Process Modelling, Simulation and Control for Chemical Engineers.

Solved
problems. Chapter 10: Laplace domain analysis of conventional feedback
control systems. Version 2.
This document contains my own solutions to the problems proposed at the end of each chapter of
the book ”Process Modelling, Simulation and Control for Chemical Engineers”, Second Edition,
by William L. Luyben. As such, I can’t guarantee that the proposed solutions are free from errors.
Think about them as a starting point for developing or as a means of checking your own solutions.
Any comments or corrections will be appreciated. Contact me at [email protected]
The computer programs developed for this chapter (Python) are available at:
https://www.dropbox.com/sh/nkt3u1na5nsl36z/AABztzYjANWSr0e N4AWJVesa?dl=0

Changes from Version 1

*Problem 10 curves where plotted as s versus arg(s), now corrected to ln|s| versus arg(s).

Problem 1
For the following system of differential equations that describe a nonisothermal CSTR:
dC
= a11 C + a12 T + a13 C0 + a15 F (1)
dt
dT
= a21 C + a22 T + a24 T0 + a25 F + a26 Tj (2)
dt
with:

F̄
a11 = − − k̄ λk̄E C̄ F̄ UA
V a22 = − − −
ρCp RT̄ 2 V V ρCp
C̄E k̄
a12 =−
RT̄ 2 F̄
a24 =
F̄ V
a13 =
V T̄0 − T̄
a25 =
C̄0 − C̄ V
a15 =
V UA
a26 =
λk̄ V ρCp
a21 =−
ρCp k̄ = αe−E/(RT̄ )

And numerical values of the parameters:

F̄ = 40 f t3 /h E = 30000 BT U/lbmol
V = 48 f t3 R = 1.99 BT U/lbmol ◦ R
C̄0 = 0.50 lbmol/f t3 U = 150 BT U/hf t2 ◦ R
C̄ = 0.245 lbmol/f t3 A = 250 f t2
T̄ = 600 ◦ R λ = −30000 BT U/lbmol
T̄0 = 530 ◦ R Cp = 0.75 BT U/lbm ◦ R
10 −1
α = 7.08 ∗ 10 h ρ = 50 lbm /f t3

(a) Use the Routh stability criterion to find the open loop stability requirements for the linearized
system.
(b) Use the numerical values of the parameters to see if the system is open loop stable.

1
(c) Add to the system a proportional controller that changes cooling jacket temperature by the
relationship:
Tj = T̄j + Kc (T set − T )

What are the closed loop stability requirements?

Solution
(a) Replacing equation (1) into equation (2) we get:

a21 a13 C0 + a24 (s − a11 )T0 + (a21 a15 + a25 (s − a11 ))F + a26 (s − a11 )Tj
T =
s2 − (a22 + a11 )s + a22 a11 − a21 a12

The characteristic polynomial is:

s2 − (a22 + a11 )s + a22 a11 − a21 a12 = 0

The Routh matrix is:  

1 a22 a11 − a21 a12
 −(a22 + a11 ) 0 
a22 a11 − a21 a12 0
For a system to be stable, we require that the elements of the first column have all the same
sign, so:

−(a22 + a11 ) > 0

a22 a11 − a21 a12 > 0

(b) The numerical values of the parameters are:

F̄
a11 = − − k̄ = (−0.8333 − 0.8672)[h−1 ] = −1.700[h−1 ]
V
λkE C̄ F̄ UA
a22 =− 2
− − = (7.118 − 0.8333 − 20.83)[h−1 ] = −14.55[h−1 ]
ρCp RT̄ V V ρCp
Rf t3
◦ 
λk̄
a21 =− = 693.8
ρCp lbmolh
 
C̄E k̄ −3 lbmol
a12 =− = −8.897 ∗ 10
RT̄ 2 f t3 h ◦ R

Replacing the values of the parameters in the conditions for stability gives:

−(a22 + a11 ) = −(−14.55 − 1.700) = 16.25 > 0

a22 a11 − a21 a12 = 14.55 ∗ 1.700 + 693.8 ∗ 8.897 ∗ 10−3 = 30.91 > 0

So the system is open loop stable.

(c) Replacing Tj into equation (2) gives:

dT
= a21 C + ā22 T + a24 T0 + a25 F + a27 T set (3)
dt

Where:

ā22 = a22 − Kc a26

a27 = a26 Kc

2
 By analogy with the open loop case, the stability conditions for the closed loop case are:

−(ā22 + a11 ) > 0

ā22 a11 − a21 a12 > 0

Replacing the value a26 = 20.83[h−1 ] we get:

Kc > −0.7801
Kc > −0.8728

So the system is closed loop stable for every value of Kc (We are assuming here that Kc is a
positive parameter).

Problem 2
Find the ultimate gain and period of a fourth order system given below. The controller is propor-
tional and the system openloop transfer function is:
0.047 ∗ 112 ∗ 2 ∗ 0.12
GM (s) =
(0.083s + 1)(0.017s + 1)(0.432s + 1)(0.024s + 1)

Solution
The closed loop characteristic equation is:

1 + GM (s) B(s) = 0

with B(s) = K. Expanding the characteristic polynomial we get:

0.047 ∗ 112 ∗ 2 ∗ 0.12K

1+ =0
(0.083s + 1)(0.017s + 1)(0.432s + 1)(0.024s + 1)
1.46 ∗ 10−5 s4 + 0.00168s3 + 0.0574s2 + 0.556s + 1 + 1.26K = 0

We are looking for a solution of the form r = wi, replacing in the previous equation we get:

1.46 ∗ 10−5 w4 − 0.00168w3 i − 0.0574w2 + 0.556wi + 1 + 1.26K = 0

Equating the real and imaginay parts of the previous equation we get:

1.46 ∗ 10−5 w4 − 0.0574w2 + 1 + 1.26K = 0 (1)

3
−0.00168w + 0.556w = 0 (2)

From equation (2) we get w2 = 331. Replacing into equation (1) we get K = 13.0. The period is
p = 2π/w = 0.345.1

Problem 3
Use the Routh stability criterion to find the ultimate gain of the closed loop three CSTR system
with a PI controller:

(a) For τI = 3.03

(b) For τI = 4.5

1
GM (s) =
8(s + 1)3
1 With w = 0 we get K = −0.794, we are looking for a positive value of K.

3
Solution
The characteristic polynomial is:
Kc τI s + 1
1+ =0
8(1 + s)3 τI s
8τI s4 + 24τI s3 + 24τI s2 + τI (8 + Kc )s + Kc = 0

The Routh matrix is:  

8τI 24τI Kc

 24τ I τI (8 + Kc ) 0 
24τI − τI (8 + Kc )/3 Kc 0
.. .. ..
 
