Derivation of the Equations

in Tables 6-8 through 6-12

In Basic and Advanced
Regulatory Control: System
Design and Application
Third Edition

By Harold L. Wade

Copyright © 2020 International Society of Automation (ISA)

All rights reserved.
Revised Edition 6-1-2020
ISBN: 978-1-64331-092-3

Wade_Derivation.indd 1 06-08-2020 22:35:17

Wade_Derivation.indd 2 06-08-2020 22:35:17
 Notice
The information presented in this publication is for the general education of the reader. Because neither the author nor the
publisher has any control over the use of the information by the reader, both the author and the publisher disclaim any and all
liability of any kind arising out of such use. The reader is expected to exercise sound professional judgment in using any of the
information presented in a particular application.
Additionally, neither the author nor the publisher has investigated or considered the effect of any patents on the ability of the
reader to use any of the information in a particular application. The reader is responsible for reviewing any possible patents that
may affect any particular use of the information presented.
Any references to commercial products in the work are cited as examples only. Neither the author nor the publisher endorses
any referenced commercial product. Any trademarks or tradenames referenced belong to the respective owner of the mark or
name. Neither the author nor the publisher makes any representation regarding the availability of any referenced commercial
product at any time. The manufacturer’s instructions on the use of any commercial product must be followed at all times, even if
in conflict with the information in this publication.

Copyright © 2020 International Society of Automation (ISA)

All rights reserved.

Printed in the United States of America.

Version: 1.0

ISBN-13: 978-1-64331-092-3 (PDF)

No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher.

ISA
67 T. W. Alexander Drive
P.O. Box 12277
Research Triangle Park, NC 27709

Library of Congress Cataloging-in-Publication Data in process

Wade_Derivation.indd 3 06-08-2020 22:35:17

Wade_Derivation.indd 4 06-08-2020 22:35:17
 Derivation of the
Equations in Tables 6-8
through 6-12

Review the figure below to see how equations in the referenced tables relate to the applicable control
loop.

Figure 6-23. Block Diagram for an Ideal Process Model Liquid-Level Control Loop

This document references equations and tables in Chapter 6 of Basic and Advanced Regulatory Control:
System Design and Application, Third Edition. Below is the legend for the equation numbers.

Equation Number Legend

= equations from Chapter 6

(6-x) 
= derivation equations
(D-x) 
= equations from the “Symbols and Reference Equations” section of this document
(R-x) 
Assume that the flow control loop is considerably faster than the level control loop (which is true in
most practical situations), so that F(s) can be approximated by F(s) = 1.

Four transfer functions can be derived from Figure 6-23.

1. The set-point–to–level transfer function:

KC K
s+ C
L(s) TL TI TL
= (6-21); (D-1)
SP(s) s2 + KC s + KC
TL TI TL
5

Wade_Derivation.indd 5 06-08-2020 22:35:17

 6 Basic and Advanced Regulatory Control, Third Edition

2. The set-point–to–flow-out transfer function:

 1
−KC s s + 
Fout (s)  TI 
= (6-22); (D-2)
SP(s) K K
s2 + C s + C
TL TI TL

3. The flow-in–to–level transfer function:

s
L(s) TL
= (D-3)
Fin (s) s2 + KC s + KC Also see (D-5)
TL TI TL


4. The flow-in–to–flow-out transfer function:
KC K
s+ C
Fout (s) TL TI TL (6-30); (D-4)
=
Fin (s) K K
s2 + C s + C
TL TI TL

Note that the set point–to–level and flow-in–to–flow-out transfer functions are identical.
L(s)
From Figure 6-23, the transfer function can be derived:
Fin (s)

L(s) = P(s)( Fin (s) − Fout (s))

Fout (s) = C(s)L(s)
L(s) = P(s)Fin (s) − P(s) C(s)L(s)
L(s) + P(s)C(s)L(s) = P(s)Fin (s) (D-5)
(1 + P(s)C(s)) L(s) = P(s) Fin (s)
L(s) P(s)
=
Fin (s) 1 + P(s)C(s)

The elements of Equation D-5 are given by:

1
P(s) =
TL s
(D-6)
 1 
C(s) = KC 1 + 

 TI s 

Substitute Equation D-6 into Equation D-5. After simplification, this gives Equation D-3:
s
L(s) T L
=
Fin (s) 2 KC K
s + s+ C
TL TI TL

Wade_Derivation.indd 6 06-08-2020 22:35:19

 Derivation of the Equations in Tables 6-8 through 6-12 7

Notice that the denominators of Equations D-1, D-2, D-3, and D-4 are identical. These denominators
can also be expressed in frequency response terms as:
K K
s2 + C s + C = s2 + 2ζωn s + ωn 2
TL TI TL

