Professional Documents
Culture Documents
ﺍﻟﻤﺮﺣﻠﺔ ﺍﻟﺮﺍﺑﻌﺔ
Temp.
indicator
Hot water
Water
Heater
Steam
Cool water
Process: Is defined as the changing or refining of raw materials that pass through or
remain in a liquid, gaseous, or slurry state to create end products.
Control: In process industries refers to the regulation of all aspects of the process.
Precise control of level, pH, oxygen, foam, nutrient, temperature, pressure and flow is
important in many process applications.
Sensor: A measuring instrument, the most common measurements are of flow (F),
temperature (T), pressure (P), level (L), pH and composition (A, for analyzer). The
sensor will detect the value of the measured variable as a function of time.
Set point: The value at which the controlled parameter is to be maintained.
Error Signal: The signal resulting from the difference between the set point
reference signal and the process variable feedback signal in a controller.
Steady-State: The condition when all process properties are constant with time,
transient responses having died out.
Disturbance: A process input (other than the manipulated parameter) which affects
the controlled parameter.
Process Time Constant( τ ): Amount of time counted from the moment the variable
starts to respond that it takes the process variable to reach 63.2% of its total change.
Block diagram: It is relationship between the input and the output of the system. It is
easier to visualize the control system in terms of a block diagram.
Block diagram
e.g. the transfer function of the above block diagram is G (s) = Y(s)/X(s)
Closed-loop control system: It is a feedback control system which the output signals
has a direct effect upon the control action.
Heater
(Heating)
Temperature
Transmitter
(Thermocouple)
Open-loop control system: It is a control system in which the output has no effect
upon the control action. (The output is neither measured nor fed back for comparison
with the input).
(Timer) Motor
Advantages:
(1) Simple construction and ease of maintenance.
(2) Less expensive than closed-loop control system.
(3) There is no stability problem.
Disadvantages:
(1) Disturbance and change in calibration cause errors; and output may be
different from what is desired.
(2) To maintain the required quality in the output, recalibration is necessary
from time to time
Note: any control system which operates on a time basis is open-loop control system,
U U
2.1 Introduction
Laplace transform techniques provide powerful tools in numerous fields of
technology such as Control Theory where knowledge of the system transfer function
is essential and where the Laplace transform comes into its own.
Definition
The Laplace transform of an expression f (t ) is denoted by L{ f (t )} and is defined as
∞
the semi-infinite integral: L{ f (t )} = ∫ f (t )e
− st
dt .
t =0
The parameter s is assumed to be positive and large enough to ensure that the integral
converge. In more advanced applications s may be complex and in such cases the real
part of s must be positive and large enough to ensure convergence.
In determining the transform of an expression, you will appreciate that the limits of
the integral are substituted for t, so that the result will be an expression in s.
∞
Therefore: L{ f (t )} = ∫ f (t )e
− st
dt = F ( s )
t =0
u = t ⇒ du = dt
− st − e − st
dv = e dt ⇒ v =
s
⇒ L{t} =
1
s2
u = t 2 ⇒ du = 2tdt
e − st
dv = e − st dt ⇒ v =
−s
∞
t 2 e − st ∞ 2te − st ∞ 2te − st
L t { }= ∫
2
∞
t e dt =
2 − st
−s
− − ∫ dt = ∫ dt
0
0 0 s 0 s
∞ ∞
∞
∫ u ⋅ dv =u ⋅ v
0
0
− ∫ v ⋅ du
0
u = 2t ⇒ du = 2dt
e − st − e − st
dv = dt ⇒ v = s
s s
∞ ∞
2te − st − 2te − st ∞ 2e − st − st
− 2e − st
{}
L t =∫ 2
∞
dt =
s2
− − ∫ 2
dt = 2
∞e
∫0 s 2 dt =
s3
=−
2
(0 − 1) = 23
0 s
3
0 s 0 0
s s
⇒ L t2 {} 2
= 3
s
⇒ Le =
at
{ } 1
( s − a)
Similarity
{ }
L e − at =
1
( s + a)
1 ( s + ia ) − ( s − ia ) 1 2ia a
= = 2 2
= 2
2i ( s − ia )( s + ia ) 2i s + a s + a 2
L{sin( at )} =
a
s + a2
2
Also
L{cos(at )} =
s
s + a2
2
L{sinh( at )} =
a
s − a2
2
Also
L{cosh(at )} =
s
s − a2
2
a
sinh at
s − a2
2
s
cosh at
s − a2
2
1
e − at
s+a
1
te − at
( s + a) 2
n!
t n e − at
( s + a ) n +1
w
e − at sin wt
( s + a) 2 + w 2
s+a
e − at cos wt
( s + a) 2 + w 2
If L{ f (t )} = F ( s) then L{tf (t )} = − F ′( s)
∞ ∞ ∞
de − st
L{tf (t )} = ∫ tf (t )e − st dt =
d
t =0
∫
t =0
f (t ) −
ds
dt = −
ds ∫
t =0
f (t )e − st dt = − F ′( s ).
{ }
L t n f (t ) = (−1) n
dn
{F ( s)} that
ds n
{ } d2
ds
d2 d2 1
L t 2 f (t ) = (−1) 2 2 {F ( s )} = 2 {F ( s )} = 2 2
ds ds s + 1
Find the first derivative:
1 a da db
z= , z = , a = 1, b = s 2 + 1, = 0, = 2s
s +1 2
b ds ds
da db
dz
b −a
= ds 2 ds =
(
s 2 + 1 (0 ) − (1)(2 s )
=
− 2s)
ds b s +1
2 2
(
s2 +1
2
) ( )
Differentiate again.
− 2s
z=
a
= 2
2
( ) (
, a = −2s, b = s 2 + 1 = s 2 + 1 s 2 + 1 = s 4 + 2s 2 + 2, )( )
b s +1 2
( )
da db
= −2 , = 4 s 3 + 4 s,
ds ds
(( )) (
da db
dz
=
b
ds
−a 2
( 3
)
ds = s + 1 (− 2) − (− 2s ) 4s + 4s = − 2 s + 1 + 8s + 8s
2 2 2 4 2
( ) )
ds b2 s2 +1
2 2
(( ))
2
s2 +1 s2 +1
2
( )( )
=
( ) (
− 2 s 4 + 2 s 2 + 2 + 8s 4 + 8s 2 ) = (− 2s 4
) (
− 4 s 2 − 4 + 8s 4 + 8s 2 ) = (6s + 4s − 4)
4 2
(s 2
)(
2
+1 s2 +1 )2
(s 2
)( 2
+1 s2 +1 )
2
(s + 1) (s + 1)
2 2 2 2
Therefore
{ d2 1
L t sin t = 2 2
2
}=
6s 4 + 4s 2 − 4
.
( )
ds s + 1 s 2 + 1 2 s 2 + 1 2 ( )( )
Theorem 3: Convolution Theorem:
U U
If L{ f (t) } = F(s) and L{ g(t) } = G(s) then the convolution of f (t) and g(t)
is denoted by (f * g)(t), is defined by
t
( f ∗ g )(=
t) ∫ f (τ ) g ( t − τ ) dτ
0
And the Laplace transform of the convolution of two functions is the product of the
separate Laplace transforms:
L { ( f ∗ g )( t ) } =
F (s) G (s)
An equivalent identity is
L −1 { F=
(s) G (s) } L −1 { F ( s ) } ∗ L −1 { G ( s ) }
Example: Find t 2 ∗ 2t
Solution:
t
τ3 τ4
( )
t t
2t 4 t 4 t 4
t ∗ 2t = ∫ τ 2(t − τ )dτ = ∫
2 2
2tτ − 2τ dτ = 2t − 2 =
2 3
− =
0 0 3 4 0 3 2 6
∫ (τe )dτ = e (− τe − e )
t t
t −τ
e ∗ t = ∫ e (τ )dτ = e
t t −τ t −τ τ − t
0
0 0
[( ) ]
= et − te − t − e − t − (0 − 1) = −et e − t (t + 1) + et = −t − 1 + et = et − t − 1
1- Step function
0 t<0 A
f (t ) =
A t≥0
.
A
f ( s) = 0 Time
s
If A=1 the change is called unit step change
0 t<0
f (t ) =
1 t≥0
1
f ( s) =
s
0 t<a
f (t ) =
A t≥a
A
A
f ( s ) = e − as
s
0 a Time
2. Pulse function
0 t<0
f (t ) = A 0≤ t ≤ a
A
0 t≥a
A A − as
f (s) = − e 0 a Time
s s
A
= (1 − e − as )
s
L{ f (t )} = ∫ Ae − st + ∫ 0e − st = −
At − st A A
e = − (e − sa − e − s 0 ) = (1 − e − sa )
0 a
s 0 s s
if A=1unit Pulse (Impulse)
3. Impulse function
0 t<0 A Area=1
f (t ) = A 0 ≤ t ≤ δt
0 t ≥ δt
Unit Impulse Time
f ( s ) = area = A × δt
This function is represented by δ(t). The unit impulse function is a special case of the
pulse function with zero width (tw →0) and unit pulse area (so a = 1/tw). Taking the
limit:
L{δ (t )} = lim
1 1
[1 − e − st w ] = lim [−e − st w ] = 1
t w →0 t s t w →0 s
w
4. Ramp function
0 t<0
f (t ) = Slope=A
At t≥0
f(t)
A
f (s) = 2
s Time
∞ ∞ ∞ ∞
A − 1 − st
L{At} = ∫ Ate − st = −
At − st A At
e + ∫ e − st dt = − e − st + e
0
s 0 s s 0 s s 0
−A A A
= (∞e − ∞ − 0e − 0 ) − 2 (e − ∞ − e − 0 ) = 2
s s s
-A
Process Control /Lec. 2 12 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
0 t<a
f (t ) =
A sin wt t≥a
Aω
f ( s) = 2
s +ω2
ω = 2πf
1
T=
f
Solution:
1. At t=0 the function looks like the very basic unit step function. But unit function
knows only about 0 and 1, here we have f(t)=2. That means we have to use 2u(t).
2. Then in time t=2 its value changes from 2 to −1 (i.e. 3 down at t=2) which means
we have to add −3u(t−2).
3. Finally the value at t=3 jumps 1 higher, which brings member u(t−3).
f(t)=2u(t)−3u(t−2)+u(t−3)
So far we collected unit step functions to express function from the graph.
L{ f (t )} = L{2u(t) - 3u(t - 2) + u(t - 3)} = L{2u(t)} - 3L{u(t - 2)} + L{u(t - 3)} =
2 3 - 2s 1 -3s
- e + e
s s s
Solution:
f (t ) = 0u (t − 0) + 1u (t − 1) + 2u (t − 3) − 1u (t − 5) − 2u (t − 6)
1 2 1 2 1
F ( s ) = e − s + e − 3 s − e − 5 s − e − 6 s = (e − s + 2e − 3 s − e − 5 s − 2e − 6 s )
s s s s s
0.5
5
Solution:
f (t ) = −0.5 + 1u (t − 0.5) − 0.5u (t − 1.5) + 0.5u (t − 2) − 1u (t − 3) + 0.5u (t − 3.5)
0.5 1 − 0.5 s 0.5 −1.5 s 0.5 − 2 s 1 − 3 s 0.5 − 3.5 s
f ( s) = − + e − e + e − e + e
s s s s s s
Solution:
f (t ) = tU (t ) − t (t − 1)U (t − 1) − (t − 1)U (t − 1) + (t − 2)U (t − 2)
= tU (t ) − 2(t − 1)U (t − 1) + (t − 2)U (t − 2)
L[ f (t )] = L[tU (t ) − 2(t − 1)U (t − 1) + (t − 2)U (t − 2)]
= L[tU (t )] − L[2(t − 1)U (t − 1)] + L[(t − 2)U (t − 2)]
1 2 −s 1 −2s
= − e + 2e
s2 s2 s
Solution:
2t 0 ≤ t <1
2 1≤ t < 4
− 2t 4≤t <5 2
Slope=2
f (t ) = 0 5≤t <6
− 2t 6≤t <7
0 1 4 5 6 7 8
2t 7≤t <8
0 t >8 -2
Solution:
f (t ) = 2tu (t ) − 2(t − 1)u (t − 1) − 2(t − 4)u (t − 4) + 2(t − 5)u (t − 5) − 2(t − 6)u (t − 6)
+ 2 × 2(t − 7)u (t − 7) − 2(t − 8)u (t − 8)
2 2 2 2 2 4 2
f ( s ) = 2 − 2 e −s − 2 e −4 s + 2 e −5 s − 2 e −6 s + 2 e −7 s − 2 e −8 s
s s s s s s s
Example:
U
1
0 1 3 4 5
T
Solution:
f (t ) = tu (t ) − (t − 1)u (t − 1) + 2(t − 3)u (t − 3) − 2(t − 4)u (t − 4) + (t − 5)u (t − 5)
1 1 − s 2 −3s 2 − 4 s 1 −5 s
f (s) = − e + 2e − 2e + 2e
s2 s2 s s s
P( s)
Often it is necessary to break down a complicated rational function of the form
Q( s)
(where P(s) and Q(s) are polynomials in s, and the degree of the top polynomial is
less than the degree of the bottom polynomial), into the sum of simpler fractions
called Partial Fractions.
The type of partial fraction that you use depends on the factors of the bottom
polynomial.
Case 1: All the factors of the bottom Q(s) are linear and non-repeating.
