠 , with unity in the • If 𝑎 , the negative of the zero, is very large, the Laplace
numerator transform of the response is approximately 𝑎𝐶 𝑠 , or a scaled
𝑅 𝑠 𝐶(𝑠) 𝑅 𝑠 (𝑠 + 𝑎)𝐶(𝑠) version of the original response
𝑇(𝑠) (𝑠 + 𝑎)𝑇(𝑠)
• If 𝑎 is not very large, the response has an additional
If adding a zero to the TF, yielding 𝑠 + 𝑎 𝑇 𝑠 , the Laplace component consisting of the derivative of the original
transform of the response will be response, 𝑠𝐶 𝑠
𝑠 + 𝑎 𝐶 𝑠 = 𝑠𝐶 𝑠 + 𝑎𝐶 𝑠 (4.70) • As 𝑎 becomes smaller, the derivative term contributes more to
the response and has a greater effect
The response of a system with a zero consists of two parts
For step responses, the derivative is typically positive at the
• the derivative of the original response, and
start of a step response. Thus, for small values of 𝑎, we can
• a scaled version of the original response expect more overshoot in second order systems because the
derivative term will be additive around the first overshoot

- An interesting phenomenon occurs if 𝑎 is negative, placing the Matlab
zero in the right half-plane deng=[1 2 9];
Ta=tf([1 3]*9/3,deng); Tb=tf([1 5]*9/5,deng);
Tc=tf([1 10]*9/10,deng); T=tf(9,deng);
From (4.70): if the derivative term, 𝑠𝐶 𝑠 , step(T,Ta,Tb,Tc)
is larger than the scaled response, text(0.5,0.6, 'no zero')
𝑎𝐶 𝑠 , the response will initially follow text(0.4,0.7, 'zero at -10')
text(0.35,0.8, 'zero at -5')
the derivative in the opposite direction text(0.3,0.9, 'zero at -3')
from the scaled response
Result
Note: the response begins to turn toward the negative direction
even though the final value is positive. A system that exhibits
this phenomenon is known as a nonminimum-phase system
𝑠 + 𝑎 𝐶 𝑠 = 𝑠𝐶 𝑠 + 𝑎𝐶 𝑠 (4.70)
11
25/02/2023

- Ex.4.9 𝑅1 TF of a Nonminimum-Phase System 𝑅1 a.𝑉1 𝑠 = 𝐼 𝑠 𝑅1 + 𝑉𝑜 𝑠
𝑅2 𝑅2 𝑉 (𝑠)
𝑉𝑜 (𝑠) a.Find the TF, 𝑉 𝑠 /𝑉 𝑠 , for the
𝑜 𝑖
1 𝑉𝑜 (𝑠) 𝑉𝑖 𝑠 − 𝑉𝑜 𝑠
𝑉𝑖 (𝑠) 𝑉𝑖 (𝑠) 𝐼(𝑠) 𝐼 𝑠 =
operational amplifier circuit 𝑉2 (𝑠)
𝑅1 + 𝑅2
𝑅3 𝐶 𝑅3 𝐶
b.If 𝑅1 = 𝑅2 , 𝑅3 𝐶 = 0.1, find the step 𝑅1 𝑉𝑖 𝑠 + 𝑅2 𝑉𝑜 𝑠
→ 𝑉1 𝑠 =
response of the filter. Show that component parts of the 𝑅1 + 𝑅2
response can be identified with those in Eq. (4.70) 𝑉2 𝑠 𝑉2 𝑠 − 𝑉𝑖 𝑠 1/𝐶𝑠
+ = 0 → 𝑉2 𝑠 = 𝑉𝑖 𝑠
Solution 𝑅1 a.𝑉1 𝑠 = 𝐼 𝑠 𝑅1 + 𝑉𝑜 𝑠 1/𝐶𝑠 𝑅3 𝑅3 + 1/𝐶𝑠
𝑅2 𝑉 (𝑠)
1 𝑉𝑜 (𝑠) 𝑉𝑖 𝑠 − 𝑉𝑜 𝑠 The amplifier’s gain 𝐴: 𝑉𝑜 𝑠 = 𝐴[𝑉2 𝑠 − 𝑉1 𝑠 ]
𝐼(𝑠) 𝐼 𝑠 =
𝑉𝑖 (𝑠) 𝑅1 + 𝑅2 𝑉𝑜 𝑠 𝐴 𝑅2 − 𝑅1𝑅3 𝐶𝑠
𝑅3
𝑉2 (𝑠)
𝐶 𝑅1 𝑉𝑖 𝑠 + 𝑅2 𝑉𝑜 𝑠 → =
𝑉𝑖 𝑠 𝑅3 𝐶𝑠 + 1 [𝑅1 + 𝑅2 1 + 𝐴 ]
→ 𝑉1 𝑠 =
𝑅1 + 𝑅2 𝑅2
𝑉𝑜 𝑠 𝑅2 − 𝑅1 𝑅3𝐶𝑠 𝑅1 𝑠 −
𝑉2 𝑠 𝑉2 𝑠 − 𝑉𝑖 𝑠 1/𝐶𝑠 𝑅1 𝑅3 𝐶
+ = 0 → 𝑉2 𝑠 = 𝑉𝑖 𝑠 → = =−
1/𝐶𝑠 𝑅3 𝑅3 + 1/𝐶𝑠 𝑉𝑖 𝑠 𝑅2 𝑅3 𝐶𝑠 + 𝑅2 1
𝑅2 𝑠 +
𝑅3 𝐶

