TELKOM
NIKA
, Vol. 13, No. 4, Dece
mb
er 201
5, pp. 1281
~1
288
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v13i4.2113
1281
Re
cei
v
ed
Jun
e
4, 2015; Re
vised Septem
ber
10, 20
15;
Accept
ed Se
ptem
ber 26, 2015
Control System for Nutrient Solution of Nutrient Film
Technique Using Fuzzy Logic
Muhammad
Nau
f
al Ra
uf Ibrahim*, Mohamad Sola
hudin, Slamet Widod
o
Dep
a
rtment of Mecha
n
ica
l
an
d Bios
ystem E
ngi
neer
in
g,
Bogor Agric
u
ltura
l
Universit
y
, Jl. Ra
ya D
a
rmag
a
Kampus IPB D
a
rmag
a
Bog
o
r 166
80, W
e
st Java, Indo
nesi
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: naufalr
auf@
y
aho
o.co.id
A
b
st
r
a
ct
Nutrie
nt F
i
l
m
T
e
chn
i
qu
e (NF
T
)
is on
e of th
e
hy
dro
pon
ic sys
tems, w
h
ich t
h
e nutri
ent is c
i
r
c
ulate
d
throug
h root of the plants.
Ele
c
trical con
ducti
vity (EC) of nutrient so
l
u
tion o
n
hydro
pon
ic culture is a cruci
a
l
poi
nt that d
e
termines
the
g
r
ow
th
rate of
the pl
ants
and
qua
lity of
th
e
prod
ucts. T
he pur
pos
e of
this
researc
h
is to design the control system
for nutrient so
lution to m
a
intain the EC
of nutrient solution us
ing
Arduino.Fu
z
z
y
logic control
system
wa
s
used with EC error and volume
of nutrient solu
tion as
inputs.
Mode
l of toma
to cultivatio
n b
a
sed o
n
certai
n refe
renc
e was simulate
d w
i
th scale 1:1440 for cultiv
a
t
io
n
time. T
h
is
mo
del w
a
s
use
d
as dist
urba
nce
of nutri
ent
sol
u
ti
on
EC
d
ue n
u
t
ri
en
t up
ta
ke
. Th
e
re
su
l
t
o
f
observ
a
tion
sh
ow
ed that syst
em cou
l
d r
eac
h setp
oint
after 130
sec
onds
w
hen the
setp
oint ch
an
ges fr
o
m
1.7
mS/c
m to
1.6
mS/c
m a
n
d
2
0
7
seco
nds
from 1.6
mS
/
c
m t
o
1.
9
mS
/
c
m.
D
u
r
i
n
g
ob
s
e
r
v
at
io
n,
RS
M
E
valu
e of the system w
h
il
e st
eady state w
a
s 0.005
mS/c
m.
Ke
y
w
ords
: EC, fuz
z
y
,
sim
u
la
t
i
on, control system
Copy
right
©
2015 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
No
wad
a
ys
h
y
dropo
nic crop p
r
od
uctio
n
ha
s
si
gnificantly increase
d
in
re
ce
nt years
worl
dwi
de, a
s
itallows a
m
o
re
efficient
u
s
e
of wa
te
r a
nd
fertili
zers, as well as
a better cont
rol
of
c
limateandpes
t fac
t
ors
[1].
Total ionic c
o
nc
entratio
n
of
a n
u
trie
nt sol
u
tion d
e
termi
nes the
growt
h
,
developm
ent
andp
ro
du
ction of pl
ants [2]. Electri
c
al
condu
ctivity (EC) of n
u
t
rient solutio
n
in
hydrop
oni
cs
can
re
pre
s
e
n
t
s the total a
m
ount of
salt
in the n
u
trie
nt solutio
n
which i
s
al
so
a
indicator of the amou
nt of ions to the plants
[3]. Th
e ideal EC is spe
c
ific for
each crop an
d
depe
ndent o
n
enviro
n
me
ntal co
nditio
n
s [4], wh
i
c
h highe
r EC hinde
rs
nut
rient upta
k
e
by
increa
sing
o
s
motic
pre
ssure a
ndlo
w
e
r
EC may se
verely affect
plant health
and yield [5].