. . .
At least one root of the characteristic polynomial will be located in the right side of the real axis
if:

24τI − τI (8 + Kc )/3 < 0

24 − (8 + Kc )/3 < 0
Kc > 64

So the ultimate gain is Kc = 64, irrespective of the value of τI .

Problem 4
Find the ultimate gain and period of a closed loop system with a proportional controller and an
open loop transfer function:
1
GM (s) =
(s + 1)(5s + 1)(0.5s + 1)

Solution
The closed loop characteristic equation is:

1 + GM (s) B(s) = 0

with B(s) = K. Expanding the characteristic polynomial we get:

2.5s3 + 8.0s2 + 6.5s + 1 + K = 0

We are searching for a solution of the form r = wi, replacing in the previous equation we get:

−2.5w3 i − 8w2 + 6.5wi + 1 + K = 0

Equating the real and imaginay parts of the previous equation we get:

−2.5w3 + 6.5w = 0 (1)

2
−8w + 1 + K = 0 (2)

Neglecting the case w = 0, from equation (2) we get w2 = 2.6. Replacing into equation (1) we get
K = 19.8. The period is p = 2π/w = 3.9.

Problem 5
Find the value of feedback controller gain that gives a closedloop damping coefficient of 0.8 for the
system with a proportional controller and an openloop transfer function:
s+4
GM (s) =
s(s + 2)

4
Solution
The characteristic equation is:
(s + 4)K
1+ =0
s(s + 2)
s2 + (2 + K)s + 4K = 0
Comparing with the standard form:
ζ 1
s2 + 2 s+ 2 =0
τp τp
We get:
2ζ/τp = 2 + K (1)
1/τp2 = 4K (2)
Replacing equation (2) into equation (1) we get:
K 2 + (4 − 16ζ 2 )K + 4 = 0
The solutions for ζ = 0.8, are: K1 = 5.515 and K2 = 0.7253. The corresponding roots (r) of the
characteristic equation, and value of τp are:

For K1 : For K2 :
r = −3.758 ± 2.818i r = −1.363 ± 1.022i
τp = 0.2129 τp = 0.5870
ζ = 0.8 ζ = 0.8

Problem 6
The liquid level h(t) in a tank is held by a PI controller that changes the flow rate F(t) out of the
tank. The flow rate into the tank F0(t) and the level setpoint hset (t) are disturbances. The vertical
2
cylindrical tank is 10 f t in cross sectional area. The transfer function of the feedback controller
plus the control valve is:
F(s)
 
1
B(s) = = −K 1 +
E(s) τI s
with units of f t3 /minf t.
(a) Write the equations describing the open loop system.
(b) Write the equations describing the closed loop system.
(c) Derive the open loop transfer functions of the system:
GM (s) = H(s) /F(s)
GL(s) = H(s) /F0(s)

(d) Derive the two closed loop transfer functions of the system
set
H(s) /H(s)
H(s) /F0(s)

(e) Make a root locus plot of the closed loop system with a value of integral time τI = 10 minutes.
(f) What value of gain K gives a closed loop system with a damping coefficient of 0.707? What
is the closed loop time constant at this gain?
(g) What gain gives critical damping? What is the time constant with this gain?

5
Solution
(a) For the open loop system we have (A: cross sectional area of the tank):

dh F0 − F
= (1)
dt A

(b) For the closed loop system we have:

 t 
dh F0 K 1
Z
hset − h + hset − h dt



= + (2)
dt A A τI 0

(c) Applying the Laplace transform to equation (1) we get:

F0 − F
h=
sA

The open loop transfer functions are:

1
GM (s) = −
sA
1
GL(s) =
sA

(d) Applying the Laplace transform to equation (2) we get:

τI s K(τI s + 1)
h= F0 + hset
τI As2 + K(τI s + 1) τI As2 + K(τI s + 1)

The closed loop transfer functions are:

set K(τI s + 1)
H(s) /H(s) =
τI As2 + K(τI s + 1)
τI s
H(s) /F0(s) =
τI As2 + K(τI s + 1)

(e) The root locus plot is shown in Figure 1. The roots start at 0, there are repeated roots K = 4
(r = −0.2).

Figure 1: Root locus plot, level control in a tank.

6
(f) The characteristic polynomial is:

K K
s2 + s+ =0
A AτI

Comparing with the standard form s2 + 2ζs/τp + 1/τp2 = 0, we get:

K/A = 2ζ/τp (3)

K/AτI = 1/τp2 (4)

Combining equation (3) and equation (4) we get:

Aζ 2
K=4
τI

For ζ = 0.707, A = 10 and τI = 10:

Aζ 2
K=4 =2
τI
ζA
τp = 2 = 7.07
K

(g) For ζ = 1, using the previous formulas, K = 4, τp = 5.

Code(s) used: P6Code.py

Problem 7
Make a root locus plot of a system with an open loop transfer function:
K τD s + 1
B(s) GM (s) =
(s + 1)(5s + 1)(0.5s + 1) (τD s/20 + 1)

(a) For τD = 2.5

(b) For τD = 5

(c) For τD = 7.5

Solution
The characteristic polynomial is:

1 + B(s) GM (s) = 0
   
2.5 8 6.5 τ
D

τD s 4 + 3
τD + 2.5 s + τD + 8 s 2 + + 6.5 + KτD s + 1 + K = 0
20 20 20 20

The root locus plot for τD = 2.5 is shown in Figure 2 for K from 0 to 5, the roots start at [-8,
-2, -1, -0.2] and end at [-8.75, -1.04+2.21j, -1.04-2.21j, -0.367]. The root locus plot for τD = 5 is
shown in Figure 3 for K from 0 to 5, the roots start at [-4, -2, -1, -0.2] and end at [-6, -0.5+0.278j,
-0.5-0.278j, -0.2]. The root locus plot for τD = 7.5 is shown in Figure 4 for K from 0 to 5, the
roots start at [-2.67, -2, -1, -0.2] and end at [-5.39, -0.169+2.91j, -0.169-2.91j, -0.139]. For each
one of the τD values examined, the roots starting at -2 and -1 converge to a single root, and then
transform into a pair of conjugated roots. At increasing values of τD , for a given value of K, the
real part of the conjugated roots decrease, so the oscillating part of the solution decreases more
slowly.
Code(s) used: P7Code.py P7RootLocus.py

7
 Figure 2: Root locus plot for τD = 2.5.

Figure 3: Root locus plot for τD = 5.