b.Letting 𝑅1 = 𝑅2 and 𝑅3 𝐶 = 1/10 1 1 1 1 1
𝐶𝑜 𝑠 = − =− +
𝑅2 1 𝑠 𝑠 + 10 10 𝑠 10 𝑠 + 10
𝑉𝑜 𝑠 𝑅1 𝑠 − 𝑠− 𝑠 − 10
𝑅1𝑅3 𝐶 𝑅3 𝐶 the response without a zero
=− =− =−
𝑉𝑖 𝑠 1 1 𝑠 + 10 1 1
𝑅2 𝑠 + 𝑠+ 𝑐𝑜 𝑡 = − + 𝑒 −10𝑡
𝑅3 𝐶 𝑅3 𝐶
10 10
For a step input, evaluate the response as suggested by (4.70) Step response of the nonminimum-phase network 𝑐 𝑡 and
𝑠 − 10 1 1 normalized step response of an equivalent network without
𝐶 𝑠 =− =− + 10
𝑠 𝑠 + 10 𝑠 + 10 𝑠 𝑠 + 10 the zero −10𝐶𝑜 𝑡
= 𝑠𝐶𝑜 𝑠 − 10𝐶𝑜 𝑠
clear all
where, 𝐶𝑜 𝑠 = −1/[𝑠 𝑠 + 10 ] is the Laplace transform of the t=0:0.001:0.5;
response without a zero c=1-2*exp(-10*t);
1 1 1 1 2 c0=-1/10+(1/10)*exp(-10*t);
𝐶 𝑠 =− + − = − plot(t,c); hold on; plot(t,-10*c0)
𝑠 + 10 𝑠 𝑠 + 10 𝑠 𝑠 + 10 text(0.1,0.4,'c’);
or the response with a zero 𝑐 𝑡 = 1 − 2𝑒 −10𝑡 text(0.1,0.8,'-10co');

- The pole-zero cancellation and its effect on our ability to make - Ex.4.10 Evaluating Pole-Zero Cancellation Using Residues
2nd-order approximations to a system For each of the following response functions, determine
• Assume a three pole system with a zero as shown in (4.85) whether there is cancellation between the zero and the pole
• If the pole term, 𝑠 + 𝑝3 , and the zero term, 𝑠 + 𝑧 , cancel closest to the zero. For any function for which pole-zero
out, we are left with cancellation is valid, find the approximate response
𝐾 𝑠+𝑧 𝐾 𝑠+𝑧 26.25 𝑠 + 4
𝑇 𝑠 = = (4.85) 𝐶1 𝑠 =
𝑠 + 𝑝3 𝑠 2 + 𝑎𝑠 + 𝑏 𝑠 + 𝑝3 𝑠 2 + 𝑎𝑠 + 𝑏 𝑠 𝑠 + 3.5 𝑠 + 5 𝑠 + 6
26.25 𝑠 + 4
as a second-order TF 𝐶2 𝑠 =
𝑠 𝑠 + 4.01 𝑠 + 5 𝑠 + 6
• From another perspective, if the zero at −𝑧 is very close to the Solution
1 1 1 1
pole at −𝑝3 , then a partial-fraction expansion of Eq. (4.85) will 𝐶1 𝑠 = + −3.5 × + 3.5 × + −1 ×
show that the residue of the exponential decay is much 𝑠 𝑠+5 𝑠+6 𝑠 + 3.5
smaller than the amplitude of the 2nd-order response The residue of the pole at −3.5, which is closest to the zero at
− 4, is equal to 1 and is not negligible compared to the other
residues. Thus, a second-order step response approximation
cannot be made for 𝐶1 𝑠
12
25/02/2023