Therefore,
m
anag
ement
of
EC
sol
u
tion i
n
hyd
r
opo
ni
c
is
ce
rtainly n
e
ce
ssary
in
o
r
de
r to
keep
the
high yiel
d
of produ
ctio
n while
preventing th
e excessive
use
of n
u
trient
sol
u
tion.
Prope
rma
nag
ement of EC of the nutrient soluti
o
n
can p
r
ovide
and effective
tool to improve
vegetable
qu
ality [6]. In this
re
sea
r
ch,
mana
gem
en
t of EC
solut
i
on in
NF
T
system by u
s
i
n
g
c
o
n
t
ro
l s
y
s
t
em is
p
r
op
os
ed
.
There a
r
e
se
veral
repo
rts
and
pape
rs a
bout effort to
manag
e the
i
m
porta
nt pa
rameter of
hydrop
oni
cs includin
g
EC by using
automatic
control syste
m
. With the same type
of
hydrop
oni
cs,
Delya
[7]
re
ported
chili
p
eppe
r th
at is
cultivated
b
y
automati
c
control
of wa
ter
conte
n
t for
pl
ant medi
um
has better re
sult tha
n
cultivated by hu
man
cont
rol,
usin
g the
sim
p
le
ON-OFF m
e
thod in
real
condition. Simi
lar p
r
eviou
s
rese
arch i
s
al
so repo
rted
by Suprijadi [
8
],
whi
c
h the
control
syste
m
is u
s
in
g
volume
a
s
i
nput pa
ram
e
ter an
d PWM value a
s
output
para
m
eter
wi
th fuzzy lo
gic as
cont
rol m
e
ch
ani
cs. In t
he othe
r ha
n
d
, Domin
gue
s [9] develo
p
ed
automated system
control for
hydr
o
poni
c with combi
ned pa
ram
e
ter by EC, temperature, p
H
of
nutrient soluti
on.
The obj
ective
of this pap
er wa
s to devel
op t
he contro
l system to m
a
intain the E
C
value
of nutrie
n
t sol
u
tion in
NFT.
In orde
r to
measur
e
the perfo
rman
ce
of
the
sy
st
em
,
cont
r
o
l
sy
st
em
perfo
rman
ce
durin
g the full time of plant cultivati
on is ob
serve
d
[7, 8, 9]. Therefore, the ai
m o
f
this
wo
rk was to p
r
op
ose t
he
way to
ob
serve
th
e
pe
rforman
c
e
of t
he
cont
rol
sy
stem
of nut
ri
ent
solutio
n
ba
se
d on nut
rient
uptake of pla
n
t with ac
cel
e
rated
cultiva
t
ion time in o
b
se
rvation. T
he
control sy
ste
m
will be ea
sier to ob
serv
e and set if
the cultivation
time was a
c
cele
rated. If the
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No
. 4, Decem
b
e
r
2015 : 128
1 – 1288
1282
system
could
co
ntrol
EC i
n
a
c
celerated
cultivatio
n
time, then
the
system
al
so
can
control
E
C
i
n
the real time.
Fuzzy logi
c
controlle
r a
ppl
ication to
im
pr
ove th
e sy
stem p
e
rfo
r
m
ance ha
s
sh
own th
e
signifi
cant re
sult. Fuzzy lo
gic ba
se
d direct torq
ue co
ntrol sy
stem
wa
s implem
e
n
ted in indu
ct
ion
motor
and had improvement in
re
duci
ng torque
rippl
es, faster to
rque
response,
stability at
very
low spe
ed compa
r
ed
to conve
n
tional dire
ct
torq
ue
techn
o
logy [
10]. Prayitno
[11] develo
ped
autonom
ou
s
flight task for quad
copte
r
to repl
a
c
e m
anual
cont
rol
by human
and capa
ble
to
following the flight task instruct
ion. P
r
eviou
s
stud
y by Prathi
staya [12] set the accel
e
rated
cultivation ti
me to 90
da
ys to 2 day
s and u
s
in
g
PID co
ntrol.
