Process Control: Designing Process

and Control Systems for Dynamic

Performance
Chapter 9. PID Tuning
Copyright Thomas Marlin 2!"
The copyright holder provides a royalty-free license for use of this
material at non-proft educational institutions
CHAPTER 9: PID TUNING
When I complete this chapter, I want to e
ale to !o the "ollowin#$

E%plain the per"ormance #oals that we

see& to achie'e 'ia t(nin#$

Appl) a t(nin# proce!(re (sin# the

process reaction c(r'e an! t(nin#
correlations$

*(rther impro'e per"ormance ) "ine

t(nin#
+(tline o" the lesson$
CHAPTER 9: PID TUNING

A trial an! error approach , wh) we

!on-t (se it

De"ine the t(nin# prolem

.ol'e an! !e'elop correlations

Appl) correlations to e%amples

*ine t(ne , the personal to(ch

PR+PERTIE. THAT WE .EE/ IN A C+NTR+00ER

Goo! Per"ormance , "ee!ac&

meas(res "rom Chapter 1

Wi!e applicailit) , a!2(stale

parameters

Timel) calc(lations , a'oi!

con'er#ence loops

.witch to3"rom man(al ,

(mplessl)

E%tensile , enhance! easil)

CHAPTER 9: PID TUNING
This chapter
Pre'io(s chapter
0ater chapters
CHAPTER 9: PID TUNING

How !o we appl) the same e4(ation to man) processes5

How to achie'e the !)namic per"ormance that we !esire5

TUNING666
The a!2(stale parameters are calle! t(nin# constants$
We can match the 'al(es to the process to a""ect the
!)namic per"ormance
I
dt
CV d
T dt t E
T
t E K t MV
t
d
I
c
+
1
]
1

0

' ) ' (
1
) ( ) (
CHAPTER 9: PID TUNING
AC
Trial 7:
(nstale,
lost 89:,;;;
0 20 40 60 80 100 120
-40
-20
0
20
40
S-LOOP plots deviation variables (IAE = 608.1005)
Time
C
o
n
tro
lle
d
V
a
ria
b
le
0 20 40 60 80 100 120
-100
-50
0
50
100
Time
M
a
n
ip
u
la
te
d
V
a
ria
b
le
Trial 9: too
slow, lost
8<,;;;
0 20 40 60 80 100 120
0
0.2
0.4
0.6
0.8
1
S-LOOP plots deviation variables (IAE = 23.0904)
Time
C
o
n
tro
lle
d
V
a
ria
b
le
0 20 40 60 80 100 120
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
ip
u
la
te
d
V
a
ria
b
le
0 20 40 60 80 100 120
0
0.5
1
1.5
S-LOOP plots deviation variables (IAE = 9.7189)
Time
C
o
n
tro
lle
d
V
a
ria
b
le
0 20 40 60 80 100 120
0
0.5
1
1.5
Time
M
a
n
ip
u
la
te
d
V
a
ria
b
le
Trial n:
+/, "inall),
(t too&
wa) too
lon#66
Is there
an easier
wa) than
trial = error5
I
dt
CV d
T dt t E
T
t E K t MV
t
d
I
c
+
1
]
1

0
' ) ' (
1
) ( ) (

CHAPTER 9: PID TUNING
>es, we can
prepare #oo!
correlations6
S-LOOP plots deviation variables (IAE = 608.1005)
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
ip
u
la
t
e
d

V
a
r
ia
b
le0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
Determine a mo!el
(sin# the process
reaction c(r'e
e%periment$
/c TI
Determine the initial
t(nin# constants "rom
a correlation$
0 20 40 60 80 100 120
0
0.5
1
1.5
S-LOOP plots deviation variables (IAE = 9.7189)
Time
C
o
n
tro
lle
d
V
a
ria
b
le
0 20 40 60 80 100 120
0
0.5
1
1.5
Time
M
a
n
ip
u
la
te
d
V
a
ria
b
le
Appl) an! "ine t(ne
as nee!e!$
De"ine the t(nin# prolem
7$ Process D)namics
9$ ?eas(re! 'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance meas(res
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
The PID controller will
"(nction s(ccess"(ll) "or
the wi!e ran#e o" "ee!ac&
process !)namics shown
here$
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
t
r
o
lle
d

