Direct control design - Analytical Method

Dr. Amin Danial

 References
❑ Katsuhiko Ogata Discrete-Time Control
Systems 2nd edition,1995
❑ M. Sami Fadali, Antonio Visioli - Digital
Control Engineering, Analysis and Design-
Second Edition_Academic Press (2019)
❑ Gene F. Franklin, J. David Powell, Michael L.
Workman - Digital Control of Dynamic
Systems-Prentice Hall (1998)
❑ A. V. OPPENHEIM, A. S. WILLSKY and S. H.
NAWAB , Signals & Systems, PRENTICE HALL,
1996.
 Direct control design - Analytical Method
• For the following system:

Where
1 − 𝑒 −𝑠𝑇
𝐺 𝑧 = 𝑍𝑡𝑟𝑎𝑛𝑠 𝐺𝑝 𝑠
𝑠
 Direct control design - Analytical Method
• In this approach, it is required to find the desired
closed loop pulse transfer function 𝐹 𝑧 (This
approach to design is known as synthesis)

• Then
𝐶 𝑧 𝐺𝐷 𝑧 𝐺 𝑧
= = 𝐹(𝑧)
𝑅 𝑧 1 + 𝐺𝐷 𝑧 𝐺 𝑧

Consequently, the controller is found as follows:

𝐹 𝑧
𝐺𝐷 𝑧 =
𝐺(𝑧) 1 − 𝐹 𝑧
 Direct control design - Analytical Method
• The designed system must be physically
realizable. The conditions for physical realizability
are as follows:
1. The order of numerator of 𝐺𝐷 𝑧 must be equal or
lower than the order of the denominator. Otherwise,
the controller will be noncausal, i.e. it needs future
input to produce the current output.
2. If the plant has a delay 𝑒 −𝑙𝑠 then the designed closed
loop system must have at least the same delay.
Otherwise, the closed loop system would have to
respond before an input was give, which is physically
impossible.
 Direct control design - Analytical Method
3. We must avoid of canceling any unstable pole of the plant be a
zero of the controller because any error in the cancellation will
cause instability.
If 𝐺(𝑧) has a pole at 𝛼 then:
𝐺1 𝑧
𝐺 𝑧 =
𝑧−𝛼
Where 𝐺1 𝑧 has no term that cancels 𝑧 − 𝛼.
Consequently,
𝐺 𝑧
𝐺𝐷 𝑧 𝑧 1− 𝛼
𝐹 𝑧 =
𝐺 𝑧
1 + 𝐺𝐷 (𝑧) 1
𝑧−𝛼
1 𝑧−𝛼
1−𝐹 𝑧 = =
𝐺 𝑧 𝑧 − 𝛼 + 𝐺𝐷 𝑧 𝐺1 𝑧
1 + 𝐺𝐷 𝑧 1
𝑧−𝛼
 Direct control design - Analytical Method
Therefore, 𝛼 is zero of 1 − 𝐹 𝑧 , hence:
𝐹 𝛼 =1
4. The poles of the controller 𝐺𝐷 (𝑧) do not cancel zeros of
𝐺 𝑧 which lies outside the unit circle.
Therefore, if there is a zero (𝑧 − 𝛽)of 𝐺(𝑧) which is outside
the unit circle. Then, we can write 𝐺 𝑧 as follows:
𝐺 𝑧 = 𝑧 − 𝛽 𝐺2 (𝑧)
Consequently, in case of (𝑧 − 𝛽) is not cancelled:
𝐺𝐷 𝑧 𝑧 − 𝛽 𝐺2 (𝑧)
𝐹 𝑧 =
1 + 𝐺𝐷 (𝑧) 𝑧 − 𝛽 𝐺2 (𝑧)
Hence,
𝐹 𝛽 =0
 Direct control design - Analytical Method
5. An additional condition can be imposed to
address steady-state accuracy
requirements.
In particular, if zero steady-state error due
to a step input is required
𝑧
lim 𝑐 𝑘𝑇 = lim 𝑧 − 1 𝐹 𝑧 . =1
𝑘→∞ 𝑧→1 𝑧−1
Therefore,
𝐹 1 =1
 Direct Control Design - Analytical Method
• Steps of design
[Link] the desired settling time Ts and the desired
maximum overshoot.
[Link] a suitable continuous-time closed-loop first-
order or second-order closed-loop system with unit
gain.
[Link] by converting the s-plane pole location to the
z-plane pole location using pole-zero matching.
[Link] that 𝐹(𝑧) meets the conditions for causality,
stability, and steady-state error. If not, modify 𝐹(𝑧)
until the conditions are met.
 Direct Control Design - Analytical Method
• EX1: Design a digital controller for a DC motor speed control
system where the (type 0) analog plant has the transfer function
1
𝐺𝑝 =
(𝑠 + 1)(𝑠 + 10)
To obtain zero steady-state error due to a unit step, 𝜔𝑛
= 1.15 𝑟𝑎𝑑/𝑠, and 𝜁 = 0.88. The sampling period is chosen as 𝑇
= 0.02 s