U
Case 1: All the bottom factors are linear (i.e. of the form x ± some number) and then
is no repeated factor (i.e. there is no factor which is squared or cubed, etc.) and there
are no irreducible quadratic terms (don’t worry about this!). In this case therefore, we
top polynomial
are talking about rational functions of the form:
( x − a )( x − b)...( x − g )
In this case we can rewrite the rational function as follows:
top polynomial A B G
= + + .... +
( x − a)( x − b)...( x − g ) x − a x − b x−g
Example
1 −1 1
Show that = 4 + 4
( s + 7)( s + 3) s + 7 s + 3
Solution
This is a case 1 partial fraction so we start with
1 A B
= +
( s + 7)( s + 3) s + 7 s + 3
Step1: Remove Fractions
U
Multiply both sides of the equation by the denominator on the left hand side
1 × ( s + 7)( s + 3) A( s + 7)( s + 3) B( s + 7)( s + 3)
= +
( s + 7)( s + 3) s+7 s+3
⇒ 1 = A( s + 3) + B( s + 7)
The equation above is true for all values of s. We can choose s values to make things
simple:
Choose s = -3 so that (s+3) = 0 and we have
1
1 = A(0) + B(4) ⇒ 1 = 4 B ⇒ B =
4
Choose s = -7 so that (s+7) = 0 and we have
1
1 = A(−4) + B (0) ⇒ 1 = −4 A ⇒ A = −
4
Step3: Substitute A and B into the original expression
U
3s − 1
Example: Write in partial fraction form
( s + 1)( s + 2)
Solution:
3s − 1 A B
write = +
( s + 1)( s + 2) s + 1 s + 2
Multiply both sides by (s+1)(s+2):
3s − 1 A B
( s + 1)( s + 2)= ( s + 1)( s + 2)+ ( s + 1)( s + 2)
( s + 1)( s + 2) s +1 s+2
⇒ 3s − 1 = A( s + 2)+ B( s + 1)
If choose s = -2
⇒ 3(− 2 ) − 1 = A(−2 + 2) + B (−2 + 1) ⇒ −7 = A.0+ B(−1) ⇒ −7 =− B
⇒B=7
If choose s = -1 ⇒ 3(− 1) − 1 = A(−1 + 2)+ B(−1 + 1) ⇒ −4 = A(1)+ B(0) ⇒ −4 = A
⇒ A = −4
3s − 1 −4 7
Therefore = +
( s + 1)( s + 2) s + 1 s + 2
Example
Find the fixed constants A, B, C so that the partial fraction decomposition can be
4s + 3 A B C
completed: = + +
s ( s − 1)( s + 3) s s − 1 s + 3
Solution:
4s + 3 A B C
= + +
s ( s − 1)( s + 3) s s − 1 s + 3
Multiplying both sides of this equation by the left hand side denominator we deduce
that,
4s + 3 A B C
s ( s − 1)( s + 3) =s ( s − 1)( s + 3) + s ( s − 1)( s + 3) + s ( s − 1)( s + 3)
s ( s − 1)( s + 3) s s −1 s+3
⇒ 4s + 3= A( s − 1)( s + 3) + Bs( s + 3)+ Cs ( s − 1)
Now in turn, put , s= 0 , s=1, s = -3. You will find at each stage that 2 of the A, B, C
terms will vanish:
If choose s = 0
4(0) + 3= A(0 − 1)(0 + 3) + B(0) + C (0) ⇒ 3=− 3 A ⇒ A = 1
If choose s = 1
7
4(1) + 3= A(0) + B(1)(1 + 3) + C (0) ⇒ 7 = 4 B ⇒ B =
4
If choose s = -3
−9
4(− 3) + 3= A(0)+ B(0) + C (−3)(−3 − 1) ⇒ −9 = 12C ⇒ B =
12
Therefore :
or ... below the line. When this occurs, you have to be careful as we have to use a
partial fraction for each of the powers. If you want, you can use the following table:
Factor in given Corresponding partial fraction
rational function
top polynomial A
s ± some number s ± some number
top polynomial A B
+
( s ± somenumber ) 2 s ± some number ( s ± some number ) 2
R R
top polynomial A B C
+ +
( s ± some number ) 3 s ± some number ( s ± some number ) 2
( s ± some number )3
Examples of case 2:
4 s 2 + 3s − 2 A B C
= + + , A, B, C = constants
( s − 1) ( s + 3) ( s − 1) s − 1 s + 3
2 2
3s 4 − 7 s 3 + 2 s 2 − 5 A B C D E F
= 3+ 2+ + + + A, B, C , D, E , F = constants
s ( s + 1) ( s + 5) s s s ( s + 1) ( s + 1) ( s + 5)
3 2 2
Example:
8s − 1
Write in terms of partial fractions
( s − 1) 2
Solution:
8s − 1 A B
In this case = +
( s − 1) 2
s − 1 ( s − 1) 2
Step1 we multiply both sides by ( s − 1) 2 to remove fractions
8s − 1 A B
( s − 1) 2 = ( s − 1) 2 + ( s − 1) 2 ⇒ 8s − 1 = A( s − 1) + B
( s − 1) 2
s −1 ( s − 1) 2
Step2: we can choose s = 1 as before so that (s-1) = 0 and we get
8 −1 = B ⇒ B = 7
We cannot choose another s value to directly find A however. There is more than one
approach to finding A but the easiest method is called “equating coefficients”. In this
case, we note that there must be the same “amount of s” on both sides of the equation.
On the left hand side we have 8s and on the right we have As, so that A must be 8.
Step3; write the answer down
8s − 1 8 7
= +
(s − 1) s − 1 ( s − 1)2
2
hand side of (1) the ( s + 2) 2 A term will contribute As2 if you multiply it out and the
P P
2
( s − 1)( s + 2) B term will contribute Bs . So we have
P P
1
0 = A+ B ⇒ B = −A = −
3
This method can involve more calculation though. Either way
2s + 1 1 1 1 B 1
= . − . +
(s − 1)(s + 2) 3 s − 1 3 s + 2 (s + 2)2
2
Example:
5s 2 − 7 s + 8
Write in partial fractions.
(s − 1)( s 2 − 2s + 5)
Note: Linear term on bottom means constant A on top. Quadratic term on bottom
means linear term Bs + C on top.
As before to find these constants multiply both sides by the denominator
( s – 1) (s2-2s+5).
P P
5s 2 − 7 s + 8 Bs + C
(s − 1)( s − 2s + 5)
2
=(s − 1)( s 2 − 2s + 5)
A
+ (s − 1)( s 2 − 2s + 5) 2
(s − 1)( s − 2s + 5)
2
(s − 1) ( s − 2s + 5)
Which gives: 5s 2 − 7 s + 8=( s 2 − 2s + 5) A+ (s − 1)(Bs + C )
The latter equation holds for all values of the variable s. So we can choose any
values for s and set up three simultaneous equations for A, B and C.
By putting s = 1 we can get one value easily:
Choose s = 1
3
⇒ 5(1) − 7(1) + 8=(12 − 2(1) + 5) A+ 0(B (1) + C ) ⇒ 6=4 A+ 0 ⇒ A=
2
2
We cannot make any more brackets = 0 by a good choice of s. As for case 2 however,
we can equate coefficients. Start with the highest power s2 first P P
Equate s2 ⇒ 5 = A + B ⇒ B = 5 − A ⇒ B = 5 − 32 = 72
P P
a
sinh at
s2 − a2
s
cosh at
s − a2
2
s − 2 s + ( −2 )
2
Example: Find the inverse Laplace transform for F ( s) = − .
3s − 4
Solution:
3s + 1
Example: Determine L−1 .
s − s − 6
2
Solution: This certainly did not appear in our list of standard transforms but if we
3s + 1
write as the sum of two simpler functions, i.e.
s −s−6
2
3s + 1 1 2
= + it makes all the difference.
s −s −6 s +2 s −3
2
This is simply the method of writing the more complex algebraic fraction in terms of
its partial fractions which we have previously seen.
We can now proceed to find
3s + 1 −1 1 2 −2t
L−1 2 =L + = e + 2e
3t
s − s − 6 s + 2 s − 3
5s + 1
Example: Determine L−1 .
s − s − 12
2
5s + 1 5s + 1
Solution: Factorise the denominator: = .
s − s − 12 ( s − 4)( s + 3)
2
9s − 8
Example: Determine L−1 .
s − 2s
2
9s − 8 −1 9 s − 8
Solution: Simplify: L−1 2 =L
s − 2s s ( s − 2)
Remember from partial fractions we have the form:
9s − 8 A B
= + × s ( s − 2)
s ( s − 2) s ( s − 2)
9s − 8 = A( s − 2) + Bs
s ( s − 2 ) s ( s − 2 )
s ( s − 2 )
13s + 11
Example: Determine L−1 .
( s − 1)( s + 3)
Solution: Remember from partial fractions we have the form:
13s + 11 A B
= + Multiple both sides by ( s − 1)( s + 3)
( s − 1)( s + 3) ( s − 1) ( s + 3)
13s + 11 = ( s + 3) A + ( s − 1) B
Let s=1 then 13 + 11 24
A= = =6
1+ 3 4
Let s=-3 then B = − 3 × 13 + 11 = − 28 = 7
− 3 −1 −4
So we now have to find
13s + 11 −1 6 7 −1 1 −1 1
L−1 =L + = 6L + 7L = 6e + 7e
t −3 t
( s − 1)( s + 3) ( s − 1) ( s + 3) ( s − 1) ( s + 3)
s+4
Example: Find the inverse Laplace transform of F ( s ) =
( s + 2)( s + 1) 2
s+4 A B C
Solution: Expand F(s) as F ( s ) = = + +
( s + 2)( s + 1) 2
s + 2 ( s + 1) ( s + 1)
2
( s + 4) = ( s + 1) 2 A + ( s + 2) B + ( s + 1)( s + 2)C
Let s=-1 then B = − 1 + 4 = 3 = 3
−1+ 2 1
Let s=-2 then A = − 2 + 42 = 2 = 2
(−2 + 1) 1
1− 7
equate s ⇒ 1 = 2 A + B + 3C ⇒ 1 = 2 × 2 + 3 + 3C ⇒ C = = −2
3
Check by cross-multiplying:
s+4 2 3 2
= + −
( s + 2)( s + 1) 2
( s + 2) ( s + 1) 2
( s + 1)
s+4 2 3 −2
= + +
( s + 1)( s + 3) 2
s + 2 ( s + 1) ( s + 1)
2
s + 4= 2( s + 2 s + 1) + 3( s + 2) − 2( s 2 + 3s + 2)
2
s 2 : 0= 2 − 2
s1 : 1 = 4 + 3 − 6
s0 : 4 = 2 + 6 − 4
1 −1 1 1
= 2 L−1 + 3L 2
− 2 L−1 − 2t −t
= 2e + 3te − 2e
−t
( s + 2) ( s + 1) ( s + 1)
s+9
Example: Find the inverse Laplace transform of F ( s) =
s + 6 s + 13
2
Solution:
s+9 s+9 s+9
f (t ) = L−1 [F ( s )] = L−1 2 = L−1 2 = L−1 2
s + 6s + 13 s + 6s + 3 − 3 + 13 ( s + 3) + 2
2 2 2
( s + 3) + 6 s+3 −1 6 s+3 −1 2
= L−1 2
= L−1 2
+L 2
= L−1 2
+ L 3 ⋅ 2
( s + 3) + 2 ( s + 3) + 2 ( s + 3) + 2 ( s + 3) + 2 ( s + 3) + 2
2 2 2 2 2
s+3 −3t
= e cos(2t ) + 3e −3t sin( 2t ) = e −3t [cos(2t ) + 3 sin( 2t )]
2
= L−1 2
+ 3L−1 2
( s + 3) + 2 ( s + 3) + 2
2 2
1
Example: Find the Laplace inverse of using
s ( s − 1)
a) partial fraction b)convolution
Solution:
a) Partial fraction
1 A B
L−1 = L−1 + Multiple both sides by s ( s − 1)
s ( s − 1) s −1 s
1 = As + B( s − 1)
Let s=1 then A=1
Let s=0 then B=-1
1 1 1 1 −1 1
L−1 = L−1 − = L−1 − L s = e − 1
t
s ( s − 1) s − 1 s s − 1
b) Convolution:
1 −1 1 1
t t
L−1 ∫ ∫
t −u
= L ⋅
s s − 1 = 1 * e t
= 1e du = e t −u du
s ( s − 1) 0 0
[ ] [ ] [ ]
t
= e ∫ e −u du = e t − e −u
t
t
= e t − e −t + e 0 = e t 1 − e −t = e t − 1
0
0
1
Example: Find the Laplace inverse of using convolution theorem
s ( s + 1) 2
2
Solution:
1 −1 1 1
t
L−1 2 ∫
−t
2
= L ⋅ 2
= t * te = u (t − u )e −(t −u ) du
s ( s + 1) s ( s + 1)
2
0
t
t
[ ] − ∫ 2ue du
t t t
= e −t ∫ u (t − u )e u du = e −t t ∫ ue u du − ∫ u 2 e u du = e −t t ∫ ue u du − u 2 e u
t u
0 0 0 0
0
0
0
0
0 0 0
= e −t (t + 2) ue
[ ] [ ]
0
[ [ ] ] [
− e − t 2 e t = e −t (t + 2) te t − e t + 1 − t 2 e t = e −t t 2 e t − te t + t + 2te t − 2e t + 2 − t 2 e t
u t
0
u t
]
−t
[ t t
]
= e te + t − 2e + 2 = t + te − 2 + 2e −t = te −t + 2e −t + t − 2
−t
du = − se − st v = ∫ f ' (t ) = f (t )
∞
L{ f ' (t )} = e [ − st
]
f (t ) t =0 + s ∫ e − st f (t )dt = (0 − f (0) ) + sF ( s )
∞
t =0
− st
Assuming e f (t ) → 0 as t → ∞
That is: L{ f ' (t )} = sF ( s) − f (0)
Thus, the Laplace transform of the derivative of f (t ) is given in terms of the Laplace
transform of f (t ) when t = 0 . The next properties is very important for the above
formula.
In general, to solve differential equation af ' (t ) + bf (t ) = g (t ) given that f (0) = k where
a, b, and k are known constants and g (t ) is a known expression in t using Laplace
transform are as follows:
i. Take the Laplace transform of both sides of the differential equation.
ii. Find the expression of F ( s) = L{ f (t )} in the form of an algebraic fraction
iii. Separate F(s) into its partial fractions.
iv. Find the inverse Laplace transform L{ f ' (t )} to find the solution f(t) to the
differential equation.
Example:
Find the solution of f ' ' (t ) + 3 f ' (t ) + 2 f (t ) = 4t where f (0) = f ' (0) = 0.
Solution:
i. Take the Laplace transform of both sides of the equation
L{ f ' ' (t )} + 3L{ f ' (t )} + 2 L{ f (t )} = 4 L{t}
d 2 y (t ) d yt
()
Example: Solve 2
+4 + 4 y (t ) = e −3t u (t ) y (0) = 1, y '(0) = 0
dt dt
1
Taking the Laplace s 2Y ( s) − sy (0 )+ 4 [ sY ( s) − y (0 ])+ 4Y ( s ) =
s+3
1
Which gives Y ( s) s 2 + 4s + 4 = s + 4 +
s+3
1
s+4+
Y (s) = s+3
( s + 2)
2
(s + 4) 1
Y(s) = +
(s + 2) 2
(s + 2) 2 (s + 3)
This gives me two different terms, but they’re simpler.