V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
ip
u
la
t
e
d

V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
Descrie the !)namics
"rom the step
chan#e !ata$
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
The PID controller will
"(nction s(ccess"(ll) "or
the wi!e ran#e o" "ee!ac&
process !)namics shown
here$
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
t
r
o
lle
d

V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
ip
u
la
t
e
d

V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
*irst or!er with
!ea! time
nth or!er with
!ea! time
(nstale
Inte#rator, see
Chapter 7B
(n!er!ampe!
Descrie the !)namics
"rom the step
chan#e !ata$
Inp(t step chan#e
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
The PID controller will "(nction s(ccess"(ll) "or a wi!e
ran#e o" "ee!ac& process !)namics
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
t
r
o
lle
d

V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
ip
u
la
t
e
d

V
a
r
ia
b
le
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
tr
o
lle
d
V
a
r
ia
b
le
We will !e'elop
t(nin# correlations
"or these !)namics$

?ost commonl)
occ(rrin#

*it mo!el (sin#

process reaction
c(r'e

+ther processes can

e controlle! with
PIDC nee! more trial
an! error
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
Realistic sit(ation: The meas(re! 'ariale will
incl(!e the e""ects o" sensor noise an! hi#her
"re4(enc) process !ist(rances$
DYNAMIC SIMULATION
Time
0 5 10 15 20 25 30 35 40 45 50
-0.5
0
0.5
1
1.5
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
Realistic sit(ation: The mo!el !oes not
represent the process e%actl)$ We will ass(me
that the mo!el has t 9:D errors in #ain, time
constant an! !ea! time, "or e%ample:
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
t
r
o
lle
d

V
a
r
ia
b
le
0 5 101520253035404550
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
ip
u
la
t
e
d

V
a
r
ia
b
le
<$1: , A$9:
7$: , 9$:
1$: ,7 9$:
#ain
Dea! time
Time constant
7 s
$
E F
E F
E F

+

10
0 2
5 s
P
e
s MV
s CV
s G
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
Realistic sit(ation: Two t)pical inp(ts will e
consi!ere!, chan#es in set point an!
!ist(rance$ *or correlations, step inp(ts, (t
controller will "(nction "or other inp(ts$
sol'ent
p(re A
AC
*
.
*
A
.P
.ol'ent D A
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
Realistic sit(ation: We will consi!er the PID
controller, which is (se! "or nearl) all sin#le,
loop F7CG, 7?GE controllers$
sol'ent
p(re A
AC
*
.
*
A
.P
I
dt
CV d
T dt t E
T
t E K t MV
t
d
I
c
+
1
]
1

0

' ) ' (
1
) ( ) (
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
CG D)namic Heha'ior:
.tale, Iero o""set, minim(m IAE
?G D)namic Heha'ior:
!ampe! oscillations an!
small "l(ct(ations !(e to
noise$
?G can e
more
a##ressi'e in
earl) part o"
transient
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
+(r primar) #oal is to maintain the CG near
the set point$ Hesi!es not wearin# o(t
the 'al'e, wh) !o we ha'e #oals "or the ?G5
AC
0 5 10 15 20 25 30 35 40
-10
0
10
20
30
40
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
.team "low
0ar#e, rapi! chan#es
to the steam "low can
!ama#e the tra)s
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
0 5 10 15 20 25 30 35 40
-10
0
10
20
30
40
Time
M
a
n
ip
u
la
t
e
d

V
a
r
ia
b
le
*(el "low
0ar#e, rapi! chan#es
to the "(el "low ca(se
thermal stress that
!ama#es t(es$
FT
1
FT
2
PT
1
PI
1
AT
1
TI
1
TI
2
TI
3
TI
4
PI
2
PI
3
PI
4
TI
5
TI
6
TI
7
TI
8
FI
3
TI
10
TI
11
PI
5
PI
6
TC
*(el
+(r primar) #oal is to maintain the CG near
the set point$ Hesi!es not wearin# o(t
the 'al'e, wh) !o we ha'e #oals "or the ?G5
CHAPTER 9: PID TUNING
De"ine the t(nin#
prolem
7$ Process
D)namics
9$ ?eas(re!
'ariale
<$ ?o!el error
@$ Inp(t "orcin#
:$ Controller
A$ Per"ormance
meas(res
C+?HINED DE*INITI+N +* TUNING
PR+H0E? *+R C+RRE0ATI+N

*irst or!er with !ea! time process mo!el

Nois) meas(rement si#nal

J 9:D parameters errors etween

mo!el3plant
PID controller: !etermine /
c
, T
I
, T
!