1 1 Matlab
𝐶2 𝑠 = 0.87 + −5.3 ×
𝑠 𝑠+5 syms s
1 1 C1=26.25*(s+4)/(s*(s+3.5)*(s+5)*(s+6));
+4.4 × + 0.033 × C2=26.25*(s+4)/(s*(s+4.01)*(s+5)*(s+6));
𝑠+6 𝑠 + 4.01 c1=ilaplace(C1);
The residue of the pole at −4.01, which is closest to the zero at c1=vpa(c1,3); 'c1', pretty (c1)
c2=ilaplace(C2);
− 4, is equal to 0.033, about two orders of magnitude below any c2=vpa(c2,3); 'c2', pretty(c2)
of the other residues
Result
Hence, we make a 2nd-order approximation by neglecting the c1
response generated by the pole at −4.01 exp(-6.0 t) 3.5 - exp(-3.5 t) 1.0 - exp(-5.0 t) 3.5 + 1.0
c2
1 5.3 4.4 exp(-6.0 t) 4.4 - exp(-5.0 t) 5.3 + exp(-4.01 t) 0.0332 + 0.873
𝐶2 𝑠 ≈ 0.87 − +
𝑠 𝑠+5 𝑠+6
and the response 𝑐2 𝑡 is approximately
𝑐2 𝑡 ≈ 0.87 − 5.3𝑒 −5𝑡 + 4.4𝑒 −6𝑡

Skill-Assessment Ex.4.7 Solution a. Expanding 𝐶 𝑠 by partial fractions
Problem Determine the validity of a 2nd-order approximation for 1 0.8942 −1.5918 −0.3023
𝐶 𝑠 = + + +
each of these two TFs 𝑠 𝑠 + 20 𝑠 + 10 𝑠 + 6.5
185.71 𝑠 + 7 −0.3023 is not an order of magnitude less than
a. 𝐺 𝑠 = nd
𝑠 + 6.5 𝑠 + 10 𝑠 + 20 residues of 2 -order terms
197.14 𝑠 + 7 → a 2nd-order approximation is not valid
b. 𝐺 𝑠 =
𝑠 + 6.9 𝑠 + 10 𝑠 + 20 b. Expanding 𝐶 𝑠 by partial fractions
1 0.9782 −1.9078 −0.0704
𝐶 𝑠 = + + +
𝑠 𝑠 + 20 𝑠 + 10 𝑠 + 6.9
−0.0704 is an order of magnitude less than
nd
residues of 2 -order terms
→ a 2nd-order approximation is valid
§9.Effects of Nonlinearities upon Time Response §9.Effects of Nonlinearities upon Time Response
- Qualitatively examine the effects of nonlinearities upon the time Effect of amplifier saturation on load angular velocity response
response of physical systems
In the following examples, we insert nonlinearities, such as
saturation, dead zone, and backlash into a system to show the
effects of these nonlinearities upon the linear responses. The
responses were obtained using Simulink
- Assume the motor and load from the Antenna Control and look
at the load angular velocity
0.2083
𝜔𝑜 𝑠 = 0.1𝑠𝜃𝑚 𝑠 = 𝐸𝑎 𝑠
𝑠 + 1.71 The effect of amplifier saturation
is to limit the obtained velocity
13
25/02/2023
§9.Effects of Nonlinearities upon Time Response §9.Effects of Nonlinearities upon Time Response
Effect of dead zone on load angular displacement response Effect of backlash on load angular displacement response
As the motor reverses direction,