Y=( s ) Y1 ( s ) + Y2 ( s )
s+4 1
=Y1 ( s ) = , Y2 ( s )
( s + 2) ( s + 2 ) ( s + 3)
2 2
First term:
1 2
= +
( s + 2 ) ( s + 2 )2
y1 (t ) = e −2t + 2te −2t
Now look at the second term:
1 C D E
Y2 ( s ) = = + +
( s + 2 ) ( s + 3) s + 3 ( s + 2 ) ( s + 2 )
2 2
1 1
=
C = = 1
( s + 2)
2
12
s = −3
1
=D = 1
( s + 3) s = −2
d 1 −1 −1
E= = 2
= = −1
ds ( s + 3) ( s + 3) s = −2 1
s = −2
Check:
1 1 −1
Y2 ( s ) = + + =
s + 3 ( s + 2) ( s + 2)
2
1( s 2 + 4 s + 4) + ( s + 3) − ( s 2 + 5s + 6) 1
=
( s + 3)( s + 2 ) ( s + 3)( s + 2 )
2 2
so
y 2 (t ) = e −3t + te −2t − e −2t
And putting the two solutions together:
y (t ) = y1 (t ) + y 2 (t ) = [2te −2t + e −2t ] + [e −3t + te −2t − e −2t ] = e −3t + 3te −2t
−t −t
Therefore, y1 (t ) = e cos t and y2 (t ) = − e sin t .
[ ]
⇔ ( s − 5) 2 − 1 y1 ( s) = 7 − 3 ( s − 5) and y2 ( s) =
7 + y1 ( s )
( s − 5)
7 s −5 7 7 3
⇔ y1 ( s) = −3 and y2 ( s ) = + −
( s − 5) − 1
2
( s − 5) − 1
2 2
( ) 2
s − 5 ( s − 5) ( s − 5) − 1 ( s − 5) − 1
7
( 2
=
A
)+
B + sC
( s − 5) ( s − 5) − 1 s − 5 ( s − 5) − 1 2
(
multiple both side ( s − 5) ( s − 5) 2 −1 )
( )
7 = A ( s − 5) − 1 + ( B + sC )( s − 5 )
2
Let s=5 ⇒ A = −7
− 168 − 7
Let s=0 ⇒ 7 = A(25 − 1) − 5 B ⇒ B = = −35
5
s2 ⇒ 0 = A + C ⇒ C = − A = 7
7 7 7 s − 35 3
y2 ( s ) = − + −
( )
s − 5 ( s − 5) ( s − 5) − 1 ( s − 5) − 1
2 2
s −5 3
y2 ( s ) = 7 − .
( s − 5) − 1 ( s − 5) − 1
2 2
s −5 −1 1
Now, L − 1 = e
5t
cosh t and L
5t
= e sinh t
( s − 5) −1 ( s − 5) −1
2 2
Example: Solve y1 ' ' = y1 + 3 y2 , y2 ' ' = 4 y1 − 4 et , y1 (0) = 2 , y1 ' (0) = 3 , y2 (0) =1 and y2 ' (0) = 2
Solution:
The subsidiary equations become
3 + 2s + 3 y2 ( s )
y1 ( s ) = and [ s 4 − s 2 − 12] y2 ( s) = s 3 + 2s 2 + 3s + 6
( s − 1)
2
3 + 2s + 3 y2 ( s ) s 3 + 2 s 2 + 3s + 6 1
y1 ( s ) = and y2 ( s) = =
( s 2 − 1) [ s − s − 12]
4 2
( s − 2)
1
3 + 2s + 3
( s − 2) (3 + 2 s )( s − 2) + 3 3s − 6 + 2 s 2 − 4 s + 3 2s 2 − s − 3
y1 ( s ) = = = =
( s 2 − 1) ( s − 2)( s 2 − 1) ( s − 2)( s 2 − 1) ( s − 2)( s 2 − 1)
s 2 + s 2 − s − 1 − 2 ( s 2 − s − 2) + ( s 2 − 1) ( s + 1)( s − 2) + ( s 2 − 1) ( s + 1) 1
= = = = 2 +
( s − 2)( s − 1)
2
( s − 2)( s − 1)
2
( s − 2)( s − 1)
2
( s − 1) ( s − 2)
1 1
= +
( s − 1) ( s − 2)
1 1 1
⇔ y2 ( s ) = and y1 ( s) = +
s−2 s −1 s − 2
Therefore, y1 (t ) = et + e 2t and y2 (t ) = e 2t .
Proof:
∞
t
t
∞
− st − 1 − st t 1
t
1
L ∫ f (t )dt = ∫ ∫ f (t ) e dt = e ∫ f (t ) + ∫ f (t )e − st dt = F ( s )
0 00 s 0 0 s 0 s
Example:
U
Find the final value of the function x(t) for which the laplace inverse is:-
1
x( s ) =
s( s 3 + 3s 2 + 3s + 1 )
s ×1
lim x( t ) = lim sx( s ) = lim
t →∞ s →0 s →0 s( s + 3s 2 + 3s + 1 )
3
1
= lim =1
s →0 ( s 3
+ 3s + 3s + 1 )
2
Example:
5s + 2
Suppose Y ( s) =
s (5s + 4)
Then steady state value of Y can be calculated by:
(5s + 2)
Y (∞) = lim Y (t ) = lim s = 0.5
t →∞ s →0
s (5s + 4)
Before studying the control system it is necessary to become familior with the
response os some of simple basic systems (i.e study the dynamic behaviour of the
first and second order systems).
y(s)
T. F = G(s) =
x(s)
This definition is applied to linear systems
Mercury Thermometer:
U
It is assumed that the thermometer is initially at steady state. This means that,
before time zero, there is no change in temperature with time. At time zero the
Process Control /Lec. 3 33 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
thermometer will be subjected to some change in the surrounding temperature x(t).
(i.e at t<0 , x(t)= y(t) =constant there is no change in temperature with time).
At t=0 there is a change in the surounding temperature x(t)
Unsteady state energy balance:
U
dy
m Cp = h A (x − y) − 0 = h A x − h A y ……………… (1)
dt
1st order differential equation
P P
Where
U
𝑑𝑑𝑦𝑦 𝑠𝑠
𝑚𝑚 𝑐𝑐𝑐𝑐 = ℎ 𝐴𝐴 (𝑥𝑥𝑠𝑠 − 𝑦𝑦𝑠𝑠 𝑠𝑠) = ℎ 𝐴𝐴𝑥𝑥𝑠𝑠 − ℎ 𝐴𝐴𝑦𝑦𝑠𝑠 ……………… (2)
𝑑𝑑𝑑𝑑
Substract eq. (2) from eq. (1)
𝑑𝑑(𝑦𝑦 − 𝑦𝑦𝑠𝑠 )
𝑚𝑚 𝑐𝑐𝑐𝑐 = ℎ 𝐴𝐴 (𝑥𝑥 − 𝑥𝑥𝑠𝑠 ) − ℎ 𝐴𝐴 (𝑦𝑦 − 𝑦𝑦𝑠𝑠 𝑠𝑠)
𝑑𝑑𝑑𝑑
𝑦𝑦 − 𝑦𝑦𝑠𝑠 = 𝑌𝑌 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑥𝑥 − 𝑥𝑥𝑠𝑠 = 𝑋𝑋
at t=0 Y(0)=0 and X(0)=0
𝑑𝑑𝑑𝑑
𝑚𝑚 𝑐𝑐𝑐𝑐 = ℎ 𝐴𝐴 𝑋𝑋 − ℎ 𝐴𝐴 𝑌𝑌
𝑑𝑑𝑑𝑑
𝑚𝑚 𝑐𝑐𝑐𝑐 𝑑𝑑𝑑𝑑
= 𝑋𝑋 − 𝑌𝑌
ℎ 𝐴𝐴 𝑑𝑑𝑑𝑑
𝑚𝑚 𝑐𝑐𝑐𝑐
Let 𝜏𝜏 = ℎ 𝐴𝐴
=time constant and has time units
𝑑𝑑𝑑𝑑
𝜏𝜏 + 𝑌𝑌 = 𝑋𝑋 taken laplace for the equation
𝑑𝑑𝑑𝑑
𝜏𝜏[𝑠𝑠𝑠𝑠(𝑠𝑠) − 𝑌𝑌(0)] + 𝑌𝑌(𝑠𝑠) = 𝑋𝑋(𝑠𝑠)
(𝜏𝜏𝜏𝜏 + 1)𝑌𝑌(𝑠𝑠) = 𝑋𝑋(𝑠𝑠)
T.F relates two variables in a physical process. One of these is (Forcing or Input
variable) and the other is the effect (Reponce or Output).
𝑌𝑌(𝑠𝑠)
𝑇𝑇. 𝐹𝐹 = = 𝐺𝐺(𝑠𝑠)
𝑋𝑋(𝑠𝑠)
If we select a particular input variation x(t) for which the L.T is x(s) then the reponce.
𝑌𝑌(𝑠𝑠) = 𝐺𝐺(𝑠𝑠). 𝑋𝑋(𝑠𝑠)
1
𝑌𝑌(𝑠𝑠) = 𝐺𝐺(𝑠𝑠). 𝑋𝑋(𝑠𝑠) = . 𝑋𝑋(𝑠𝑠)
𝜏𝜏𝜏𝜏 + 1
x(s) G(s) y(s)
1
𝑌𝑌(𝑠𝑠) = . 𝑋𝑋(𝑠𝑠)
𝜏𝜏𝜏𝜏 + 1
Y(t)=? For different types of x(t)
1-Step Change
U
A
X (s) =
s
A ----------
1 A αo α x(t)
Y (s) = . = + 1
τs + 1 s s τs + 1
A = α o ( τs + 1 ) + α1s t
S = 0 ⇒ αo = A
1
s = −1 / τ ⇒ A = α o (−τ / τ + 1) − α1 then α1 = − Aτ
τ
A Aτ A Aτ 1 / τ A A
A --------Ultimat value
Y ( s) = − = − = −
s τs + 1 s τs + 1 1 / τ s τ + 1/τ y(t)
Process Control /Lec. 3 35 Writtent by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
Y (t ) = A − Ae − t τ = A(1 − e − t τ )
Several features of this response, worth remembering, are
• The value of y(t) reaches 63.2 % of its ultimate value when the time elapsed is
equal to one time constant 𝜏𝜏.
• When the time elapsed is 2𝜏𝜏, 3 𝜏𝜏, and 4 𝜏𝜏, the percent response is 86.5%, 95%,
and 98%, respectively.
Where the ultimate value is final steady state value
U .V = lim y( t ) = lim sy( s )
t →∞ s →0
Example:
A thermometer having a time constant of 0.1 min is at a steady state temperature of
90 Fo. At time t = 0, the thermometer is placed in a temperature bath maintained at
P P
100°F. Determine the time needed for the thermometer to read 98 Fo. P P
Solution:
At s.s. x s =y s =90 Fo
R R R R P
A
Step change X (s) =
s
A=100-90=10
10
X( s ) =
s
1 A 1 10 10 10 A B
Y( s ) = = = = = +
τs + 1 s 0.1s + 1 s s( 0.1s + 1 ) 0.1s( s + 10 ) 0.1s s + 10
A( s + 10 ) + B( 0.1s ) = 10
10
s = 0 ⇒ A = =1
10
s = −10 ⇒ B = −10
1 10 10 10
Y( s ) = − = −
0.1s s + 10 s s + 10
By taken laplace inverse for the equation
Y ( t ) = 10 − 10e −10t = 10( 1 − e −10t )
Substitute Y(t)=y(t)-y s =98-90 R R
Y(t)=8
8 = 10( 1 − e −10t )
0.8 = 1 − e −10t
ln( e −10t ) = ln( 0.2 )
− 10t = ln( 0.2 )
t = − ln(0.2) × 0.1
t=0.161 min
2-Impulse Input
x(t)
Process Control /Lec. 3 36 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
Area
X ( s) = A
1 A
Y (s) = ×A=
τs + 1 τs + 1
Aτ
Y (s) =
s +1 τ
A
Y (t ) = y (t ) − ys = e−t τ
τ
A y(t)
y (t ) = e −t τ + y s
τ
t
3-Sinsoidal input
x( t ) = xs + A sin wt t >0
x( t ) − xs = A sin wt
X (t ) = x(t ) − xs = A sin wt
Aw
X( s ) = 2
s + w2
Aw 1 1
Y( s ) = 2 × = Aw[ 2 ]
s + w ( τs + 1 )
2
( s + w )( τs + 1 )
2
This equation can be solved for y(t) by means of a partial fraction expansion as
described in previous lectures.
1 α s + α1 α2
Y ( s ) = Aw[ 2 ] = Aw[ o2 + ]
( s + w )( τs + 1 )
2
( s + w ) ( τs + 1 )
2
( α o s + α1 )( τs + 1 ) + α 2 ( s 2 + w 2 ) = 1
α o τs 2 + α o s + α1τs + α1 + α 2 s 2 + α 2 w 2 = 1
s0 α1 + α 2 w 2 = 1 (4)
s1 α o + α1 τ = 0 ⇒ α o = − α1 τ (5)
s2 α o τ + α 2 = 0 ⇒ α 2 = −α o τ (6)
By substitution eq.(5) in eq.(6)
α 2 = α1 τ 2 (7)
By substitution eq.(7) in eq.(4)
α1 + α1 τ 2 w 2 = 1
1
α1 =
1 + τ2 w2
Aw tτ 1 + w2τ 2
Y (t ) = [τe + sin( wt + φ )]
1 + τ 2 w2 w
Awτ t τ A
Y (t ) = e + sin( wt + φ )]
1 + τ 2 w2 1+ w τ
2 2
where
φ = tan −1 (− wτ )
A X(t) _______T_______
-A
By comparing Eq. for the input forcing Y(t) function with Eq. for the ultimate
periodic response X(t), we see that
1. The output is a sine wave with a frequency w equal to that of the input signal.
1
2. The ratio of output amplitude to input amplitude is < 1.
1 + w2 τ2
3. The output lags behind the input by an angle φ . It is clear that lag occurs, for the
sign of φ is always negative.
φ<0 phase lag
φ>0 phase load
Example:
A mercury thermometer having a time constant of 0.1 min is placed in a temperature
bath at 100°F and allowed to come to equilibrium with the bath. At time t= 0, the
temperature of the bath begins to vary sinusoidally about its average temperature of
l00oF with an amplitude of 2°F If the frequency of oscillation is 10/𝜋𝜋 cycles/min, plot
P P P P
the ultimate response of the thermometer reading as a function of time. What is the
phase lag?