This
pape
r a
l
so o
b
served
and
comp
ared th
e perfo
rma
n
ce of the sy
ste
m
if the cu
ltiv
ation time is
set by faste
r
and u
s
in
g fuzzy
logic controlle
r
2. Rese
arch
Metho
d
In this pape
r,
control syste
m
of nutrient
so
lutio
n
wa
s
built based o
n
nutrient u
p
take of
tomato. The
r
e are
several
wo
rks
about
nutrie
n
t upta
k
e and
E
C
setpoints of
n
u
trient solutio
n
in
hydrop
oni
cs
esp
e
ci
ally nutrient f
ilm tech
nique, on
e of them is
the report
s
abo
ut nutrient upta
k
e
of tomato o
n
nutrie
n
t film
tech
niqu
e [1
3]. This pap
e
r
explai
ned
the n
u
trient
u
p
take
of tom
a
to
durin
g 6
3
da
ys of tom
a
to
cultivation
se
parately
by e
a
ch
ion
in
th
e nut
rient
sol
u
tion a
nd
on
e of
them is Nitra
t
e or NO
3
-
. Nutrient sol
u
tio
n
called AB mix which i
s
two se
parate
nutrient sto
c
ks
solutio
n
nam
ely stock
solu
tion of A and
B [14].
Usin
g ion u
p
take
informatio
n a
nd contain
ed
ion
in AB mix, model of nutrie
n
t uptake by
certai
n ion ca
n be set.
Based o
n
the
model of nut
rient upta
k
e,
contro
l sy
ste
m
wa
s built with co
nsi
deration of
the sy
stem
can
control the
EC o
r
n
o
t while nut
rient
u
p
take
is
occu
rre
d. Co
ntrol
system i
s
te
st
ed
by simulatio
n
with sh
orte
ne
d cultivation a
nd init
ial nutri
ent volume. Nutrie
nt upta
k
e on
cultivati
o
n
results d
e
cli
n
ation of EC value which can be re
p
r
e
s
ented by ad
d
i
ng wate
r to
nutrient tan
k
or
called EC di
sturbance.
With this
, control
system can
be evaluated
by observing
its capability to
control EC of nutrient soluti
on.
2.1. EC Distu
r
bance
EC distu
r
ba
n
c
e will b
e
ap
plied by addi
ng wate
r to the nutrie
n
t solution an
d mimics the
EC de
clini
n
g
due
to nut
ri
ent upta
k
e
d
u
ring
60
day
s of tom
a
to
cultivation. Th
e nutri
ent u
p
take
rate is ap
plie
d based on the previo
us
resea
r
ch [13]. Based on the refe
ren
c
e
[13], model of
tomato cultiv
ation is set. The set mo
del
has a tom
a
to plant with initi
a
l nutrient
sol
u
tion 200 liters.
This mo
del u
s
ed io
n upta
k
e NO
3
-
as n
u
trient upta
k
e a
nd interp
ret it to EC value. Figure 1 sh
o
w
s
the EC decli
n
i
ng rate of to
mato cultivation due n
u
trie
nt uptake.
Figure 1
.
Model of EC decling rate of tomato cultivation due n
u
trie
nt uptake
Thus, thi
s
m
odel of E
C
d
e
clini
ng rate
of tomato cul
t
ivation wa
s
simulate
d
with scale
1:1440
(from
60 days to 6
0
minute
s
) fo
r cultivation ti
me and 1:2
5
(4 liters to 10
0 liters) for in
tial
nutrient volu
me. An automated pump
was u
s
ed to mimics th
e EC declini
ng rate at the
simulatio
n
. T
h
is
sim
u
lation
used
thre
e
setpoi
nts
that depe
nd on
th
e
cultivation pha
se ba
sed
on
the referen
c
e
[13]. Figure 2 sho
w
s setp
oint of EC during simul
a
tion
.
1400
1500
1600
1700
1800
0
1
02
03
04
05
06
0
EC(µS/cm)
Cultiv
ation time
(day
)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Control Syste
m
for Nutrient
Solution of Nutrient Film
Tech
niqu
e Using …
(M.N.R.
Ibrahim)
1283
Figure 2. EC setpoi
nt on tomatocultivatio
n
2.2. Contr
o
l Sy
stem Hard
w
a
r
e
Control
syste
m
for this nu
trient
solutio
n
of
NFT
u
s
ed Arduin
o
microcontroll
er
and
2
s
e
ns
or
s
.