Problem 8
Find the ultimate gain and period of the closed loop three CSTR system with a PID controller
tuned at τI = τD = 1. Make a root locus plot of the system.

Solution
For an open loop transfer function:
1
GM (s) =
(s + 1)3
and a controller with transfer function:
(1 + s + s2 )
B(s) = K
s
The characteristic equation is:

1 + GM (s) B(s) = 0 (1)

4 3 2
s + 3s + (3 + K)s +(1 + K)s + K = 0 (2)

8
 Figure 4: Root locus plot for τD = 7.5.

Figure 5: Root locus plot.

We are looking for a solution of the form r = iw, replacing r in equation (2) we get:

w4 − 3iw3 − (3 + K)w2 + (1 + K)iw + K = 0 (3)

Equating the real and imaginary parts of equation (3) we get:

w4 − (3 + K)w2 + K = 0 (4)
3
−3w + (1 + K)w = 0 (5)

Replacing equation (5) into equation (4) we get:

2K 2 + K + 8 = 0

The solution is Kc = −0.25 ± 1.98i, because there is no real solution, we conclude that the system
is stable for every value of K. The locus plot is shown in Figure 5, for K from 0 to 10, the roots
start at [-1, -1, -1, 0] and end at [-1.06+3.00j, -1.06-3.00j, -0.439+0.890j, -0.439-0.890j].
Code(s) used: P8RootLocus.py

9
 Figure 6: Root locus plot.

Problem 9
Make a root locus plot of a system with an open loop transfer function:

Ke−2s
B(s) GM (s) =
s+1
Use the first order Pade approximation of deadtime. Find the ultimate gain.

Solution
The first order Pade approximation for the exponential function is:
1 + 0.5x
exp(x) ≈
1 − 0.5x
The characteristic polynomial is:

1 + B(s) GM (s) = 0
2
s + (2 − K)s + 1 + K = 0

Replacing the root r = iw, we get the following equations:

−w2 + 1 + K = 0 (1)
(2 − K)w = 0 (2)
√
From equation (2) we get K = 2, replacing in equation (1) we get w = 3. The root locus plot is
shown in Figure 6, for K from 0 to 2, the roots start at -1 and end at [1.73j, -1.73j].
Code(s) used: P9RootLocus.py

Problem 10
When a system has poles that are widely different in value, it is difficult to plot them all on a
root locus plot using conventional rectangular coordinates in the s plane. It is sometimes more
convenient to make the root locus plots in the log s plane. Instead of using the conventional axis
Re s and Im s, an ordinate of the arg s and an abscissa of the log|s| are used (Figure 7), since the
natural logarithm of a complex number is defined:

ln s = ln |s| + iarg(s)

10
(a) Show that the region of stability becomes a horizontal band in the log s plane.
(b) Show that lines of constant time constant in the s plane transform into vertical lines in the log
s plane.
(c) Show that lines of constant damping coefficient in the s plane transform into horizontal lines
in the log s plane.
(d) Make log s plane root locus plots of systems with the open loop transfer functions B(s) GM (s) :
(i)
Kc
8(s + 1)3
(ii)
Kc
(s + 1)(5s + 1)(0.5s + 1)
(iii)
Kc (−3s + 1)
(s + 1)(5s + 1)

Figure 7: arg s - log|s| plane.

Solution
(a) The stability region is the left half side of the real/imaginary plane, that is equivalent to an
argument value in the interval ]π/2, 3π/2[.
(b) For a characteristic polynomial of the form:
ζ 1
s2 + 2 s+ 2
τp τp
The roots are:
p
ζ 1 − ζ2
r=− ± i
τp τp
The norm of the root is:
1
|r| =
τp
So a constant value of τp is equivalent to a constant value of |s|.
(c) The angle between the horizontal axis and the root is:
arg(r) = cos−1 ζ
So a constant value of ζ is equivalent to a constant value of arg(s).
(d) The root locus plots are shown in i) Figure 8 for Kc between 0 and 64, the roots start at [(1,π)],
ii) Figure 9 for Kc between 0 and 19.8, the roots start at [(2,π),(1,π),(0.2,π)], iii) Figure 10 for
Kc between 0 and 2, the roots start at [(1,π),(0.2,π)].
Code(s) used: P10RootLocus.py

11
 Figure 8: Root locus plot for B(s) GM (s) = Kc /8(s + 1)3 .

Figure 9: Root locus plot for B(s) GM (s) = Kc /(s + 1)(5s + 1)(0.5s + 1).

Figure 10: Root locus plot for B(s) GM (s) = Kc (−3s + 1)/(s + 1)(5s + 1).

12
Problem 11
A two tank system with recycle is sketched in Figure 11. Liquid levels are held by proportional
controllers F1 = K1 h1 and F2 = K2 h2 . Flow into the system F0 and recycle flow FR can be varied
by the operator.
(a) Derive the four closed loop transfer functions relating the two levels and the two load distur-
bances:
(H1 /F0 )(s)
(H1 /FR )(s)
(H2 /F0 )(s)
(H2 /FR )(s)

(b) Does the steady state level in the second tank vary with the recycle flow rate FR ? Use the
final value theorem of Laplace transforms.

Figure 11: Two tank system with recycle.

Solution
(a) For tank 1 we have (A1 : the cross sectional are of tank 1):
dh1
A1 = F0 + FR − K1 h1
dt
Applying the Laplace transform to the previous equation we get:
F0 FR
h1 = + (1)
A1 s + K 1 A1 s + K 1
For tank 2 we have:
dh2
A2 = K 1 h1 − K 2 h2 − FR
dt
Applying the Laplace transform to the previous equation we get:
K1 h1 FR
h2 = − (2)
A2 s + K 2 A2 s + K 2
K 1 F0 K 1 FR FR
= + − (3)
(A1 s + K1 )(A2 s + K2 ) (A1 s + K1 )(A2 s + K2 ) A2 s + K2
K 1 F0 A1 sFR
= − (4)
(A1 s + K1 )(A2 s + K2 ) (A1 s + K1 )(A2 s + K2 )

13
 From equation (1) and equation (4) we have:
1
(H1 /F0 )(s) =
A1 s + K1
1
(H1 /FR )(s) =
A1 s + K1
K1
(H2 /F0 )(s) =
(A1 s + K1 )(A2 s + K2 )
A1 s
(H2 /FR )(s) =−
(A1 s + K1 )(A2 s + K2 )

(b) The steady state value of the level in tank 2 for a step change in FR is:

∆h2 A1 s
= lim − =0
∆FR s→0 (A1 s + K1 )(A2 s + K2 )

Problem 12
The system of problem 4 is modified by using the cascade control system sketched in Figure 12.
(a) Find the value of the gain Ks in the proportional controller that gives a 0.707 damping coeffi-
cient for the closed loop slave loop.