Determining PI Controller Tuning Parameters for

Underdamped (ζ <1) Control Loop
Assuming that the tank holdup time, TL, is known, the next step is to determine the desired loop
decay ratio following a set-point change. Book Equation 6-25 then converts this ratio to a damping
factor, ζ. The next step is to determine a target value for the maximum deviation of level from the set
point, ∆Lmax , following an assumed worst-case step disturbance of inflow, ∆Fin. Note that the val-
ues for ∆Lmax and ∆Fin depend upon the particular operating situation. (See Chapter 6, “Tuning for
Integrating Processes,” subsection “Derived Equations for Liquid Level Control Tuning Parameters—
Ideal Model” for further discussion of these parameters.)

The numerical values for ∆Lmax and ∆Fin are used in derivation D-9. Using ­parameters TL and ∆Fin
and the relation for a step change in inflow, Fin (s) = (∆Fin /s), Equation D-4 can be written and further
developed as:

s
∆Fin TL
L(s) =
s K K
s2 + C s + C
TL TI TL
∆Fin 1
=
TL s2 + KC s + KC
TL TI TL
∆Fin 1
=
TL s2 + 2 ζ ωn s + ωn2
∆Fin ω
=
TL ω (s + ζ ωn )2 + ω 2

From the last equation, the time response of level to a step change in inflow is given by:

∆Fin −ζ ωn t
L(t) = e sin(ω t) (D-7)
TL ω

d L(t)
Extremities occur whenever = 0:
dt

d L(t) ∆Fin −ζ ωn t
= e (−ζ ωn sin(ωt) + ω cos(ωt))
dt TL ω

d L (t )
= 0 whenever ζ ωn sin (ωt ) = ω cos (ωt )
dt

Wade_Derivation.indd 7 06-08-2020 22:35:21

 8 Basic and Advanced Regulatory Control, Third Edition

sin(ωt) ω 1−ζ 2
Using R-1: = =
cos(ωt) ζ ωn ζ
1−ζ 2
tan(ωt) =
ζ

The time of the first extremity is given by:

 
1  1 − ζ 2 
TaF = t = tan −1  (D-8)
ω  ζ 
 

Substitute TaF from Equation D-8 into Equation D-7 to find Lmax , the deviation of level at the first
extremity.
  
 1 − ζ 2 
−ζ ωn  1 tan −1     
∆Fin  ω  ζ 
  1  1 − ζ 2  
Lmax = L(TaF ) = e  
sin ω tan−1   
TL ω   
 ω  ζ

 
ζ  1 − ζ 2 
− tan−1    
∆Fin 1−ζ2  ζ   1 
 
Lmax = L(t) = e   (D-9)
TL ω  1− ζ 2 

 
ζ  1 − ζ 2 
− tan−1  
∆Fin 1−ζ2  ζ 
 
Lmax = e
2
TL ω 1 − ζ

From R-3:

(Note: Equations R-1 through R-8 are located in the “Symbols and Reference Equations” section of
this document.)

KC
TL =
2ζωn

 
ζ  1 − ζ 2 
− tan−1 
Lmax 1 1−ζ2  ζ 
 
= e
∆Fin K C ω 1−ζ 2
2ζωn
 
ζ  1 − ζ 2 
− tan−1  
Lmax 2ζ 1−ζ2  ζ 
 
= e
∆Fin KC (1 − ζ 2 )

Wade_Derivation.indd 8 06-08-2020 22:35:23

 Derivation of the Equations in Tables 6-8 through 6-12 9

Solve for KC :

 
ζ  1 − ζ 2 
− tan−1 
1−ζ2  ζ 
 
2ζ e (D-10)
KC =
 Lmax 
 ∆F (1 − ζ )
  2

in

Define f (ζ ) :
 1 − ζ 2 
1  
f (ζ ) = tan−1  Tables 6-8, 6-9, and 6-10,
1−ζ 2  ζ 
 Last Row; (D-11)

−ζ f (ζ )
2ζ e
KC = Table 6-8, Cell B1; (D-12)
L 
( )
1 − ζ 2  max 
 ∆Fin 

From R-4:

T 
TI = 4ζ 2  L  Table 6-8, Cell B2; (D-13)
 KC 

From Equations D-12 and D-13:

ζ f (ζ )  Lmax 
TI = 2ζ (1 − ζ 2 )TL e   (D-14)
 ∆Fin 

Predicted Behavior
Level Response to Set-Point Change
Repeating Equation D-1, written in frequency response terms:

