Internati
o
nal
Journal of Ele
c
trical
and Computer
Engineering
(IJE
CE)
V
o
l.
5, N
o
. 5
,
O
c
tob
e
r
201
5, p
p
. 1
075
~108
2
I
S
SN
: 208
8-8
7
0
8
1
075
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJECE
Perform
a
nce E
v
aluati
on of Th
ree PID Controller Tuning
Algorith
m on a P
r
ocess Plant
Oladimeji Ibr
a
him
*
, Sulym
an A.
Y. Amuda**,
Ol
atunji O.
Mohamm
ed*, Ganiyu
A. Kareem
***
*Departm
ent
of
Ele
c
tri
cal
and
E
l
ectron
i
cs
Eng
i
ne
ering, University of Ilor
in, Ilor
in
Nigeria
**Department of
Computer Eng
i
neering
,
Univ
ersity
of
Ilorin
,
Ilor
in Niger
i
a
***
Nationa
l Ag
enc
y
for
Sci
e
n
ce and
Engin
eerin
g Infrast
ru
cture, Abuja
Niger
i
a
Article Info
A
B
STRAC
T
Article histo
r
y:
Received Apr 29, 2015
Rev
i
sed
Jun
28,
201
5
Accepte
d
J
u
l 12, 2015
Accurate tun
i
ng
of contro
ller
in
industria
l p
r
ocess
operation
is prerequisite to
s
y
stem smooth operation wh
ich d
i
rectly
redu
ce pr
o
cess variab
ility
, improved
effic
i
enc
y
,
red
u
ced energ
y
cos
t
s
,
and in
creas
ed produ
c
tion rat
e
s
.
Performance evaluation of a model base
d PID controller tuning algorithm on
a ch
em
ical
proc
es
s
plant
is
pre
s
ented
in
this p
a
per.
The
contr
o
l ac
tion o
f
three d
i
ffer
e
nt PID controller tu
ni
ng algorithms namely
; Hagglu
nd-Astrom,
Cohen and Coon, and Zieg
ler-N
ichols on
the process plant was examined in
a closed loop
co
ntrol configur
ation
under normal operating
condition and in
the face of distu
r
bance. LabVIEW soft
ware was
us
ed to m
odel a chem
ic
al
process plant fr
om open loop control
test data. The
tim
e domain response
anal
ysis of
the
c
ontrolle
rs
shows that each
tuning
algor
ithm exhib
it diff
eren
t
time response.
Ziegler-Nic
ho
ls
algorithm shows the best p
e
rfor
m
ance with
fastest r
i
se t
i
m
e
, set
tling
tim
e
a
nd was abl
e
to
restore
the s
y
ste
m
back t
o
normal operatin
g condition
in a short tim
e when subjected
to
disturbance
compare to Cohen & Coon controller and Hagglund-Astrom algorithm
settings.
Keyword:
Co
n
t
ro
l algorith
m
PI
D con
t
ro
ller
Plan
t m
o
d
e
l
Tim
e
resp
onse
Tu
ni
n
g
param
e
t
e
r
Copyright ©
201
5 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
Oladim
eji Ibra
him
Depa
rt
m
e
nt
of
El
ect
ri
cal
and
El
ect
roni
cs
E
n
gi
nee
r
i
n
g,
Un
i
v
ersity of
Ilo
r
in,
PMB 15
15
, I
l
or
inN
i
g
e
r
i
a.
Em
a
il: reachol
aibra
h
im
@g
mail.com
1.
INTRODUCTION
C
ont
r
o
l
sy
st
em
reg
u
l
a
t
e
s fl
o
w
of e
n
er
gy
or
m
a
t
t
e
r and i
t
s
i
m
port
a
nce can
not
be o
v
e
r
em
pha
si
sed i
n
all facets of hum
a
n activitie
s from
dom
est
i
c operations to
industrial applications.
Industrial process plant
co
m
p
rises of series of
p
r
o
cess un
its in
tercon
n
ected
an
d
t
h
e ab
ility to
con
tin
uou
sly m
e
asu
r
e and
co
n
t
ro
l th
e
pr
ocess
vari
abl
e
s (PV
)
i
s
pre
r
equi
si
t
e
t
o
sm
oot
h r
u
nni
ng a
n
d o
p
t
i
m
i
zat
i
on of t
h
e sy
st
em
.
The m
o
st
co
m
m
only
use
d
co
nt
r
o
l
sy
st
em
i
n
i
ndust
r
i
a
l
appl
i
cat
i
on
i
s
t
h
e pr
o
p
o
r
t
i
o
nal
i
n
t
e
g
r
al
an
d de
ri
vat
i
v
e c
o
nt
r
o
l
l
e
r (P
ID
)
due t
o
i
t
s
sim
p
li
ci
t
y
and
ro
b
u
st
nes
s
[1-
4
]
.
Acc
u
r
a
t
e
t
uni
ng
of
t
h
e cont
rol
sy
st
em
i
s
necessary
fo
r sy
st
em
best
p
e
rform
a
n
ce wh
ich
d
i
rectly red
u
ce
p
r
o
cess
v
a
riab
ility,
ma
x
i
mize syste
m
efficien
cy, m
i
n
i
mize en
erg
y
co
sts,
and i
n
c
r
ease
d
pr
o
duct
i
o
n rat
e
s. The t
u
ni
n
g
of a cont
r
o
l
l
e
r i
n
v
o
l
v
es set
t
i
ng t
h
e t
a
rget
ed per
f
o
rm
ance by
speci
fy
i
n
g
des
i
red
out
put
t
h
at
can be
m
a
int
a
i
n
t
h
ro
u
g
h
o
u
t
t
h
e
pr
ocess
op
erat
i
o
n i
r
re
spect
i
v
e
of
p
r
oces
s
v
a
riab
ility an
d
su
rroun
d
i
ng
con
d
ition
.