In terms of the symbols used in this chapter
τ = 0.1
t<0 xs = y s = 100
t≥0 x( t ) = 100 + 2 sin( wt )
10
f =
π
Process Control /Lec. 3 39 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
Solution
U
10
w = 2πf = 2π × = 10 rad / min
π
1 10
T= = min/ cycle
f π
X ( t ) = x( t ) − x s = 100 + 2 sin 20t − 100
X ( t ) = 2 sin 20t
2 × 20
X( s ) = 2
s + 20 2
Ultimate response t → ∞ then e −t τ = 0
A
Y( t ) = sin( wt + φ )]
1 + w2 τ 2
φ = tan −1( − wτ ) = tan −1( −20 × 0.1 ) = tan −1( −2 )
φ = −63.5o
Ultimate response at the above angle
2
Y( t ) = sin( 20t − 63.5 )]
1 + ( 0.1 × 20 )2
2
Y( t ) = sin( 20t − 63.5 )]
5
Y ( t ) = 0.896 sin( 20t − 63.5 )] Ultimate response
10 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 10
A frequency of means that a complete cycle occurs in ( )−1 𝑚𝑚𝑚𝑚𝑚𝑚. since cycle
𝜋𝜋 𝑚𝑚𝑚𝑚𝑚𝑚 𝜋𝜋
o o
is equivalent to 360 and lag is 63.5
P P P
How to calculate the time constant (𝝉𝝉) for first order system
U
1) Mathematical method
U
0.63 A
τ t
Y (t ) = A(1 − e −t τ )
as t → ∞ Y (∞) = A(1 − e −∞ ) = A
as t → τ Y (τ ) = A(1 − e −1 ) = A(1 − 0.3678) ≈ 0.63 A
Time constant (𝜏𝜏)is the time required for the response to reach 63% of the its utimate
value.
3. Third method
U
Y ( t ) = A( 1 − e −t τ ) A
dy A
= − Ae −t τ ( −1 τ ) = e −t τ
dt τ
Y(t)slope
dy A A
lim = e 0 =
t →0 dt τ τ
A t
Slope of the tangent at t=0 is τ
A
Therfore τ =
slope
Y ( t ) = A( 1 − e −t τ )
Y ( t ) = A − Ae −t τ
Ae −t τ = A − Y ( t )
A − Y( t )
e −t τ =
A
A − Y( t )
− t τ = ln
A
A − Y( t )
Let B =
A
Process Control /Lec. 3 41 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
−1
ln B = t
τ
let y = ln B and x = t
1
slope = −
τ
Y(t) A − Y( t ) ln B t
B= lnB
A 1
slope = −
τ
1
slope = −
τ
1
τ =−
slope
1.Mercury Thermometer
Find the T.F. that relates the head to the input flowrate q(t).
We can analyze this system by writing a transient mass balance around the tank:
Mass flow in - mass flow out = rate of accumulation of mass in the tank
dvρ
qρ − qo ρ = (v = Ah)
dt
dh
q − qo = A
dt
h dh
q− = A ………………(2)
R dt
At staedy state
hs dh
qs − = A s =0 ………………(3)
R dt
Substracting Eq(3) from Eq. (2)
h − hs d (h − hs )
( q − qs ) − =A
R dt
let Q = q − q s , H = h − hs
R
H (s) = Q( s) …………………(5)
(τs + 1)
Comparing the T.F. of the tank with thr T.F. of the thermometer we see that eq. (5)
contain the factor (R) which is relate H(t) to Q(t) at s.s. as s → 0 , t → ∞
R
For this reason, a factor R in the transfer function is called the steady state gain
τs + 1
To show that
1
Take Q( s) =
s
R 1
Therefore H ( s) = .
τs + 1 s
Final value theorm
R 1 R
lim H (t ) = lim sH ( s ) = lim s . = lim = R s.s gain
t →∞ s →0 s →0 τs + 1 s s →0 τs + 1
3.Mixing Tank
Example:
U
R R
When F 2 constant ⇒ F2 ( s ) = 0 ⇒ h ( s ) = F1 ( s )
(τs + 1) (τs + 1)
R R
Subtract equations
dT d (T − Ts )
VρC = VρC = wC[(Ti − Tis ) − (T − Ts )] + (Q − Qs )
dt dt
Define some important new variables(Deviation variables).
T = T − Ts , Ti = Ti − Tis , Q = Q − Qs
By substituting deviation variables for variables.
dT
VρC = wC (Ti − T ) + Q
dt
Vρ dT
= Ti − T + Q / wC
w dt
Let k = 1 wC ,τ = Vρ w
Apply Laplace Transform.
τsT ( s) = (Ti ( s ) − T ( s )) + kQ ( s )
(τs + 1)T ( s ) = Ti ( s ) + kQ ( s )
1 k
T (s) = Ti ( s ) + Q (s)
τs + 1 τs + 1
T ( s) k
If Ti ( s) = 0 G1 ( s) = =
Q ( s ) τs + 1
T (s) 1
If Q ( s) = 0 G2 ( s) = =
Ti ( s ) τs + 1
dC A
V = FC Ai − FC A − VKC A
dt
dC A,ss
at steady state V = FC Ai,ss − FC A,ss − VKC A,ss
dt
by subtracting both equations
d (C A − C A,ss )
V = F (C Ai − C Ai ,ss ) − F (C A − C A,ss ) − VK (C A − C A,ss )
dt
dC A
V + ( F + VK )C A = F CAi ÷ ( F + VK )
dt
V dC A F
+ CA = CAi
( F + VK ) dt ( F + VK )
Let τ = V F
R
and K =
( F + VK ) ( F + VK )
dC A
τ + C A = K CAi
dt
Taking Laplace Transform:
(τs + 1)C A ( s ) = K CAi ( s )
K CAi ( s )
C A ( s) =
(τs + 1)
K w b cs + d
C A (s) = = + 2
(τs + 1) s + w
2 2
(τs + 1) s + w 2
Multiply both sides by (τs+1)(s2+w2) gives: P P P P
+ w2 w 2τ 2 + 1
τ 2
Kw 3τ 2 Kw 3τ 2 + Kw - Kw 3τ 2 Kw
Let s = 0 ⇒ d = Kw - bw 2 = Kw − = = 2 2
R R
w τ +1
2 2
w τ +1
2 2
w τ +1
Equating the coefficients of each power of s2, yields: P P
-b - Kwτ
s2 : b + cτ = 0 ⇒ c = =
τ w 2τ 2 + 1
Substituting the constants in the main equation:
Kw τ2 τs 1 Kw τ τs 1
C A ( s) = 2 2 [ − 2 + 2 ]= 2 2 [ − 2 + 2 ]
w τ + 1 τs + 1 s + w 2
s +w 2
w τ +1 s +1 τ s + w 2
s + w2
Using Laplace Table:
−t
Kw 1
C A (t ) = [τe τ − τ cos( wt ) + sin( wt )]
w τ +1
2 2
w
h1
In fig (5.1 a) variation of h 2 does not effect on q 1 then q1 =
R R R R
R1
h1 − h2
In fig (5.1 b) variation of h 2 does effect on q 1 then q1 =
R R R R
R1
R2 R1 R
H 2( s ) = × Qi ( s ) 1
( τ2 s + 1 ) τ1s + 1
R2
H 2( s ) = Qi ( s )
( τ1s + 1 )( τ 2 s + 1 )
H 2( s ) R2
=
Qi ( s ) ( τ1s + 1 )( τ 2 s + 1 ) Non-interacting system
In the case of three non-interacting tanks in sereies the transfer function of the system
will be as below:-
H 3( s ) R3
=
Qi ( s ) ( τ1s + 1 )( τ 2 s + 1 )( τ3 s + 1 )
Process Control /Lec. 5 49 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
Example:
Two non-interacting tanks are connected in series as shown in Fig. 5.1 a. The time
constants are τ 2 =1 and τ 1 =0.5; R 2 =1. Sketch the response of the level in tank 2 if a
R R R R R R
The transfer function for this system is found directly from Equation above thus
R2
H 2( s ) = Qi ( s )
( τ1s + 1 )( τ 2 s + 1 )
1
Substituting Qi ( s ) = Unit step change in Q i R
s
R2 1
H 2( s ) =
( τ1 s + 1 )( τ 2 s + 1 ) s
αo α1 α2
= + +
s ( τ1 s + 1 ) ( τ 2 s + 1 )
R2 = α o ( τ1s + 1 )( τ 2 s + 1 ) + α1s( τ 2 s + 1 ) + α 2 s( τ1s + 1 )
let s = 0 ⇒ α o = R2
1 1 1 τ 1 τ −τ
let s = − ⇒ α1( − )( τ 2 ( − ) + 1 ) = R2 ⇒ α1( 22 − ) = R2 ⇒ α1( 2 2 1 ) = R2
τ1 τ1 τ1 τ1 τ1 τ1
τ12
∴ α1 = R2 ( )
τ 2 − τ1
1 1 1 τ 1 τ −τ
let s = − ⇒ α 2 ( − )( τ1( − ) + 1 ) = R2 ⇒ α 2 ( 12 − ) = R2 ⇒ α 2 ( 1 2 2 ) = R2
τ2 τ2 τ2 τ2 τ2 τ2
τ 22
∴ α 2 = R2 ( )
τ1 − τ 2
R2 τ12 1 τ 22 1
H 2( s ) = + R2 ( ) + R2 ( )
s τ 2 − τ1 ( τ1s + 1 ) τ1 − τ 2 ( τ 2 s + 1 )
1 ττ τ 1 ττ τ 1
H 2 ( s ) = R2 [ + ( 1 2 ) 1 +( 1 2 ) 2 ]
s τ 2 − τ1 τ 2 ( τ1s + 1 ) τ1 − τ 2 τ1 ( τ 2 s + 1 )
1 ττ 1 1 ττ 1 1
H 2 ( s ) = R2 [ − ( 1 2 ) +( 1 2 ) ]
s τ1 − τ 2 τ 2 ( s + 1 / τ1 ) τ1 − τ 2 τ1 ( s + 1 / τ 2 )
τ 1τ 2 1 1
H 2 (t ) = R2 (1 − ( )( e −t / τ − e −t / τ ))
1 2
τ1 −τ 2 τ 2 τ1
1 × 0.5 1 −t / 0.5 1 −t / 1
H 2( t ) = 1 − ( )( e − e )
0.5 − 1 1 0.5
H 1( t ) = 1( 1 − e −t / 0.5 )
H 1( t ) = 1 − e −2t
Example:
U
dt
Laplace transform →τ 1 s C1 (s) + C1 (s) = Ci (s)
C1 ( s ) 1
= 1
Ci ( s ) τ 1 s + 1 …. (1) Ci (s)
τ 1s + 1
C1 (s)
dC 2
Tank 2: F1 C1 + Fii C ii - F2 C 2 = V2
dt
V2 dC 2 F F
+ C 2 = 1 C1 + ii C ii
F2 dt F2 F2
dC 2
τ2 + C 2 = K 1C1 + K 2 C ii
dt
Process Control /Lec. 5 51 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
V2 F1 Fii
τ2 = , K1 = , K2 =
F2 F2 F2
K1 K2
C 2 ( s) = Ci (s) + C ii ( s )
(τ 1 s + 1)(τ 2 s + 1) τ 2s +1
2.Interacting System
dh
qi − q1 = A1 1
dt
h −h dh
qi − 1 2 = A1 1
R1 dt
Steady state
h − h2 s dh
qis − 1s = A1 1s = 0
R1 dt
By substracting both equations
h −h h − h2 s d ( h1 − h1s )
( qi − qis ) − 1 1s + 2 = A1
R1 R1 dt
H 2 H1 dH
[ Qi + = + A1 1 ] × R1
R1 R1 dt
dH 1
Qi R1 + H 2 = H 1 + A1 R1
dt
dH 1
τ1 + H 1 = Qi R1 + H 2
dt
( τ1s + 1 )H 1( s ) = R1Qi ( s ) + H 2 ( s )
R1 1
H1( s ) = Qi ( s ) + H 2( s ) …………..(1)
( τ1s + 1 ) ( τ1s + 1 )
R R1 1
( τ 2 s + 1 )H 2 ( s ) = 2 [ Qi ( s ) + H 2 ( s ) − H 2 ( s )]
R1 ( τ1s + 1 ) ( τ1s + 1 )
R2 Qi ( s ) R2 H 2 ( s ) R2
[( τ 2 s + 1 )H 2 ( s ) = + − H 2 ( s )] × ( τ1 s + 1 )
( τ1 s + 1 ) R1 ( τ1 s + 1 ) R1
R2 ( τ s + 1 ) R2
( τ 2 s + 1 )( τ1 s + 1 )H 2 ( s ) = R2 Qi ( s ) + H 2( s ) − 1 H 2 ( s )]
R1 R1
τ1 R2 s
( τ1τ 2 s 2 + τ1s + τ 2 s + 1 )H 2 ( s ) + H 2 ( s ) = R2Qi ( s )
R1
τ1 R2 A1 R1 R2
Let = = A1 R2 = τ12
R1 R1
( τ1τ 2 s 2 + ( τ1 + τ 2 + τ12 )s + 1 )H 2 ( s ) = R2Qi ( s )
R2
H 2( s ) = .Qi ( s ) Interacting system
τ1τ 2 s 2 + ( τ1 + τ 2 + τ12 )s + 1
R2
H 2( s ) = .Qi ( s ) Non- Interacting system
τ1τ 2 s 2 + ( τ1 + τ 2 )s + 1
The difference between the transfer function for the non-interacting system, and that
for the interacting system, is the presence of the cross-product term A 1 R 2 in the R R R R
coefficient of s. τ12 = A1 R2
τ 1 = τ 2 = τ 12 =τ
R R R R R R
1
Qi ( s ) =
s
Solution:
Non-interacting system
H 2( s ) R2
= τ1 = τ 2
Qi ( s ) τ1τ 2 s 2 + ( τ1 + τ 2 )s + 1
H 2( s ) R H (s)
= 2 2 2 but Q2 ( s ) = 2
Qi ( s ) τ s + 2τs + 1 R2
Q2 ( s ) 1 1 1 2
= 2 2 = =( )
Qi ( s ) τ s + 2τs + 1 ( τs + 1 )( τs + 1 ) τs + 1
1
If Qi ( s ) =
s
1 1 αo α1 α2
Q2 ( s ) = . = + +
( τs + 1 )2 s s ( τs + 1 )2 τs + 1
By multiplying both sides by s( τs + 1)2 and expanding, we get
α o ( τs + 1 )2 + α1s + α 2 s( τs + 1 ) = 1
α o ( τ 2 s 2 + 2τs + 1 ) + α1s + α 2 ( τs 2 + s ) = 1
s 2 ( α o τ 2 + α 2 τ ) + s( 2τα o + α1 + α 2 ) + α o = 1
s0 ⇒ αo = 1
s2 ⇒ α o τ 2 + α 2 τ = 0 ⇒ τ 2 + α 2 τ = 0 ⇒ α 2 = −τ
s1 ⇒ 2α o τ + α1 + α 2 = 0 ⇒ 2τ + α1 − τ = 0 ⇒ α1 = −τ
1 τ τ
Q2 ( s ) = − −
s ( τs + 1 ) 2
τs + 1
1 τ τ
Q2 ( s ) = − −
s ( τs + 1 )2 τs + 1
1 1 1 1
Q2 ( s ) = − −
s τ ( s + 1 / τ )2 s + 1 / τ
t
Q2 ( t ) = 1 − e −t / τ − e −t / τ for non-interacting
τ
If the tanks are interacting, the overall transfer function, according to Equation of
interacting system (assuming further that A 1 =A 2 ) R R R R
1 1
Q2 ( s ) = 2 2 .