?inimiIe IAE with ?G insi!e o(n!

We achie'e the #oals )
a!2(stin# /c, TI an! T!$
Details in chapter an!
Appen!i% E$
CHAPTER 9: PID TUNING
C+?HINED DE*INITI+N +* TUNING
*irst or!er with !ea! time process
mo!el
Nois) meas(rement si#nal
J 9:D parameters errors etween
mo!el3plant
PID controller: !etermine /
c
, T
I
, T
!
?inimiIe IAE with ?G insi!e o(n!
/p K 7
K :
K :
TC
'
7
'
9
05101520253035404550
-0.5
0
0.5
1
1.5
05101520253035404550
0
0.2
0.4
0.6
0.8
1
TC
'
7
'
9
/c K ;$1@
TI K 1$:
T! K ;$9;
Process
reaction c(r'e
.ol'e the t(nin#
prolem$ Re4(ires a
comp(ter pro#ram$
Appl), is the
per"ormance
#oo!5
CHAPTER 9: PID TUNING
0 20 40 60 80 100 120
-5
0
5
10
15
C
V
0 20 40 60 80 100 120
0
5
10
15
20
25
time
M
V
0 20 40 60 80 100 120
-5
0
5
10
15
C
V
0 20 40 60 80 100 120
0
10
20
30
40
time
M
V
0 20 40 60 80 100 120
-5
0
5
10
15
C
V
0 20 40 60 80 100 120
0
10
20
30
time
M
V
Plant K mo!el Plant K L 9:D
Plant K , 9:D
The t(nin# is not the est "or an) in!i'i!(al case, (t it is
the est "or the ran#e o" possile !)namics , it is ro(st6
?G o(n! ?G o(n! ?G o(n!
CHAPTER 9: PID TUNING
C+?HINED DE*INITI+N +* TUNING
*irst or!er with !ea! time process
mo!el
Nois) meas(rement si#nal
J 9:D parameters errors etween
mo!el3plant
PID controller: !etermine /
c
, T
I
, T
!
?inimiIe IAE with ?G insi!e o(n!
/p K 7
K :
K :
TC
'
7
'
9
05101520253035404550
-0.5
0
0.5
1
1.5
05101520253035404550
0
0.2
0.4
0.6
0.8
1
TC
'
7
'
9
/c K ;$1@
TI K 1$:
T! K ;$9;
0 20 40 60 80 100120
-5
0
5
10
15
C
V
0 20 40 60 80 100120
0
10
20
30
time
M
V
Goo!
Per"ormance
Process
reaction c(r'e
.ol'e the t(nin#
prolem$
Re4(ires a comp(ter
pro#ram$
CHAPTER 9: PID TUNING
We co(l! sol'e each prolem in!i'i!(all), (t this wo(l!
e too time cons(min#$ We wo(l! li&e to !e'elop a
correlation ase! on man) sol(tions$
Dimensionless T(nin#
Constants
In!epen!ent 'ariale
Recall that M3FL EN L M 3FL EN K 7