• Solution
Since in digital control system a ZOH is always inserted between the
digital controller and the system to be controlled, hence
1 − 𝑒 −𝑠𝑇 −1
𝐺𝑝 𝑠
𝐺 𝑧 = 𝑍𝑡𝑟𝑎𝑛𝑠 𝐺𝑝 𝑠 = (1 − 𝑧 )𝑍𝑡𝑟𝑎𝑛𝑠
𝑠 𝑠
 Direct Control Design - Analytical Method
• EX1 (cont.)
−4
𝑧 + 0.9293
𝐺 𝑧 = 1.86 × 10
(𝑧 − 0.8187)(𝑧 − 0.9802)
Then, based on the requirements 𝐹 𝑠 is:
𝜔𝑛2 1.322
𝐹 𝑠 = 2 2 = 𝑠 2 + 2.024𝑠 + 1.322
𝑠 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛
Using zero-pole matching
−3
𝑧+1
𝐹 𝑧 = 0.25921 × 10
𝑧 2 − 1.95981𝑧 + 0.96033
 Direct Control Design - Analytical Method
EX1(cont.):
Then from

𝐹 𝑧
𝐺𝐷 𝑧 =
𝐺(𝑧) 1 − 𝐹 𝑧
We get

1.3932(𝑧 − 0.8187)(𝑧 − 0.9802)(𝑧 + 1)

𝐺𝐷 𝑧 =
(𝑧 − 1)(𝑧 + 0.9293)(𝑧 − 0.9601)
 Direct Control Design - Analytical Method
• Ex2: Design a digital controller for type zero plant
1
𝐺 𝑠 =
𝑠 − 1 (𝑠 + 2)
Such that it has a zero steady state error due to a unit step.
Damping ratio of 0.7 and settling time of about 1 sec. Sampling
time is 𝑇 = 0.01𝑠𝑒𝑐.

Solution:
From the settling time
4
𝑡𝑠 = =1
𝜁𝜔𝑛
Therefore,
4
𝜔𝑛 = = 5.71 𝑟𝑎𝑑/𝑠𝑒𝑐
0.7
 Direct Control Design - Analytical Method
Ex2(cont.):

−1
𝐺 𝑠
𝐺 𝑧 = 1−𝑧 𝑍𝑡𝑟𝑎𝑛𝑠
𝑠
𝑧 + 0.997
𝐺 𝑧 = 0.0000498 2
𝑧 − 1.99025𝑧 + 0.99005
𝑧 + 0.997
𝐺 𝑧 = 0.0000498
(𝑧 − 0.98)(𝑧 − 1.01)
 Direct Control Design - Analytical Method
• Ex2 (cont.)
Using pole-zero matching. Note that there is unstable pole in
the system, hence the closed loop TF is chosen to be:
𝑧−𝛽
𝐹 𝑧 =𝛼 2
𝑧 − 2𝑒 −𝜁𝜔𝑛 𝑇 cos 𝜔𝑑 𝑇 𝑧 + 𝑒 −2𝜁𝜔𝑛 𝑇
𝑧−𝛽 𝐴(𝑧)
𝐹 𝑧 =𝛼 2 =𝛼
𝑧 − 1.92𝑧 + 0.923 𝐵(𝑧)
For avoiding unstable pole cancellation
𝐴(1.01)
𝐹 1.01 = 𝛼 =1
𝐵(1.01)
𝐵(1.01) 𝐵(1.01)
𝛼= = (1)
𝐴(1.01) 1.01 − 𝛽
 Direct control design - Analytical Method
• Ex2 (cont.)
For zero steady state error for unit step
𝐴(1) 1−𝛽
𝐹 1 =𝛼 =𝛼 =1 2
𝐵(1) 𝐵(1)
Substitute (1) in (2)
𝐵(1.01) 1 − 𝛽
=1
1.01 − 𝛽 𝐵(1)
Then
𝛽 = 0.967

From (1) or (2) we get 𝛼

𝛼 = 0.09
 Direct control design - Analytical Method

Ex2 (cont.)
𝐹(𝑧)
𝐺𝐷 𝑧 =
𝐺 𝑧 1−𝐹 𝑧
𝐺𝐷 𝑧
(𝑧 − 0.967)(𝑧 − 1.01)(𝑧 − 0.98)
= 1807.23
(𝑧 + 0.99667)(𝑧 − 1)(𝑧 − 1.01)

(𝑧 − 0.967)(𝑧 − 0.98)
𝐺𝐷 𝑧 = 1807.23
(𝑧 + 0.99667)(𝑧 − 1)
Direct control design - Analytical Method
 Direct Method
Digital implementation of analog controller design
• Ex3: design a controller 𝐶𝑎 (𝑠) for the following unstable system
1
𝐺𝑎 𝑠 =
𝑠 − 0.5
Such that :
𝐶𝑎 (𝑠)𝐺𝑎 𝑠 𝑠+𝛼
1. The closed loop 𝑇𝐹 = has the form 𝐾 2 2
1+𝐶𝑎 𝑠 𝐺𝑎 𝑠 𝑠 +2𝜁𝜔𝑛 𝑠+𝜔𝑛