(b) Using this value of Ks in the slave loop, find the maximum closed loop stable value of the master
controller gain Km . Compare this with the ultimate gain found without cascade control in
problem 4. Also compare ultimate periods.

Figure 12: Cascade control system.

Solution
(a) For the slave loop we have:
xD xm
x1 = + (1)
(0.5s + 1)(s + 1) (0.5s + 1)(s + 1)
= −Ks x1 + Km xset


xm 2 − x2 (2)

Combining equation (2) and equation (1) we get:

xD Km (xset
2 − x2 )
x1 = +
(0.5s + 1)(s + 1) + Ks (0.5s + 1)(s + 1) + Ks

The characteristic polynomial for the slave loop is:

0.5s2 + 1.5s + 1 + Ks

14
 Comparing with the standard polynomial s2 + 2ζs/τp + 1/τp2 we get:

3 = 2ζ/τp
2(1 + Ks ) = 1/τp2
√
For ζ = 0.5 the solution is Ks = 1.25, τp = 0.471.

(b) For the master loop we have:

x1 xD Km (xset
2 − x2 )
x2 = = +
5s + 1 (5s + 1)((0.5s + 1)(s + 1) + Ks ) (5s + 1)((0.5s + 1)(s + 1) + Ks )
xD Km xset
2
= +
(5s + 1)((0.5s + 1)(s + 1) + Ks ) + Km (5s + 1)((0.5s + 1)(s + 1) + Ks ) + Km

The master loop characteristic equation is (with Ks = 1.25):

(5s + 1)((0.5s + 1)(s + 1) + 1.25) + Km = 0

2.5s3 + 8s2 + 12.75s + 2.25 + Km = 0

Replacing a solution of the form r = iw, we get, after separating the real and imaginary parts
of the characteristic equation:

−2.5w3 + 12.75w = 0
−8w2 + 2.25 + Km = 0

The solution is w = 2.26, p = 2.78, Km = 38.6. In problem 4 we got at the stability limit
K = 19.8, p = 3.9, so in this case using a master slave configuration allows the use of a greater
value of K, and increases the period of oscillation at the stability limit.

Problem 13
Repeat problem 6 using a proportional feedback controller [parts (b) and (d)]. Will there be a
steady state error in the closed loop system for (a) a step change in setpoint hset or (b) a step
change in feed rate F0 ?

Solution
The closed loop equations are:
dh F0 − F
= (1)
dt A
F = −Kc (hset − h) (2)

Combining equation (1) and equation (2) we get, after applying the Laplace transform:

F0 Kc hset
h= +
sA + Kc sA + Kc

The steady state values for changes in hset and F0 are:

Kc ∆hset
h(∆hset ) = lim = ∆hset
s→0 sA + Kc
F0 F0
h(∆F0 ) = lim =
s→0 sA + Kc Kc

15
Problem 14
We would like to compare the closed loop dynamic performance of two types of reboilers.
(a) In the first type, there is a control valve on the steam line to the reboiler and a steam trap on
the condensate line leaving the reboiler. The flow transmitter acts like a first order lag with
a 6 second time constant. The control valve also acts like a 6 second lag. These are the only
dynamic elements in the steam flow control loop. If a PI flow control is used with τI = 0.1
minutes, calculate the closed loop time constant of the steam flow loop when a closed loop
damping coefficient of 0.3 is used.
(b) In the second type of reboiler, the control valve is on the condensate line leaving the reboiler
(Figure 13). There is no valve in the steam line so the tubes in the reboiler see full steam
header pressure.
Changes in steam flow are achieved by increasing or decreasing the area used for condensing steam
in the reboiler. This variable area flooded reboiler is used in some processes because it permits the
use of lower pressure steam. However, as you will show in your calculations (I hope), the dynamic
performance of this configuration is distinctly poorer than direct manipulation of steam flow.
The steam flow meter still acts like a first order lag with a 6 second time constant, but the smaller
control valve on the liquid condensate can be assumed to be instantaneous.
The condensing temperature of the steam is 300◦ F. The process into which the heat is transferred is
at a constant temperature of 200◦ F. The overall heat transfer coefficient is 300 BT U/h ◦ F f t2 . The
reboiler has 509 tubes that are 10 ft long and 1 inch inside diameter. The steam and condensate are
inside the tubes. The density of the condensate is 62.4 lbm /f t3 and the latent heat of condensation
of the steam is 900BT U/lbm . Neglect any sensible heat transfer.
Derive a dynamic mathematical model of the flooded condenser system. Calculate the transfer
function relating steam flow rate to condensate flow rate. Using a PI controller with τI = 0.1
minute, calculate the closed loop time constant of the steam flow control loop when a closed loop
damping coefficient of 0.3 is used. Compare this with the result found in (a).

Figure 13: Control valve on the condensate alternative.

Solution
(a) The block diagram for part is shown Figure 14.
From the block diagram, we have:
  
6s + 1 set F 1
F =K F −
6s 6s + 1 6s + 1
 
K F
= F set −
6s 6s + 1
(6s + 1)K
= F set
36s2 + 6s + K

16
 Figure 14: Block diagram, control valve on the steam line.

From the standard form of the characteristic equation we get:

2ζ/τp = 0.6/τp = 1/6

1/τp2 = K/36

The solution is τp = 3.6, K = 2.778.

