Professional Documents
Culture Documents
Wade - Derivation - 978 1 64331 092 3 - 6 1 20
Wade - Derivation - 978 1 64331 092 3 - 6 1 20
By Harold L. Wade
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
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.
1
−KC s s +
Fout (s) TI
= (6-22); (D-2)
SP(s) K K
s2 + C s + C
TL TI TL
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)
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
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
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
sin(ωt) ω 1−ζ 2
Using R-1: = =
cos(ωt) ζ ωn ζ
1−ζ 2
tan(ωt) =
ζ
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 )
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
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 )
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
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
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
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
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
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
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
Maximum outflow change for step change in inflow (analogous to Equation D-21):
T f (ζ )
TaF = t = I Table 6-10, Row 2; (D-21)
ζ
−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
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
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:
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)
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):
F (ω )
AR = G( jω) = out = [ Re(G( jω))]2 + [ Im(G( jω))] 2 (D-27)
Fin (ω)
ζ Damping factor
DR = 0 no oscillation
− ln(DR) −2πζ
ζ= DR = exp
2
4π 2 + (ln(DR))2 1 − ζ
ω = ωn 1 − ζ 2 (R-1)
2ζ 1 − ζ 2
ω = (R-2)
TI
KC
= 2ζωn (R-3)
TL
KC
= ωn 2 (R-4)
TI TL
2ζ
ωn = (R-5)
TI
4ζ 2 TL
TI = (R-6)
KC
From R-5:
1 KC TI
ζ= (R-7)
2 TL
2π π
P = = TI ×
ω ζ 1−ζ 2 (R-8)