Applications

62 Process Control: Princlples and

temperature drops by
heat, and its
A water tank is gradually losing at 4'K per
Example 3.4 gains temperature
When the heater is on, the system zone of
a neutral
2 per miute. control lag and
controller has a 0.5 min. and find
mimte. A wo position versus time graph
temperature
4% about a set point
of 323°K. Plot the and the heater
temperature
is at set point
Assume at t =0,
theoscillation period.
is oft.

Soktion will drop

value with heater off, then temperature
Assume we start at the set point

linearly at 2°K per minute of

Neutral zone of 4% plies temperature band

x 323 323 + 13 336K, 310°K

323 (4/100)
temperature will
start dropping linearly at
From the set point (323°K), the
till it reaches 310°K.
2°Kmin
323-2 3 1 0

I6.5 min
Control lag is of 0.5 min, so the temperature will further drop by I degree to

309°K
When the temperature reaches 309°K, heater becomes active and the
temperature starts building up at the rate of 4°K/min till it reaches 336°K.
309+4 (-7)=336
-13.75 min.
Because of control lag, the time required will increase up to 14.25 min.,
temperature would become 338°K, and the heater would be off.

323 338-2(t-14.25)
I= 21.75 min.
Oscillation period is 21.75 min. There is an undershoot of 1°K and over-
shoot of 2°K around neutral zone.
If neutral zone is
reduced,
time period will also reduce. The
time response is shown in temperature vs.
Fig. E3.4.
lemperature
338K
336K

21.75
14.25 r (min)
31OK
309K

Fig. E3.4 Response of system

 Theory of Controllers 67

Valve A is lincar with a Now scale factor of i0per percent controller output.
The controller output is nominally So% with &, = 10% per %. A load change

occurs when the flow through valve B changes from 500to 550Set
hr hr
point is at 6 m. and the liquid level can vary from 1-13m. Calculate the new level
upon load change.

Solution
When flow through the outlet pipe increases to 550 m'/hr, the inflow must in-
crease up to 550 m°/hr; otherwise, the tank will become empty
550 m'hr of inflow can be obtained by new controller output of 55%. So E,
corresponding to 55% controller output, with Po as 50% can be found (55%
controller output inflow 550 m'/hr).
=

E- 5% = 0.5%
10%/%
Now, if inflow is increased to 550 m'/hr, the liquid level also changes accord-
ngly.
r-b x 100 = 6-bX100 0.5
13-
bmaxmin
b 5.94 m
IfK, was more, the error would be less and 'b (representing new level) will be
closer tor (set point= 6 m)

Example 3.6 A proportional controller is used to control the height of water in a

tank where the water level can vary from 0-9 m. Calculate the value of PB and
K, that will set the water level at a desired height of 5.0 m. The controller is
required to fully close the inlet valveA when water level rises to 5.5 m and fully
open the inlet valve A when the water level falls down to 4.5 m.o

Solution
A variation of 1 m (5.5 m-4.5 m) in a span of9.0 m will change the controller
output by 100%.

Therefore, PB x 100 11.2% and

K 100/11.2 8.9
Another way of finding PB:
5-5.5
5.5 m
corresponds to an error of
9
x 100-5.6%

4.5 m
corresponds to an error of
9 X100= 5.6%
PB (5.6--5.6)] =
11.2%
 Applications
Process Control: Principles end
100

; and N=
A=pmestcapacitance; G, -G,
s

The
controller equation is: () K,|
m

Puting expressions for G, and Gz in characteristic equation (1 + G, G),

=0
Characteristic equation -T,+2T,s+1
where

Tcharacteristic time constant

= damping factor =

For adamping ratio of

1/2

-k,7=
Example 4.6 For the process shown in Fig. E4.1 (a), inflow to tank is the
manipulated variable "m'. Outflow 'u' through a pump on outflow line acts as
disturbance. Controller is PI having a proportional gain of 4, area of the tank is
2 m, measuring element has
unity transfer function.
(a) Draw the block diagram
of the process control loop.
(b) Find o, and the value of K, which will result in
having a damping factor of 0.5.
a
closed-loop response

Solution
Block diagram of the controlled
system is shown in Fig. E4.6.