KC K
s+ C
L(s) TL TI TL
=
SP(s) s2 + C s + KC
K
TL TI TL

∆SP
For a step set-point change, SP(s) =
s

1
s+
K ∆SP TI
L(s) = C

TL (
s s2 + 2ζωn s + ωn 2 )

Wade_Derivation.indd 9 06-08-2020 22:35:25

 10 Basic and Advanced Regulatory Control, Third Edition

By partial fraction expansion:

K ∆SP  A Bs + C 

L(s) = C + 2 (D-15)
TL  s s + 2ζωn s + ωn 2 

Note that:

1
As2 + 2ζωn A s + Aωn 2 + Bs2 + Cs = s +
TI

Equating powers of s gives:

1
A=
ωn 2TI
1
B=− 2
ωn TI
2ζω 2ζ
Utilize R-5: C = 1− 2 n = 1− =0
ωn TI 2 ζ
TI
TI

Substitute these values into D-15 and simplify:

 1 s 
 2 2

KC ∆SP  ωn TI ωn TI 
L(s) =  − 2 2
 (D-16)
TL  s s + 2ζωn s + ωn 
 
 
 
1 s 
= ∆SP  − 2 
 s (s + ζωn ) + ω 2

 
1 s + ζωn − ζωn 
= ∆SP  − 
 s (s + ζωn )2 + ω 2 
 
 
1 ζωn ω s + ζωn 
= ∆SP  + 2
− 2 
s ω (s + ζωn ) + ω 2 ( s + ζω n ) + ω 2
 
 
1 ζ ω s + ζωn 
= ∆SP  + 2
− 2 
s 1 − ζ 2 (s + ζω ) + ω 2
n ( s + ζω n ) + ω 2
 

From Equation D-16, the time response of level to a set-point change is given by:

  
 −ζωnt  ζ 
l(t) = ∆SP 1 + e  sin(ωt) − cos(ωt)
  1 − ζ 2 

 

Wade_Derivation.indd 10 06-08-2020 22:35:26

 Derivation of the Equations in Tables 6-8 through 6-12 11

  
   2   2  

= ∆SP 1 + e − ζ ωnt   ζ  + 1 sin ωt − tan−1  1 − ζ 
     ζ 
   1 − ζ 2    
  

    2 
  ζ 2   1 − ζ 
−ζω t
= ∆SP 1 + e n  + 1 sin ωt − tan 
−1

  1 − ζ 2   ζ 
    
 
  
 − ζ ωn t  1 
= ∆SP 1 + e  sin (ωt − φ ) (D-17)
  1 − ζ 2 
 

 
−1  1 − ζ 2 
where φ = tan  
 ζ 


Peaks and valleys of l(t) occur wherever the derivative is equal to 0.

From Equation D-17, the derivative is:

dL(t) e− ζωn t
= ∆SP [−ζωn sin(ωt − φ) + ω cos(ωt − φ)] (D-18)
dt 1−ζ 2

dL(t)
= 0 whenever
dt

−ζωn sin(ωt − φ) + ω cos(ωt − φ) = 0

ζωn sin(ωt − φ) = ω cos(ωt − φ)

sin(ωt − φ) ω
=
cos(ωt − φ) ζωn
ωn 1 − ζ 2
=
ζωn
1−ζ 2
tan(ωt − φ) =
ζ
 2
−1  1 − ζ 
ωt − φ = tan  
 ζ 
 
ω t − φ = φ1

Wade_Derivation.indd 11 06-08-2020 22:35:27

 12 Basic and Advanced Regulatory Control, Third Edition

d L(t)
Because φ = φ1 then = 0 whenever
dt
 
 1 − ζ 2 
2tan−1  
 ζ 

t=
ω
 
TI  1 − ζ 2 
TaL = t = tan−1   (D-19)
ζ 1−ζ 2  ζ 


T f (ζ )
TaL = t = I Table 6-9, Cell B3; (D-20)
ζ

Equation D-19 utilizes R-2 in the previous equation and is written only for the first extremity. Equation
D-20 utilizes Equation D-11 in Equation D-19 for the reduction of algebraic symbols.

