Model Answer Control 2 - 2018 - 2019

Electrical Engineering Department
]txep TpyT[ ECOMP 423: Automatic Control (2)

4th year Computers And Systems Section
Final Examination 1st Term 2018/2019
Date: 02/01/2019 Time: 3.0 Hours
Model Answer (100 Marks) Faculty of Engineering
Aswan University
Answer 1: (10 Marks)
Put (√) to the correct sentence and (x) to the mistake Sentence
a) The phase-lead controller is generally more effective if the uncompensated system
is very unstable to begin with. (x)
b) The minimum phase that is available from a single-stage phase-lead controller is
90°. (x)
c) The phase-lead controller may not be effective if the negative slope of the
un-compensated process transfer function is too steep near the gain-crossover
frequency. (√)
d) For a phase-lag controller, if the value of T is large and the value of β is large , it
is equivalent to adding a pure attenuation of (1/β) to the original uncompensated
system at low frequencies. (√)
e) If a PD controller is designed so that the characteristic-equation roots have better
damping than the original system, then the maximum overshoot of the system is
always reduced. (√)
f) A system compensated with a PI controller is usually less error than the system
compensated with a PD controller. (√)
g) lag compensation permits a high gain at low frequencies and reduces gain in the
higher critical range of frequencies (√)
h) Phase lead occurs in the high -frequency region and phase lag occurs in the low -
frequency region. (√)
i) Physically, the addition of a zero in the feed forward transfer function means the
addition of integral control to the system. (x)
j) Lead Compensation essentially yields an appreciable improvement in transient
response and a small change in steady state accuracy. (√)
Question 2: (30 Marks)

a) A unity feedback control system has an open loop transfer function G(s)
( )
( )
Design a lead compensator such that the dominant closed-loop poles are located at
√ . (14 marks)
Answer 2-a):
The angle deficiency at the closed loop poles √ is
180o - 120o - 90o = - 30o

The lead compensator must contribute 30o.
Page 1 of 10
Aswan University
– First, draw a horizontal line passing through point P, A P
the desired location for one of the dominant closed-
loop poles. This is shown as line PA in following figure. 15 15
– Draw also a line connecting point P and the origin O.
C
– Bisect the angle between the lines PA and PO,
x B D
– Draw two lines PC and PD that make angles Pc Zc O
with the the bisector PB.
– The intersections of PC and PD with the negative real axis give the necessary
locations for the pole and zero of the lead network.
– Thus, Gc(s) can be given as ( ) =
– For this compensator value of is →
– Also →
– Determine the value of Kc of the lead compensator from the magnitude
( )
condition. ( ) ( ) ( )
( )( )
( )
| |
( )( ) √
– The Kc is calculated as
– Hence, the lead compensator Gc(s) just designed is given by
( )
b) A unity feedback control system has an open loop transfer function G(s)
( )
( )( )
Design a lag compensator to achieve the following specification: Velocity error constant Kv = 8
and phase margin PM = 40o. (16 marks)
Answer 2-b):
We shall use a lag compensator of the form
( ) ̂ = ̂ , ( )
• Define ̂
• Define also ( )= ( ) ( )( )
• The first step in the design is to adjust the gain K to meet the required static
velocity error constant. Thus,
( ) ( ) ( ) ( )
( )( )
Or K= 8
Page 2 of 10
Aswan University
• With K=8, the compensated system satisfies the steady-state performance
requirement.
• We shall next plot the Bode diagram of ( ) ( )( )
• The magnitude curve and phase-angle curve of G1(jω) are shown in the
following figure.
Bode Diagram
100
20 dB
50
Magnitude (dB)
-50
-100
-90
-128o
-135
Phase (deg)
-140o
-180
-186.02o
-225
-270
-2 -1 0 1 2
10 10 10 10 10
Frequency (rad/s)
• From this plot, the phase margin is found to be –-6.02°, which means that
the gain-adjusted but uncompensated system is unstable.
• Noting that the addition of a lag compensator modifies the phase curve of the
Bode diagram, we must allow 5° to 12° to the specified phase margin to
compensate for the modification of the phase curve.
• Since the frequency corresponding to a phase margin of 40° is 0.856 rad/sec,
the new gain crossover frequency (of the compensated system) must be chosen
near this value.
• To avoid overly large time constants for the lag compensator, we shall choose
the corner frequency ω = 1/T (which corresponds to the zero of the lag
compensator) to be 0.1 rad/sec.
• Since this corner frequency is not too far below the new gain crossover
frequency, the modification in the phase curve may not be small. Hence, we
Page 3 of 10
Aswan University
• add about 12° to the given phase margin as an allowance to account for the
lag angle introduced by the lag compensator.
• The required phase margin is now 52°.The phase angle of the uncompensated
open-loop transfer function is –128° at about ω =0.603 rad/sec. So we
choose the new gain crossover frequency to be 0.603 rad/sec.
• To bring the magnitude curve down to 0 dB at this new gain crossover
frequency, the lag compensator must give the necessary attenuation, which in
this case is –20 dB. Hence,
• Or β = 10
• The other corner frequency ω=1/(βT), which corresponds to the pole of the lag
compensator, is then determined as
• Thus, the transfer function of the lag compensator is
( ) ̂( ) = ̂
• Since the gain K was determined to be 8 and β was determined to be 10, we