(b) Assuming that the mass of liquid inside the tubes is mostly liquid, his variation can be described
by (h: level of liquid inside the tubes [f t], At : cross sectional area of the tubes [f t2 ], Fs : steam
flow [lbm /s], Fc : condensate flow [lbm /s]):

dh
At ρ = F s − Fc (3)
dt

The steam flow is related to the heat transfer rate (Ah : heat transfer area [f t2 ], N : number
of tubes):

Q U ∆T Ah U ∆T N πD(L − h)
Fs = = = (4)
λ λ λ

Replacing equation (4) into equation (3) we get:

At ρλ dFs
− = F s − Fc (5)
U ∆T N πD dt
Dρλ dFs
− = F s − Fc (6)
4U ∆T dt
Fc
Fs = (7)
140.4s + 1

The condensate flow is set by the controller:

  
Fs 6s + 1
Fc = K Fsset −
6s + 1 6s

Replacing equation (7) into equation (8) we get:

Fs 6s(140.4s + 1) = K(Fsset (6s + 1) − Fs )

K(6s + 1)Fsset
Fs =
842.4s2 + 6s + K

17
 Again comparing with the standard form, we get:
2ζ/τp = 6/842.4
1/τp2 = K/842.4

The solution is τp = 84.2 and K = 0.119. Comparing the respective values of τp , we observe
that the second option, for identical damping coefficient, takes longer to stabilize.

Problem 15
A chemical reactor is cooled by both jacket cooling water and condenser cooling water (Figure 15).
A mathematical model of the system has yielded the following open loop transfer function (time
is in minutes):
T 1
=−
Fc s+1
T 5
=−
Fj 10s + 1
The span of the temperature transmitter is 100-200◦ F. Control valves have linear trim and constant
pressure drop, and are half open under normal conditions. Normal condenser flow is 30 gpm.
Normal jacket flow is 20 gpm. A temperature measurement lag of 12 seconds is introduced into
the system by the thermowell.
If a proportional feedback temperature controller is used, calculate the controller gain K that
yields the closed loop damping coefficient of 0.707 and calculate the closed loop time constant of
the system when:
(a) Jacket water only is used.
(b) Condenser water only is used.
Derive the closed loop characteristic equation for the system when both jacket and condenser water
are used.

Figure 15: Reactor cooled by jacket cooling and condensation.

Solution
For both jacket (j) and condenser (c) we have:
√
(Cv ∆P )c = 30/0.5 = 60
√
(Cv ∆P )j = 20/0.5 = 40

18
For the jacket, the relation between Fj and reactor temperature is:

T(t−0.2) − T set
Fj = 40 K
100
0.4 − 0.04s
Fj = 0.4e−0.2s (T − T set )K ≈ (T − T set )K
1 + 0.1s
Replacing in the expression for the transfer function between Fj and T , we have:
5 0.4 − 0.04s
T =− (T − T set )K
10s + 1 1 + 0.1s
(2 − 0.2s)K
T = 2 T set
s + (10.1 − 0.2K)s + 1 + 2K
The characteristic equation is:
s2 + 2ζs/τp + 1/τp2 = 0
So, we have:

10.1 − 0.2K = 2ζ/τp

1 + 2K = 1/τp2

The solution is: K=13.32, τp =0.19.

For the condenser, we have:
0.6 − 0.06s
Fc ≈ (T − T set )K
1 + 0.1s
The relation between T and T set is:
(0.6 − 0.06s)K
T = T set
0.1s2 + (1.1 − 0.06K)s + 1 + 0.6K
The values of K and τp are obtained from:

1.1 − 0.6K = 2ζ/τp

10 + 6K = 1/τp2

The solution is: K=4.27, τp =0.17.

When both the condenser and jacket are used we have after replacing the expression for Fc and
Fj :
5 1
T =− Fj − Fc
10s + 1 s+1
(2 − 0.2s)(T set − T )K (0.6 − 0.06s)(T set − T )K
T = +
(10s + 1)(1 + 0.1s) (s + 1)(1 + 0.1s)
2
(−0.8s + 7.74s + 2.6)K
T = 3 T set
s + (11.1 − 0.8K)s2 + (11.1 + 7.74K)s + 1 + 2.6K

Problem 16
Oil and water are mixed togheterand then decanted (Figure 16). Oil flow rate is ratioed to water
flow rate FW . Interface is controller by oil flow FL from the decanter with a proportional level
controller. Water flow (FH ) from the decanter, which is liquid full, is on pressure control (PI).
Steadystate flow rates are:

Oil = 177.8 gpm

Water = 448.2 gpm

Water and oil holdups in the decanter are each 1130 gallons at steady state.

19
(a) Derive the closed loop transfer function between FH and FW in terms of level controller gain
K.
(b) Determine the transient response of FH for a step change in FW of 1 gpm when level controller
gain is 6.36 gpm/gal.

Figure 16: Oil water decanter.

Solution
1. The equations that describe the system are (r: ratio between oil inlet flow and water inlet
flow, w: water hold up, o: oil hold up):

sw = FW − FH (1)
so = FW r − FL (2)
FL = Ko (3)
 
1 + τs
FH = K p (w + o) (4)
τs

Replacing equation (1) and (2) into equation (3) and (4) we get:
KFW r
FL = (5)
s+K
 
1 + τs
FH = Kp (FW (1 + r) − FH − FL ) (6)
τ s2

Finally, replacing equation (5) into equation (6) we get:

FW Kp (1 + τ s)(s(r + 1) + K)
FH =
(s + K) (τ s2 + Kp (1 + τ s))

2. A simplified numerical can be obtained, assuming that the PI controller acts perfectly, so
instead of equation (4), we use:

0=w+o (7)

From equation (1), (2) and (7) we get:

FW (1 + r) − FH − FL = 0 (8)

20
 Combining equation (3), (1) and (7) we get:

K(FW − FH )
FL = − (9)
s
Finally, replacing equation (9) into equation (8) we get:

K(FW − FH )
FW (1 + r) − FH + =0
s
((1 + r)s + K)
FH = FW
s+K
((1 + r)s + K)
= ∆FW
s(s + K)
 
1 r
= ∆FW +
s s+K
= ∆FW (1 + re−Kt )
= 1 + 0.3967e−6.36t

Problem 17
The open loop transfer function relating steam flow rate to temperature in a feed preheater has
been found to consist of a steady state gain Kp and a first order lag with the time constant τp . The
lag associated with temperature measurement is τm . A proportional only temperature controller
is used.
(a) Derive an expression for the roots of the closed loop characteristic equation in terms of the
parameters τp , Kp , τm and Kc .
(b) Solve for the value of controller gain that will give a critically damped closed loop system when
Kp = 1, τp = 10 and
(i) τm = 1
(ii) τm = 5

Solution
(a) The characteristic equation is:
Kp Kc
1+ =0
(τp s + 1) (τm s + 1)
τ p + τm 1 + Kc Kp
s2 + s+ =0
τp τm τp τ m

The roots are:

p
τ p + τm (τp + τm )2 − 4(1 + Kp Kc )τp τm
r=− ±
2τp τm 2τp τm

(b) We have:
2ζ 2 τp + τ m 10 + τm
= = =
τ τ τp τ m 10τm
1 1 + Kc Kp 1 + Kc
= =
τ2 τp τm 10τm

(i) The solution is τ = 1.818, Kc = 2.025, the corresponding root is -0.55.

