21/11/2022
Instrumentation & Control Systems 1
Analysis of Control System Performance
Jacob Morewood MEng (Hons) PGCHE AMRAeS FHEA RAFAC
Instrumentation & Control Systems 2
Learning Objectives
• Understand the concept of Steady State Error
• Identify the behaviour differences for open loop and unity feedback systems
• Understand the philosophy behind frequency response analysis
• Understand the time response of a system to a sinusoidal input
• Appreciate Gain & Phase Shift
1
21/11/2022
Instrumentation & Control Systems 3
Unity Feedback
Unity feedback allows us to take a copy of the output and include it at the start of the control system to offer
an improvement to the response of the system.
If there is still an error and we do not achieve the desired response we require, a controller may be used to
change the input to the system we wish to control. This is sometimes known as the “plant”
Instrumentation & Control Systems 4
Example – Tracking Antenna
The order of the system is second order, therefore we can
see the typical behaviour of the system:
2
21/11/2022
Instrumentation & Control Systems 5
Example – Tracking Antenna
The order of the system is second order, therefore we can
see the typical behaviour of the system:
Instrumentation & Control Systems 6
Example – Tracking Antenna
If we placed a display to show the error value, it is shown as a -0.002585 which can be assumed
as zero
3
21/11/2022
Instrumentation & Control Systems 7
Classification of Unity Feedback Systems
We can classify unity feedback systems by their order and type number:
• The order of a system is given by the order of the open loop characteristic equation.
• The type number for a system is given by the number of roots of the open loop characteristic equation
that are zero.
For example, if the open loop characteristic equation of a unity feedback system is:
𝑠 𝑠 + 4𝑠 + 3 = 0
the system is said the be a fourth order type 2 system
Instrumentation & Control Systems 8
Classification of Unity Feedback Systems
For example:
Open loop characteristic equation is 1+2s = 0 Open loop characteristic equation is 2s2+s = 0
s0(1+2s) = 0 s1(1+2s) = 0
Hence the system is Type 0 Hence the system is Type 1
4
21/11/2022
Instrumentation & Control Systems 9
Steady State Error for Unity Feedback Systems
Instrumentation & Control Systems 10
Position Error Constant
If we assume the output of a unity feedback system is position then the reference signal must be the
desired position.
If the input is a unit step, the steady state error will be a positional error.
1
𝑒 =
1+𝐾
10
5
21/11/2022
Instrumentation & Control Systems 11
Position Error Constant - Example
If the input is a unit step, the steady state error will be a positional error.
Desired angle is fixed (i.e ϕd is a step input)
The Open Loop Transfer Function is:
0.1
𝑠(1 + 10𝑠)
Type 1 System hence, from the table: ess = 0
11
Instrumentation & Control Systems 12
Velocity Error Constant
If the input is a ramp function, the steady state output will be a constant velocity assuming the system is
stable.
So if u(t) = t
1
𝑒 =
𝐾
Where KV is the Velocity Error Constant
12
6
21/11/2022
Instrumentation & Control Systems 13
Acceleration Error Constant
If the input is a parabolic function the steady state output will be a constant acceleration assuming the
system is stable.
So if u(t) = 0.5t2 1
𝑒 =
𝐾
Where Ka is the Acceleration Error Constant
13
Instrumentation & Control Systems 14
Response of a System
The following Step Response is presented. Assuming a constant input of 10, determine the overshoot and
Steady State Error
𝑦 −𝑦
𝑂𝑣𝑒𝑟𝑠ℎ𝑜𝑜𝑡 =
𝑦
Max Amplitude = 11
11 − 10
= = 0.1
10
Overshoot Percentage = 10%
Steady State Error = 10-10 = 0
Steady State Output = 10
14
7
21/11/2022
Instrumentation & Control Systems 15
Response of a System
The following second order transfer function is presented. Calculate the Natural Frequency & Damping
Constant.
10 𝐾𝜔
G ( s) 𝐺 𝑠 =
s 2 s 11
2 𝑠 + 2ζ𝜔 𝑠 + 𝜔
ω = 11 ω = 3.317
2ζω = 2 2ζ × 3.317 = 2 ζ = 0.3015
15
Instrumentation & Control Systems 16
Response of a System
The following second order transfer function is presented. Calculate the Damping Constant.