the output shaft remains
The effect of dead zone is lower stationary while the motor
the output amplitude begins to reverse. When the
gears finally connect, the output
shaft itself begins to turn in the
reverse direction
§9.Effects of Nonlinearities upon Time Response §10.Laplace Transform Solution of State Equations
Skill-Assessment Ex.4.8 - Consider the state and output equations
Problem Use MATLAB’s Simulink to reproduce the following 𝒙ሶ = 𝑨𝒙 + 𝑩𝒖 (4.92)
figure 𝒚 = 𝑪𝒙 + 𝑫𝒖 (4.93)
- Taking the Laplace transform of the state equation
𝑠𝑿 𝑠 − 𝒙 0 = 𝑨𝑿 𝑠 + 𝑩𝑼 𝑠 (4.94)
→ 𝑠𝑰 − 𝑨 𝑿 𝑠 = 𝒙 0 + 𝑩𝑼 𝑠
→ 𝑿 𝑠 = 𝑠𝑰 − 𝑨 −1 𝒙 0 + 𝑠𝑰 − 𝑨 −1 𝑩𝑼 𝑠
adj 𝑠𝑰 − 𝑨
= [𝒙 0 + 𝑩𝑼 𝑠 ] (4.96)
det 𝑠𝑰 − 𝑨
- Taking the Laplace transform of the output equation
𝒀 𝑠 = 𝑪𝑿 𝑠 + 𝑫𝑼 𝑠 (4.97)
§10.Laplace Transform Solution of State Equations §10.Laplace Transform Solution of State Equations
Eigenvalues and Transfer Function Poles - Ex.4.11 Laplace Transform Solution; Eigenvalues and Poles
- The poles of the TF determine the nature of the transient Given the system
response of the system → Is there an equivalent quantity in 0 1 0 0
the state-space representation that yields the same information? 𝒙ሶ = 0 0 1 𝒙 + 0 𝑒 −𝑡
- Consider the SISO system with the output, 𝑌 𝑠 , the input, −24 −26 −9 1
𝑈 𝑠 and assumed that 𝑥 0 = 0 𝑦= 1 1 0𝒙
𝑌𝑠 adj 𝑠𝑰 − 𝑨 𝑪adj 𝑠𝑰 − 𝑨 𝑩 + 𝑫det 𝑠𝑰 − 𝑨 1
=𝑪 𝑩+𝑫= (4.98) 𝒙 0 = 0
𝑈𝑠 det 𝑠𝑰 − 𝑨 det 𝑠𝑰 − 𝑨 2
The roots of the denominator of Eq. (4.98) are the poles of the do the following
system. Since the denominators of Eqs. (4.96) and (4.98) are a.Solve the preceding state equation and obtain the output for
identical, the system poles equal the eigenvalues. Hence, if a the given exponential input
system is represented in state-space, the poles can be found
b.Find the eigenvalues and the system poles
from
det 𝑠𝑰 − 𝑨 = 0
14
25/02/2023
Solution Find 𝑿 𝑠
a.Find 𝑠𝑰 − 𝑨 −1 𝑿 𝑠 = 𝑠𝑰 − 𝑨 −1 [𝒙 0 + 𝑩𝑼 𝑠 ]
𝑠 0 0 0 1 0 𝑠 −1 0 𝑠 2 + 9𝑠 + 26 𝑠+9 1
𝑠𝑰 − 𝑨 = 0 𝑠 0 − 0 0 1 = 0 𝑠 −1 −24 𝑠 2 + 9𝑠 𝑠
0 0 𝑠 −24 −26 −9 24 26 𝑠 + 9 −24𝑠 − 26𝑠 + 24 𝑠 2 1 0 1
adj 𝑠𝑰 − 𝑨 = 0 + 0
𝑠𝑰 − 𝑨 −1 = 𝑠 3 + 9𝑠 2 + 26𝑠 + 24 𝑠+1
det 𝑠𝑰 − 𝑨 2 1
𝑠 2 + 9𝑠 + 26 𝑠+9 1 𝑠 2 + 9𝑠 + 26 𝑠+9 1 1
−24 𝑠 2 + 9𝑠 𝑠 −24 𝑠 2 + 9𝑠 𝑠 0
−24𝑠 − 26𝑠 + 24 𝑠 2 2𝑠 + 3
−24𝑠 − 26𝑠 + 24 𝑠 2 𝑠 + 1
= =
𝑠 3 + 9𝑠 2 + 26𝑠 + 24 𝑠+2 𝑠+3 𝑠+4
Find 𝑼 𝑠
1 1 𝑠 3 + 10𝑠 2 + 37𝑠 + 29
𝑼 𝑠 = ℒ 𝑒 −𝑡 = = 2𝑠 2 − 21𝑠 − 24
𝑠+1 𝑠+1 𝑠+2 𝑠+3 𝑠+4
𝑠 2𝑠 2 − 21𝑠 − 24
The output equation Run ch4sp1 in Appendix F
𝑋1 𝑠 Learn how to use the Symbolic Math Toolbox to
𝑌 𝑠 = 1 1 0 𝑋2 𝑠 • solve state equations for the output response using
𝑋3 𝑠 the Laplace transform
= 𝑋1 𝑠 + 𝑋2 𝑠 • solve Ex.4.11
𝑠 3 + 12𝑠 2 + 16𝑠 + 5
=
𝑠+1 𝑠+2 𝑠+3 𝑠+4
−6.5 19 11.5
= + −
𝑠+2 𝑠+3 𝑠+4
→ 𝑦 𝑡 = −6.5𝑒 −2𝑡 + 19𝑒 −3𝑡 − 11.5𝑒 −4𝑡
b.The denominator of the system’s transfer function det 𝑠𝑰 −
𝑨 = 0 furnishes both the poles of the system and the
eigenvalues −2; −3, and −4
Skill-Assessment Ex.4.9 1
𝑼 𝑠 = ℒ 𝑒 −𝑡 =
Problem Given the system represented in state space 𝑠+1
0 2 0 −𝑡 The state vector
𝒙ሶ = 𝒙+ 𝑒 ,𝑦 = 1 3 𝒙 𝑿 𝑠 = 𝑠𝑰 − 𝑨 −1 [𝒙 0 + 𝑩𝑼 𝑠 ]
−3 −5 1
2 𝑠+5 2
𝒙 0 =
1 −3 𝑠 0 0 1
= 2 +
do the following 𝑠 + 5𝑠 + 6 1 1 𝑠+1
a.Solve for 𝑦 𝑡 using state-space equation and 𝑠+5 2
0
Laplace transform techniques −3 𝑠 𝑠+2
=
𝑠+2 𝑠+3
b.Find the eigenvalues and the system poles 𝑠+1
𝑠 0 0 2 𝑠 −2 1 2 𝑠 2 + 7𝑠 + 7
Solution a. 𝑠𝑰 − 𝑨 = − = =
0 𝑠 −3 −5 3 𝑠+5 𝑠+1 𝑠+2 𝑠+3 𝑠 2 − 4𝑠 − 6
𝑠+5 2
adj 𝑠𝑰 − 𝑨 −3 𝑠
𝑠𝑰 − 𝑨 −1 = = 2
det 𝑠𝑰 − 𝑨 𝑠 + 5𝑠 + 6
15
25/02/2023
The output Matlab
𝑋 𝑠 syms s
𝑌 𝑠 = 1 3 1 A= [0 2;-3 -5];B=[0;1];C= [1 3];X0=[2;1];
𝑋2 𝑠 U=1/(s+1);I= [1 0;0 1];
5𝑠 2 + 2𝑠 − 4 X=((s*I-A)^-1)*(X0+B*U);
= Y=C*X;Y=simplify(Y);
𝑠+1 𝑠+2 𝑠+3 y=ilaplace(Y);pretty(y)
0.5 12 17.5 eig(A)
=− − + Result
𝑠+1 𝑠+2 𝑠+3
→ 𝑦 𝑡 = −0.5𝑒 −𝑡 − 12𝑒 −2𝑡 + 17.5𝑒 −3𝑡 exp(-3 t) 35 exp(-t)
--------------- - exp(-2 t) 12 - --------
b. 2 2
The eigenvalues are given by the roots of det 𝑠𝑰 − ans =
-2.0000
𝑨 = 𝑠 2 + 5𝑠 + 6 = 0 , or −2 and −3 -3.0000
§11. Time Domain Solution of State Equations §11. Time Domain Solution of State Equations
- The solution in the time domain is given directly by - Ex.4.12 Time Domain Solution