(ii) The solution is τ = 6.667, Kc = 0.125, the corresponding root is -0.15.

21
Problem 18
The liquid flow rate from a vertical cylindrical tank, 10 feet in diameter, is flow controlled. The
liquid flow into the tank is manipulated to control liquid level in the tank. The control valve on
the inflow stream has linear installed characteristics and can pass 1000 gpm when wide open. The
level transmitter has a span of 6 feet of liquid. A proportional controller is used with a gain of 2.
Liquid density is constant.
(a) Should the control valve be A-O or A-C.
(b) Should the controller be reverse or direct acting?
(c) What is the dimensionless open loop system transfer function relating liquid height and inflow
rate?
(d) Solve for the time response of the inflow rate to a step change in the outflow rate from 500 to
750 gpm with the tank initially half full.

Solution
(a) In case of failure, we want to store the liquid into the tank, so the valve is AC.
(b) If the level increases, the signal to the valve must increase, so the controller is direct acting
(the gain is negative).
(c) The equation that describe the level inside the tank is (A: Cross sectional area of the tank, i:
inflow, o: outflow):
dh
A = F i − Fo (1)
dt
So, the transfer function between inlet flow and level is:
h 1
=
Fi sA

(d) The signal to the valve (CO) and inlet flow are (B: Bias, i, m: Maximum inlet flow):

CO = B − 2(hset − h) (2)
Fi = (1 − CO)Fi,m (3)

Replacing equation (2) and equation (3) into equation (1) we get:
dh
A = (1 − B + 2(hset − h))Fi,m − Fo (4)
dt
Expressing equation (4) in terms of perturbation variables we get2 :

dh̄
A = (1 − B + 2(h̄set + hset
ss − h̄ − hss ))Fi,m − F̄o − Fo,ss (5)
dt

The variables at steady state satisfy:

0 = (1 − B + 2(hset
ss − hss ))Fi,m − Fo,ss (6)

Replacing equation (6) into equation (5) we get:

dh̄
A = 2Fi,m (h̄set − h̄) − F̄o
dt
2Fi,m F̄o
h̄ = h̄set −
A(s + 2Fi,m /A) A(s + 2Fi,m /A)
2v = v̄ + vss , v: generic variable, v̄: perturbation variable, vss : steady state value

22
 Replacing the values Fi,m = 1000[gpm] = 133.7[f t3 /min], A = 78.54[f t2 ], hset = 0, Fo =
250[gpm]/s = 33.42[f t3 /min]/s we can obtain h̄ as a function of time.
0.4255
h̄ = −
s(s + 3.405)
 
1 1
h̄ = 0.1250 −
s + 3.405 s
h̄ = 0.1250(e−3.405t − 1)

With t in [min]. The expression for h is (hss = 3, the tank is initially half full):

h = 3 + 0.1250(e−3.405t − 1)

At t = 0, hset = 3, and (1 − CO) = 0.5, so B = 0.5. Replacing into equation (3) we get:

Fi = (1 − B + 2(hset − h))Fi,m
= (1 − 0.5 + 2(3 − (3 + 0.1250(e−3.405t − 1))))1000[gpm]
= (0.5 − 0.250(e−3.405t − 1))1000[gpm]
= (750 − 250e−3.405t )[gpm]

Problem 19
A process has a positive pole located at (+1, 0) in the s plane (with time in minutes). The process
steady state gain is 2. An addition lag of 20 seconds exits in the control loop. Sketch root locus
plots and calculate controller gains which give a closed loop damping coefficient of 0.707 when:
(a) A proportional feedback controller is used.
(b) A proportional derivative feedback controller is used with the derivative time set equal to the
lag in the control loop. The ratio of numerator to denominator time constants in the derivative
unit is 6.

Solution
2
G(s) =
s−1
(a)

3K
B(s) =
s+3
The characteristic equation is:

1 + GB = 0
2
s + 2s + 6K − 3 = 0

From√ the standard form of the characteristic equation we solve for both τp and K, given
ζ = 0.5:

2ζ/τp = 2
1/τp2 = 6K − 3
√
The solution is: τp = 0.5, K = 5/6. A√root locus plot is shown in Figure 17 for K from 0 to
1. The dashed lines correspond to ζ = 0.5, the roots start at (1, -3).

23
 Figure 17: Root locus plot for proportional control.

Figure 18: Root locus plot for proportional derivative control.

(b)
3K(6s + 18)
B(s) =
(s + 3)(s + 18)
The characteristic equation is:
1 + GB = 0
s3 + 20s2 + (36K + 33)s + 108K − 54 = 0
A root locus plot is shown in Figure 18 for K from 0 to 5, the roots start at (1, -3, -18). The
damping coefficient is 0.707 for K = 4.5139 with roots (-3, -8.5+8.5j, -8.5-8.5j).
Code(s) used: P19FindKPI.py P19Figure.py

Problem 20
A process has an open loop transfer function that is a first order lag with a time constant τp and
a steady state gain Kp . If a PI feedback controller is used with a reset time τI , sketch root locus
for the cases where:

24
(a) τp > τI
(b) τp < τI
(c) τp = τI
√
What value of the τp /τI ratio gives a closed loop system that has a damping coefficient of 0.5 2
for only one value of the controller gain?