G
Fig EA.6 Block diagram of the process of
Fig. E4.1(a) with Pl controller
Characteristic equation =1+ G,
2+4s+ 4K, = 0
G2- 0; K,=4, A 2m
 Closed-loop Response 101

+ 2 s + 2, 0

Comparing with standard form

s+2g+ a -0, 2go, -2; for -0.5
2 rad/see
2
2K or K,== 2 sec

A process control loop has G, as Pl controller. Process is inte-

Example 4.7 an

grating element process gain of K. Control valve and measuring element

with a

Find the relationship among K, and K, for a Kp,

have unity transfer füunction.
damping factor of 0.5.

Solution

HersG-K,1.G
Characteristic equation (CE) = 1 + G , =0

1+K1 -
K, K, K=0
+sK, K+standard have:
Comparing with the equation, s +2os + =0, we

a,-A,K,K
2o,=K, K
Given that =0.5

K, K=K,
and a time constant of
Example 4.8 The servo otor has a gain of 4 Volt

0-8 V for a speed

20 seconds. The speed sensor changes
its output linearly from
The controller is of PI mode with 0-10 V output and 50%
change of 0-80 rpm.
of integration is 6o,%mn, Output
of the controller goes
PB. The constant

through an amplifier of gain 2 to drive themotor. Set point is 55 each

rpm.
block and
Draw the process control loop
with transfer function inside
)
units outside each block.
range of values with appropriate
(1) Find natural frequency of the closed-loop system.
Solution values
of controlled system with proper
Figure E4.8 shows the block diagram
and units and transfer functions.
 102 Process Control: Principles and Applications

5.5to Controller Motor

Amplifier
-20V 0-80 rpm
20s+1
Gp
Measuring element

0-8V
Om

Fig. E4.8

100% output 10V

2.5
50% input 50
100 *8V

6
100*10
K =sec
8
60sec
Characteristic equation of the closed-loop transfer function is:

1+G.G, G, G,=0

1++),
1 8s (20s+1)
0

8s (20s + 1)+16s + 2=0

160s+8s+ 16s +2 0
160s + 24s+2 0
80s +125+1 =0

or , 0.11 rad/sec
Example 4.9 The liquid level in a tank (cross-sectional area of tank is 9 m,
steady state height of tank is 4 m.) is controlled
by
valve
controlling the inflow through
a on
inlet pipe which in turm is acted upon by a PI controller. The
outflow
is equal 2h. Transfer function of the control valve is
to
unity and
function of the measuring element is a first-order transter
with steady state gain
system
and time constant as unity. Draw the
process control loop. Find the value of
that will make the closed loop oscillate at a constant Aj
(ii) K,= 10. Find the frequency of oscillation in both the cases.
frequency, when (i) K,=
 Closed-loop Response 103

Soluition

9-9 A

-2-4
Using Taylor's series,
95 4o+(2h Yh-h, (h -h,) A, =4m
4, 4o+3h, (h-h,)
In terms of deviation variables,
.6H
Q, ( ) - 0 . ( ) = AsH (8)

0 ) - 6 H (9) = 9sH (5)

H()
2,() As+3h, 6+9s FRLIONTM0O O9
to sslloo C
function=
Here, G Controller transfer

Process transfer function =

G, 6+9

G= Measuring element transfer function =

1+s
Characteristic equation = 1 +G. G, G, - 0

9g+15s +s(6 +K) +K, K, =0

Case I: K =5
The characteristic equation becomes 9s +15s+11s + 5K, = 0

Applying Routh's stability criterion,

9 11

5 SK
165-45K o 15
SK
To oscillate at constant frequency, 165-45
11
=0

Using direct substitution method characteristic equation (CE)