τ s + 3τs + 1 s
By application of the quadratic formula, the denominator of this transfer function can
be written as
1 1
Q2 ( s ) =
s ( 0.38τs + 1 )( 2.62τs + 1 )
α α1 α2
Q2 ( s ) = o + +
s 0.38τs + 1 2.62τs + 1
let s = 0 ⇒ α o = 1
1 1
let s = − ⇒ α1 = −0.38τ = 0.0664τ
0.38τ 2.62τ( −
1
) +1
0.38τ
1 1
let s = − ⇒ α1 = −2.62τ = −3.664τ
2.62τ 0.38τ( −
1
) +1
2.62τ
1 0.0664τ 3.0664τ
Q2 ( s ) = + −
s 0.38τs + 1 2.62τs + 1
1 0.0664τ / 0.38τ 3.0664τ / 2.62τ
Q2 ( s ) = + −
s s + 1 / 0.38τ s + 1 / 2.62τ
1 0.17 1.17
Q2 ( s ) = + −
s s + 1 / 0.38τ s + 1 / 2.62τ
1-Linearization method
Making the non-linear function as linear using Taylar series and give approximate
results.
2-Non-linear solution
It is difficult and give exact solution
Linearization Technique
y = 2t
Linear(all terms to power =1)
y =t + x
y = t2
y=2 t Non-Linear (power ≠1)
y = ln x
Neglacting the non linear terms because their value are very small.
Then
d f ( x)
f ( x ) = f ( xo ) + ( x − xo )
dx x = xo
The flow of water through a valve or other construction usually follow a square-root
law.
h qi
If qo = Linear valve
R
qo = c h Non-linear valve
c is a constant
h
qo
d ( ρV ) dh
qi ρ − q o ρ = = ρA
dt dt 𝒒𝒒𝒐𝒐 = 𝒄𝒄√𝒉𝒉
dh
qi − ch1 / 2 = A ……..(1)
dt
q o may be expanded around the s.s. value h s using linearization method
R R R R
1 1
h1 / 2 = hs1 / 2 + ( h − hs )
2 h1S / 2
1 1 dh
qi − c hs1 / 2 + 1/ 2
( h − hs = A
) …….. (2)
2 h s dt
[ ] dh
qis − c hs1 / 2 = A s = 0
dt
at s.s h=h s …….. (3)
R R
1 1 d ( h − hs )
( qi − qis ) − c 1 / 2 ( h − hs ) = A
2 hs dt
c 1 dH
Qi − 1 / 2 H = A
2 hs dt
c 1
assume =
2 hs R
H dH
∴ Qi − =A
R dt
dH
RA + H = RQi
dt
By taking laplace transform
( τs + 1 )H ( s ) = RQi ( s )
H( s ) R
= ………….. 1st order system
Qi ( s ) τs + 1 P P
2 hs
Where R = and τ = AR
c
Process Control /Lec. 6 57 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
1-Transfer function is similar to linear.
2- R depends on the s.s. condition (at steady state the flow entering the tank equals to
the flow leaving the tank,then q o =q os ).
R R R R
V: Constant=L V
F: Constant=L/min 2A → B + C
Solution:
In - out - rate of reaction = accumulation
dc
Fci − Fc − Vkc 2 = V Un-steady state
dt
c 2 = cs2 + 2cs ( c − c s )
dc
Fci − Fc − Vk [ cs2 + 2cs ( c − c s )] = V
dt
dc s
Fcis − Fcs − Vkcs2 = V = 0 Steady state c=c s R
dt
d ( c − cs )
F ( ci − cis ) − F ( c − cs ) − Vk [ 2cs ( c − c s )] = V
dt
dC
V + ( F + 2Vkcs )C = FCi ÷ ( F + 2Vkcs )
dt
dC
τ + C = RCi
dt
C( s ) R
= 1st order system
Ci ( s ) τs + 1
Where C = c − cs , Ci = ci − cis
V F
τ= , R=
F + 2Vkcs F + 2Vkcs
th
Consider the n order nonlinear differential equation f ( x1 , x2 ..., xn ) = u( t )
Expanding the nonlinear function in a Taylor series about an operating point that
xio , i =1,2...,n satisfies the original differential equation and retaining only the linear
terms yields:-
df ( x1 , x2 ,...xn )
f ( x1 , x2 ,..., xn ) ≈ f ( x1 , x2 ,..., xn ) x = x0 + ( x1 − x1o )
i i dx1 x = x0i i
df ( x1 , x2 ,...xn ) df ( x1 , x2 ,...xn )
+ ( x2 − x2o ) + + ( xn − xno )
dx2 x = x0
dxn x = x0
i i i i
Example:
U
Mixing tank F, ci
C( s )
G( s ) = =? F, c
F( s )
c : Variable (kg/L)
F: Variable (L/min)
C i : Constant
R R
V: Constant
Solution:
dc
Fci − Fc =V
non linear term dt
∂f ∂f
f ( x , y ) = f ( xs , y s ) + x = xs ( x − xs ) + x = xs ( y − ys )
∂x y = ys ∂y y = ys
∴ Fc = Fs cs + cs ( F − Fs ) + Fs ( c − cs )
dc
Fci − [ Fs cs + cs ( F − Fs ) + Fs ( c − cs )] = V Un s.s
dt
dcs
Fs ci − Fs cs = V =0 s.s
dt
d ( c − cs )
ci ( F − Fs ) − cs ( F − Fs ) − Fs ( c − cs ) = V
dt
dY
ci X − cs X − FsY = V
dt
Process Control /Lec. 6 59 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
Y = c − cs
where
X = F − Fs
dY
[V + FsY = ( ci − c s ) X ] ÷ Fs
dt
V dY ( c − cs )
+Y = i X
Fs dt Fs
dY
τ
+ Y = RX
dt
Where
V ( c − cs )
τ= , R= i
Fs Fs
Y( s ) R
= st
1 order system
X ( s ) τs + 1 P P
Time Delay
The most commonly used model to describe the dynamics of chemical process is
First-Order Plus Model Delay Model. By proper choice τ d , this model can be
U U R R
• Time delay or dead time between inputs and outputs are very common
industrial procsses, engineering systems, economical, and biological systems.
qi(t) qo(t)
Dead
Time
(constant slope), the approximation would be perfect for any value of τ d . When the R R
slope of q i (t) varies rapidly, only smal τ d ’s will give a good approximation.
R R R R
If the variation in x(t) were some arbitrary function, as shown in figure below, the
response y(t) at the end of the pipe would be identical with x(t) but again delayed by t
T1( s ) K
=
V ( s ) τs + 1
Due to the delay time the temperature T 2 represented by:
R R
T2 ( s ) K e − τ d s
=
V( s ) τs + 1
F, ci
F, c1 F, c2
F, V : Constants
V L
c1
Time delay
C2 ( s ) − τ d s
=e
C1( s )
Volume of tube AL AL L
τd = = = =
Volumetric flow rate q uA u
C1( s ) R
=
Ci ( s ) τs + 1
C2 ( s ) C2 ( s ) C1( s )
= ×
Ci ( s ) C1( s ) Ci ( s )
C2 ( s ) R −τd s L
=
∴ C ( s ) τs + 1
e Where
τd =
i u (Time Units)
A linear second order system under dynamic condition is given by the differential
equation:-
1 d 2Y 2ξ dY
+ + Y = kX
ωn2 dt 2 ωn dt
1
=τ
ωn
d 2Y dY
∴ τ2 + 2ξτ + Y = kX
dt 2 dt
Where:-
k : Steady state gain
Y : Response value
X : Input disturbing variable
ω n : Natural frequency of oscillation of the system.
R R
Y (0) = Y (0) = 0
ξ : Damping factor (damping coefficient)
By taking laplace for the above second order differential equation
τ 2 s 2Y ( s) + 2ξτsY ( s) + Y ( s) = kX ( s)
(τ 2 s 2 + 2ξτs + 1)Y ( s ) = kX ( s )
Y ( s) k
G ( s) = = 2 2 T.F. of second order system
X ( s ) τ s + 2ξτs + 1
If x is sudden force, such as, step change inputs, Y will oscillate depending on the
value of damping coefficient ξ .
ξ <1 Response will oscillate (Under damped)
ξ>1 Response will oscillate (Over damped)
ξ=1 Response critical oscillation (critical damped)
ξ 2 −1
−ξ −ξ ξ 2 −1
s1 = − and s 2 = + ……(2)
τ τ τ τ
k A α0 α1 s + α 2
Y ( s) = × = +
τ 2 s 2 + 2ξτs + 1 s s τ 2 s 2 + 2ξτs + 1
α o (τ 2 s 2 + 2ξτs + 1) + α1s 2 + α 2 s = kA
s0 α 0 = kA
s1 2α 0ξτ + α 2 = 0 ⇒ α 2 = −2kAξτ
s2 α 0τ 2 + α 1 = 0 ⇒ α 1 = −kAτ 2
1 τ 2 s + 2ξτ
∴ Y ( s ) = kA[ − 2 2 ]
s τ s + 2ξτs + 1
1− ξ 2
w=
τ
ξ
Y (t ) = kA[1 − e ( −ξ / τ )t (cos wt + sin wt ]
1− ξ 2
ξ 1
r= p2 + q2 = 1+ ( )2 =
1− ξ 2 1− ξ 2
p 1 1− ξ 2
φ = tan −1
= tan −1 = tan −1
q ξ ξ
1− ξ 2
𝒚𝒚
𝒌𝒌𝒌𝒌
1.05 b
0.95 b
tp
1. Overshoot (OS)
U U
Overshoot is a measure of how much the response exceeds the ultimate value (new
𝐴𝐴
steady-state value) following a step change and is expressed as the ratio in the Fig.
𝐵𝐵
(8-3).
− πξ
OS = exp
1− ξ 2 )
OS % = 100 × OS
The decay ratio is defined as the ratio of the sizes of successive peaks and is given by
𝐶𝐶
in Fig. (8.3). where C is the height of the second peak
𝐴𝐴
− 2πξ
DR = exp = (OS ) 2
1−ξ 2
3. Rise time(t r )
U U R R
This is the time required for the response to first reach its ultimate value and is
labeled in Fig. (8.3).
1− ξ 2
π − tan −1
ξ
tr =
w
Process Control /Lec. 7 67 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
4. Response time
U
This is the time required for the response to come within ±5 percent of its ultimate
value and remain there. The response time is indicated in Fig. (8.3).
The radian frequency (radians/time) is the coefficient of t in the sine term; thus,
2πτ
T=
1−ξ 2
1
wn =
τ
The corresponding natural cyclical frequency f n and period T n are related by the
R R R R
expression:-
1 1
fn = = Thus, τ has the significance of the undamped period.
Tn 2πτ
Is the time required for the output to reach its first maximum value.
π πτ
tp = =
w 1− ξ 2
1-Over shoot
wt + φ = φ + nπ
nπ
t= max or min n=1, 2, 3 ……
w
nπ π
t= =
w w
−ξπ
1 π
y (t ) = kA[1 − e τw
sin ( w + φ )]
1− ξ 2 w
−ξπ
1 1−ξ 2
y max = kA[1 − e (− sin φ )]
1−ξ 2
−ξπ
1−ξ 2
y max = kA[1 + e ]
A max − B
Overshoot = =
B B
−ξπ
1−ξ 2
kA[1 + e ] − kA
Overshoot =
kA
− ξπ
Overshoot = exp
1−ξ 2 )
3. Rise time (t r )
U U R R
It is the time required for the response to first tauch the ultimate line.
−ξt
1
y (t ) = kA[1 − e τ
sin (tw + φ )]
1− ξ 2
At t r y(t)=kA
R R
−ξt
1
kA = kA[1 − e τ
sin (t r w + φ )]
1−ξ 2
0 = sin (t r w + φ )]
sin −1 (0) − φ
tr =
w
1− ξ 2
nπ − tan −1
nπ − φ ξ
tr = = =
w w
1−ξ 2
π − tan −1
ξ
tr = for n=1
w
1−ξ 2
f =
2πτ
2πτ
∴T =
1−ξ 2
6-Response time(t s ) R R
The time required for the response to reach (±5%) of its ultimate value and remain
there.
Is the time required for the output to reach its first maximum value.
nπ
t=
w
First peak is reached when n=1
nπ π πτ
tp = = =
w w 1− ξ 2
If impulse δ(t) is applied to second order system then transfer response can be written
as.
k
Y (s) = 2 2 X (s)
τ s + 2ξτs + 1
X ( s ) = Area = A
k
Y (s) = .A
τ 2 s 2 + 2ξτs + 1
kA / τ 2 kA / τ 2
Y (s) = =
2ξ 1 2ξ 1 ξ ξ
s +
2
s + 2 s2 + s + 2 + ( )2 − ( )2
τ τ τ τ τ τ
kA / τ 2 kA / τ 2
= =
2ξ ξ 2 1 ξ 2 ξ 2 1− ξ 2
s +
2
s+( ) + 2 −( ) (s + ) + 2
τ τ τ τ τ τ
i) ξ>1
kA τ ξ 2 −1
τ ξ 2 −1 τ
2
kA / τ 2 kA / τ 2
Y (s) = = =
ξ 2 1− ξ 2 ξ ξ 2
−
2
ξ ξ 2
−
2
(s + ) + 2 (s + ) 2 −
1 (s + ) 2 −
1
τ τ τ τ τ τ
−ξt
kA
Y (t ) = e τ
sinh wt
τ ξ −1
2
ξ 2 −1
w=
τ
ii)ξ<1
kA τ 1− ξ 2
kA / τ 2 kA / τ 2 τ 2 1− ξ 2 τ
Y ( s) = = =
ξ 1−ξ 2 ξ 2 1 − ξ 2
2
ξ 2 1 − ξ 2
2
(s + ) 2 + 2 (s + ) + (s + ) +
τ τ τ τ τ τ
−ξt
kA
Y (t ) = e τ
sin wt
τ 1− ξ 2
iii)ξ=1
kA / τ 2 kA / τ 2 kA / τ 2
Y (s) = = =
ξ 1 −ξ 2 1 1 − 12 1
(s + ) 2
(s + ) 2 + 2 (s + ) 2 + 2
τ τ τ τ τ
kA
Y (t ) = te −t /τ
τ 2
Example A step change from 15 to 31 psi in actual pressure results in the measured
response from a pressure indicating element shown in Fig. E5.14.