have
• The open-loop transfer function of the compensated system is
( )
( ) ( )
( )( )( )
Question 3: (18 marks)
a) A unity feedback control system has a feed-forward transfer function
( )
( )( )
Design PID controller using Ziegler-Nichols tuning rules. (10 marks)
Answer 3-a):
The PID controller has the transfer function
( ) ( )
Since the plant has an integrator, we use the second method of Ziegler-Nichols
tuning rules. By setting Ti = ∞ and Td = 0, we obtain the closed-loop transfer
function as follows: C ( s) Kp

R( s) s( s  1)(s  5)  K p
The value of Kp that makes the system marginally stable so that sustained oscillation
occurs can be obtained by use of Routh's stability criterion. Since the characteristic
equation for the closed-loop system is
the Routh array becomes as follows
Page 4 of 10
Aswan University
Examining the coefficients of the first column of the Routh table, we find that
sustained oscillation will occur if Kp = 30. Thus, the critical gain Kcr is
With gain K, set equal to , the characteristic equation becomes
To find the frequency of the sustained oscillation, we substitute s = jw into this

characteristic equation as follows:
( ) ( ) ( )
or ( ) ( )
from which we find the frequency of the sustained oscillation to be or √ .
Hence, the period of sustained oscillation is
√
Referring to Table 1, we determine Kp, Ti and Td as follows:
The transfer function of the PID controller is thus

( )
( ) ( ) ( )
The PID controller has a pole at the origin and double zero at s = -1.4235
( )
b) Consider the following transfer function system: ( )
. Obtain the state-space
representation of this system in Controllable canonical, observable canonical and diagonal
forms. (8 marks)
Answer 3-b):
To obtain the state-space representation in controllable canonical form we rewrite
the previous equation in the following form:
( ) ( )
( ) ( )
From this last equation, the following two equations may be obtained:
( ) ( ) ( )
( ) ( )
Now define state variables as follows:
( ) ( )
( ) ( )
Page 5 of 10
Aswan University
Then, clearly, ( ) ( )
Which may be rewritten as ̇ ( ) ( )
Noting that ( ) ( ) ( ) ( )
Or ̇ ( ) ( ) ()
̇ ( ) ( )
[ ] [ ][ ] [ ] ( )
̇ ( ) ( )
( )
( ) [ ][ ]
( )
To obtain the state-space representation in observable canonical form we rewrite
the previous equation in the following form: ( ) ( ) ( )
By dividing the entire equation by s2 and rearranging, we obtain
( ) ( ) ( )
Now define state variables as follows
( ) [ ( ) ( )]
( ) ( )
( ) ( )
By substituting the last equation into previous equations and multiplying both
sides of the equations by s, we obtain
( ) ( ) ( )
( ) ( )
Taking the inverse Laplace transforms of the preceding n equations and writing
them in the reverse order, we get
̇ ( ) ( ) ( )
̇ ( ) ()
( ) ( )
̇ ( ) ( )
Or [ ] [ ][ ] [ ] ( )
̇ ( ) ( )
( )
( ) [ ][ ]
( )
To obtain the state-space representation in diagonal form we use partial fraction
expansion then the previous equation will be in the following form:
( )
( ) ( ) ( )
Last equation may be written as ( ) ( ) ( )
( ) ( )
Define the state variables as follows:
( ) ( )
Page 6 of 10
Aswan University
( ) ( )
( )
Which may be rewritten as
( ) ( )
( ) ( ) ( )
The inverse Laplace transforms of these equations give
̇ ( ) ()
̇ ( ) ( ) ()
( ) ( ) ( )
̇ ( ) ( )
Or [ ] [ ][ ] [ ] ( )
̇ ( ) ( )
( )
( ) [ ][ ]
( )
Consider a system defined by:
 x1(t )   5 0  x1(t )  1  x (t ) 