(CE)- -9ja-15 +11jo+ SK =0
 and Applications
104 Process Control: Principles

11 rad
K,-3af and w=
3 sCC

So
3

Case I1: K, - 10
becomes
The characteristic equation
9+ 15s+ 16s + 10K,=0
rad
equating real and imaginary parts, *
Substituting s
-jo and sec

K
Observation: High value of K, results in reduced Kp

4.6 PD CONTROLLER IN CLOSED LOOP

PD controllers offer the advantages of stability of proportional action and the
reduction of stabilization time due to the derivative action. The derivative action
is useful for control of systems having large dead time and a very large number
of storage elements. Consider the PD control of a two capacitance process as
shown in Fig 4.13.

sa

C
Fig.4.13 Two capacltance
non-interacting process with PD controller
The system cquations are:
dh
m-9 C2
dt
h

4-u-
dt
Capacitance C and C2 are equivalent to areas
variable and then taking A, and A.
Introducing deviat
Laplace transform,
 Hydraulic and Pneumatic Controllers 139

Example 5.1 Consider the pneumatic pressure system shown in Fig. ES5.1, wherein
bellows are assumed to be identical. The output ends of the two bellows
the two
link from which the mechanical (lincar) movement x is picked
connected by a
are
to the differential force of the bellows. The zero position
up: x being proportional
to steady input pressure. Relate the changes
of the output
motion
() corresponds motion
with the output () in the form of transfer function
in input pressure P, What is (steady state) for unit
of the two bellows is &,. x

AS PS). Stiffness
in input pressure?
step change

Bellows Bellows 2
w

P2

RX
TP
Fig. E5.1 Pneumatic system

Solution
valve and bellows 1, we can write, ) (E5.1.1)
For the P(s) R Cs +1

Similarly, for the valve and bellows 2, we can write, P(S) ES.1.2)
R Cs+1
Movement (X) of the bellows is given by:
(K = constant of proportionality)
A (P-P)=KX
Or, A (PS) - P2s)) = K X() (E5.1.3)
Substituting Eq. (ES.1.1) and (ES.1.2) in Eq. (E5.1.3):

K X(G)
RCS+1 Rcs+i
Or, X)A (R-R)C
K[(R,Cs+10R, Cs+ )]

For PAs)=1/s

Mo)- (R-R)Cs
Ks (RCs +1XR, Cs+ 1DJ
 140 Process Control: Principles and Applications

Final value is given by:

=0
Asteady-stateA s X(s)],
At steady state, presSures inside the
The above result can be easily visualized.
and bellows being identical, the netet
two bellows (Pi and P>) become equal;
mechanical force is zero. So, X(o) =0.
bellows with a restrictor. Find the transfer
Example 5.2 Figure ES.2 shows a

function
function P($)of the system.
ww
P
www
C

Fig. E5.2 Bellow system

Solution
For the bellows with resistance R and capacitance C, the relation between P, and

P can be written as:

(E5.2.1)
RCS+1
Characteristic equation of the bellow is:

PA=(K, + K) X
where K, is stiffness of spring and K, is stiffiness of [Link]

Or, Ph= +).X dt 3o (3 (E5.2.2)

Substituting Eq (ES.2.2) in Eq. (ES.2.1):

X(K, +K,) =
AP RCs+1

Or,
RCs+K, +K,)
Or, Ts+1 (K, +K,) AIK
1+Ts where K-,+K, and T= RC
 Applications
160 Process Control: Principles and

(SV) 100%

(2.5V) 50%

(OV) 0%

+50% 0%-50%

100 125 150

pm pm pm
(2V) (3.5V) (5V)

Error -1.5 V 0V 1.5 V

Fig E6.80b) Controller output vs. motor speed characteristics

The thre values 3.7 V-3.7 in and 2.5 V are to be added. Accordingly, resis.
tor values the factor. Since the
are chosen provide
to
corresponding scale
op-amp configuration used is an inverting type circuit, a negative input is given to
adder
nullify the inverting effect and get a positive sum.