Figure E5.14
Assuming second-order dynamics, calculate all important parameters and write and
approximate transfer function in the form
R' ( s) K
= 2 2
P' ( s ) τ s + 2ζτs + 1
where R' is the instrument output deviation (mm), P' is the actual pressure deviation
(psi).
Solution:
11 .2 mm − 8 mm
Gain = = 0.20 mm / psi
31 psi − 15 psi
12.7 mm − 11 .2 mm
Overshoot = = 0.47
11 .2 psi − 8 psi
− πζ
Overshoot = exp = 0.47
1−ζ 2
ζ = 0.234
2πτ
Period = = 2.3 sec
1−ζ 2
Solution:
Given
2
X ( s) =
s
Time constant τ = 0.5 min
Radian frequency w = 1.9 rad / min
1−ξ 2 1−ξ 2
w= ⇒ 1.9 = ⇒ ξ = 0.312
τ 0.5
i) Rise time
1−ξ 2 1 − 0.312 2
π − tan −1 3.1416 − tan −1
ξ 0.312
tr = = = 1.0 min
w 1.9
C − 2πξ − 2π 0.312
ii) Deacay ratio = = exp( ) = exp( )
A 1−ξ 2 1 − 0.312 2
∴Decay ratio = 0.127
iii) Ultimate value of the response Yultimate (B) at t → ∞
Y ( s) 5
=
X ( s ) 0.25s + 0.316 s + 1
2
2
X ( s) =
s
10
Y ( s) =
s (0.25s + 0.316s + 1)
2
Yultimate (B) = 10
B
Maximum value of response = B(1 + )
A
B − πξ
Overshoot= = exp( )
A 1−ξ 2
Decay ratio=Overshoot2 P
0.127=Overshoot2 P
B
overshoot = 0.356 =
A
Maximum value of response =10(1+0.356)=13.56
τ
iv) Response time ts =3 =4.8077 min for ± 5% of ultimate value
ξ
flow rate w (mass/time). It is desired to maintain (or control) the temperature in the
tank at T R by means of the controller. If the indicated (measured) tank temperature T m
R R R
differs from the desired temperature T R , the controller senses the difference or error,
R R R
E = TR- Tm
R
R R R
have a positive feedback systems which is inherently unstable. To see that this
is true, again assume that be system is at steady state and that T=T R =T i . R R R R
If T i were to increase, T and T m would increas which would cause the signal
R R R R
from the compartor to increase, with the result that the heat to the system
would increse.
At s.s. T=T R =T in R R R
E=T R +T m R R R
Servo Problem
There is no change in load T i , and that we are interested in changing the bath
R R
temperature (change in the desired value (set point) with no disturbance load).
Requlating problem
The desired value T R is to remain fixed and the purpose of the control system is
R R
Tmor Ym
Measuring
device
Gm
Process
U
The procedure for developing the transfer function remain the same.
An unsteady-state energy balance around the heating tank gives.
dT
W Cp (Ti − To ) + q − W Cp (T − To ) = ρ Cp V
dt
Where T o is the reference temperature
R R
dT
At steady state, dt =0
dT
W Cp( Tis − To ) + q s − W Cp( Ts − To ) = ρ Cp V =0
dt
By substracting both equations
d ( T − Ts )
W Cp(( Ti − Tis ) − ( T − Ts )) + q − qs = ρ Cp V
dt
Note that the refernece temperature T o cancels in the subtraction. If we introduce
R R
thedeviation variables.
Ti = Ti − Tis
T = T − Ts
Q = q − qs
dT
W Cp( Ti − T ) + Q = ρ Cp V
dt
Taking the laplace transform gives
W Cp (Ti ( s ) − T ( s )) + Q( s ) = ρ Cp VsT ÷ W Cp
ρV Q( s)
sT + T ( s ) = + Ti ( s )
W WCp
The last expression can be written as
1 Q( s ) Ti ( s )
T (s) = +
(τs + 1) WCp τs + 1
Where
ρV
τ=
W
T (s) or Y(s) = controlled var iable
Q(s) or m(s) = manipulated var iable
Ti (s) or d (s) = disturbance var iable
If there is a change in Q(t) only then Ti ( t ) = 0 and the transfer function relating
Ti to Q is
Gd
d( s ) 1
Ti ( s ) τs + 1
+
Gp
m( s ) 1 / WCp + Y( s )
Qi ( s ) τs + 1 T(s)
Measuring Element
U
T(s) Km
Gm = Tm (s )
τm s + 1
Controller
Tsp(or TR)+ E P(s)
Gc
-
Tm
Control valve
Control valve that can control the rate of flow of a fluid in proportion to the
amplitude of a pressure (electrical) signal from the controller. From experiments
conducted on pneumatic valves, the relationship between flow and valve-top pressure
for a linear valve can often be represented by a first-order transfer function:
Air supply Air supply
τv = Time lag
τv ≤ 10 sec (Good)
Where:
constants of other components of the control system. A typical pneumatic valve has a
time constant of the order of 1 sec. Many industrial processes behave as first-order
systems or as a series of first-order systems having time constants that may range
from a minute to an hour. So the lag of the valve is negligible and the T. F. of the
valve sometimes is approximated by:
Q(s)
= KV
P(s)
The time constant of lag valve depends on the size of valve, air supply characteristics,
whether a valve positioner is used, etc.
Control Action
U
It is the manner, in which the automatic controller compares the actual value of the
process output with the actual desired value, determines the deviations and produce a
control signal which will reduce the deviation to zero or to small value.
On-Off Control
On-Off control is a special case of proportional control.
If the gain K C is made very high, the valve will move from one extreme position to
R R U U
the other if the set point is slightly changed. So the valve is either fully open or fully
closed (The valve acts like a switch).
The P.B. of the on-off controller reaches a zero because the gain is very high
P.B ≈ 0
Taking L.T
A
Prob(10-1): PI controller with step change in error E(s) =
U U
s
1 A
P ( s ) = K C (1 +
)
τIs s E(t) p(t)
KC.A
K A A KC A τi
∴ P(t ) = K C A + C t
τI 0 ps
0 t 0 t
Y=c+mX Response of a PI controller (lineaer)
t E(t) P(t)
Example:
A
For Ramp Error E(t) = At (Ramp) E(s) =
s2
A AK K Aτ
P(s) = K C (1 + τ D s ) × E(s) = K C (1 + τ D s ) × 2 = 2 C + C D
s s s
P(t) = K C At + K C AτD
P( s) 1
= K C (1 + + τ s)
E (s) τIs D
The value of the controlled variable is seen to rise at time zero owing to the
disturbance. With no control, this variable continues to rise to a new steady-state
value.
With control, after some time the control system begins to take action to try to
maintain the controlled variable close to the value that existed before the
disturbance occurred.
With proportional action only, the control system is able to arrest the rise of the
controlled variable and ultimately bring it to rest at a new steady-state value.
The difference between this new steady-state value and the original value (the
set point, in this case) is called offset.
The addition of integral action eliminates the offset; the controlled variable
ultimately returns to the original value. This advantage of integral action is
balanced by the disadvantage of a more oscillatory behavior.
The addition of derivative action to the PI action gives a definite improvement
in the response. The rise of the controlled variable is arrested more quickly,
and it is returned rapidly to the original value with little or no oscillation.
Solution:
U
Example: Consider the 1st order T. F. of the process with control valve
U U P P
Valve process
P(s) KV KP Y(s)
τV s + 1 τ s +1
Process Control /Lec. 9 86
Fourth Class
If we assume no interaction;
The T. F. from P(s) to Y(s) is
Y (s) KV K p
= For a unit step input in P
P( s ) (τ v s + 1)(τ s + 1)
1 KV K p
Y(s) =
s (τ v s + 1)(τ s + 1)
ττ 1 −t / τ v 1 −t / τ
y(t) = K v K p 1 − v e − e
τv −τ τ τv
R R R R
Y ( s) KK
If τ>>τ v then the T. F. is = v p
P( s ) (τ s + 1)
R R
Example: a pneumatic PI controller has an output pressure of 10 psi when the set
point and pen point are togather. The set point is suddenly displaced by 1.0 in (i.e a
step change in error is introduced) and the follwing data are obtained.
Time (s) 0- 0+ 20 60 80
Psi 10 8 7 5 3.5
Determine the actual gain (psi/inch displacement) and the integral time
E
1.0 in
10 psi
Kc
8psi
Kc
τI
For PI control
K
p (t ) = K c + c ∫ Edt + p s
τI
For E=1
K
p (t ) = K c + c t + ps
τI
From the above figure
Kc = 2
Kc 7−5 2
= =
τ I 60 − 20 40
τ I = 20 K c = 20 × (2) = 40 sec
Process Control /Lec. 9 87
Fourth Class
Example: (A) a unit-step change in error is introduced into a pid controller, if K c =10
U U R R
(B) if the error changed with a ratio of 0.5 in/min plot the response of p(t).
Solution:
U
K dE
p (t ) = K ct + c ∫ Edt + K cτ D + ps
τI dt
For a unit step change in error E(t)=1
At t=0 p (0) = K c + p s
Kc
t>0 p (t ) = K c + t + ps
τI
P = p − ps = 10 + 10t
Kc
1 10 τI
E P
0 0
0 0
dE
(B) E=0.5 t = 0.5 and ∫ dEdt = ∫ 0.5dt
dt
∴ p(t ) = 10 × 0.5 t + 10∫ 0.5 tdt + 10 × 0.5 × 0.5 + ps
p (t ) − p s = 5 t + 2.5 t 2 + 2.5
P( t ) = p( t ) − p s = 2.5 + 5 t + 2.5 t 2 t P(t)
0 2.5
1 10
E 0.5 P(t) 2 22.5
3 40
0 2.5 4 62.5
5 90
0 t 0 t
d(s) Gd
Comparator
ysp E(s) C(s) m(s) y(s)
GC Gf Gp
ym(s) y (s)
Gm
Process balance
y ( s ) = G p m( s ) + Gd d ( s ) Measuring Device
Measuring device
y m ( s ) = Gm y ( s )
Controller system
E ( s ) = ysp ( s ) − ym ( s ) Comparator
C ( s ) = Gc E ( s ) Controller
Final control element
m( s ) = G f C ( s )
Algebra manipulation of the above equations and arrange then
y ( s ) = G p m( s ) + Gd d ( s )
y ( s ) = G p G f C ( s ) + Gd d ( s )
y ( s ) = G p G f Gc E ( s ) + Gd d ( s )
y ( s ) = G p G f Gc ( ysp ( s ) − ym ( s )) + Gd d ( s )
y ( s ) = G p G f Gc ( ysp ( s ) − Gm y ( s )) + Gd d ( s )
y ( s) = G p G f Gc ysp ( s ) − G p G f GcGm y ( s ) + Gd d ( s )
(1 + Gc G f G pGm ) y ( s ) = Gc G f G p ysp ( s ) + Gd d ( s )
Gc G f G p Gd
y(s) = ySP ( s ) + d (s)
1 + Gc G f G p Gm 1 + Gc G f G p Gm
Let G = G C G f G P
R R R R R
G Gd
∴ y( s) = y SP ( s ) + d (s)
1 + GGm 1 + GGm
G Gd
= GSP = Gload
1 + GGm 1 + GGm
The distubance does not change (i.e. d (s) = 0 ) while the set point undergoes change.
The feedback controller act in such away as to keep y close to the changing ysp . The
T.F. of closed loop system of this type is:
G p G f Gc
y ( s) = ySP ( s )
1 + Gc G f G p Gm
2) Regulated systems:
In these systems the set point (desired value) is constant ( ysp (s) = 0 ) and the change
occurring in the load. The T.F. of closed loop control system of this type is:
Gd
y ( s) = d (s)
1 + Gc G f G p Gm
y ( s ) = Gload d ( s )
The feedback controller tries to eliminate the impact of the load change d to keep y at
the desired setpoint.
Kp Kd
put G p = and Gd =
τ ps + 1 τ ps + 1
Then by substitution in eqn. (**) and take the closed loop reponse as
Kp Kd
Kc
τ ps + 1 τ ps + 1
y( s ) = ySP ( s ) + d( s )
Kp Kp
1 + Kc 1 + Kc
τ ps + 1 τ ps + 1
K P KC Kd 1 + K p KC
y ( s) = [ ySP ( s ) + d ( s )] ×
τ p s + 1 + K p KC τ p s + 1 + K p KC 1 + K p KC
K P KC Kd
1 + K p KC 1 + K p KC
y ( s) = ySP ( s ) + d (s)
τ ps 1 + K p KC τ ps 1 + K p KC
+ +
1 + K p KC 1 + K p KC 1 + K p KC 1 + K p KC
Rearrange the last eqn.
KP Kd
y (s) = y SP ( s ) + d ( s)
τ ps +1 τ ps +1
Where
τP
τp = Closed loop time constant
1 + K p KC
K p KC
Kp = Closed loop ststic gain
1 + K p KC
Kd
Kd = Closed loop ststic gain
1 + K p KC
The close-loop response has the follwing charactrstics:-
1- It remains first order with respect to load and setpoint change
2- The time constant has been reduced (τ p < τ P ) which mean that the closed-loop
response has become faster than the open loop response, to change in set point
or load.
3- The static gain have been decreased.
1
ysp(t) 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 =
1+𝐾𝐾𝑃𝑃 𝐾𝐾𝐶𝐶
1
y(t)
0
2- Effect of Integral Control
U
1- Effect of PI control
U
Combination of propertional and integral control modes lead to the follwing effects
on the response of closed-loop system.
1- The order of the response inceases ( effect of I mode).
2- The offset is eliminated (effect of I mode).
3- As K c increses, the response becomes fater ( effect of P and I modes) and more
R R
oscillatory to set point changes [ovesrshoot and decay ratio increase (effet of I
mode)].
Large value of K c create a very sensitive response and may lead to instability.
R R
4- Asτ I decreases, for constant K c , the reponse become faster but more oscillatory
R R
To increase the speed of the closed loop response, increase the value of the controller
gain K c . But increasing enough K c in order to have acceptable speed, the response
R R R R
3
G P (s) = (process)
10 s + 1
R R
1
G d (s) = (Load)
10 s + 1
R R
G m (s) = 1
R R (measuring device) if not given take 1
G C (s) = 2
R R (controller)
G f (s) = 1.5
R R (valve)
Gd(s)
θd(s)
Determine the system response for a unit step in load 1
10 s + 1
θSP(s) θo(s)
2 1.5 3
Set point
10 s + 1
GC(s) Gf (s) GP(s)
Solution:
U
Gm(s)
θo ( s) Gd ( s )
Regulator loop: =
θ d ( s ) 1 + Gm ( s )G ( s )
3 9
G(s) = G C (s) G f (s) G P (s) = 2×1.5× =
10 s + 1 10 s + 1
R R R R R R
1
θo ( s )
= 10 s + 1
1
=
θd ( s ) 1 + 9 10 s + 10
10 s + 1
1
θ d (s) =
R R 1/10
s
1
θ o (s) = 10 θ o (t) off-set
s ( s + 1)
R R
in the equation.