L(t) at first peak is L(taL ). Substitute Equation D-19 into Equation D-17:
 T f (ζ )    2 
 −ζ ωn I
ζ 
 1  TI −1  1 − ζ  

L(taL ) = ∆SP 1 + e  sin  ω tan   − φ 
   ζ 
  1 − ζ 2  ζ 1 − ζ
2
  

  2 
 TI −1  1 − ζ 
sin ω tan   − φ
 ζ 1 − ζ 2 
 ζ  
 
  2 
 2 −1  1 − ζ 
= sin ω tan   − φ
 ω  ζ  
  
    
  1 − ζ 2   1 − ζ 2 
= sin 2tan−1   − tan−1  
  ζ   ζ 
   
  
  1 − ζ 2 
= sin tan−1  
  ζ 
 

= 1−ζ 2

 2ζ
 
 1 − ζ 2 

 − tan −1   

 
 1− ζ 2 
 ζ  
 1 2 


L(taL ) = ∆SP 1 + e 
 × 1−ζ 
  1 − ζ 2 
 
 
 

Substitute f (ζ) from Equation D-11 into the previous equation:

∆Lmax = L(taL ) = ∆SP 1 + e

− 2ζ f (ζ ) 
  Table 6-9, Cell B1

Wade_Derivation.indd 12 06-08-2020 22:35:29

 Derivation of the Equations in Tables 6-8 through 6-12 13

L(taL ) − ∆SP −2ζ f (ζ )

The overshoot ratio = =e Table 6-9, Cell B2
∆SP
2π
See R-8: Period =
ω
Table 6-9, Cell B4
π
Utilize R-2: Period = TI ×
ζ 1−ζ

Outflow Response to Step Inflow Change

L(s) F (s)
Because = out (see Equations D-1 and D-4), the first two items will require no new deriva-
SP(s) Fin (s)
tion, only a change of symbols.

Maximum outflow change for step change in inflow (analogous to Equation D-21):

∆Fout−max = ∆Fin 1 + e

− 2ζ f (ζ ) 
  Table 6-10, Cell B1

Outflow arrest time (analogous to level arrest time):

T f (ζ )
TaF = t = I Table 6-10, Row 2; (D-21)
ζ

Maximum Rate of Change of Outflow

The maximum rate of change of outflow is analogous to Equation D-17:

  
 −ζ ωn t  1 
Fout (t) = ∆Fin 1 + e  sin (ωt − φ ) (D-22)
  1 − ζ 2 

 

 2
−1  1 − ζ 
where φ = tan  
 ζ 


dFout (t) e−ζωn t

= ∆Fin [−ζωn sin(ω t − φ) + ω cos(ω t − φ)]
dt 1− ζ 2
e−ζωn t  2 2 2 
= −∆Fin  ζ ωn + ω sin(ω t − φ − φ 1)
1− ζ 2  

ζ 2 ωn 2 + ω 2 = ζ 2 ωn 2 + ωn 2 − ζ 2 ωn 2 = ωn

 
−1  1 − ζ 2 
φ1 = tan   = φ
 ζ 


dFout (t) ωn
= −∆Fin e− ζωn t sin(ωt − 2φ)
dt 1−ζ 2

Wade_Derivation.indd 13 06-08-2020 22:35:32

 14 Basic and Advanced Regulatory Control, Third Edition

2ζ
From R-5: ωn =
TI
Therefore:
d Fout (t) 2ζ
= − ∆Fin e− ζωn t sin(ω t − 2φ) (D-23)
dt TI 1 − ζ 2

d Fout (t)
The maximum rate of change of Fout , the maximum of , occurs where its derivative is 0.
dt
d 2 Fout (t) ∆Fin 2ζ d  − ζωn t
= − e sin(ω t − 2φ)
dt 2
TI 2 dt  
1−ζ
∆Fin 2ζ
= − e− ζωn t [ − ζ ωn sin(ω t − 2φ) + ω cos(ω t − 2φ)]
TI 1−ζ 2

∆Fin 2ζ
= e− ζωn t [ ζ ωn sin(ω t − 2φ) − ω cos(ω t − 2φ)]
TI 1−ζ 2

∆Fin 2ζ    ω 
= e− ζωn t  ζ 2 ωn 2 + ω 2 sin ω t − 2φ − tan −1  
TI 1−ζ 2
 
 
 ζ ω n 
 
∆Fin 2 ζωn − ζωn t
= e sin (ω t − 3φ )
TI 1−ζ2

d 2 Fout (t)
= 0 whenever
dt 2
sin(ω t − 3φ) = 0

ω t − 3φ + n π = 0
3φ − n π
t = , for n = ± 0, 1, 
ω
Change φ in the above equation to ψ, for consistency with Table 6-10.