 x (t )     x (t )  0u (t ), y(t )  0 100  1 
 2   1 0   2     x2 (t )
a) By use of the state feedback control u=-Kx, it is desired to place the closed loop poles at
with damped ratio of and natural undamped frequency is ⁄ . Determine the
values of feedback gains k1 and k2 by applying transformation matrix method. (9 marks)
Answer 4-a):
Checking the controllability
[ ] [ ] | |
The rank of matrix M is , then the system is completely state controllable.
The desired dominant closed loop poles for the system is √
( ) √ ( )
For the defined system by Equation, the characteristic polynomial is
| | | |
Thus, ,
The desired characteristic polynomial for the compensated system is
( )( ) ( )( )
Thus, ,
For the determination of the state-feedback gain matrix, we use
[ ]
Where [ ] [ ]
Hence s the state-feedback gain matrix K is
[ ][ ][ ] [ ][ ] [ ]
Page 7 of 10
Aswan University
Using this state-feedback gain matrix K, the control signal u is given by
( )
[ ][ ]
( )
Suppose that we use the observed-state feedback control instead of the actual-state
feedback control, or
̃ ( )
[ ][ ]
̃ ( )
b) Design a full order state observer, the desired eigenvalues for the observer matrix are
by applying Ackerman’s formula method. (9 marks)
Answer 4-b):
Checking the observability

[ ] [ ] | |
The rank of matrix N is 2 = n, then the system is completely observable.

For the defined system by Equation, the characteristic polynomial is
| | | |
Thus, ,
The desired characteristic polynomial for the system is
| | ( )( )
Hence, ,
Ackerman's formula ( )[ ] [ ]
( ) [ ] [ ] [ ] [ ]
The observer gain matrix Ke is
[ ][ ] [ ] [ ]
The observer equation is given by:
̇
( ̃) ( ̇) ̃
Since [ ]
̃
Then ̇
( ̃) ( ̇ )̃
Or
( ̃)̇ ̃
[ ] {[ ] [ ] [ ]} [ ] [ ]
̇
( ̃) ̃
( ̃)̇ ̃
[ ] [ ][ ] [ ]
( ̃)̇ ̃
Page 8 of 10
Aswan University
a) Consider the following system: (12 marks)
̇
̇
̇
Let Liapunov function is:
( )
I) Show that ̇ ( ) ( )
Answer 5-a-I):
̇( ) ̇ ̇ ̇ ̇ ̇ ̇ ̇
( )
II) For what range of is this system stable?
Answer 5-a-II):
This system will be stable if ( ) positive definite ̇ ( ) is negative definite, then
( ) [ ][ ][ ]
We find that the origin of state space is equilibrium state.

Then A and Q matrices to be
[ ] [ ]
To be matrix Q positive semi-definite real symmetric matrix then the rank of
√ √
[ ]
√ √ √
[ ]
Then and
To be matrix P positive definite real symmetric matrix must
then and | | then then and
| | then then , then
III) For what range of is ̇ ( ) negative definite?

Answer 5-a-III):
̇ ( ) is negative definite if
̇( ) ( ) then then the range of is
Page 9 of 10
Aswan University
b) The state equations for the harmonic oscillator have the form (12 marks)
̇ () () ̇ () () ( ).
It is desired to find the optimal control signal u such that the performance index is minimized:
∫ ( )
Answer 5-b):
From the given system:
( ) ( ) ( ) ( )
The optimal feedback gain matrix K can obtained by the following

Riccati equation for a positive-definite matrix P.
[ ][ ] [ ][ ] [ ] [ ] [ ][ ][ ] [ ]
[ ]
[ ] [ ] [ ] [ ] [ ]
[ ] [ ]
By solving the last equation we obtain the matrix P as following:
[ ]
Substituting this P matrix into the following equation gives the optimal
K matrix
[ ][ ][ ] [ ]
Thus, the optimal control signal is given by
Examiner: Dr. Mountasser M. Ramadan
Page 10 of 10

Model Answer Control 2 - 2018 - 2019

Uploaded by

Copyright:

Available Formats

Model Answer Control 2 - 2018 - 2019

Uploaded by

Document Information

Original Title

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

Model Answer Control 2 - 2018 - 2019

Uploaded by

Copyright:

Available Formats

Electrical Engineering Department

]txep TpyT[ ECOMP 423: Automatic Control (2)

Question 2: (30 Marks)

180o - 120o - 90o = - 30o

– Thus, Gc(s) can be given as ( ) =

– For this compensator value of is →

• Since the gain K was determined to be 8 and β was determined to be 10, we

the Routh array becomes as follows

To find the frequency of the sustained oscillation, we substitute s = jw into this

The transfer function of the PID controller is thus

 x1(t )   5 0  x1(t )  1  x (t ) 

Checking the observability

The rank of matrix N is 2 = n, then the system is completely observable.

The observer gain matrix Ke is

We find that the origin of state space is equilibrium state.

To be matrix Q positive semi-definite real symmetric matrix then the rank of

then and | | then then and

| | then then , then

III) For what range of is ̇ ( ) negative definite?

From the given system:

The optimal feedback gain matrix K can obtained by the following

By solving the last equation we obtain the matrix P as following:

Thus, the optimal control signal is given by

Examiner: Dr. Mountasser M. Ramadan

You might also like