ψ
Y ( t ) = kA[ 1 − e( −ψ / τ )t (cos wt + sin wt ]
1 −ψ 2
1 −ψ 2
w=
τ
2) The offset=New set point- Ultimate value
=0.1-0.088889=0.01111
2πτ 2π × 0.471
3) Period of oscillation = = = 3.1640
1 −ψ 2
1 − (0.3538) 2
Example: Consider the figure below, a unit step change in load enters at either
location 1 or location 2.
What is the offset when the load enters at location 1 and when it enters at location 2
Process Control /Lec. 10 96 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
U1 G1 U2 G2
Gc
2 1
R Kc = 5 C
(2 s + 1) (2 s + 1)
1
U 1( s ) = , U2( s ) = 0
s
G1G2
C ( s) = U1 ( s )
1 + GcG1G2
2 1
⋅
C( s ) = 2s + 1 2s + 1 U ( s ) = 2
U 1( s )
+ + +
1 2
2 1 4 s 4 s 1 10
1+ ⋅ ×5
2s + 1 2s + 1
2 2 / 11
= 2 U 1( s ) = U 1( s )
4 s + 4 s + 11 4 2 4
s + s+1
11 11
2
K = = 0.1818
11
4
τ= = 0.6030
11
4 4 1
2ψτ = ⇒ ψ = × = 0.3015
11 11 2τ
Ultimate value=A.K=1*0.1818=0.1818
Offset=0 - 0.1818= -0.1818
G2
C ( s) = U 2 (s)
1 + GcG1G2
1
2s + 1 2s + 1
C ( s) = U 2 ( s) = 2 U 2 (s)
1+
2
⋅
1
×5 4 s + 4 s + 1 + 10
2s + 1 2s + 1
2s + 1 2s + 1
= 2 U 2 (s) =
4 s + 4 s + 11 4 2 4
s + s + 11
11 11
2s + 1 1
C (∞) = lim = = 0.091
s →0 4 2 4
s + s + 11 11
11 11
Offset= 0 - 0.0.091= -0.091
1 1 C
R K c (1 + + τ s)
τIs D τ 1s + 1
Forτ D = τ I = 1 andτ I = 2
a- Calculate ψ when K c =0.5 and K c =2
R R R R
C ( s) GcG p
=
R( s ) 1 + GcG p
1 1 1 1
+ τ D s)
K c (1 + K c (1 + + s )
C (s) τIs τ 1s + 1 s 2s + 1
= =
R ( s ) 1 + K (1 + 1 1 1 1
+ τ D s) 1 + K c (1 + + s )
c
τIs τ 1s + 1 s 2s + 1
s + 1 + s2 1
Kc ( )
s 2s + 1 = Kc (s + 1 + s 2 ) Kc (s + 1 + s 2 )
= =
s + 1 + s2 1 2s 2 + s + K c ( s + 1 + s 2 ) (2 + K c ) s 2 + (1 + K c ) s + K c
1 + Kc ( )
s 2s + 1
(s + 1 + s )
2
=
(2 + K c ) 2 (1 + K c )
s + s +1
Kc Kc
a-1) K c =0.5
R R
2 + Kc 2 + 0.5
τ= = = 2.2361
Kc 0.5
(1 + K c ) 1 + 0.5 3 3
2ψτ = = = 3 ⇒ψ = = = 0.6708
Kc 0.5 2τ 2 × 2.2361
a-2) K c =2R R
2+2
τ= = 1.4142
2
(1 + K c ) 1 + 2 1.5 1.5
2ψτ = = = 1.5 ⇒ ψ = = = 0.5303
Kc 2 2τ 2 × 1.41421
Gp
B) C ( s ) = = U (s)
1 + G pGc
1 1
τ 1s + 1 2s + 1
C ( s) = U (s) = U (s)
1 1 1 1
1 + K c (1 + + τ s) 1 + K c (1 + + s )
τ I s D τ 1s + 1 s 2s + 1
1
= 2
4 s + 3s + 1
τ =2
3
2ψτ = 3 ⇒ ψ = = 0.75
4
1
C (∞) = lim s 2 =0
s → 0 4 s + 3s + 1
Offset=0-0=0
K c ( 1 + τ D s )( τ m s + 1 )
( 1 + Kc )
C= R( s )
τ 1τ m s 2
( τ 1 + τ m + K cτ D )
+ s+1
( 1 + Kc ) ( 1 + Kc )
Process Control /Lec. 10 99 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
τ 1τ m
τ=
1 + Kc
τ 1 + τ m + K cτ D
2ψτ =
1 + Kc
τ 1 + τ m + K cτ D 1 + K c
ψ=
2(1 + K c ) τ 1τ m
τ + τ + K cτ D
∴ψ = 1 m
2 (1 + K c )τ 1τ m
τ1 + τ m + K cτ D
b)∴ψ = for ψ=0.7
2 (1 + K c ) τ 1τ m
1) τ D =0
R R
60 + 10 + 0 35
∴ 0.7 = =
2 (1 + K c ) 60 × 10 600 + 600 K c
600 + 600 K c = 50
600 + 600 K c = 2500
K c = 3.166
2) τ D =3 sec
R R
60 + 10 + 3K c 70 + 3K c
∴ 0.7 = =
2 (1 + K c ) 600 2 (1 + K c ) 600
70 + 3K c = 34.292 (1 + K c )
2.04(1 + 0.042 K c ) = (1 + K c )
4.1616 + 0.355K c + 0.0075K c2 = (1 + K c )
0.0075K c2 − 0.0645K c + 3.1616 = 0
∴ K c = 80.73 or K c = 5.266
The transfer function of a block diagram is defined as the output divided by its input
when represented in the Laplace domain with zero initial conditions. The transfer
function G(s) of the block diagram shown in Fig. (1).
Y(s)
= G (s)
X(s)
Here the path of the signals X(s) and Y(s) is a forward path.
Consider the block diagram of cascaded elements shown in Fig. (2a). Form the
definition of a transfer function we have:
X 2 (s)
= G1 (s)
X1 (s)
X 3 (s)
= G 2 (s)
X 2 (s)
Y(s)
= G 3 (s)
X 3 (s)
And substitution yields
Y(s) = G 3 (s)X 3 (s) = G 3 (s)[G 2 (s)X 2 (s)] = G 3 (s)G 2 (s)G1 (s)X1 (s)
Which can be written as
Y(s)
= G 3 (s)G 2 (s)G1 (s) = G (s)
X1 (s)
(a) (b)
The overall transfer function then is simply the product of individual transfer
functions.
For applications where it is required to generate a signal which is the sum of two
signals we define a summer or summing junction as shown in Fig. (3a). If the
Process Control /Lec. 11 101 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
difference is required, then we define a subtractor as shown in Fig. (3b). Subtractors
are often called error detecting devices since the output signal is the difference
between two signals of which one is usually a reference signal.
X1(s) + +
X1(s)+X2(s) X1(s) X1(s)-X2(s)
+ -
X2(s) X2(s)
(a) (b)
Summer Subtractor (Error detecting device)
Fig. (3) Addition or subtraction of signals
The combination of block diagrams in parallel is shown in Fig. (4a). Form the
definition of the transfer function we have
Y1 (s) = G1 (s)X(s)
Y2 (s) = G 2 (s)X(s)
Y3 (s) = G 3 (s)X(s)
And the summer adds these signals,
Y(s) = Y1 (s) + Y2 (s) + Y3 (s)
or
Y(s) = [G1 (s) + G 2 (s) + G 3 (s)]X(s)
The overall transfer function shown in Fig.(4b) is
Y(s)
= G (s)
X(s)
where
G (s) = G1 (s) + G 2 (s) + G 3 (s)
Y1(s)
G1(s)
X(s) Y2(s) +
G2(s) Y(s) X(s) G(s) Y(s)
+
+
Y3(s)
G3(s)
(b)
(a)
Fig. (4) Parallel combination of elements
In summary, we observe that for cascaded elements the overall transfer function is
equal to the product of the transfer function of each element, whereas the overall
(5).
H1(s)
B1(s)
+ E1(s) C1(s) - E2(s) C2(s) C(s)
R(s) G1(s) G2(s) G3(s)
- +
B2(s) H2(s)
Y(s)
loop transfer function
R (s)
R(s) +
2( s + 1)( s + 3) Y(s)
G(s)=
- s ( s + 2)( s + 4)
Solution
Y(s) G
Transfer function =
R (s) 1 + GH
2(s + 1)(s + 3) 2(s + 1)(s + 3)
Y(s) s(s + 2)(s + 4) s(s + 3)(s + 4)
= =
R (s) 1 + 2(s + 1)(s + 3) * 1 s(s + 2)(s + 4) + 2(s + 1)(s + 3)
s(s + 2)(s + 4) s(s + 2)(s + 4)
2(s + 1)(s + 3) 2s 2 + 6s + 2s + 6
=
s(s + 2)(s + 4) + 2(s + 1)(s + 3) s3 + 4s 2 + 2s 2 + 8s + 2s 2 + 6s + 2s + 6
Y(s) 2s 2 + 8s + 6
∴ =
R (s) s 3 + 8s 2 + 16s + 6
When the block diagram representation gets complicated, it is advisable to reduce the
diagram to a simpler and more manageable form prior to obtaining the overall
transfer function. We shall consider only a few rules for block diagram reduction. We
have already two rules, viz. Cascading and parallel connection.
Consider the problem of moving the starting point of a signal shown in Fig. (6a) from
behind to the front of G(s). since B(s)=R(s) and R(s)=C(s)/G(s), then B(s)=C(s)/G(s).
therefore if the takeoff point is in front of G(s), then the signal must go through a
transfer function 1/G(s) to yield B(s) as shown in Fig. (7b).
(a) (b)
in Fig. (7b). A table of the most common reduction rules is given in Table 1.
R(s) + C(s)
G1(s)
Addition or + R(s) C(s)
G1(s)+G2(s)
subtraction
G2(s)
R(s) C(s)
G(s)
R(s) C(s)
G(s)
Moving a
starting point 1
B(s) G (s)
B(s)
R(s) + C(s)
Closed loop G(s) R(s) G (s) C(s)
-
system 1 + H ( s )G ( s )
H(s)
Consider the transfer function of the system shown in Fig. (8a). The final transfer
function is shown in Fig. (8d).
R(s) + +
G1(s) G2(s) C(s)
+ G4(s)
- +
-
H1(s)
H2(s)
R(s) + C(s)
G1(s) G2(s)+G3(s) G4(s)
+
- -
H 1 (s)
G4 (s )
H2(s)
(b)
H2(s)
(c)
(d)
Fig.8 Obtaining transfer function by block diagram reduction
Example: Obtain the transfer function C/R of the block diagram shown in Figure
U U
below.
below.
G4(s)
R(s) +
+ + +
G1(s) G2(s) G3(s) C(s)
- +
H1(s)
H2(s)
Solution:
U U
R(s)
+ +
G1(s) G2(s) G3(s) + G4(s) C(s)
- +
H1(s)
H2(s)
R(s)
+ G1(s) x G2(s)
G3(s) + G4(s) C(s)
1- H1(s)x(G1(s) x G2(s))
-
H2(s)
A stable system is one where the controlled variable will always settle near the set
point. An unstable system is one where, under some conditions, the controlled
variable drifts away from the set point or breaks into oscillations that get larger and
larger until the system saturates on each side.
H
G1G 2 G2
C= R (s) + U(s)
1 + G1G 2 H 1 + G1G 2 H
1 + G1G 2 H = 0 (Characterstic Equation)
(s − r1 )(s − r2 )(s − r3 )......... = 0 Im
s-p
s
p ×
Re
A linear control system is unstable if any roots of its characterstic equation are to the
right of imaginary axis.
If this equation has some roots with positive real parts, then the system is unstable, or
some roots equal to zero, the system is marginally stable (oscillatory), therefore it is
unstable.
Then for stability the roots of characteristic equation must have negative
real parts.
Process Control /Lec. 12 109 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
Example:if
0.5s + 1
G1 = 10 PI control
s
1
G2 = Stirred tank
2s + 1
H =1 Mesuring element without lag
1 + G = 1 + G1G 2 H = 0
10(0.5s + 1)
1+ =0
s(2s + 1)
s(2s + 1) + 5s + 10 = 0
2s 2 + 6s + 10 = 0
s 2 + 3s + 5 = 0
−3 9 − 20
s=
2 2
−3 11 −3 11
∴ s1 = +j and s 2 = −j
2 2 2 2
3
Since the real part in s1 and s2 in -ve ( − )∴The system is stable
2
2-Routh’s Method
P P
Where ao is positive
It is necessary that ao, a1, a2,…. an-1, an be positive. If any coeff. is negative, the
system is unstable.
If all of the coeff. are positive, the system may be stable or unstable. Then apply the
next step.
b. Routh array:
Arrange the coeff. of eqn. (*) into the first two rows of the Routh array shown
below.
Row
1 ao a2 a4 a6
2 a1 a3 a5 a7
3 A1 A2 A3
4 B1 B2 B3
n+1 C1 C2 C3
Process Control /Lec. 12 110 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
a1a 2 − a o a 3 a1a 4 − a o a 5 a1a 6 − a o a 7
A1 = , A2 = , A3 =
a1 a1 a1
A1a 3 − a1A 2 A a −a A
B1 = , B2 = 1 5 1 3
A1 A1
B1A 2 − A1B2 B A − A1B3
C1 = , C2 = 1 3
B1 B1
Examine the elements of the first column of the array ao, a1, A1, B1,C1……….W1
a) If any of these elements is negative, we have at least one root on the right of
the imaginary axis and the system is unstable.
b) The number of sign changes in the elemnts of the first column is equal to the
number of root to the right of the imaginary axis.
∴The system is stable if all the elements in the first column of the array are
positive
Solution:
Row 3 × 5 − 4 × 11 11
A1 = =
1 1 5 2 3 3
2 3 4 0 3× 2 − 0
A2 = =2
3 11 3 2 0 3
4 2.36 0 11 3 × 4 − 6
B1 = = 2.36
5 2 11 3
2.36 × 2
C1 = =2
2.36
∴The system is stable
s4 + 2s3 +3s2 + 4s + 5 = 0
P P P P P P
Solution:
s4 1 3 5
P P
s3 2 4
P P
s2 1 5
P P
s1 -6 0
P P
s0 5
P P
Example: Consider the feedback control system with the characteristic equation.