A PID C
o
n
t
ro
l
l
er is a feed
b
a
ck
au
to
m
a
tic co
n
t
ro
l syst
e
m
th
at
in
teg
r
ates p
r
o
portio
n
a
l (P), in
teg
r
al (I)
and
deri
vat
i
v
e
(D
) m
odes w
h
i
c
h ca
n be a
r
ran
g
e
d
i
n
seri
es, i
d
eal
o
r
pa
ral
l
e
l
st
ruct
ure
s
[5]
.
P
I
D c
o
n
t
rol
l
e
r
o
p
e
rates
b
y
summin
g
th
e con
t
ro
l actio
n of th
e
p
r
op
ortio
nal, th
e in
teg
r
al
and
d
e
ri
v
a
tiv
e actio
n
t
o
p
r
odu
ce a
com
m
on cont
r
o
l
si
g
n
al
t
h
at
i
s
appl
i
e
d t
o
t
h
e sy
st
em
und
er co
nt
r
o
l
[
6
,
7]
. T
h
e
pr
op
o
r
t
i
onal
co
nt
r
o
l
m
ode
chan
ges
t
h
e
co
nt
r
o
l
l
e
r
out
put
i
n
pr
op
o
r
t
i
o
n
t
o
t
h
e e
r
r
o
r
)
(
e
and
th
e ad
ju
stab
le
settin
g
is called
t
h
e
p
r
op
ortion
a
l
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE Vo
l. 5
,
N
o
. 5
,
O
c
tob
e
r
20
15
:
107
5
–
10
82
1
076
gai
n
݇
s
o
m
e
times re
ferred to as
proportional setting.
The
tim
e
and La
place dom
a
in represe
n
tations
of
pr
o
p
o
r
t
i
onal
co
nt
r
o
l
l
e
r
i
s
gi
ve
n by
eq
uat
i
o
n (
1
)
an
d (2
).
Tim
e
dom
ai
n,
)
(
)
(
t
e
k
t
u
p
c
(1
)
Laplace dom
ain,
)
(
)
(
s
e
k
s
U
p
c
(2
)
Whe
r
e,
)
(
t
u
c
and
)
(
t
e
are
the control a
n
d error signals
The i
n
t
e
gral
c
ont
rol
m
ode
of
a PI
D c
o
nt
r
o
l
l
e
r p
r
od
uces
a l
o
ng te
rm
corre
ctive cha
n
ge i
n
c
ont
roller
out
put
by
d
r
i
v
i
ng t
h
e er
r
o
r
of
f
s
et
t
o
zer
o.
It
appea
r
s as a
ra
m
p
of w
h
i
c
h t
h
e sl
ope i
s
det
e
r
m
i
n
ed by
t
h
e
s
i
ze of
th
e erro
r and
th
e adj
u
stab
le settin
g
is ter
m
e
d
in
teg
r
al ti
m
e
i
T
called
th
e I-settin
g
o
f
th
e co
ntro
ller. Th
e ti
me
and La
place domain represe
n
tations
for i
n
tegral co
ntroller is
presente
d
by e
quation
(3) a
nd (4).
Tim
e
dom
ai
n:
t
i
p
t
I
c
dt
t
e
T
K
dt
t
e
K
t
u
0
0
)
(
1
)
(
)
(
(3
)
Laplace dom
ain:
)
(
)
(
s
e
s
k
s
U
I
c
(4
)
Th
e
d
e
ri
v
a
tiv
e con
t
ro
l is
rarely u
s
ed
i
n
con
t
ro
ller
app
licati
o
n as it is v
e
ry sen
s
itiv
e to measu
r
em
en
t
noi
se a
n
d ca
n
m
a
ke t
uni
ng
v
e
ry
di
f
f
i
c
ul
t
b
u
t
i
t
has ad
vant
a
g
e
of m
a
ki
ng c
ont
rol
l
o
o
p
re
s
p
o
n
d
fa
st
er
wi
t
h
l
e
ss
o
v
e
rsh
o
o
t
.
Its
ad
ju
stab
le settin
g is called
d
e
riv
a
tiv
e tim
e
d
T
. The
tim
e
and Laplace dom
ain represe
n
tations
are
gi
ve
n
by
eq
uat
i
o
n
(
5
)
an
d
(6
).
Tim
e
dom
ai
n
dt
t
de
T
k
dt
t
de
k
t
u
d
p
D
c
)
(
)
(
)
(
(5
)
Laplace dom
ain:
)
(
)
(
s
e
s
k
s
U
D
c
(6
)
Th
e effectiv
e co
n
t
ro
l si
g
n
a
l
prov
id
ed
b
y
th
e PID co
n
t
ro
ller is su
mm
a
tio
n
o
f
t
h
e three con
t
ro
l term
s
represe
n
ted i
n
tim
e
and La
place dom
a
in as [1,
6, 8].
dt
t
de
k
dt
t
e
K
t
e
k
t
u
D
t
I
p
)
(
)
(
)
(
)
(
0
(7
)
)
(
)
(
s
e
s
k
s
k
k
s
U
D
I
p
(8
)
Th
e
b
e
st con
t
ro
ller setting
s
is ex
p
ected
to
g
i
v
e
fastest
respon
se in
t
e
rm
s o
f
system rise ti
m
e
,
m
i
nim
u
m
set
t
l
i
ng t
i
m
e
, l
east
ove
rs
ho
ot
, an
d
zero st
eady
st
ate error. T
h
e classical
cont
r
o
l
l
e
r desi
g
n
e
m
pl
oy
s
sy
st
em
m
odel for st
u
d
y
i
n
g
cont
rol
l
e
r pe
rf
orm
a
nce un
de
r di
ffe
re
nt
op
erat
i
ng c
o
n
d
i
t
i
on
bef
o
re real
t
i
m
e
i
m
p
l
e
m
en
tatio
n
.