1 𝐾𝜔
𝐺(𝑠) = 𝐺 𝑠 =
3𝑠 + 1 10𝑠 + 1 𝑠 + 2ζ𝜔 𝑠 + 𝜔
1
𝐺(𝑠) =
30𝑠 + 13𝑠 + 1
0.033
𝐺(𝑠) =
𝑠 + 0.433𝑠 + 0.033
𝜔 = 0.033 → 𝜔 = 0.182 𝑟𝑎𝑑/𝑠
0.433 0.433
2ζ𝜔 = 0.433 → ζ = = = 1.19
2𝜔 2 × 0.182
16
8
21/11/2022
Instrumentation & Control Systems 17
Response of a System
The damping constant can determine the general behaviour of a system’s ability to stabilise:
OVERDAMPED 1 UNDERDAMPED 0 1 UNDAMPED 0
17
Instrumentation & Control Systems 18
Frequency Response Philosophy
From Fourier series, we know that any waveform can be made up from a series of sinusoidal waveforms
of different frequencies and amplitudes.
18
9
21/11/2022
Instrumentation & Control Systems 19
Frequency Response Philosophy
From Fourier series, we know that any waveform can be made up from a series of sinusoidal waveforms
of different frequencies and amplitudes.
19
Instrumentation & Control Systems 20
Frequency Response Philosophy
From Fourier series, we know that any waveform can be made up from a series of sinusoidal waveforms
of different frequencies and amplitudes.
20
10
21/11/2022
Instrumentation & Control Systems 21
Frequency Response Philosophy
From Fourier series, we know that any waveform can be made up from a series of sinusoidal waveforms
of different frequencies and amplitudes.
21
Instrumentation & Control Systems 22
Frequency Response Philosophy
From Fourier series, we know that any waveform can be made up from a series of sinusoidal waveforms
of different frequencies and amplitudes.
22
11
21/11/2022
Instrumentation & Control Systems 23
Frequency Response Philosophy
• An input function can always be regarded as part of a waveform
• So given the Fourier Series for an input, the response of a system can be determined from the
response to a combination of sinusoidal inputs
• Hence to model a system, measure the response of the system to all sine waves with frequency
between 0 and infinity and we can determine the time response to any input.
• If the system is stable, the steady state output will be a sine wave of the same frequency as the input
except that the amplitude may have changed and it will usually have been delayed in time.
u(t) = sin(wt) y(t) = Rsin(wt+f)
G(s)
23
Instrumentation & Control Systems 24
Gain – R vs K
• The Gain is defined as the output divided by the input. For a linear system with a step response (i.e.
input is fixed) the Gain is a constant called K.
• In frequency response the input is constantly changing so use the amplitude of the output and input
wave forms.
• Since the amplitude of the output wave form is a function of the frequency, so will the Gain be. So to
avoid confusion with K we use the symbol R.
Input Output
U sin(wt ) Y sin(wt f)
output amplitude Y
R
input amplitude U
24
12
21/11/2022
Instrumentation & Control Systems 25
Phase Shift
• Phase shift is a measure of the time delay between the input
sine wave and the output sine wave.
• Since this time delay is a function of frequency so will the
phase shift be a function of frequency.
Input
delay
U sin(wt )
f 360
T
Output
Y sin(wt f)
25
Instrumentation & Control Systems 26
Nyquist Diagrams
Since Frequency Response data is in the form of a “vector” with
magnitude and direction, (gain, R, and phase shift, f), Nyquist argued
that the logical way to present the data graphically is using a polar
plot. The resulting plot is called a Nyquist Diagram.