,
_

+ +

,
_

+ +
+
+ +

,
_

+ +

,
_

+ +
+
+

+
+
EE 3F F O
E 3F F O
E 3F F O
EE 3F F O
E 3F F O
E 3F F O
E F
E F
E 3F O
E 3F O








s
e
T s
T s
K K
s
e
T s
T s
K K
s MV
s CV
s
d
I
p c
s
d
I
p c
1
1
1 1
1
1
1
CHAPTER 9:
PID TUNING
T(nin# Charts "or PID
*ee!ac& Controllers
(See page 281 in the textbook fo !age p!ot"#
These were !e'elope! )
s(mmariIin# a lar#e
n(mer o" case st(!ies
in these !imensionless
charts5
!ist(rance
.et point chan#e
CHAPTER 9:
PID TUNING
T(nin# Charts "or PI
*ee!ac& Controllers
These were !e'elope! )
s(mmariIin# a lar#e
n(mer o" case st(!ies
in these !imensionless
charts5
(See page 28$ in the textbook fo !age p!ot"#
!ist(rance .et point
CHAPTER 9: PID TUNING
sol'ent
p(re A
AC
*
.
*
A
Lets apply the tuning charts to the three-tank miing process!
"hich is not #irst or$er "ith $ea$ time%
T(nin# "rom chart
/c K 55
TI K 55
T! K 55
Process reaction c(r'e
/p K ;$;<9 DA3Dopen
K :$: min
K 7;$: min
CHAPTER 9: PID TUNING
sol'ent
p(re A
AC
*
.
*
A
Lets apply the tuning charts to the three-tank miing process!
"hich is not #irst or$er "ith $ea$ time%
T(nin# "rom chart
/c K 7$93;$;<9 K <; Dopen3DA
TI K ;$A9F7AE K 77 min
T! K ;$;:F7AE K ;$B; min
Process reaction c(r'e
/p K ;$;<9 DA3Dopen
K :$: min
K 7;$: min
CHAPTER 9: PID TUNING
0 20 40 60 80 100 120 140 160 180 200
25
30
35
40
45
50
time
m
a
n
i
p
u
l
a
t
e
d

f
o
w
0 20 40 60 80 100 120 140 160 180 200
3
3.1
3.2
3.3
3.4
time
c
o
n
c
e
n
t
r
a
t
i
o
n
Concentration !ist(rance
Gal'e D
open
E""l(ent concentration
sol'ent
p(re A
AC
*
.
*
A
Goo!
Per"ormance
50 &0 % 0 ' ) ' (
11
1
) ( '0
0
+
1
]
1

t
dt
CV d
dt t E t E %
CHAPTER 9: PID TUNING
*INE TUNING: Process reaction c(r'e an! t(nin# charts
pro'i!e a #oo! metho! "or t(nin# man) Fnot allE PID
loops$ We nee! to learn how to "ine t(ne loops to "(rther
impro'e per"ormance ase! on c(rrent loop eha'ior
,WH>5

.ome loops co(l! ha'e !i""erent per"ormance o2ecti'es

.ome loops co(l! ha'e !)namics !i""erent "rom "irst or!er

with !ea! time

Co(l! ha'e een error in the process reaction c(r'e,

perhaps a !ist(rance occ(rre! !(rin# the e%periment$

Plant !)namics can chan#e !(e to chan#es in "ee! "low

rate, reactor con'ersion, an! so "orth$
CHAPTER 9: PID TUNING
What is the e""ect o" chan#in# the controller #ain on the
control per"ormance o" a PID loop5
0et-s !o an e%periment ) chan#in# /c an! monitorin#
the per"ormance$
I
dt
CV d
T dt t E
T
t E K t MV
t
d
I
c
+
1
]
1

0

' ) ' (
1
) ( ) (
CHAPTER 9: PID TUNING
PID controller with /c chan#in#, TI K 7;, T! K ;$

Wh) !oes IAE

increase "or
small /c5

Wh) !oes IAE

increase "or
lar#e /c5
0 0.5 1 1.5 2
0
20
40
60
controller gain
c
o
n
t
r
o
l

p
e
r
f
o
r
m
a
n
c
e
,

I
A
E
Ha!
5
TC
'7
'9
0 50 100 150 200
-1
-0.5
0
0.5
1
time
c
o
n
t
r
o
l
l
e
d

v
a
r
i
a
b
l
e
0 50 100 150 200
-1
-0.5
0
0.5
1
time
c
o
n
t
r
o
l
l
e
d

v
a
r
i
a
b
l
e
0 50 100 150 200
-1
-0.5
0
0.5
1
time
c
o
n
t
r
o
l
l
e
d

v
a
r
i
a
b
l
e
Is this the PestQ5
/c K ;$A9
/c K 7$7@ /c K 7$:9
CHAPTER 9: PID TUNING
What is the e""ect o" chan#in# the inte#ral time on the
control per"ormance o" a PID loop5
Is the answer !i""erent "rom /c5 What is !i""erent5
I
dt
CV d
T dt t E
T
t E K t MV
t
d
I
c
+
1
]
1