A syste
m
mo
d
e
l is ob
tain
ed
fro
m
ex
istin
g
m
o
d
e
l, d
e
v
e
lop
e
d
fro
m
n
e
w math
e
m
atica
l
relatio
n
or
usi
n
g m
odel
l
i
ng soft
ware
t
a
ken i
n
t
o
c
o
nsi
d
e
r
at
i
on al
l
t
h
e obse
r
va
bl
e vari
abl
e
s o
f
t
h
e sy
st
em
[8]. The
m
e
t
hods f
o
r t
u
ni
n
g
PI
D co
nt
r
o
l
l
e
r are b
r
oa
dl
y
cl
assi
fi
ed t
o
ope
n l
o
op a
n
d cl
osed l
o
o
p
t
echni
que
. I
n
o
p
e
n
l
o
o
p
m
e
t
hod, t
h
e cont
rol
l
e
r
para
m
e
t
e
rs are ob
t
a
i
n
ed m
a
nual
l
y
from
open
l
oop t
e
st
dat
a
of t
h
e
pl
ant
un
de
r
con
s
i
d
erat
i
o
n
.
In
cl
ose
d
l
o
o
p
m
e
t
hod, t
h
e
cont
rol
l
e
r
pa
ra
m
e
t
e
rs i
s
aut
o
m
a
t
i
call
y
t
uned
whe
n
t
h
e
pl
ant
i
s
ope
rat
e
d i
n
cl
ose
d
l
o
o
p
m
o
d
e
. The m
o
st
com
m
onl
y
used cl
ose
d
l
o
o
p
m
e
t
hods i
n
cl
ude
s Zi
egl
e
r
-
N
i
c
h
o
l
s
m
e
thod, Tyreus-L
uybe
n m
e
thod a
nd
da
m
p
ed oscillation m
e
thod, while open
loop m
e
th
od a
r
e the ope
n
loop
Zi
egl
e
r-
Ni
ch
ol
s m
e
t
hod, C
o
h
e
n a
n
d
C
o
o
n
m
e
t
h
o
d
,
Fert
i
k
m
e
t
h
o
d
a
n
d
Ha
g
g
l
u
nd
-
A
st
r
o
m
m
e
t
hod
[
9
]
.
In t
h
i
s
pa
pe
r,
t
h
ree
di
f
f
ere
n
t
PI
D c
ont
r
o
l
l
e
r t
u
ni
n
g
al
g
o
r
i
t
h
m
s
, nam
e
l
y
; Ha
ggl
un
d-
Ast
r
om
, C
ohe
n
an
d Coon
, and Zieg
ler-
N
i
ch
ols ar
e
u
s
ed
t
o
d
e
sign
PI
con
t
ro
ller settin
g
s
for a ch
em
ical proces
s
plant. The
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Perf
or
ma
nce E
v
al
u
a
t
i
o
n
of
Th
ree PI
D C
ont
r
o
l
l
e
r T
uni
ng
Al
gori
t
hm
o
n
a P
r
ocess
Pl
a
n
t
(Olad
i
meji Ib
rah
i
m)
1
077
cont
rol
l
e
r
desi
gn
pr
ocess i
n
v
o
l
v
es
devel
o
p
m
ent
of p
r
oces
s pl
ant
m
odel
from
l
a
borat
o
r
y
ope
n l
o
o
p
t
e
st
dat
a
o
f
t
h
e pl
ant
usi
n
g
L
abor
ato
r
y
V
irtu
al
I
nstrumen
tatio
n
E
ngin
eer
ing
W
or
k
b
enc
h
(
L
ab
VI
E
W
) s
o
ft
ware
.
The PI
cont
rol
l
e
r
gai
n
param
e
t
e
rs were cal
cul
a
t
e
d
fr
om
t
h
ree PI
D t
uni
ng c
o
nt
r
o
l
l
e
r al
go
ri
t
h
m
s
and i
m
pl
em
ent
e
d i
n
LabVIEW co
ntro
l sim
u
lato
r to
stud
y th
e perfo
r
m
a
n
ce of th
e ob
tain
ed
settin
g
s
. Th
e
resu
lts shows th
at th
e
tran
sien
t resp
on
se o
f
th
e
t
h
ree
con
t
ro
ller d
i
ffers with
Zieg
ler-Nicho
l
s m
e
th
od
sho
w
i
n
g th
e
fastest tran
sient
r
e
spon
se.
2.
PLANT MODELLING
The
p
r
o
cess
pl
ant
ope
n l
o
o
p
st
ep
resp
o
n
set
e
st
dat
a
was
u
s
ed t
o
m
odel
t
h
e
pl
ant
fr
om
whi
c
h
pl
ant
p
a
ram
e
ters were ob
tain
ed
for con
t
ro
ller settin
g
s
calcu
lation
s
.
Th
e test d
a
ta was logg
ed
in
ex
cel spread
th
en
upl
oade
d t
o
L
a
bV
IE
W c
ont
r
o
l
t
o
ol
b
o
x
t
o
gene
rat
e
t
h
e
p
l
ant
res
p
o
n
se
gra
p
h as s
h
ow
n i
n
Fi
g
u
re
1
.