R
If each data point is plotted on polar paper, a smooth line can be
drawn through the points to give the frequency response locus. f
26
13
21/11/2022
Instrumentation & Control Systems 27
Nyquist Diagram Example
10
Draw the Nyquist Diagram for the following first order system: G ( s)
20 s
ω Φ
R
rad/s degrees
0.1 0.50 -0.3
0.2 0.50 -0.6 0.5
0
0.5 0.50 -1.4
1 0.50 -2.9
2 0.50 -5.7 -10
5 0.49 -14.0
10 0.45 -26.6
-20
20 0.35 -45.0
50 0.19 -68.2
100 0.10 -78.7
-30
200 0.05 -84.3
500 0.02 -87.7
1000 0.01 -88.8 -110 -100 -90-80 -70 -60 -50 -40
27
Instrumentation & Control Systems 28
Nyquist Stability Criterion
• Harry Nyquist (1889 – 1976) was a Swedish-born American electronic engineer particularly
known in engineering for his contributions to communication theory
• Worked for the Bell Telephone Laboratories in New Jersey which became owned by Nokia
• Determined a graphical technique for determining the stability of a BIBO System which has
become a general fundamental for determining general stability of a control system
28
14
21/11/2022
Instrumentation & Control Systems 29
Nyquist Stability Criterion
A Closed loop system is only
stable if the open loop gain is
less than 1 when the phase
shift is 180 degrees
29
Instrumentation & Control Systems 30
Justification for Nyquist Stability Criterion
No Change
30
15
21/11/2022
Instrumentation & Control Systems 31
Justification for Nyquist Stability Criterion
Stable
31
Instrumentation & Control Systems 32
Justification for Nyquist Stability Criterion
Unstable
32
16
21/11/2022
Instrumentation & Control Systems 33
Nyquist Stability Criterion
A Closed loop system is only stable if the open loop gain is less than 1 when the phase shift is 180 degrees
33
Instrumentation & Control Systems 34
Example
Take the following Transfer Function. Determine if the
system is stable according to Nyquist Stability Criterion
90
𝐺 𝑠 =
𝑠 + 9𝑠 + 18
34
17
21/11/2022
Instrumentation & Control Systems 35
Example
Take the following Transfer Function. Determine if the
system is stable according to Nyquist Stability Criterion
20
𝐺 𝑠 =
𝑠 + 5𝑠 + 6𝑠
35
Instrumentation & Control Systems 36
Example
Take the following Transfer Function. Determine if the
system is stable according to Nyquist Stability Criterion
90 .
𝐺 𝑠 = 𝑒
𝑠 + 9𝑠 + 18
36
18
21/11/2022
Instrumentation & Control Systems 37
Example
Take the following Transfer Function. Determine if the
system is stable according to Nyquist Stability Criterion
90 .
𝐺 𝑠 = 𝑒
𝑠 + 9𝑠 + 18
37
Instrumentation & Control Systems 38
Gain Margin
• The gain margin is one of the stability margin measurements that can be used to determine
how much more a system can be modified before it becomes unstable.
• More specifically, how much Gain can be introduced into the system before it becomes
unstable
• The greater the gain margin, the more stable the system.
• To determine the gain margin, measure the distance between the origin and where the plot
intercepts the negative real axis on a Nyquist Plot.
38
19
21/11/2022
Analysis of Applied Control Systems 39
Example – Determining Gain Margin
Take the following Transfer Function. Determine the gain margin (gm)
90 .
𝐺 𝑠 = 𝑒
𝑠 + 9𝑠 + 18
Distance between origin and intercept is approximately 0.5
1
∴ =2
0.5
39
Instrumentation & Control Systems 40
Phase Margin
• The phase margin is another of the stability margin measurements that can be used to
determine how much more a system can be modified before it becomes unstable.
• More specifically, how much Phase Shift Angle can be introduced into the system before it
becomes unstable
• The greater the phase margin, the more stable the system.
• To determine the phase margin, measure the angle between the origin and the intersection of
the locus with the -1, 0j Unit Circle
40
20
21/11/2022
Instrumentation & Control Systems 41
Example – Determining Phase Margin
Take the following Transfer Function. Determine the phase margin (ϕm)
90 .
𝐺 𝑠 = 𝑒
𝑠 + 9𝑠 + 18
Angle between these points is approximately 30°
∴ 𝜙 = 30°
41
Instrumentation & Control Systems 42
Learning Objectives
• Understand the concept of Steady State Error
• Identify the behaviour differences for open loop and unity feedback systems
• Understand the philosophy behind frequency response analysis
• Understand the time response of a system to a sinusoidal input
• Appreciate Gain & Phase Shift
42
21
21/11/2022
Instrumentation & Control Systems 43
Analysis of Control System Performance
Jacob Morewood MEng (Hons) PGCHE AMRAeS FHEA RAFAC
43
22