T
h
e s
e
n
s
o
r
s
is
co
n
s
is
te
d o
f
EC
s
e
ns
or
Atla
s Sci
entific K
1
.0 to me
asu
r
e the
EC value
of nutrie
n
t so
lution a
nd ult
r
asoni
c
sen
s
or
HC-SR
04 to
mea
s
u
r
e
t
he
h
e
ight of nutrient sol
u
tion,
whi
c
h later
converter into v
o
lume in calculations.
For
output
si
gnal, Arduin
o
is
co
nne
cted
to a fo
ur cha
nnel
s relay to
co
ntrol t
he a
c
tuator.
The co
ntrol
system hardware is
sho
w
n
by Figure 3.
Figure 3. Hardwa
re of co
ntrol syst
em
Four A
C
pu
mps that u
s
e
d
as
actu
ators ar
e two
nut
rient pu
mp
s (for A and B
nutrient
), a
water pu
mp
and di
stu
r
ba
nce
pum
p. T
he nut
rient
p
u
mps an
d
water p
u
mp
are u
s
ed fo
r
co
ntrol
purp
o
se
whil
e the
distu
r
b
ance p
u
mp
is u
s
e
d
fo
r
repre
s
e
n
ting t
he E
C
d
e
clin
ing
rate
due
to
nutrient upta
k
e by adding t
he wate
r. Fig
u
re 4
sho
w
s
schemati
c
dia
g
ram of control system.
Figure 4. Sch
e
matic dia
g
ra
m of control
system
1.5
1.6
1.7
1.8
1.9
2
0123
456789
EC Setpoint (mS/cm)
n
th
W
eek
EC
Ultrason
ic
Ar
dui
n
o
Rela
y
PC
Pum
p
s
I
npu
t
Out
put
Mo
n
itoring
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No
. 4, Decem
b
e
r
2015 : 128
1 – 1288
1284
2.3. Fuzzy
Logic Con
t
rol
l
er
The de
sign
ed
fuzzy logic
controlle
r ha
s two
input vari
able
s
and on
e output varia
b
le. The
input variable
s
are e
rro
r of EC (EEC) an
d volume of
nutrient sol
u
tio
n
(V).Error of EC is acq
u
ire
d
by differe
nce
between
EC setp
oint
(
EC
s
) an
d a
c
tual
EC
(
EC
a
)
.
V
o
lume
ofnutri
ent solution
i
s
acq
u
ire
d
by
height of
nutrient sol
u
tion
(h) time
s
a
r
ea
of nutrie
n
t solution tan
k
(A).The fo
rmul
as
are:
EEC =
EC
a
EC
s
(
1
)
V
=
A
*
h
(2)
Ran
ge of membershi
p
functionfor e
r
ror
of EC
and volume of nutri
ent solution
are [-0.4,
+0.4] mS/cm
and [3.5,
9.5] liters. Maxim
u
m
capa
city
of
nutri
ent sol
u
tion
tan
k
on simulatio
n
i
s
10
liters.
The output variabl
e is pu
mp activation
time with membe
r
ship function rang
e
[-11.8
+11.8]
se
con
d
s. All fuzzy
membe
r
ship
is se
t by tri
a
l and
error.
Figure 5, 6,
and 7
sh
o
w
s
membe
r
ship functio
n
error
of EC and volume of nutrie
n
t solution.
Figure 5. Membershi
p
fun
c
tion of error
of EC
Figure 6. Membershi
p
fun
c
tion of volu
me
E M
F
LN
N
Z
P
LP
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Control Syste
m
for Nutrient
Solution of Nutrient Film
Tech
niqu
e Using …
(M.N.R.
Ibrahim)
1285
Figure 7. Membershi
p
fun
c
tion of pump
activation time
The impli
c
ati
on rule
s that
use
d
in fu
zzy rule
s a
r
e IF
error E
C
‘x m
S
/cm’ AND volume ‘y
liters’ T
H
EN
pump a
c
tivation ‘z second
s’ with
Mam
d
ani infere
nce. Approp
riate
pump a
c
tivation
time is expe
cted to be the
output that can maintai
n
the EC of nut
rient solutio
n
. Fuzzy rule
ba
se
for pump a
c
ti
vation time is sho
w
n by Ta
ble 1.