Usi
n
g
co
n
tinuo
us sing
le in
pu
t sin
g
l
e o
u
t
pu
t (
S
I
S
O
)
arr
a
y b
l
o
c
k. Th
e LabV
IEW
sof
t
w
a
re was u
s
ed
to
an
alyse th
e
syste
m
respons
e
-to-step i
n
put
)
(
t
u
sti
m
u
l
u
s
to
esti
m
a
te th
e p
l
an
t
tran
sfer fun
c
tio
n.
Fig
u
r
e
1
.
Plan
t
r
espon
se t
o
step
inpu
t
Th
e ti
m
e
respo
n
s
e in
Figu
re1
shows th
at th
e p
l
an
t
is a first-order syste
m
characterized with tim
e-
del
a
y
at
t
r
ansi
e
n
t
.
T
h
e t
r
an
sfe
r
fu
nct
i
o
n o
f
fi
r
s
t
or
der sy
st
em
pl
us
del
a
y
i
s
g
i
ven
by
t
h
e ex
pressi
o
n
o
f
eq
uat
i
o
n
(9
) [1]
.
Tran
sfe
r
F
u
nct
i
on,
1
)
(
s
ke
s
G
s
t
d
(9
)
Whe
r
e,
k
is p
l
ant g
a
in
and
is time co
n
s
tan
t
(s)
Th
e tim
e resp
on
se to
a step
i
n
p
u
t
o
f
a first
o
r
d
e
r system
wit
h
resp
ect to
t
h
e g
a
in
am
p
litu
de isex
press
as;
t
e
t
Y
1
)
(
(1
0)
At tim
e
t
=
,t
he
p
l
ant
res
p
o
n
se
a
m
pli
t
ude
)
(
t
Y
is expected t
o
ha
ve
reach 63
.2% of
its
final value
[10].
A
t
63
.2
% of
)
(
t
Y
, t
h
e
pl
ant
c
o
r
r
es
po
n
d
i
n
g
gai
n
a
m
pli
t
ude i
s
1.
2
6
at
a t
i
m
e of
3
.
9
6
sec
o
nds
.
The dead
(
d
t
) time associated
to the
plant
resp
on
se is
estimated
to
b
e
1
sec as obse
rve
d
from
the
pl
ant
resp
o
n
se gra
p
h(Fi
gu
re 1
)
,
t
h
ere
f
ore the
plant tim
e constant (
) is estimated
to
b
e
(3.96 –
1) seco
nd
s.
Pl
ant
par
amet
ers:
Plan
t g
a
in
k
is 2
,
th
e tim
e c
o
n
s
tan
t
is 2
.
96 sec.
The estim
ated plant m
odel
of
the pr
ocess
pl
a
n
t
ba
sed
o
n
(9
)
y
i
el
ds:
Plan
t tran
sfer
fu
n
c
tion
,
1
96
.
2
2
)
(
1
s
e
s
G
s
(1
1)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE Vo
l. 5
,
N
o
. 5
,
O
c
tob
e
r
20
15
:
107
5
–
10
82
1
078
Th
e
ob
tain
ed
plan
t transfer
fun
c
tio
n sh
ows th
at th
e syst
em
is a first ord
e
r
syste
m
with
time d
e
lay. Th
e
m
o
d
e
l
param
e
t
e
rs wereuse
d
i
n
t
h
e sect
i
on t
h
ree (
3
) f
o
r cont
rol
l
e
r set
t
i
ngs cal
cul
a
t
i
ons a
n
d si
m
u
l
a
t
i
on
i
m
p
l
e
m
en
tatio
n
.
Th
e ex
pon
en
tial ter
m
(
e
) in
th
e p
l
an
t m
o
d
e
l (11
)
is as a
resu
lt o
f
tim
e d
e
lay asso
ciated
with
th
e p
l
an
t respon
se
[10
]
.
3.
CONTROLLER DE
SIGN
AND IMPLE
M
ETATION
The c
ont
r
o
l
l
e
r
t
uni
n
g
p
aram
eters we
re cal
cul
a
t
e
d f
o
r
bot
h t
h
e p
r
op
ort
i
o
na
l
and i
n
t
e
gral
g
a
i
n
base
d
o
n
t
h
ree t
u
ni
ng al
go
ri
t
h
m
nam
e
ly
;
Hag
g
l
u
nd
-A
st
rom
t
uni
n
g
a
l
go
ri
t
h
m
,
C
ohe
n a
nd C
o
on t
u
ni
n
g
al
g
o
ri
t
h
m
,
an
d
Ziegler-Nichol
s tuning
algorit
h
m
as follows.
3.
1 H
a
ggl
u
n
d
-
Astr
om
C
o
n
t
r
o
l
l
er Set
t
i
n
g
s
The t
uni
ng
al
g
o
ri
t
h
m
or
Ha
g
g
l
u
nd
-
A
st
r
o
m
tuni
ng
set
t
i
ngs
i
s
p
r
ese
n
t
e
d i
n
Tabl
e
1 [
7
]
.
Tabl
e 1. Ha
ggl
un
d
-
Ast
r
om
de
si
gn
pa
ram
e
t
e
r
s
Plant Transf
er Fun
c
tion G(s)
Pr
opor
tional Gain (
)
T
i
m
e
Constant (
࣎
ࡵ
)
s
Ke
s
K
35
.
0
7
1
s
Ke
s
K
K
28
.
0
14
.
0
10
8
.
6
33
.
0
From
t
h
e pl
ant
m
odel
(1
1),
Plan
t d
e
lay
d
t
=
= 1sec,
gai
n
am
pl
i
t
ude
k
=2
, a
n
d t
i
m
e
const
a
nt
= 2.