𝑡
Given the state equation and initial state vector
𝒙 𝑡 = 𝑒 𝑨𝑡 𝒙 0 + න 𝑒 𝑨 𝑡−𝜏 𝑩𝒖 𝜏 𝑑𝜏
0 1 0
0 𝒙ሶ = 𝒙+ 𝑢 𝑡
𝑡 −8 −6 1
=𝚽 𝑡 𝒙 0 + න 𝚽 𝑡 − 𝜏 𝑩𝒖 𝜏 𝑑𝜏 1
0 𝒙 0 =
0
where 𝑢 𝑡 is a unit step, find the state-transition matrix and
𝚽 𝑡 : the state-transition matrix, 𝚽 𝑡 = 𝑒 𝑨𝑡 then solve for 𝑥 𝑡
- For the unforced system Solution
ℒ 𝒙 𝑡 = ℒ 𝚽 𝑡 𝒙 0 = 𝑠𝑰 − 𝑨 −1 𝒙 0 The eigenvalues
→ ℒ 𝚽 𝑡 = 𝑠𝑰 − 𝑨 −1 det 𝑠𝑰 − 𝑨 = 0
→ 𝚽 𝑡 = ℒ −1 𝑠𝑰 − 𝑨 −1 𝑠 0 0 1 𝑠 −1
det − = det = 𝑠2 + 6𝑠 + 8 = 0
adj 𝑠𝑰 − 𝑨 0 𝑠 −8 −6 8 𝑠+6
= ℒ −1 or 𝑠1 = −2 and 𝑠2 = −4
Since each term of the state-transition matrix is the sum of Therefore
responses generated by the poles (eigenvalues), we assume a −2𝑡 −4𝑡 −2𝑡 −4𝑡
𝚽 𝑡 = 2𝑒 −2𝑡− 4𝑒 −4𝑡 0.5𝑒 −2𝑡 − 0.5𝑒
state-transition matrix of the form −4𝑒 + 4𝑒 −𝑒 + 2𝑒 −4𝑡
𝐾 𝑒 −2𝑡 + 𝐾2 𝑒 −4𝑡 𝐾3 𝑒 −2𝑡 + 𝐾4 𝑒 −4𝑡 −2 𝑡−𝜏 −4 𝑡−𝜏 −2 𝑡−𝜏 −4 𝑡−𝜏
𝚽 𝑡 = 1 −2𝑡 → 𝚽 𝑡 − 𝜏 𝑩 = 2𝑒 −2 𝑡−𝜏 − 4𝑒 −4 𝑡−𝜏 0.5𝑒 −2 𝑡−𝜏 − 0.5𝑒
𝐾5 𝑒 + 𝐾6 𝑒 −4𝑡 𝐾7 𝑒 −2𝑡 + 𝐾8 𝑒 −4𝑡 −4𝑒 + 4𝑒 −𝑒 + 2𝑒−4 𝑡−𝜏
In order to find the values of the constants, we make use of the 0
×
properties of the state-transition matrix 1
−2 𝑡−𝜏 −4 𝑡−𝜏
𝚽 0 = 𝑰 → 𝐾1 + 𝐾2 = 1 𝐾3 + 𝐾4 = 0 = 0.5𝑒 −2 𝑡−𝜏 − 0.5𝑒
−𝑒 + 2𝑒 −4 𝑡−𝜏
𝐾5 + 𝐾6 = 0 𝐾7 + 𝐾8 = 1 −2𝑡 −4𝑡 −2𝑡 −4𝑡 1
→ 𝚽 𝑡 𝒙 0 = 2𝑒 −2𝑡− 4𝑒 −4𝑡 0.5𝑒 −2𝑡 − 0.5𝑒
𝚽ሶ 0 = 𝑨 → −2𝐾1 − 4𝐾2 = 0 −2𝐾3 − 4𝐾4 = 1 −4𝑒 + 4𝑒 −𝑒 + 2𝑒 −4𝑡 0
−2𝑡 −4𝑡
−2𝐾5 − 4𝐾6 = −8 −2𝐾7 − 4𝐾8 = −6 = 2𝑒 −2𝑡 − 𝑒 −4𝑡
→ 𝐾1 = 2, 𝐾2 = −1, 𝐾3 = 1/2, 𝐾4 = −1/2 −4𝑒 + 4𝑒
𝐾5 = −4, 𝐾6 = 4, 𝐾7 = −1, 𝐾8 = 2
16
25/02/2023
𝑡 −2 𝑡−𝜏 𝑡
−4 𝑡−𝜏 The final result
→ න 𝚽 𝑡 − 𝜏 𝑩𝒖 𝜏 𝑑𝜏 = න 0.5𝑒 −2 𝑡−𝜏 − 0.5𝑒 𝑑𝜏 𝑡
0 0 −𝑒 + 2𝑒 −4 𝑡−𝜏
𝑡 𝑡 𝒙 𝑡 = 𝚽 𝑡 𝒙 0 + න 𝚽 𝑡 − 𝜏 𝑩𝒖 𝜏 𝑑𝜏
−2𝑡 2𝜏 −4𝑡 4𝜏 0
0.5𝑒 න 𝑒 𝑑𝜏 − 0.5𝑒 න 𝑒 𝑑𝜏
0 0
1 1 −2𝑡 1 −4𝑡
= − 𝑒 + 𝑒
𝑡 𝑡 2𝑒 −2𝑡 − 𝑒 −4𝑡 + 8 4
= 8
−2𝑡 2𝜏 −4𝑡 4𝜏
−𝑒 න 𝑒 𝑑𝜏 + 2𝑒 න 𝑒 𝑑𝜏 −4𝑒 −2𝑡 + 4𝑒 −4𝑡 1 −2𝑡 1 −4𝑡
0 0 𝑒 − 𝑒
1 1 −2𝑡 1 −4𝑡 2 2
− 𝑒 + 𝑒 1 7 −2𝑡 7 −4𝑡
+ 𝑒 − 𝑒
= 8 4 8
1 −2𝑡 1 −4𝑡 →𝒙 𝑡 = 8 4 8
𝑒 − 𝑒 7 7
2 2 − 𝑒 −2𝑡 + 𝑒 −4𝑡
2 2
- Ex.4.13 State-Transition Matrix via Laplace Transform Expanding each term in the matrix on the right by partial
Find the state-transition matrix of Ex.4.12, using 𝑠𝑰 − 𝑨 −1 fractions yields
2 1 1/2 1/2
Solution − −
𝑠 0 0 1 𝑠 −1 𝑠𝑰 − 𝑨 =−1 𝑠 + 2 𝑠 + 4 𝑠 + 2 𝑠 +4
𝑠𝑰 − 𝑨 = − = −4 4 −1 2
0 𝑠 −8 −6 8 𝑠+6 + +
adj 𝑠𝑰 − 𝑨 𝑠+2 𝑠+4 𝑠+2 𝑠+4
𝑠𝑰 − 𝑨 −1 = Finally, taking the inverse Laplace transform of each term
−2𝑡 −4𝑡 −2𝑡 −4𝑡
𝑠+6 1 𝚽 𝑡 = 2𝑒 −2𝑡− 4𝑒 −4𝑡 0.5𝑒 −2𝑡 − 0.5𝑒
−8 𝑠 −4𝑒 + 4𝑒 −𝑒 + 2𝑒 −4𝑡
= 2
𝑠 + 6𝑠 + 8
𝑠+6 1
2 2
= 𝑠 + 6𝑠 + 8 𝑠 + 6𝑠 + 8
−8 𝑠
𝑠 2 + 6𝑠 + 8 𝑠 2 + 6𝑠 + 8
Run ch4sp2 in Appendix F Run ch4p3 in Appendix B
Learn how to use the Symbolic Math Toolbox to Learn how to use MATLAB to
• solve state equations for the output response using • simulate the step response of systems represented
the convolution integral in state space
• solve Ex.4.12 and Ex.4.13 • specify the range on the time axis for the plot
17
25/02/2023
Skill-Assessment Ex.4.10 𝑠 0 0 2 𝑠 −2
Solution a. 𝑠𝑰 − 𝑨 = − =
Problem Given the system represented in state space 0 𝑠 −2 −5 2 𝑠+5
𝑠 +5 2 𝑠+5 2
0 2 0 −2𝑡 adj 𝑠𝑰 − 𝑨
𝒙ሶ = 𝒙+ 𝑒 −2 𝑠 −2 𝑠
−2 −5 1 𝑠𝑰 − 𝑨 −1 = = 2 =
det 𝑠𝑰 − 𝑨 𝑠 + 5𝑠 + 4 𝑠 + 1 𝑠 + 4
𝑦= 2 1𝒙
4/3 1/3 2/3 2/3
1 − −
𝒙 0 =
2 = 𝑠+1 𝑠+4 𝑠+1 𝑠+4
−2/3 2/3 −1/3 4/3
do the following + +
𝑠+1 𝑠+4 𝑠+1 𝑠+4
a.Solve for the state-transition matrix
Taking the Laplace transform of each term, the state
b.Solve for the state vector using the convolution transition matrix is given by
integral 4 −𝑡 1 −4𝑡 2 −𝑡 2 −4𝑡
c.Find the output, 𝑦 𝑡 𝑒 − 𝑒 𝑒 − 𝑒
𝚽 𝑡 = 3 3 3 3
2 −𝑡 2 −4𝑡 1 −𝑡 4 −4𝑡
− 𝑒 + 𝑒 − 𝑒 + 𝑒
3 3 3 3
4 − 𝑡−𝜏 1 𝑡
b.
𝑒 − 𝑒 −4 𝑡−𝜏 න 𝚽 𝑡 − 𝜏 𝑩𝒖 𝜏 𝑑𝜏
𝚽 𝑡−𝜏 = 3 3 0
2 2 2 −𝜏 −𝑡 2 2𝜏 −4𝑡
− 𝑒 − 𝑡−𝜏 + 𝑒 −4 𝑡−𝜏 𝑡𝑒 𝑒 − 𝑒 𝑒
3 3 3 3
2 − 𝑡−𝜏 2 =න 𝑑𝜏
𝑒 − 𝑒 −4 𝑡−𝜏 1 4
0 − 𝑒 −𝜏 𝑒 −𝑡 + 𝑒 2𝜏 𝑒 −4𝑡
3 3 3 3
1 4 2 −𝑡 𝑡 −𝜏 2 𝑡
− 𝑒 − 𝑡−𝜏 + 𝑒 −4 𝑡−𝜏 𝑒 න 𝑒 𝑑𝜏 − 𝑒 −4𝑡 න 𝑒 2𝜏 𝑑𝜏
3 3 3 3
0 0 0
𝑩𝒖 𝜏 = −2𝜏 = 𝑡 𝑡
𝑒 1 4
− 𝑒 −𝑡 න 𝑒 −𝜏 𝑑𝜏 + 𝑒 −4𝑡 න 𝑒 2𝜏 𝑑𝜏
2 −𝜏 −𝑡 2 2𝜏 −4𝑡 3 0 3 0
𝑒 𝑒 − 𝑒 𝑒 2 −𝑡 1
𝚽 𝑡 − 𝜏 𝑩𝒖 𝜏 = 3 3
𝑒 − 𝑒 −2𝑡 + 𝑒 −4𝑡
1 −𝜏 −𝑡 4 2𝜏 −4𝑡
− 𝑒 𝑒 + 𝑒 𝑒 = 3 3
3 3 1 −𝑡 −2𝑡
2
− 𝑒 +𝑒 − 𝑒 −4𝑡
3 3
§11. Time Domain Solution of State Equations §12.Performance Indices