Solution
The characteristic equation is:
Kp Kc (τI s + 1)
1+ =0
(τp s + 1) τI s
(1 + Kp Kc ) Kp Kc
s2 + s+ =0
τp τI τ p

(a) The root locus plot for the case τp = 1.03, τI = 0.26τp , and Kp Kc from 0 to 6 is shown in
Figure 19, the roots start at [-0.971, 0] and end at [-3.40+3.19j, -3.40-3.19j].
(b) The root locus plot for the case τp = 1.03, τI = 1.27τp , and Kp Kc from 0 to 6 is shown in
Figure 20, the roots start at [-0.971, 0] and end at [-6.06, -0.735].
(c) The root locus plot for the case τp = 1.03, τI = τp , and Kp Kc from 0 to 6 is shown in Figure
21, the roots start at [-0.971, 0] and end at [-5.83, -0.971].
√
Comparing with the standard form of the characteristic equation we get (with ζ = 0.5 2):
√
2/τ = (1 + Kp Kc )/τp (1)
Kp Kc /(τI τp ) = 1/τ 2 (2)

Replacing equation (1) into equation (2) we get:

τp
(Kp Kc )2 + Kp Kc (2 − 2 )+1=0
τI
For this equation to have only one solution, its determinant must equal zero:
τp 2
(2 − 2 ) −4=0
τI
The solution is τp /τI = [0, 2], a root locus plot for the case τp√= 2τI is shown in Figure 22 for
Kp Kc from 0 to 6, together with a line corresponding to ζ = 0.5 2, the root locus curve touch the
line at the value Kp Kc = 1.
Code(s) used: P20Figure.py

Problem 21
The open loop process transfer functions relating the manipulated and load load variables (M and
L) to the controlled variable (X) are first order lags with identical time constants (τ ) but with
different gains (KM and KL ). Derive equations for the closed loop steady state error and the closed
loop time constant for step disturbances in load if a proportional feedback controller is used.

Solution
The controlled variable X and manipulated variable are, respectively:
KL KM
X= L+ M (1)
τs + 1 τs + 1
M = Kc (X set − X) (2)

25
Figure 19: τp > τI .

Figure 20: τp < τI .

Figure 21: τp = τI .

26
 Figure 22: τp = 2τI .

Replacing equation (2) into equation (1) gives the closed loop equation for X:
KL KM Kc
X= L+ X set
τ s + 1 + K c KM τ s + 1 + Kc K M
The steady state error is (E):
KL KL
E = lim ∆L = ∆L
s→0 τ s + 1 + Kc KM 1 + Kc KM
The closed loop time constant is:
τ
τp =
1 + Kc KM

Problem 22
The liquid level in a tank is controlled by manipulating the flow out of the tank, using a PI
controller. The outflow rate is a function of only the valve position. The valve has linear installed
characteristics and passes 20 f t3 /min wide open.
The tank is vertical and cylindrical with a cross sectional area of 25f t2 and a 2 ft level transmitter
span.
(a) Derive the relationship between the feedback controller gain K and the reset time τI which
gives a critically damped closed loop system.
(b) For a critically damped system with τI = 5 minutes, calculate the closed loop time constant.

Solution
(a) For the valve, we have:
p
Cv ∆P/s.g. = 20

The level inside the tank (h) can be described by (A: cross sectional area of the tank):
(τI s + 1)
sAh = 20(hset − h)K
τI s
set
h 20K(τI s + 1)/(τI s)
h=
sA + 20K(τI s + 1)/(τI s)

27
 The characteristic polynomial is:
sA + 20K(τI s + 1)/(τI s) = 0
τI As2 + 20KτI s + 20K = 0

For a critically damped system we have:

1/τp2 = 20K/(τI A) (1)
2/τp = 20K/A (2)

Replacing equation (2) into equation (1) we get:

KτI = A/5 = 5[f t2 ] (3)

(b) From equation (3) we get K = 1[f t2 /min]. Replacing into equation (2) we obtain τp = 2.5
[min].

Problem 23
A process has an open loop transfer function GM relating controlled and manipulated variables
that is a first order lag τp and steady state gain Kp . There is an additional first order lag τm in
the measurement of the controlled variable. A proportional only feedback controller is used.
Derive an expression relating the controller gain Kc to the parameters τp , τm and Kp such that
the closed loop system damping coefficient is 0.707. What happens to Kc as τm gets very small or
very large? What is the value of τm that provides the smallest value of Kc ?

Solution
The closed loop characteristic equation is:
Kp Kc
1+ =0
(τp s + 1) (τm s + 1)
τp τm s2 + (τp + τm )s + Kp Kc + 1 = 0
√
For ζ = 0.5 we have:
1 Kp Kc + 1
2
= (1)
τ τp τ m
√
0.5 τp + τm
2 = (2)
τ τ p τm
Replacing equation (2) into equation (1) we get:
!
1 τp2 + τm
2
+ 2τp τm
Kc = −1
Kp 2τp τm
 
1 τp
lim Kc = lim −1 =∞
τm →0 τm →0 Kp 2τm
 
1 τm
lim Kc = lim −1 =∞
τm →∞ τm →∞ Kp 2τp
To find the smallest value of Kc we take its derivative with respect to τm :
dKc (τm + τp ) (τp2 + τm
2
+ 2τp τm )
Kp = − 2τp
dτm τp τ m (2τp τm )2
dKc τ m − τp
Kp =
dτm 2τp τm
The minimum corresponds to a derivative value equal to zero, so τm = τp :
Kc,min = 1/Kp

28
Problem 24
A fixed gain relay is used as a low base level override controller on a distillation column. The
column is 7 ft in diameter and has a base level transmitter span of 3 ft. The density of the liquid
in the base is 50 lbm /f t3 . Its heat of vaporization is 200 BT U/lbm .
The reboiler steam valve has linear installed characteristics and passes 30000 lbm /h when wide
open. Steam latent heat is 1000 BT U/lbm .
There is a first order dynamic lag of τ minutes between a change in the signal to the steam valve
and vapor boilup. The low base level override controller pinches the reboiler steam valve over the
lower 25 percent of the level transmitter span.
Solve for the value of the gain that should be used in the relay as a function of τ to give a closed
loop damping coefficient of 0.5 for the override level loop.

Solution
When the override control is acting, the flow of steam is (K: relay gain, Fs : steam flow):
hK
CO = (3)
0.75
CO
Fs = M (4)
τs + 1
Where the constant M is equal to 30000 lbm /h. On the other hand the relation between the level
and the steam flow rate is (L: load variable, A: cross sectional area of the tank):

dh λs
ρA = −Fs +L
dt λ
Replacing the expression for Fs we get:
dh hKM λs
ρA =− +L
dt 0.75(τ s + 1) λ
L(τ s + 1)
h=  
KM λs
ρAτ s2 + τs + 0.75ρAτ λ

L(τ s + 1)
h=
1924τ (s2 + s/τ + 1.733K/τ )
Comparing with the standard form of the characteristic equation we get:
2ζ/τp = 1/τ
1/τp2 = 1.733K/τ
The solution is τp = τ , K = 1.733/τ .