K
s3 + 2s 2 + (2 + K c )s + c = 0
τI
Solution:
The corresponding Routh array can now be formed
Row
1 1 2 + Kc 0
2 2 Kc 0
τI
3 2(2 + K c ) − K c τ I 0 0
2
4 Kc τ I 0 0
The elemnets of the first-column are positive except the third, which can be positive
or negative depending on Kc and τI.
So state the stability
Put 2(2 + K c ) − K c τ I > 0 ⇒ 2(2 + K c ) > K c
2 τI
If Kc and τI satisfy the condition, then the system is atable
Critical stability
Put the third element=0
i.e 2(2 + K c ) = K c
τI
For τI=0.1
2(2 + K c ) = 10K c ⇒ 4 = 8 K c
K c = 0.5
Process Control /Lec. 12 112 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
1) if K c < 0.5 , the system is stable (all of the elements in the 1st column is +ve)
2) if K c > 0.5 , the third element of the 1st column is negative. We have two sign
change in the elements of the first column.
Example:
U
R 1
Kc C
(τ 1s + 1)(τ 2 s + 1)
1
(τ 3s + 1)
1 1
Ifτ 1 = 1, τ 2 = , τ 3 =
2 3
Determine Kc for a stable system
Solution:
The char. Eqn.
1
1 + Kc =0
1 1
(s + 1)( s + 1)( s + 1)
2 3
1 1
(s + 1)( s + 1)( s + 1) + K c = 0
2 3
1 3 1
( s 2 + s + 1)( s + 1) + K c = 0
2 2 3
s3 s 2 s s 2 3s
+ + + + + 1 + Kc =0
6 2 3 2 2
1 3 2 11
s + s + s + 1 + Kc = 0
6 6
Row
1 1/6 11/6
2 1 1+Kc
3 10 − K c 0
6
4 1+Kc
Since Kc>0
∴The system will be stable
If 10-Kc>0
Example: U
R 3
K c (1 + ) 2
s
C
1
0.2s 2 + 0.4s + 1
Study the stability for Kc=2
Solution:
3 1
1 + K c (1 + ) × 2 × =0
s 0.2s + 0.4s + 1
2
s+3 2
1 + Kc ( )× =0
s 0.2s + 0.4s + 1
2
sK + 3K c 2
1+ ( c )× =0
s 0.2s + 0.4s + 1
2
+ controller ( s + 20)
R(s) Y(s)
k s ( s + 10) 2 Heading
-
Solution
k (s + 20) ks + 20k ks + 20k
G(s) = = 2 = 3
s(s + 10) 2
s(s + 20s + 100) s + 20s 2 + 100s
Characteristic equation,
1+GH = 0
ks + 20k
1+ 3 *1 = 0
s + 20s 2 + 100s
s 3 + 20s 2 + 100s +ks+20k =0
s3 + 20s 2 + (100 + k )s + 20k = 0
The corresponding Routh array can now be formed
Row
1 1 100+k
2 20 20k
3 a 0
4 b 0
It is how the output response (amplitude, phase shift) change with the frequency of
the input sinusoidal.
The input frequency is varied, and the output characteristics are computed or
represented as a function of the frequency. Frequency response analysis provides
useful insights into stability and performance characteristics of the control system.
Figure below shows the hypothetical experiment that is conducted.
K p Aωτ p − K Aωτ p K A
C1 = 2 2 , C2 = 2 p 2 , C3 = 2 p2
τp ω +1 τp ω +1 τp ω +1
Kp Aωτ p − t /τ p Kp Aωτ p Kp A
y( t ) = e − cos(ωt ) + sin(ωt )
τ p2 ω 2 + 1 τ p2 ω 2 + 1 τ p2 ω 2 + 1
−t /τ p
As t→∞, then e → 0 , and the first term disappers.
q =1 p = −ω τ p
r = (−τ pω ) 2 + (1) = τ 2p ω 2 + 1
2
p − ω τp
φ = tan −1 = tan −1 ( ) = tan −1 (−ω τ p )
q 1
Then eq.(3) yield
Kp A
yss ( t ) = 2 2 [( τ 2p ω 2 + 1) sin(ωt + φ )]
τp ω +1
Kp A
yss ( t ) = sin(ω t + φ ) (4)
τ p2 ω 2 + 1
G (s) = e − t d s
Put s=jω
∴ G ( jω ) = e − j t d ω
AR = 1
φ = − t dω
That is aphase lag since φ < 0
1- Propertional controller:
U
∴ AR = K c
φ =0
2- PI controller:
U
1
The transfer function is G (s) = K c (1 + )
τI s
1
∴ AR = K c 1 +
(ωτ I ) 2
−1
φ = tan −1 ( ) <0
ωτ I
3- PD controller:
U
AR = K c 1 + τ 2D ω 2
φ = tan −1 (τ D ω ) >0
The positive phase shift is called phase lead and implies that the controller output
lead the input.
4-PID controller:
U
1
The transfer function is G (s) = K c (1 + + τ D s)
τI s
1
AR = K c (τ Dω − )2 + 1
τ Iω
1
φ = tan −1 (τ Dω − )
τ Iω
Kp
Amplitude ratio AR = (*)
1 + τ p2ω 2
Phase lag= φ = tan −1 − τ p ω
AR 1
log = − log(1 + τ p2 ω 2 ) (**)
Kp 2
The plot can be carried by considering its asymptotic behaviour as ω→0 and as
ω→∞ . Then
1. As ω→0, then τ p ω → 0 and from eq.(*)
AR AR
log → 0 or = 1.This is the low-frequency asymptote. It is a horizontal line
Kp Kp
AR
passing through the point = 1.
Kp
as ω → 0 , φ →0
1
as ω → , φ → tan −1 (−1) = −45o
τp
as ω → ∞ φ → tan −1 (−∞) = −90 o
,
Process Control /Lec. 13 120
Fourth Class
𝐴𝐴𝐴𝐴
𝐾𝐾𝑝𝑝
𝜙𝜙
or
Phase lag
𝜏𝜏𝑝𝑝 𝜔𝜔
Figure:Bode diagram for first-order system.
Kp − 2ψτω
AR = φ = tan −1 ( )
(1 − τ ω ) + (2ψτω )
2 2 2 2 1 − τ ω
2 2
ψ=0.1
𝐴𝐴𝐴𝐴
𝐾𝐾𝑝𝑝
ψ=0.1
𝜙𝜙
or
Phase lag
𝜏𝜏𝑝𝑝 𝜔𝜔
1
Figure: Block diagram for second-order system
τ 2s 2 + 2ψτs + 1
Process Control /Lec. 13 121
Fourth Class
AR 1
log = − log[(1 − τ 2ω 2 ) 2 + (2ψτω ) 2 ]
Kp 2
AR 1
1) as ω → 0 , then log = − log(1) = 0
Kp 2
AR
∴ = 1 stright line of a slope=0 (L.F.A)
Kp
−0
φ = tan −1 =0
1
AR 1
2) as ω → ∞ , then log = − log (τω ) 4 = −2 log (τω ) (H.F.A)
Kp 2
It is a straight line with a slope -2 passing through the pointAR=1 and τω=1
1
3) ω = ωc =
τ
Pure dead-time system
U
as ω → 0 , φ = 0
as ω → ∞ , φ = ∞
1.0
AR
ϕ -180
-360 | | |
1 10 100
ω
Example: Two systems in series
1 6
G1 (s) = and G 2 (s) =
2s + 1 5s + 1
The overall T.F. is
1 6
G (s) = ⋅
2s + 1 5s + 1
1 6
∴ AR = ⋅
1 + 4ω 2 1 + 25ω 2
log AR = log 6 + log(AR )1 + log(AR ) 2
-45
𝜙𝜙 -90
-135
-180 | |
0.2 ω 0.5
Feedback Controller
U
1-Propertional controller
AR = K c φ = 0
𝐴𝐴𝐴𝐴
1.0
𝐾𝐾𝑐𝑐
ϕ 0
ω
Process Control /Lec. 13 123
Fourth Class
2-Propertional Integral controller (PI)
1
AR = K c 1 +
(ωτ I ) 2
1
φ = tan −1 (− )<0
ωτ I
AR 1 1
log( ) = log(1 + )
Kc 2 (ω τ I ) 2
AR 1 AR
as ω → 0 , log( ) = log(τ D2 ω 2 ) = 0 ⇒ =1 (L.F.A) slope=0
Kc 2 Kc
φ = tan −1 0 = 0o
2) High frequency asymptote
AR 1
as ω → ∞ , log( ) = log(τ D2 ω 2 ) = log(τ Dω ) (H.F.A) slope=+1
Kc 2
φ = tan −1 0 = 90o
as ω = 0 , φ = 0
as ω = ωc , φ = +45o
as ω = ∞ , φ → +90o
ω=1/τD
AR 1 2
1) as ω → 0 then = 1+ ( ) PI Controller
Kc τ Iω
AR
2) as ω → ∞ then = 1 + (τ Dω ) 2 PD Controller
Kc
1 AR
3) as ω → then = 1 + (τ Dω − 1) 2
τI Kc
1 AR 1 2
4) as ω → then = 1 + (1 − )
τD Kc τ Iω
Frequency Response of non-interacting capactine in sereies
U
φ t = φ1 + φ 2 + φ3 + ............ + φ n
Process Control /Lec. 13 126
Fourth Class
Example: Bode Digram of PID Controller
U U
1
G1 (s) = 10(1 + + 5s)
10s
1
ωc1 (s) = = 0.1 signal(−1)
10
1
ωc1 (s) = = 0.2 signal(+1)
5
𝐴𝐴𝐴𝐴
𝐾𝐾𝑐𝑐 1.0
1� 1�
𝜏𝜏𝐼𝐼 𝜏𝜏𝐷𝐷
90
+45
ϕ
0
-45
-90
1�
𝜏𝜏𝐼𝐼 ω 1�
𝜏𝜏𝐷𝐷
Example:
U U
Gc Gf Gp
YSP(s) 1 10 5e −0.2s
K c (1 + ) Y(s)
τ Is 0.1s + 1 (2s + 1)(s + 1)
Ym(s)
2
0.5s + 1
The Open loop T.F. of the feedback control
G OL = G c G f G p G m
1 1 1 1
G OL = 100K c (1 + ) ⋅ ⋅ e − 0.2s
τ Is 0.1s + 1 (2s + 1)(s + 1) (0.5s + 1)
With K c =4 and τ I =0.25
R R R R
∴K=400
𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = −1 1
1
0.1𝑠𝑠 + 1
𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = −2
𝐴𝐴𝐴𝐴 1
𝐾𝐾𝑐𝑐 0.5𝑠𝑠 + 1
1
𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = −3 𝑠𝑠 + 1
1
2𝑠𝑠 + 1
𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = −4
0.1 0.5 1 2 4 10
0 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = −3
-45
𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = −4
-90
𝜙𝜙 .0
-135
-180
𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂
Example: Plot the B.D. for the open loop T.F. for the fig. below
U U
G U Gp
c
R K c (1 + τ D s) 1
(s + 1) (0.1s + 1)
2 C
Ym(s) s
−
e 10
s
−
10(0.5s + 1)e 10
G OL (s) =
(s + 1)2 (0.1s + 1)
Overall Bode diagram
1 → 1
G1 (s) = ωc1 = = 1 stright line slope=-1
1s + 1 1
1 → 1
G 2 (s) = ωc 2 = = 1 stright line slope=-1
1s + 1 1
G 3 (s) = 0.5s + 1 → ωc3 = 1 = 2 stright line slope=+1
0.5
1 → 1
G 4 (s) = ωc 4 = = 10 stright line slope=-1
0.1s + 1 0.1
s
−
G 5 (s) = e 10 stright line slope=0
Figure: Block diagram for: (a) Amplitude ratio; (b) phase angle.
Instrumentation
The example level-control problem had three critical pieces of instrumentation: a
sensor (measurement device), actuator (manipulated input device), and controller.
The sensor measured the tank level, the actuator changed the flow rate, and the
controller determined how much to vary the actuator, based on the sensor signal.
Each device in a control loop must supply or receive a signal from another device.
Figure 14.2 The feedback control of the liquid level in a boiler drum.
The feedforward control scheme in Fig. 14.3 can provide better control of the liquid
level. Here the steam flow rate is measured, and the feedforward controller adjusts
the feedwater flow rate.
2)Pressure Control
• Pressure control is usually by venting a gas or vapor.
• In hydrocarbon processes, off-gas is often vented to fuel.
• In other processes, nitrogen may be brought in to maintain pressure and vented
via scrubbers.
• Most common arrangement is direct venting.
• Several vessels that are connected together may have a single pressure
controller.
PV
PIC
PT
FV
PI FT
FIC
Vapor
LIC
FV
FT
LV
LT
Feed
Steam
Trap
Condensate
Figure 14.6 Vaporizer control system
Process
TIC
TE TT
TV
TIC
TE TT
TV
TO VARIABLE
TIC SPEED MOTOR TIC
TE TT TE TT
CONTROL CIRCUIT
(a) (b)
Figure 14.9 Temperature control of air coolers
Feed
TC
Cooling
water
TT
Product
Distillation Control
Distillation control is a specialized subject in its own right.
In addition to controlling condenser pressure and level in the sump, a simple
distillation column has two degrees of freedom.
Material balance (split) and energy balance (heat input or removed).
Therefore needs two controllers.
Therefore has the possibility that the controllers will interact and “fight”
each other.
Side streams, intermediate condensers & reboilers, pump-arounds, etc. all add
extra complexity and degrees of freedom.
Figure 14.12 Schematics of L/D-Q, L/D-B, and D-Q/B distillation column control
structures.
Process Control /Lec. 14 138 Written by Assoc. Prof.
Fourth Class Dr. Zaidoon M. Shakoor
Batch Distillation
• Reflux flow control adjusted based on temperature (used to infer composition)
Heat Exchangers
Heat exchangers process used to transfer heat between two process streams. The flow
of these process streams is usually set elsewhere in the plant so that adjusting the
flowrate of one of the process streams to regulate the amount of heat transferred is
not possible.
To provide a control degree-of-freedom for regulating the heat transferred, a small
by-pass (~5-10%) of one of the process streams around the heat exchanger is
provided. The outlet temperature of this process stream or the other process stream
can be controlled by manipulating the by-pass rate. These two schemes are illustrated
in Figure 14.14. In the former, tight temperature control is possible as the amount of
heat transferred is governed by the bypass. In the latter, a thermal lag of the order of
0.5 to 2 minutes exists between the manipulated and controlled variable.