0

' ) ' (
1
) ( ) (
CHAPTER 9: PID TUNING
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
S-LOOP plots deviation variables (IAE = 9.6759)
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
*INE TUNING: 0et-s appl) o(r (n!erstan!in# to (il! "ine
t(nin# #(i!elines$
This is P#oo!Q control
per"ormance$
E%plain the shape o"
the CG an! ?G
responses$
CHAPTER 9: PID TUNING
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
S-LOOP plots deviation variables (IAE = 9.6759)
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 5 10 15 20 25 30 35 40 45 50
0
0.5
1
1.5
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
Note: this is a step chan#e to the set point , #oo! "or !ia#nosis6
?G
;
K /c F.PE sho(l! e close to the
nee!e! chan#e at stea!) state$
?Gss
Constant slope
EFtE K constant
CG !oes not chan#e eca(se o" !ea! time
?G o'ershoot mo!erate RK ;$:F?GssE
CG limite! set point o'ershoot, "ast !ampin#, an!
ret(rn to the set point
CHAPTER 9: PID TUNING
Appl) the "ine t(nin# #(i!elines to the response elow an!
s(##est speci"ic chan#es "or impro'ement$
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
S-LOOP plots deviation variables (IAE = 19.3873)
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
CHAPTER 9: PID TUNING
Appl) the "ine t(nin# #(i!elines to the response elow an!
s(##est speci"ic chan#es "or impro'ement$
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
S-LOOP plots deviation variables (IAE = 19.3873)
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
The CG response is 'er)
slow, not a##ressi'e
eno(#h
The initial chan#e in the ?G is too small, less
than @;D o" the "inal, stea!),state chan#e$
This is poor control
per"ormance$
Controller not
a##ressi'e eno(#h$
.mall ?G
;
, increase
controller #ain,
/
c
) ao(t %9
CHAPTER 9: PID TUNING
Appl) the #(i!elines to the response elow an! s(##est
speci"ic chan#es "or impro'ement$
0 10 20 30 40 50 60 70 80 90 100
0
0.5
1
1.5
2
S-LOOP plots deviation variables (IAE = 20.1754)
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 10 20 30 40 50 60 70 80 90 100
0
0.5
1
1.5
2
2.5
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
CHAPTER 9: PID TUNING
Appl) the #(i!elines to the response elow an! s(##est
speci"ic chan#es "or impro'ement$
0 10 20 30 40 50 60 70 80 90 100
0
0.5
1
1.5
2
S-LOOP plots deviation variables (IAE = 20.1754)
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 10 20 30 40 50 60 70 80 90 100
0
0.5
1
1.5
2
2.5
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
This is poor control
per"ormance$
Controller too
a##ressi'e$
?G
;
is +/$ There,
"ore, increase
inte#ral time,
T
I
) ao(t %9
CG too oscillator)
?G o'ershoot too lar#e
?G
;
CHAPTER 9: PID TUNING, W+R/.H+P 7
Ima#ine that )o( are shipwrec&e! on an islan! an! that
)o( !o not ha'e )o(r te%too& or lect(re notes6 Nat(rall),
)o( want to t(ne some PID controllers$
Re'iew the t(nin# charts an! !e'elop some ro(#h
#(i!elines "or t(nin# that )o( will rememer "or the rest o"
)o(r li"e$
Tropical para!ise (t
no te%too& or internet
connection$
CHAPTER 9: PID TUNING, W+R/.H+P 9
TC
'7
'9
The controller #ain has een positi'e "or the e%amples in
the notes$ Is /
c
alwa)s #reater than Iero5 In )o(r answer,
!isc(ss the temperat(re control s)stem in the pict(re
elow$
What are the (nits o" the controller #ain5
CHAPTER 9: PID TUNING, W+R/.H+P <
0 5 10 15 20 25 30 35 40 45 50
-1
0
1
2
3
4
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 5 10 15 20 25 30 35 40 45 50
0
5
10
15
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
The !ata elow is a process reaction c(r'e "or a process,
plotte! in !e'iation 'ariales$ Determine the t(nin# "or a
PID controller$
TC
'7
'9
CHAPTER 9: PID TUNING, W+R/.H+P @
0 5 10 15 20 25 30 35 40 45 50
-0.5
0
0.5
1
1.5
S-LOOP plots deviation variables (IAE = 6.1515)
Time
C
o
n
t
r
o
l
l
e
d