96.
Usi
n
g P
I
t
e
rm
s o
f
Ta
bl
e 1
,
t
h
e
p
r
o
po
rt
i
onal
ga
i
n
p
k
= 0.484
,
I
=
1.
88
3,
an
d
I
k
=0
.257
Tran
sfe
r
f
u
nct
i
o
n
o
f
P
I
c
ont
rol
l
er,
s
k
s
k
s
U
i
p
)
(
(1
2)
Th
erefo
r
e,
th
e co
n
t
ro
ller settin
g
yield
s
tran
sferfun
c
tion
o
f
eq
u
a
tion
(1
3).
s
s
s
U
257
.
0
484
.
0
)
(
(1
3)
3.
2 Cohen an
d
Coon Contr
o
ller
Settin
g
Th
eCoh
en
an
d
Co
on
con
t
ro
ller settin
g fo
r a
first ord
e
r syst
em
p
l
u
s
d
e
ad
time is p
r
esen
ted
in Tab
l
e 2 [11
,
12
].
Tabl
e
2. C
o
h
e
n
an
d C
o
o
n
desi
gn
pa
ram
e
t
e
rs
Plant Transf
er Fun
c
tion G(s)
Pr
opor
tional Gain (
)
I
n
tegr
al T
i
m
e
Constant (
࣎
ࡵ
)
s
Ke
s
12
9
.
0
1
K
20
9
3
30
The pr
o
p
o
r
t
i
o
n
a
l
and
i
n
t
e
gral
gai
n
pa
ram
e
t
e
r
s
we
re obt
ai
ne
d base
d on
t
h
e Tabl
e 2
al
g
o
r
i
t
h
m
:
p
k
= 1
.
37
3,
I
=
1.968
, and
I
k
=0
.6
97
The t
r
a
n
s
f
er
f
u
nct
i
o
n
f
o
r
t
h
e
C
ohe
n a
n
d C
o
on
PI
co
n
t
ro
ller settin
g
s
is presen
ted
i
n
eq
uatio
n (14
)
s
s
s
U
697
.
0
373
.
1
)
(
(1
4)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Perf
or
ma
nce E
v
al
u
a
t
i
o
n
of
Th
ree PI
D C
ont
r
o
l
l
e
r T
uni
ng
Al
gori
t
hm
o
n
a P
r
ocess
Pl
a
n
t
(Olad
i
meji Ib
rah
i
m)
1
079
3.3 Z
i
egler-Ni
chols
Contr
o
ller Settings
The Zie
g
ler-Ni
chols m
e
thod
was
ba
sed on process reaction
curve
m
e
thod with the ass
u
m
p
tion that
pr
ocess c
ont
r
o
l
has o
p
en l
o
o
p
st
ep res
p
o
n
se l
i
ke -S-
s
ha
pe as
sho
w
n i
n
Fi
g
u
r
e 2. T
h
e PI
D
cont
rol
l
e
r
para
m
e
t
e
r
settin
g
s
were ob
tain
ed using
t
h
e Zieg
ler-Nich
o
l
s algo
rith
m
p
r
esen
ted in
Tab
l
e 3 [8
, 13
].
Tabl
e
3.
Zi
egl
e
r-
Ni
ch
ol
s st
ep
res
p
o
n
se
t
u
ni
n
g
param
e
t
e
rs
Controller S
t
ruc
t
ure
P
r
oport
i
onal G
a
in (
K
P
)
Int
e
gral Tim
e
Const
a
nt
(
࣎
ࡵ
)
Derivative
Ti
m
e
Constant
(
࣎
ࡰ
)
Case (i)
P
L
R
N
1
Case (ii) P
I
L
R
N
9
.
0
3L
Case (iii) P
I
D
L
R
N
2
.
1
2L
0.5L
u
T
y
R
N
,
T
y
is th
e slo
p
e
o
f
po
in
t of po
in
t o
f
in
flex
ion o
f
th
e p
r
o
cess reactio
n
curve an
d
u
is th
e
hei
g
ht
o
f
t
h
e
re
act
i
on c
u
r
v
e.
Fi
gu
re 2.
Pl
ant
r
esp
o
n
se
Fr
o
m
th
e r
e
spon
se cu
rv
e i
n
Fi
g
u
r
e
2
8
.
0
1
5
.
3
0
2
T
y
, and
2
k
u
,
4
.
0
2
8
.
0
N
R
Co
n
s
i
d
eri
n
g
case (ii) wh
ich
is PI term
o
f
th
e Zieg
ler-
Nichols tu
n
i
ng
algo
ri
th
m
,
th
e co
n
t
roller p
a
ram
e
ters was
esti
m
a
ted
as follo
ws:
Th
e p
r
op
or
tional
g
a
in
p
k
=2.25,
I
=3
, a
n
d
I
k
= 0.
75
Th
e
con
t
ro
ller tran
sfer fu
n
c
tion
s
s
s
U
75
.
0
25
.
2
)
(
(1
5)
Th
e
p
e
rform
a
n
ce o
f
t
h
e ob
tain
ed
settin
g
s
for th
e threeP
I co
n
t
ro
ller setting
s
on
t
h
e pro
c
ess p
l
an
t
u
n
d
e
r
n
o
rm
al
ope
rat
i
n
g co
n
d
i
t
i
on a
nd
wi
t
h
di
st
ur
ba
nce
s
was i
n
v
e
st
i
g
at
ed. T
h
e cl
ose
d
l
o
o
p
sy
s
t
em
was sim
u
lat
e
d i
n
L
a
bV
I
E
W
s
o
f
tw
a
r
e
w
ith
a s
t
ep
ch
a
n
g
e
in
th
e set-po
in
t
fo
ll
owed b
y
a
un
it step
d
i
st
u
r
b
a
n
c
e after
40
seco
nd
s as
prese
n
t
e
d
i
n
t
h
e
si
m
u
l
a
ti
on
bl
o
c
k
of
Fi
g
u
re
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE Vo
l. 5
,
N
o
. 5
,
O
c
tob
e
r
20
15
:
107
5
–
10
82
1
080
Fig
u
re
3
.