V 3.7 kEy +S-3.5V) +3(25 v)

3.7 k2
The output of this op-ampconfiguration out 1s fed to the motor. As per the
input, thespeed of the motor is altered. The modified
speed is again converted
into an electric signal through the transducer and fed back
to the op-amp (V)
The circuit in (C)Fig. E6.8(©) shows the electronic
controller
in a closed-loop.
lkQ 3.7 k2
-3.5 Vww ww
Ik2
ww

3.7 ka Motor Transducer Vin

-2.5 V w Dut

Fig. E6.8 (c) Controller circuit for motor

speed control
Example 6.9 A
proportional controller is used to control the
tank, which varies from 20°C to 100°C.
The controller is
temperature ma
valve on steam supply when
temperature rises to
required to close n
80°C and open the steam sup
ply when temperature drops to 20°C. Set
(a) Calculate proportional band and point is 60°C.
controller output when
tank is 50°C. temperature in
 Electronic Controllers 161

Fot the above process, design an clectronic

(b) controller around a single op
amp. Controller output Is 0-5 V; voltage varies
linearly from 1 V to 5 V
when temperature varies from 20C to 100°C. Zero error
50% controller output. corresponds to

c)Draw and label the closed loop with clectronic controller.

Solton

(a) PB 80-20 100 75%

100-2000=759%

When temperature exceeds the set point (error is negative), the controller
otout should increase to close the valves. Assuming AC valve, K, is negative.
So. P-,E+ Pi K5
When temperature
=
80°C, E. = OUC-80°C -25%, and controller output,
100°C 20°C
Pis 100%.

P--K, Ep+P
100 -25) +
P, or P= 66.6%mlo anom
When temperature = 50°C, E, 60°C-50°Cx 100 12.5%, P,= 66.6%
100°C-20°C
So, P-K,E, + R
Ps0c 50%

6) K, 75 5 100% of output
75% of input
x4
100

V 2.5V

SP 3V corresponding to 60°C
igures E6.9a) and E6.96) show the control loop and the electronic circuit
1agram of the controller, respectively.

2V to-2Electronic
Controller -5 FCE+Process
20-100

Sensor
1-SV

Fig E6.9a) Process control loop

164 Process Control: Principles and Applications

a proportional-integral
controller with a proportional L.
Example 6.10 Design
of 30% and an integration time of 10 seconds. The 4-20 mA input conver. nd
verts into
0-2 V eTor signal. The output is to be in between 0-10 V.

Solution
A proportional band of 30% means, when the error changes by 30% of full.
(i.e, 0.6 V), the output must change by 100% (i.e, 10 V). This gives a s -scale
gain o
100%of cotrolleroutput - 16.77 2
p30% of controller input (error 0.6 R

0.1
of 10 seconds implies K,
= =

An integration time sec

RC
I f C - 100 uE, it will require R, = l100 kS2, which in turn would require:

R o 6 k2 30R
16.67
Here, the reason why we fix capacitors values instead of resistor values is tha
the variable capacitors are costly and not available easily
so as compared to
vaniable resistors.

Example 6.11 The heat exchanger response to steam flow has a gain of l0°C
and time constant of 20 sec. The
a
kg//sec
temperature sensor changes its
output linearly
from 0-10V for temperature
change of 0-200°C. The controller is of electronic
PI mode with 0-10 V
output, having a PB of 40% and an integration constant of
2%I
Tminh e control valve linearly outputs steam, giving 20 kg/sec for 100%
controller output. Set point is
100°C.
) Draw the process control loop, with transfer
function inside each block
and range of values with
appropriate units outside each block.
i) Find natural frequency and damping factor of the closed-loop system.
Solution

Gy p
250 2-2 0-20kg/sec10
20s+1
0-200C
Controller Valve Heat exchanger

Gm
0-10V 101
200
Measuring
element
Fig. E8.11 Process control
loopP
 Electronic Controllers 165

100% output 10V

40% PB K, 2.5
40% input 40 10V
100
2
2 1
K, 60x10 60 30 sec

100
Characteristic equation of closed-loop transfer function (CLTF) is:

1+G.G,G,G,-0
1+25 1+
30s (2s + 1) +2.5(30s +1) =0
6 0 s +30s + 75s +2.5 = 0

60s + 105s +2.5 =0

+42s +1 0
24s
=

0.204 rad/sec

= 42; 5=4.284

6.7 ELECTRONIC PD cONTROLLER

The proportional and derivative mode combination is implemented using the cir-
cuit shown in Fig. 6.18.