Table 1. Fu
zzy rule base for cont
rol sy
stem
Fuz
z
y
Rul
e
Volume of Nutr
ient Solution (
liter
)
E M F
Error o
f
EC
(mS/c
m
)
LN SAB
MAB
LAB
N SAB
SAB
MAB
Z Z
Z
Z
P SW
SW
MW
LP SW
MW LW
Defu
zzifi
catio
n
is ca
rrie
d
out by using cent
er of gravi
t
y method [15] for determi
ning the
cri
s
p value of
pump activat
i
on time in se
con
d
s.
Cente
r
of gravity (Z) can b
e
cal
c
ulated by:
Z
=
∑
∑
(
3
)
Data a
c
qui
siti
on by both E
C
an
d ultra
s
onic
sen
s
o
r
s
are p
r
o
c
e
s
se
d per
se
co
nd
durin
g
simulatio
n
, while fuzzy lo
gic
cal
c
ulatio
n and
pum
p
activation
are processe
d
per
35
th
s
e
ns
or
readi
ng. Thi
s
is due to the
EC se
nsor ha
rdware, whi
c
h is stabl
e after 15
-25
th
se
nso
r
re
adin
g
and
also mixing p
r
ocess until n
u
trient sol
u
tio
n
is uniform e
noug
h.
3. Results a
nd Discu
ssi
on
3.1. EC Profile During Ob
serv
a
tion
Control sy
ste
m
is tested
and ob
se
rve
d
durin
g 60
minutes
with
3 different setpoints.
Control me
ch
anismth
at used on
syste
m
is pum
p O
N
-OFF
control with pu
mp
activation time i
s
based on fu
zzy inferen
c
e.
Figure 12 sho
w
s EC
pr
ofile
of nutrient so
lution duri
ng
control.
LAB
MAB
SAB
Z
SW
MW
LW
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No
. 4, Decem
b
e
r
2015 : 128
1 – 1288
1286
Figure 9. EC profile of nutri
ent solutio
n
d
u
ring
cont
rol
Based o
n
Fig
u
re 12, the system can m
a
intain the EC duri
ng ste
a
d
y state con
d
i
tion and
control the E
C
of n
u
trient
solutio
n
to t
he
setpoint whe
n
the se
tpoints cha
n
ge.
Tho
ugh
t
he
system
can
maintain the
EC, oversho
o
t and un
dersh
o
o
t is still
occurre
d
du
ri
ng co
ntrol. T
h
is is
due to
the
sy
stem
used
O
N
-OFF
co
ntrol me
ch
ani
sm
, as the
on
e of
ch
ara
c
te
ristic of
ON-OFF
control me
ch
anism
is
occurri
ng ove
r
sh
oot and
un
d
e
r
sh
oot d
u
rin
g
cont
rol [16].
Figure 13
sh
ows
error value d
u
ring
cont
rol.
Figure 10. Error of EC valu
e durin
g co
ntrol
Two
errors a
t
35
th
and
50
th
minutes
are
cont
rol
syst
em l
ag
wh
en
setpoi
nt ch
a
nge
s is
occurre
d
. Based
on Fig
u
re 13, the
system
can
rea
c
h the
setp
oi
nt within 1
3
0
se
con
d
s
wh
en
setpoi
nt ch
an
ges f
r
om
1.7
mS/cm to 1.6
mS/c
m an
d
207
se
con
d
s
whe
n
setpoin
t
chan
ge
s fro
m
1.6 mS/cm
to
1.9 mS/cm.
Du
ri
ng stea
d
y
state con
d
ition
,
the
high
e
s
t po
sitive e
r
ror i
s
0.02 mS
/cm
and high
est n
egative error
is -0.01
8
mS/cm. RSME
value of the system
co
ntrol i
s
0.005 mS/cm
durin
g the ste
ady state.