PI con
t
ro
ller clo
s
ed
lo
op
sim
u
latio
n
4.
R
E
SU
LTS AN
D ANA
LY
SIS
The t
i
m
e respo
n
se
param
e
t
e
rs fo
r
Hag
g
l
un
d-
Ast
r
om
, C
ohe
n a
nd C
o
o
n
, a
n
d Zi
e
g
l
e
r-
Ni
ch
ol
s
co
n
t
ro
llers tun
i
n
g
setting
on
th
e ch
em
ical p
r
o
cess p
l
an
t und
er clo
s
ed
loop o
p
e
ratio
n
is presen
ted
in
Tab
l
e 4
.
It
sho
w
s eac
h co
nt
r
o
l
l
e
r t
i
m
e
respo
n
se
per
f
o
r
m
a
nce on t
h
e
pl
ant with res
p
ect to the desirable specifications
of
rise ti
m
e
, settli
n
g
ti
m
e
, p
e
rcen
tag
e
s ov
ershoo
t, an
d p
e
ak
v
a
lu
e.
Tabl
e
4. Ti
m
e
resp
o
n
se
para
m
e
t
r
i
c
dat
a
Co
n
t
ro
llers
Rise Ti
m
e
(s
)
Over
shoot (
%
)
Steady St
ate G
a
in
Settling Ti
m
e
(s)
Peak Value
Hagglund-
Astr
o
m
4.
210
3.
115
1
16.
237
1.
031
Cohen &
C
oon
1.
896
3.
902
1
10.
426
1.
039
Ziegler-
Nichols
1.
172
0
1
4.
689
0.
999
Th
e Zieg
ler-Ni
ch
o
l
s con
t
ro
ller h
a
s
a fastest rise ti
m
e
o
f
1
.
17
2
sec, settlin
g ti
m
e
o
f
4
.
68
9
an
d
with
no
ove
rs
ho
ot
. T
h
e
C
ohe
n an
d C
o
on c
ont
rol
l
e
r e
xhi
bi
t
a
m
oder
a
t
e
l
y
sl
ow resp
ons
e, t
h
e ri
se t
i
m
e i
s
1.89
6 se
c wi
t
h
settlin
g
ti
m
e
o
f
10
.42
6
sec
an
d
3
.
90
2% ov
ershoo
t. Th
e
Hag
g
l
u
nd-Ast
r
o
m
co
n
t
ro
ller sett
in
g
s
resp
ond
ed with
lo
ng
est ti
m
e
d
e
lay o
f
4
.
2
1
0
sec, settlin
g
ti
me o
f
16
.23
7
sec b
u
t
it h
a
s lesser ov
ersho
o
t
o
f
3
.
1
1
5
%
com
p
are to
t
h
e C
ohe
n a
nd
C
o
o
n
res
p
onse
.
The res
p
o
n
se param
e
t
r
i
c
dat
a
sho
w
s t
h
at
t
h
e Zi
egl
e
r-
Ni
ch
ol
s t
uni
ng m
e
tho
d
i
s
m
u
ch bet
t
e
r fo
r desi
gni
ng c
o
nt
r
o
l
l
e
r fo
r a f
i
rst
or
der sy
st
e
m
pl
us dea
d
t
i
m
e co
m
p
ared
t
o
ot
he
rs ot
her
t
w
o
tu
n
i
ng
m
e
th
ods
h
a
v
e
n
d
e
m
o
n
s
trated
fastest p
r
o
cess
re
spo
n
se ti
m
e
, sh
ortest settlin
g
tim
e
with
n
o
ov
ersho
t
.
In
o
r
d
e
r to
furth
e
r i
n
v
e
sti
g
ate th
e ro
bu
stness o
f
each
co
n
t
ro
ller settin
g, th
e clo
s
ed
loop syste
m
w
a
s
subj
ecte
d
to disturba
nce
at interval to
re
ve
al each
c
o
ntroller disturba
nc
e re
j
ection ca
pability. The
syste
m
respon
se presen
ted
in
Figu
re 4
show
ing
the b
e
h
a
v
i
our of th
e p
l
an
t under no
rm
al o
p
e
ratin
g
con
d
ition
and
whe
n
s
u
b
j
ect
ed t
o
di
st
ur
banc
e at
4
0
sec
o
nds
.
Fi
gu
re
4.