) 1+R,C) VA) (6.14)

R
, R
Ka=T= R,C
f , is not zero, then add V, volts to the second op-amp through R in adder

mode
R K2
ww-
ww- ww

Fig.6.18 Circuit of electronic PD controller

 100 Process Control: Prinaiplas andd Auplicatiuons

Example 6.12 Derive the output voltge V or the cireut surwn in Fig
and dentity the type of cotroller and the comtroler constants, 12

Fhg. K612

Soludion
Applying Kirchhoff's current law at point A (point B is at virtual groundy

(E6.12.1)

Applying KCL at point B, we get,

- a - (since V, -ov)
R

(E6.12.2)
Combining (E6,12.1) and (E6,12.2),

(E6.12.3)

Equation (E6.12.3) is in standard PD 2R RC

form, where K,
-and K

6.8 ELECTRONIC PID CONTROLLER

The circuit for oftctronic PID controller is given in Fig. 6.19.

+Voun (0) (6.15)

,-KyR and ARC

R
 Electronic Controllers 167

[Link]
R2

R
ww
R
ww.
www
R
Rd
ww
Vout

Fig. 6.19 Circuit of electronic PlID controller

Example 6.13 Design a PID temperature control system which inputs error in
0-4 V range. The output to final control element is 0-8V. Given,
K, 2.4% per %
=

K,=9.0 (%/min)/ o go rojhiee d

K 0.7 %/ (%%/min)

Solution
K, = 2.4% per % implies 1% of error gives 2.4% of output, i.e.,

24 x8 aaioaax3
]00 = 4.8
x4
100 rnoiae0 oive

k 9%/ min)2 an error of 1% would cause the output to change

gdfed at a rate of )
min

9% output 0.3 secremul

60 sec (1% of input) 60 sec (0.04)
0.7%
K i.e., when errorchanges by 1% per minute, output will
min
change by 0.7%)

ETor of % 0.04 6.67 x 104

sec
min 60 sec

K (0.007)(8VV -84 sec

6.67 x 104
 Digital Controllers 185

Cross-mu
multiplying and takin inverse Z-transform,

A)-1.2 uk 1) 0.5 uk 2) etk) etk- 1)

+ = +

Mk)=etk) + etk 1)
+ 1.2
ulk - 1) 0.5 u (k -2)
-

This
example demonstrates how difference equation is obtained from discrete
and denominator of discrete transfer
fer function by multiplying numeratorand then
takinge inverse z-transform.
nction by highest negative power of z,
vample 7.10 A process can be represented as an integrating clement and a first
der element in cascade,having all characterizing parameters as unity. It is
order
PD controller having a proportional gain K and derivative time
antrolled by a
transfer functions in the closed unity transfer
loop have
U) sec. All other is summation of disturbance D{s) and output of
nction. Input to the process
1s Y(®). R{s) is the reference input.
controller. Output of process
the process control loop.
(a) Draw
for unit step change in reference input.
b) For K= 32, find Y(:)
Find (A) as k> o

(c)
Soluton
D(s)

(a

Fig. E7.10 Process control loop

(b) Y) KtIS for K 32 and R(s)= (step input)

R(s) s+12s+K

32+1s 32+11s ,C
Y6) +12s+32 s (s+4 s+8)

Yo)= s 4(s+4) 4(s+8)

Ye) 3z
T* 4(2 -e) 4( -eT)
(c) Using final value theorem:
Ye) =1