By evaluation of the performance in the
contro
l
sy
st
e
m
,
t
h
is cont
r
o
l sy
st
em ha
d les
s
e
r
error
comp
ared to the pre
v
ious work b
y
Prathi
sthay
a [12], which the cont
rol sy
stem ha
d RS
ME
0.023 mS/cm. This also
means eve
n
with fast
er accel
e
rated
cultivation time, fuzzy logic
controlle
r wa
s abl
e to b
e
more
accu
rat
e
than PI
D control m
e
cha
n
ism.
Wide
r range
of setp
oint
and fa
ster re
spo
n
se
comp
ared
to
Domi
ngue
s [9], all
o
w th
e
syste
m
to be
ap
pl
icabl
e in
oth
e
r
comm
odity of
hydro
poni
cs. The l
e
sse
r
e
rro
r al
so
allo
ws the
syste
m
to save the
more n
u
trien
t
up
durin
g the co
ntrol.
1.4
1.5
1.6
1.7
1.8
1.9
2
0
1
02
0
3
04
0
5
06
0
EC (mS/cm)
Observ
a
tion tim
e
(m
inutes)
Reference
Controlled
Uncontrolled
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0
1
02
03
04
05
0
6
0
Error
of EC (mS/cm)
Observ
a
tion time
(minutes)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Control Syste
m
for Nutrient
Solution of Nutrient Film
Tech
niqu
e Using …
(M.N.R.
Ibrahim)
1287
3.2. Scaling up to Re
al Dimension of
Con
t
rol Sy
st
em
Thoug
h the simulation is a
n
accele
rated
cult
ivation time (from 6
0
days to 60 m
i
nutes),
the control
system ma
na
ge to
maintai
n
an
d
co
ntro
l EC
of n
u
tri
ent solution.
With a
c
cepta
b
le
error i
n
result, the syste
m
is
supp
osed t
o
be
appli
c
a
b
le in
real
sit
uation. Real
situation
of the
model
ha
s 6
0
day
s of
cul
t
ivation and
100 lite
r
s of
nutrient ta
nk
cap
a
city. In
orde
r to tu
ne
the
cont
r
o
l sy
st
e
m
in sim
u
lat
i
on t
o
c
ont
rol
sy
st
em f
o
r re
al situatio
n, scale
up i
s
set. Scaling u
p
can
be don
e by modifying fuzzy membershi
p
of pump activ
a
tion time or flow rate of pu
mp.
Ho
wever, m
o
difying those
para
m
eters
a
r
bitra
r
y can
result
an u
nde
sira
ble im
pa
ct on the
control
syste
m
. Modifying
only fuzzy m
e
mbe
r
ship
of pump
a
c
tivation time
re
sul
t
in lon
g
settlin
g
time wh
en
se
tpoint is chan
ged, b
u
t this
modificati
o
n
has sta
b
le n
u
trient solution
mixing p
r
o
c
e
ss
and le
sser n
o
ise
s
while control. In the
other
ha
nd,
modifying only flow rate
of pump ca
use
unsta
ble nutri
ent solution
mixing pro
c
e
s
ses tho
ugh i
t
has faster
settling time.
This p
r
oble
m
will
cau
s
e u
nne
cessary a
dditi
on of wate
r o
r
nutri
ent sol
u
tion in control pro
c
e
s
s. It also o
c
cu
re
d in
the simulatio
n
, describ
ed
by fuzzy inference of
pum
p activation while control. Figure 11
sh
ows
the fuz
z
y
inferenc
e s
t
atus
.
Figure 11. Fu
zzy infe
ren
c
e
stat
us du
ring
control
syste
m
Duri
ng the
st
eady state
conditi
on,
the nutrient pum
p
is ex
pecte
d
to be the
on
ly pump
that open
ed
becau
se the
disturban
ce i
s
only E
C
d
e
clin
e du
e water a
ddition.
Sensor
rea
d
i
ng
durin
g in
com
p
lete mixing
pro
c
e
ss
co
ul
d give an ina
c
curate
eithe
r
nutrie
n
t or
water
additio
n
to
achi
eve the setpoint, and u
nde
si
re
d wa
ste of nutrient usa
ge.