(3ψ − n π )
t = , for n = ± 0, 1, 
ω

Without derivation but by chart record observation on the PC-ControLAB simulation, we will use
d F (t)
n = −1. Hence out achieves its maximum value at:
dt
1
TaR = t = (3ψ − π ) Table 6-10, Cell C5; (D-24)
ω
Substitute TaR from Equation D-24 into Equation D-23 to find the maximum rate of change of outflow:

d Fout (TaR ) ∆Fin 2ζ      

=− exp−ζωn  1 (3ψ − π) ×sinω 1 (3ψ − π)− 2ψ 
dt TI    ω 
  ω 
1−ζ 2  
    
∆Fin 2ζ   − 3ζζωn − 1  1 − ζ 2  
=− exp tan  
 + ζωn π ×sin(ψ − π)
 
TI 1 − ζ 2   ωn 1 − ζ 2
  ζ  
 

Wade_Derivation.indd 14 06-08-2020 22:35:34

 Derivation of the Equations in Tables 6-8 through 6-12 15

  2 
 −1  1 − ζ  2
Note that sin(ψ − π ) = −sin (ψ ) = −sin tan   = − 1 − ζ .
  ζ 

Substituting this into the previous equation and simplifying yields:

 ζπ 
d Fout (TaR ) ∆Fin  −ζ f (ζ ) 3  1 − ζ 2 
= × 2 ζ e  ×e 
dt TI     Table 6-10, Cell B3; (D-25)
 

Sinusoidal Fin: Amplitude Ratio, Fout /Fin

Let:

KC K
s+ C
F (s) TL TI TL
G(s) = out = (6-30); (D-26)
Fin (s) K K
s2 + C s + C
TL TI TL

KC K
s+ C
TL TI TL
= 2
s + 2ζωn s + ωn 2
KC 1 + TI s
=
TI TL (s + 2ζωn s + ωn 2 )
2

KC 1 + TI ( jω )
G ( jω ) =
TI TL (( jω )2 + 2ζωn ( jω ) + ωn2 )
KC 1 + j ωTI
=

TI TL ((ωn 2
)
− ω 2 + j 2ζωn ω )
Convert to standard complex number format (i.e., A + jB):

KC (1 + j ωTI ) ((ωn2 − ω 2 ) − j 2ζωn ω )

G ( jω ) =
TI TL ((ωn2 − ω2 ) + j 2ζωnω)((ωn2 − ω2 ) − j 2ζωnω)
 
( ) (
 ωn 2 − ω 2 + ωTI (2ζωnω) + j (ωTI ) ωn 2 − ω 2 − 2ζωn ω 
= ωn 2   )
 
2 2 2
 (
2
ωn − ω + (2ζωnω) ) 

 
Re (G ( jω )) = ωn 2 
( )
 ωn 2 − ω 2 + ωTI (2ζωn ω) 

 2 2 2 
( 2
 ωn − ω )
+ (2ζωnω) 
 
Im (G ( jω )) = ωn  2  ( )
 (ωTI ) ωn 2 − ω 2 − 2ζωn ω 


 2 2 2 

( 2
 ωn − ω )+ (2ζωnω) 

Wade_Derivation.indd 15 06-08-2020 22:35:36

 16 Basic and Advanced Regulatory Control, Third Edition

G( jω) = Re(G( jω)) + j Im(G( jω))

F (ω )
AR = G( jω) = out = [ Re(G( jω))]2 + [ Im(G( jω))] 2 (D-27)
Fin (ω)

Symbols and Reference Equations

KC = control gain, dimensionless
TI = integral time, minutes (per repeat)
TL = process time constant (tank time constant). See (corrected) Equation 6-20.

ζ Damping factor

ζ = 0 sustained oscillation

ζ = 1 no oscillation
DR Decay ratio

DR = 1 sustained oscillation

DR = 0 no oscillation

Relationship between ζ and DR:

 
− ln(DR)  −2πζ 
ζ= DR = exp 
 2 
4π 2 + (ln(DR))2  1 − ζ 

ωn  = undamped natural frequency

ω = damped frequency, radians per minute (or other time units)

ω = ωn 1 − ζ 2 (R-1)

Combining R-1 and R-5 yields:

2ζ 1 − ζ 2
ω = (R-2)
TI

From corresponding components of Equation 6-21 and 6-23:

KC
= 2ζωn (R-3)
TL

KC
= ωn 2 (R-4)
TI TL

Wade_Derivation.indd 16 06-08-2020 22:35:37

 Derivation of the Equations in Tables 6-8 through 6-12 17

Combining R-3 and R-4 yields:

2ζ
ωn = (R-5)
TI

Combine R-4 and R-5 and solve for TI :

4ζ 2 TL
TI = (R-6)
KC

From R-5:

1 KC TI
ζ= (R-7)
2 TL

Period (utilizing R-2):

2π π
P = = TI ×
ω ζ 1−ζ 2 (R-8)

Wade_Derivation.indd 17 06-08-2020 22:35:38

Wade_Derivation.indd 18 06-08-2020 22:35:38