𝑡
- Performance index: A quantitative measure of the performance
𝒙 𝑡 = 𝚽 𝑡 𝒙 0 + න 𝚽 𝑡 − 𝜏 𝑩𝒖 𝜏 𝑑𝜏
0
of a system and is chosen so that emphasis is given to the
4 −𝑡 1 −4𝑡 2 −𝑡 2 −4𝑡 important system specifications
𝑒 − 𝑒 𝑒 − 𝑒
1 - A system is considered an optimum control system when the
= 3 3 3 3
2 −𝑡 2 −4𝑡 1 −𝑡 4 −4𝑡 2 system parameters are adjusted so that the index reaches an
− 𝑒 + 𝑒 − 𝑒 + 𝑒
3 3 3 3 extremum value, commonly a minimum value
2 −𝑡 1
𝑒 − 𝑒 −2𝑡 + 𝑒 −4𝑡
+ 3 3
1 2
− 𝑒 −𝑡 + 𝑒 −2𝑡 − 𝑒 −4𝑡
3 3
10 −𝑡 4
𝑒 − 𝑒 −2𝑡 − 𝑒 −4𝑡
= 3 3
5 8
− 𝑒 −𝑡 + 𝑒 −2𝑡 + 𝑒 −4𝑡
3 3
c. 𝑦 𝑡 = 2 1 𝒙 = 5𝑒 −𝑡 − 𝑒 −2𝑡
18
25/02/2023
§12.Performance Indices §12.Performance Indices