V
a
r
i
a
b
l
e
0 5 10 15 20 25 30 35 40 45 50
-5
0
5
10
15
20
Time
M
a
n
i
p
u
l
a
t
e
d

V
a
r
i
a
b
l
e
Dia#nose the close!,loop !ata in the "i#(re an! s(##est
mo!i"ications, i" necessar)$
TC
'7
'9
0 5 10 15 20 25 30 35 40 45 50
0
0.2
0.4
0.6
0.8
1
DYNAMIC SIMULATION
Time
C
o
n
t
r
o
lle
d

V
a
r
ia
b
le
0 5 101520253035404550
0
0.2
0.4
0.6
0.8
1
Time
M
a
n
ip
u
la
t
e
d

V
a
r
ia
b
le
9$: , 1$:
7$; , <$;
:$; ,7 :$;
#ain
Dea! time
Time constant
CHAPTER 9: PID TUNING, W+R/.H+P :
E'en with the most care"(l e%periments, )o( are ale to
!etermine the mo!el parameters with t :;D (ncertaint)$
Recommen! initial t(nin# constant 'al(es "or a PID
controller$
7 s
$
E F
E F
E F

+

10
0 2
5 s
P
e
s MV
s CV
s G
When I complete this chapter, I want to e
ale to !o the "ollowin#$
CHAPTER 9: PID TUNING
0ot-s o" impro'ement, (t we nee! some more st(!)6

Rea! the te%too&

Re'iew the notes, especiall) learnin# #oals an! wor&shop

Tr) o(t the sel",st(!) s(##estions

Nat(rall), we-ll ha'e an assi#nment6

E%plain the per"ormance #oals that we see& to

achie'e 'ia t(nin#$

Appl) a t(nin# proce!(re (sin# the process

reaction c(r'e an! t(nin# correlations$

*(rther impro'e per"ormance ) "ine t(nin#

CHAPTER 9: 0EARNING RE.+URCE.

.ITE PC,EDUCATI+N WEH

, Instr(mentation Notes
, Interacti'e 0earnin# ?o!(le FChapter 9E
, T(torials FChapter 9E

.earch the WEH an! "in! a Pa(tomatic PID t(nin#Q

so"tware pro!(ct$ Prepare a critical re'iew o" the
techni4(e$
CHAPTER 9: .UGGE.TI+N. *+R .E0*,.TUD>
7$ *in! some process reaction c(r'e plots in Chapters <,:
an! !etermine the t(nin# "or PID an! PI controllers
(sin# the t(nin# charts$
9$ Usin# .S0++P, repeat the sim(lation res(lts "or the
three,tan& mi%er (n!er PID control$ Then !etermine
the sensiti'it) to chan#es in t(nin# ) chan#in# /
C
an!
T
I
Fone at a time, D chan#es "rom the asis case t(nin#EC
,:;D, ,7;D, L:;D$ Disc(ss )o(r res(lts$
<$ Usin# .S0++P, a!! noise to the meas(rement in
s(men( 7, /n K ;$;: $ .im(late with the ori#inal
t(nin# an! other 'al(es "or T!$ What happens to the
per"ormance5
CHAPTER 9: .UGGE.TI+N. *+R .E0*,.TUD>
@$ *orm(late 4(estions similar to those in the Interacti'e
0earnin# ?o!(les, one each "or Chec& >o(r Rea!in#,
.t(!) T(estions an! Tho(#ht T(estions$
:$ In chapters <,:, "in! e%amples o" processes "or which the
t(nin# "rom the t(nin# charts wo(l! e F7E applicale
an! F9E not applicale$
A$ +n ?on!a), we t(ne! the three,tan& mi%er composition
controller$ +n *ri!a), we anticipate re!(cin# the "ee!
"low rate ) :;D F"rom 1 to <$: m
<
3minE$ When this
occ(rs, sho(l! we chan#e the t(nin# o" the controller5 I"
)es, which constants an! ) how m(ch5
FHint: Three,tan& mi%er mo!el is in E%ample 1$9 on Pa#e 99< o" te%too&$E