𝟏 O/P
I/P
𝒔 − 𝟎. 𝟓
𝐶𝑎 𝑠 𝐺𝑎 𝑠

2. Maximum overshoot is 5%
1
3. The settling time using 2% criteria is sec
3
4. Zero steady state error for unit step.
Then implement it digitally using bilinear transfotmation where 𝑇𝑠
= 0.01𝑠. Then draw the whole system block diagram.
 Direct Method
Digital implementation of analog controller design

• Ex3 (cont.)
• From the requirements
𝜁𝜋
−
1−𝜁2
Maximum Overshoot(𝑀𝑃 )= 𝑒 =0.05
Then
2
ln 𝑀𝑝 ln 0.05 2
𝜁= 2 = 2 2
≅ 0.7
ln 𝑀𝑝 + 𝜋2 ln 0.05 + 𝜋
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)
Setting time according to the 2% criteria is:
4 1
𝑡𝑠 = =
𝜁𝜔𝑛 3
Therefore
12
𝜔𝑛 = = 17.14
0.7
Consequently, the required transfer function is
𝑠+𝛼 𝐴(𝑠)
𝑇𝐹 𝑠 = 𝐾 2 =𝐾
𝑠 + 23.996𝑠 + 293.78 𝐵(𝑠)
Where
𝐴 𝑠 =s+𝛼
𝐵 𝑠 = 𝑠 2 + 23.996𝑠 + 293.78
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)
• To determine 𝛼 and 𝐾 we use two conditions
1. Avoiding zero- pole cancellation
– It is not allowed to get a zero in the controller 𝐶𝑎 𝑠
at the unstable pole of the controlled system 𝐺𝑎 𝑠 .
– We have unstable pole of the controlled system
located at 𝑠 = 0.5.
– Therefore, 𝐶𝑎 𝑠 must have no zero at 𝑠 = 0.5.
– To be sure that this condition is satisfied the following
equation must be satisfied:
𝑇𝐹 𝑠 ቚ =1
𝑠=𝑢𝑛𝑠𝑡𝑎𝑏𝑙𝑒 𝑝𝑜𝑙𝑒𝑠
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)
• Therefore
𝑇𝐹 𝑠 ቚ =1
𝑠=0.5

𝐴(0.5)
𝑇𝐹 0.5 = 𝐾 =1
𝐵(0.5)
Then
𝐵(0.5) 𝐵(0.5)
𝐾= =
𝐴(0.5) 0.5 + 𝛼
0.00327𝐾 0.5 + 𝛼 = 1 (1)
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)
2. Steady state error 𝑒𝑠𝑠
𝑠𝑡𝑒𝑎𝑑𝑦 𝑠𝑡𝑎𝑡𝑒 𝑒𝑟𝑟𝑜𝑟
1
= lim 𝑠𝐸 𝑠 = lim
𝑠→0 𝑠→0 1 + 𝐶𝑎 𝑠 𝐺𝑎 (𝑠)
= lim 1 − 𝑇𝐹 𝑠 = 0
𝑠→0
Therefore
𝛼
𝐾 =1
𝐵 0
0.0034𝐾𝛼 = 1 (2)
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)
• Solving the two equations (1) and (2)
we get
𝐾 = 24.496
𝛼 = 11.993
Consequently,
𝑠 + 11.993
𝑇𝐹 𝑠 = 24.496 2
𝑠 + 23.996𝑠 + 293.78
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)
• To get the 𝐶𝑎 𝑠 , from the closed loop 𝑇𝐹(𝑠)
𝐶𝑎 𝑠 𝐺𝑎 (𝑠)
𝑇𝐹(𝑠) =
1 + 𝐶𝑎 𝑠 𝐺𝑎 (𝑠)
Then by rearranging the equation of the T.F., we get
1 𝑇𝐹 𝑠
𝐶𝑎 𝑠 =
𝐺𝑎 𝑠 1 − 𝑇𝐹 𝑠
Then
𝑠 + 11.993
1 24.496 2
𝐶𝑎 𝑠 = 𝑠 + 23.996𝑠 + 293.78
1 𝑠 + 11.993
1 − 24.496 2
𝑠 − 0.5 𝑠 + 23.996𝑠 + 293.78
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)
Therefore
24.496𝑠 + 293.78
𝐶𝑎 𝑠 =
𝑠
Getting the digital controller using bilinear
transformation:
𝐶 𝑧 = 𝐶𝑎 𝑠 ቚ 2 𝑧−1
𝑠=
𝑇𝑠 𝑧+1
2 𝑧−1
24.496 + 293.78
𝑇𝑠 𝑧 + 1
=
2 𝑧−1
𝑇𝑠 𝑧 + 1
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)
Therefore,

25.9649𝑧 − 23.0271
𝐶 𝑧 =
𝑧−1
 Direct Method
Digital implementation of analog controller design

• Ex8 (cont.)

I/P
O/P

Block Diagram
 Direct Method
Digital implementation of analog controller design
• Ex3 (cont.)

Analog Controller

Digital Controller
Thanks