1.The ISE Index 2.The IAE Index
𝑇
Integral of the Absolute magnitude of the Error
𝐼𝑆𝐸 = න 𝑒 2 𝑡 𝑑𝑡 𝑇
0 𝐼𝐴𝐸 = න |𝑒 𝑡 |𝑑𝑡
𝐼𝑆𝐸 : Integral of the Square of the Error 0
𝑇 : the upper limit 𝑇 is a finite time chosen somewhat 3.The ITAE Index
arbitrarily so that the integral approaches a steady- Integral of Time multiplied by the Absolute magnitude of the Error
𝑇
state value. It is usually convenient to choose 𝑇 as the
𝐼𝑇𝐴𝐸 = න 𝑡|𝑒 𝑡 |𝑑𝑡
settling time, 𝑇𝑠 0
𝑒 𝑡 : error 4.The ITSE Index
- The ISE performance index is easily adapted for practical Integral of Time multiplied by the Square of the Error
𝑇
measurements because a squaring circuit is readily obtained
𝐼𝑇𝑆𝐸 = න 𝑡𝑒 2 𝑡 𝑑𝑡
- Furthermore, the squared error is mathematically convenient 0
for analytical and computational purposes
§12.Performance Indices §13.Case Studies

5.The ISTAE Index
Integral of the Squared Time multiplied by the Absolute
magnitude of the Error
𝑇
𝐼𝑆𝑇𝐴𝐸 = න 𝑡 2 |𝑒 𝑡 |𝑑𝑡
0
6.The ISTSE Index
Integral of the Squared Time multiplied by the Square of the Error
𝑇
𝐼𝑆𝑇𝑆𝐸 = න 𝑡 2 𝑒 2 𝑡 𝑑𝑡
0
- The performance indices ITSE, ISTAE and ISTSE have not