C
o
nt
r
o
l
l
e
r re
sp
o
n
se t
o
di
st
ur
ba
nce
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Perf
or
ma
nce E
v
al
u
a
t
i
o
n
of
Th
ree PI
D C
ont
r
o
l
l
e
r T
uni
ng
Al
gori
t
hm
o
n
a P
r
ocess
Pl
a
n
t
(Olad
i
meji Ib
rah
i
m)
1
081
The Zie
g
ler-Ni
chols
c
on
tro
ller d
e
m
o
n
s
trated strong
est ab
ility to
restore the syste
m
b
ack
to
no
rm
al
ope
ration in the face of
disturba
nce at
short
e
st time of 2 s
ec. It took
about 3 sec for the Cohe
n and
Coon
cont
rol
l
e
r t
o
r
e
st
ore t
h
e pl
a
n
t
and
Hag
g
l
u
n
d
-
Ast
r
om
cont
rol
l
e
r co
ul
d o
n
l
y
bri
n
g t
h
e sy
st
em
t
o
norm
a
l
ope
rat
i
n
g co
n
d
i
t
i
on aft
e
r
5 se
c, as sh
o
w
n
Fi
gu
re
4. A
co
nt
r
o
l
l
e
r t
u
ni
n
g
o
b
j
ect
i
v
e i
s
t
o
fee
d
set
t
i
ngs
pa
ra
m
e
t
e
rs
th
at will pro
v
i
d
e
t
h
e
b
e
st con
t
ro
l actio
n for sm
o
o
t
h
p
r
o
c
ess op
eratio
n
un
d
e
r
no
rm
al o
p
e
rating
con
d
i
t
i
o
n
and
wh
en
t
h
ere is
d
i
stu
r
b
a
n
ce. Th
e ab
ility o
f
Zieg
ler-Nicho
l
s
co
n
t
ro
ller to
rej
ect th
e d
i
st
u
r
ban
ce i
n
earliest ti
m
e
and t
h
e fast
t
r
ansi
ent
res
p
on
se dem
onst
r
at
ed sh
o
w
s i
t
s
be
t
t
e
r
m
e
t
hod f
o
r t
uni
ng
pr
oce
ss pl
ant
pl
us d
e
l
a
y
as
com
p
ared t
o
C
ohe
n a
n
d C
o
on and
Hagglund-Astrom
m
e
thod.
5.
CO
NCL
USI
O
N
A
pr
ocess
pl
a
n
t
has
bee
n
m
odel
l
e
d
f
r
om
ope
n l
o
o
p
t
e
st
dat
a
, a
n
d t
h
r
ee vari
ou
s P
I
D c
ont
r
o
l
l
e
r
al
go
ri
t
h
m
were
use
d
t
o
desi
g
n
e
d c
ont
rol
l
e
r
p
a
ram
e
t
e
r fo
r the syste
m
. The
plant tra
n
s
f
er function re
veals
that
t
h
e sy
st
em
i
s
a fi
rst
o
r
de
r
pl
us
del
a
y
s
.T
hre
e
di
f
f
ere
n
t
c
ont
rol
l
e
r
t
u
ni
n
g
a
l
go
ri
t
h
m
were
use
d
t
o
cal
cul
a
t
e
PI
cont
rol
l
e
r set
t
i
ngs a
nd i
m
pl
em
ent
e
d i
n
LabV
IE
W
c
o
nt
r
o
l
-
t
o
ol
ki
t
.
Th
e sy
st
em
cont
i
n
u
o
u
s
t
i
m
e
dom
ai
n
respon
se sh
ows th
estab
ility a
n
d
robu
stn
e
ss
o
f
each
co
n
t
ro
ll
eron
th
e
p
l
an
t
u
n
d
e
r no
rm
al
o
p
e
rating
conditio
n
and when subjected
todist
urbances.
T
h
e t
i
m
e dom
ai
n resp
ons
e sh
o
w
s t
h
a
t
Zi
egl
e
r-
Ni
ch
ol
s co
nt
r
o
l
l
e
r e
xhi
bi
t
s
th
e b
e
st performan
ce with
fastest rise ti
m
e
,
settlin
g
ti
m
e
an
d
ab
ility to
restore th
e syste
m
b
ack
to
no
rm
al
ope
rating condition in earliest tim
e
in the face of dist
urbanc
e. The Cohe
n& Coon
controller perform
a
nce was
m
o
d
e
rately b
e
tter as co
m
p
are
to
Hagg
lun
d
-Astro
m
s
ettin
g
s
.
REFERE
NC
ES
[1]
K.J. Åström and
T. Hägglund,
A
d
vanced PID co
ntr
o
l
: ISA-The Instrumentation
,
S
y
st
ems, and Automation Society;
Research
Tr
iang
le Park
, NC 277
09, 2006
.
[2]
J. Basilio
and S.
Matos, "Design
of PI and PID
controll
ers with
transien
t perfor
m
ance speci
fic
a
t
i
on",
Educa
tion
,
IEEE Transactio
ns on,
vol. 45
, p
p
. 364-370
, 200
2.
[3]
S.N.S. Salim,
M. Rahmat, A.
Faudzi,
Z. Ismail, N. S
unar,
and S.A. Samsu
d
in, "R
obust Control Strateg
y
for
P
n
eum
a
tic Driv
e S
y
s
t
em
via
Enhanc
ed Nonli
n
ear P
I
D Contr
o
ller",
International Journal of Electrica
l
and
Computer Engin
eering (
I
JEC
E
)
,
vol. 4
,
pp
. 658-6
67, 2014
.
[4]
M.R.R. Kho
y
gani, S. Hajigha
semi, and D. Sanaei, "Design
i
ng and Simula
tion fo
r Vertical Movin
g
Control of UAV
S
y
stem using PID, LQR and Fuzzy
logic",
In
tern
ational Journal of
Electrical
and
Computer Engineering (
I
JECE)
,
vol. 3
,
pp
. 651-6
59, 2013
.
[5]
K.H. Ang, G. Chong, and Y.
Li, "PID control s
y
stem anal
y
s
is, design
, an
d technolog
y
"
,
Control Systems
Technology, I
E
EE Transactions
on,
vol. 13
, pp
. 5
59-576, 2005
.
[6]
J. Smuts, "PID Controllers
Explained",
Canadia
n
Process Equip
m
ent
&
Control News,
vol. 2
,
20
02.