Problem 25
A process has GM and GL open loop transfer functions that are:
KM
GM =
(τ s + 1)2
KL
GL =
(τ s + 1)2
If a PI controller is used with τI set equal to τ , calculate:
(a) The value of the controller gain that gives a closed loop damping coefficient of 0.707.
(b) The closed loop time constant, using this value of gain.
(c) The closed loop transfer function between the load variable and the output variable.
(d) The steady state error for a step change in the load variable.

29
Solution
(a) With a controller transfer function:
 
τs + 1
B(s) = Kc
τs

The closed loop characteristic equation is:

 
KM τs + 1
1+ Kc =0
(τ s + 1)2 τs
(τ s + 1)2 τ s + KM Kc (τ s + 1) = 0
τ 2 s 2 + τ s + KM K c = 0

From the standard form of the characteristic equation we get:

√
2 0.5/τp = 1/τ
1/τp2 = KM Kc /τ 2
√
The solution is τ /τp = 0.5 and KM Kc = 0.5.
√
(b) τp = τ / 0.5.

(c) x: response variable, L: load variable.

x GL
=
L 1 + GM B
KL τ s
=
(τ s + 1)2 τ s + KM Kc (τ s + 1)

(d) The steady state error is:

KL τ s∆L
lim x = lim =0
t→∞ s→0 (τ s + 1)2 τ s + KM Kc (τ s + 1)

Problem 26
Two tanks are connected by a pipe (Figure 23) through which liquid can flow in each direction,
depending on the difference in liquid levels.

Fc = Kb (h1 − h2 )

where KB is a constant with units of f t2 /min. Disturbances are the flow rates F01 , F02 and F2 .
The manipulated variable is the flow rate F1 . Cross sectional areas of the vertical cylindrical tanks
are A1 and A2 .
(a) Derive the open loop transfer function between the height of liquid in tank 1 (h1 ) and F1 .
(b) What are the poles and zeros of the open loop transfer function? What is the open loop
characteristic equation?

(c) If a proportional only level controller is used, derive the closed loop characteristic equation
and sketch a root locus plot for the case where A1 = A2 .

30
 Figure 23: Two tanks connected by a pipe.

Solution
(a) The equations that describe the level on each tank are:

dh1
A1 = F01 − Kb (h1 − h2 ) − F1
dt
dh2
A2 = F02 + Kb (h1 − h2 ) − F2
dt
Applying the Laplace transform we get:

h1 (sA1 + Kb ) = F01 + Kb h2 − F1 (1)

h2 (sA2 + Kb ) = F02 + Kb h1 − F2 (2)

Replacing equation (2) into equation (1) we get:

(F01 − F1 )(sA2 + Kb ) Kb (F02 − F2 )

h1 = + (3)
A1 A2 s2 + Kb (A1 + A2 )s A1 A2 s2 + Kb (A1 + A2 )s

The transfer function between h1 and F1 is:

h1 (sA2 + Kb )
=− 2
F1 A1 A2 s + Kb (A1 + A2 )s

(b) The open loop characteristic equation is:

A1 A2 s2 + Kb (A1 + A2 )s = 0

The zero is s = −Kb /A2 . The poles are: s = 0 and s = −Kb (A1 + A2 )/(A1 A2 ).

(c) With proportional control, we have F1 = −K(hset − h), replacing into equation (3) and with
A1 = A2 we get:

(F01 + Khset )(sA + Kb ) Kb (F02 − F2 )

h1 = 2 2
+ 2 2
A s + A(2Kb + K)s + Kb K A s + A(2Kb + K)s + Kb K

The closed loop characteristic equation is:

A2 s2 + A(2Kb + K)s + Kb K = 0

A root locus plot for the case A = 5, Kb = 3.1 is shown in Figure 24, both roots are real and
negative, and start at 0 and -1.24.

Code(s) used: P26Code.py

31
 Figure 24: Root locus plot for A = 5, Kb = 3.1.

Problem 27
Our old friend the three isothermal CSTR process has the following open loop transfer function
relating controlled and manipulated variables:
1
GM (s) =
8(s + 1)3

When a proportional feedback

√ controller is used, this process has an ultimate gain of 64 and an
ultimate frequency of 3.
(a) Find the ultimate gain and ultimate frequency when a proportional derivative (PD) controller
is used with derivative time set equal to 1.
τD s + 1
B(s) = K
0.1τD s + 1

(b) Sketch root locus plots using:

(i) A proportional controller.
(ii) A PD controller with τD = 1.
(iii) A PD controller with τD much less than 1.
(iv) A PD controller with τD much greater than 1.

Solution
(a) The characteristic equation is:

1 K(s + 1)
1+ 3
=0
8(s + 1) (0.1s + 1)
0.8s4 + 10.4s3 + 26.4s2 + (24.8 + K)s + 8 + K = 0

Replacing a root of the form iw and separating the real and imaginary parts we get:

0.8w4 − 26.4w2 + 8 + K = 0
−10.4w3 + (24.8 + K)w = 0

The solution is: w = 4.58, K = 193.

32
 Figure 25: Root locus plot for proportional only control.

(b) The root locus plot for the proportional controler is shown in Figure 25, for K from 0 to 64,
the roots start at [-1, -1, -1] and end at [-3.0, 1.73j, -1.73j]. The characteristic equation is:

K
1+ =0
8(s + 1)3
8s3 + 24s2 + 24s + K + 8 = 0

The root locus plot for τD = 1 is shown in Figure 26, for K from 0 to 193. The roots start at
[-10, -1, -1, -1] and end at [-12.0, 4.58j, -4.58j, -1]. The characteristic equation is:

0.8s4 + 10.4s3 + 26.4s2 + (24.8 + K)s + 8 + K = 0

The root locus plot for τD << 1 is shown in Figure 27, this is equivalent to a proportional
controller so the maximum value of K is 64. The characteristic equation is:

8s3 + 24s2 + 24s + K + 8 = 0

The root locus plot for τD >> 1 is shown in Figure 28, for K from 0 to 6.4. The roots start
at [-1, -1, -1] and end at [-3, 1.73j, -1.73j]. The characteristic equation is:

K
1+ =0
0.8(s + 1)3
0.8s3 + 2.4s2 + 2.4s + K + 0.8 = 0

Code(s) used: P27Code.py

Conversion factors

1[gal] = 0.1337[f t3 ]

33
Figure 26: Root locus plot for τD = 1.

Figure 27: Root locus plot for τD << 1.

Figure 28: Root locus plot for τD >> 1.

34