been applied to any great extent in practice because of the
increased difficulty in handling them
19

Ch.04 Time Response

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ch.04 Time Response

Uploaded by

Copyright:

Available Formats

25/02/2023

§1.Introduction §2.Poles, Zeros, and System Response

output transform 𝐶(𝑠) =

§3.First Order Systems §3.First Order Systems 𝑐(𝑡)

§3.First Order Systems 𝑐(𝑡)

§3.First Order Systems 𝑐(𝑡)

§3.First Order Systems §3.First Order Systems

§4.Second Order Systems: Introduction §4.Second Order Systems: Introduction

[r,p,k]=residue([9],[1,9,9,0]), 𝐶 𝑠 = 1/𝑠 → 𝑐 𝑡 = 1, 𝐶 𝑠 = 1/ 𝑠 + 𝑎 → 𝑐 𝑡 = 𝑒 −𝑎𝑡

§4.Second Order Systems: Introduction §4.Second Order Systems: Introduction

• Underdamped 𝑐(𝑡) 𝑐(𝑡)

Natural response: 𝑐𝑛 𝑡 = 𝐴𝑒 −𝜎𝑑 𝑡 cos 𝜔𝑑 𝑡 − 𝜙

§4.Second Order Systems: Introduction §4.Second Order Systems: Introduction

• Undamped 𝑐(𝑡) • Critically damped

§4.Second Order Systems: Introduction §4.Second Order Systems: Introduction

§4.Second Order Systems: Introduction §4.Second Order Systems: Introduction

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

𝑒 −ζ𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 1 − ζ 2 𝑡 − 𝜙 𝑒 −ζ𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 1 − ζ 2 𝑡 − 𝜙

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

𝜔𝑑 : damped frequency of oscillation

𝜎𝑑 : exponential damping frequency

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

pole motion pole motion

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

The damping ratio of torque 𝑇 𝑡

§6.Underdamped Second Order Systems §6.Underdamped Second Order Systems

§6.Underdamped Second Order Systems §7.Systems Response with Additional Poles

• a third pole on the real axis, −𝛼𝑟 𝑝3 𝑝3 𝑝3

§8.Systems Response with Zeros §8.Systems Response with Zeros

§8.Systems Response with Zeros §8.Systems Response with Zeros

§8.Systems Response with Zeros §8.Systems Response with Zeros

§8.Systems Response with Zeros §8.Systems Response with Zeros

§8.Systems Response with Zeros §8.Systems Response with Zeros

§8.Systems Response with Zeros §8.Systems Response with Zeros

§8.Systems Response with Zeros §8.Systems Response with Zeros

§8.Systems Response with Zeros §8.Systems Response with Zeros

As the motor reverses direction,

§11. Time Domain Solution of State Equations §12.Performance Indices

§12.Performance Indices §12.Performance Indices

§12.Performance Indices §13.Case Studies

- The performance indices ITSE, ISTAE and ISTSE have not

You might also like