[7]
K.J. Aström and
T. Hägg
lund, "PID cont
rollers: theor
y
, d
e
sign
an
d tuning", 1995.
[8]
R. Burns,
Advan
ced contr
o
l
engi
neer
ing
: Butterworth-Heinemann
,
2001.
[9]
P. Cominos and
N. Munro, "PID
controllers: recent tuning meth
o
d
s and design to
specification",
I
EE Proceedings
-
Control Theory
and Applications,
vol. 149
, pp
. 46
-53, 2002
.
[10]
K.S. Pati
l and
D.
Patil
, "Ef
f
ec
tive
Te
aching
L
earn
i
ng Pro
cess for
PID Controller
Based
on
Experimental Setup with
LabVIEW
"
.
[11]
G. Cohen
and
G. Coon, "
T
heor
etical
consideration
of r
e
tard
ed
control",
T
r
ans
. As
m
e
,
vo
l. 75, pp. 82
7-834, 1953
.
[12]
J.F. Sm
uts, "Process contro
l for
pract
ition
e
rs",
L
e
ague C
ity, TX:
OptiControls Inc,
2011
.
[13]
J.G. Z
i
egl
e
r
and
N.B. Ni
chols,
"Optim
um
setting
s
for autom
a
tic
c
ontrolle
rs",
trans
. ASM
E
,
vol. 64
, 1942.
BIOGRAP
HI
ES OF
AUTH
ORS
Oladimeji Ibrahim
rec
e
ived
th
e B.
Eng degr
ee
in E
l
e
c
tr
i
cal
E
ngineer
ing from
Univers
i
t
y
o
f
Ilorin, Niger
i
a in 2005, Masters degree in
Instrumentatio
n and Control from
Glasgow
Caledonian University
, UK in 2009, and currently
pursuing a Ph.D. at Elect
rical and
Electronics
Engineering Department, Univ
ersiti Teknologi
PETRONAS (UTP), Malay
s
ia.
His research
inter
e
sts includ
e digital contro
l o
f
power electr
on
ics, switch
i
ng power c
onverters, and
renewab
l
e
energ
y
. He is a
m
e
m
b
er of The Institut
e
of
Measurem
ent a
nd Control (MInstMC, UK), ),
member of The
Nigerian
Society of Eng
i
neer
s (MNSE) and stud
en
t member of
the
IEEE.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE Vo
l. 5
,
N
o
. 5
,
O
c
tob
e
r
20
15
:
107
5
–
10
82
1
082
Suly
m
an A.
Y.
Amuda
is a Ph.
D
. holder in Electr
i
cal Engin
e
ering from University
of Ilorin
,
Nigeria
,
where
he als
o
e
a
rned h
i
s
M
.
Eng degr
ee
. His
res
ear
ch in
teres
t
is
in s
p
ee
ch proces
s
i
ng
,
renewable en
er
g
y
conv
ersion,
computer simula
tion and
contr
o
l. He is
a Fulbright Scholar,
res
earch
er,
and
pres
entl
y
le
cture
s
in the
Com
puter
Engin
eer
ing
Department,
Unive
r
si
ty
of Il
ori
n
,
Nigeria
.
He was
a visit
i
ng rese
a
r
cher
at th
e Ce
nter for Robust
Speech S
y
s
t
em
, Universit
y
of
Texas, D
a
llas, in
2009–2010 for nine
months. He is a reg
i
stered
p
r
ofessional eng
i
neer in Nig
e
ria
(COREN), m
e
m
b
er of
the
Niger
i
an S
o
ci
et
y
of En
gineers (NSE)
,
and IEEE Member.
Olatunji Obalo
w
u Mohammed
rece
ived B
.
Eng degr
ee i
n
Elec
tri
cal
a
nd Elec
troni
cs
Engineering fro
m Bay
e
ro Univer
sity
, Niger
i
a in 2010, Master
s degree in
Electrical
and
Electronics from Coventr
y
Un
iversity
, UK in
201
4.
His
r
e
s
earch
i
n
teres
t
s
includ
e
power s
y
s
t
em
control and
ren
e
wable energ
y
integr
ation and c
ontrol
.
He is
a m
e
m
b
er of t
h
e Institu
te of
Ele
c
tri
cal
and E
l
ec
tronics Engin
eers (MIEEE) a
nd m
e
m
b
er of
Nigerian S
o
ci
et
y of Engin
eers
(M
NS
E). He is
curent
l
y
a
le
ctur
er in
the d
e
part
m
e
nt of El
ectr
i
c
a
l and
El
ect
roni
cs
Engine
ering
,
Universit
y
of
Ilo
r
in, Nig
e
ri
a.
Kareem Aduagba Ganiy
u
r
e
ceiv
e
d th
e B
.
E
ng degre
e
in
M
echani
cal
En
gineer
ing from
University
of
Ilorin, Nig
e
ria in
2004, and
curr
en
tly
pursuing a Master'
s
d
e
gree at
Mechanical
Engineering Department, Feder
a
l University
o
f
Technolog
y
,
Minna Nigeria. His research
inter
e
sts includ
e corrosion an
d green inhibitors
, Materi
al sele
ction
and applications
and
renewabl
e m
a
t
e
rials
and
en
erg
y
s
ources
. He
is
a m
e
m
b
er of
Nigerian
S
o
cie
t
y of
Engin
eers
(MNSE), a m
e
mber, Nigeri
an Institution of Mec
h
anic
al Engin
e
e
r
s (MNIm
echE) and a m
e
m
b
er
of the coun
cil fo
r the Regu
lation
of
Engin
eerin
g in Nigeria (COR
EN)
Evaluation Warning : The document was created with Spire.PDF for Python.