Con
s
id
erin
g the re
al ap
plication, Pra
s
ta
ya [
17] rep
o
rt
ed at the
red
spin
ach cultivation in
hydrop
oni
c fa
rm of
Joy Fa
rm in Indo
ne
si
a, ther
e
is
an
occu
rre
nce t
hat within
an
hour EC val
u
e
of nutrient
so
lution drop b
y
0.1 mS/cm. Based
on
Fi
gure
9, the control
system
can m
ana
ge
to
control the differen
c
e of 0.1 mS/cm by 130 se
co
nd. Fr
om this
situ
ation, it is safe to chang
e only
the fuzzy m
e
mbe
r
ship
s
whi
c
h
re
sult
in long
er
set
t
ling time bu
t stable
mixing p
r
o
c
e
ss
and
pr
ec
ise
fuzzy dec
is
ion. N
e
w
fuz
z
y
member
ship
for
pump activation time c
an be s
e
t
b
y
cal
c
ulate
s
th
e fuzzy mem
bership time
s 25. T
hough
so, a faster settling time and less no
ise
setting is
still possibl
e to set if ce
rtain model of nutrient
solution mixi
ng process is known.
4. Conclusio
n
Control syste
m
of nutrient solutio
n
is bu
ilt
by using Arduin
o
integrated with fuzzy logic
controlle
r. Model of tomat
o
cultivation
based o
n
ce
rtain refe
ren
c
e from the
previous
re
sea
r
ch
wa
s simul
a
te
d
with
scale 1:1440
from
60
d
a
ys
to
6
0
min
u
tes of
cultivation tim
e
. Setpoint
s t
hat
set for contro
l are
1.7 mS/
c
m, 1.6 mS/
c
m, and
1.
9
mS/cm. Actu
ator i
s
contro
lled by O
N
-O
FF
mech
ani
sm with time activation is de
cide
d by fuzzy log
i
c co
ntrolle
r.
Based
on E
C
value du
ring
simulatio
n
, sy
stem can
rea
c
h the
de
sire
d setp
oint wit
h
in 13
0
se
con
d
s
wh
e
n
setp
oint
ch
ange
s from 1
.
7 mS/cm
to 1.6
mS/cm
a
nd
20
7 se
con
d
s whe
n
setp
oint
cha
nge
s from
1.6 mS/cm
to 1.9 mS
/
c
m. Du
ring
stea
d
y
state
condit
i
on
,
the hi
ghe
st po
sitive e
r
ror
-4
-3
-2
-1
0
1
2
0
1
02
03
04
05
06
0
Fuzzy
inference
status
Sim
u
lation tim
e
(m
inutes)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No
. 4, Decem
b
e
r
2015 : 128
1 – 1288
1288
is 0.02 mS/
c
m and hi
ghe
st negative error is
-0.0
18
mS/cm. RSM
E
value of du
ring
cont
rol
while
in stea
dy sta
t
e is 0.00
5
mS/cm.
Overall, the co
ntrol system
ca
n maintain
a
nd control E
C
of
nutrient sol
u
tion
b
a
sed o
n
de
cid
ed setpoints.
Co
mpared to
PID control m
e
ch
ani
sm, fu
zzy
control had le
ss e
r
ror.
In ord
e
r to
ap
ply this
cont
rol sy
stem for
real
appli
c
ati
on, scal
e up
i
s
set. Set sca
l
e up i
s
based on rati
o
of
mo
del a
nd simulatio
n
initial
nut
rie
n
t sol
u
tion. T
w
o
paramete
r
s that
can
b
e
scaled
up
is fuzzy mem
bership
of p
u
mp
activati
on
time an
d flo
w
rate
of pum
p. Modifying
o
n
ly
fuzzy mem
b
e
r
shi
p
of pum
p activation ti
me re
sult in l
ong settling time wh
en setpoint is
chan
ged
but it has sta
b
le nutrie
n
t solution mixin
g
pro
c
e
s
s an
d lesse
r
noi
ses while cont
rol. Modifying
only
flow rate of
pump
ca
use
unsta
ble n
u
trient solu
tion
mixing proce
s
ses but it h
a
s fa
ster settling
time. Base
d o
n
real
pro
b
le
m, it is
su
gge
sted th
at
cha
nging
fuzzy
membe
r
ship
of pum
p a
c
tivatio
n
time is
safe i
n
order to a
c
hieve p
r
e
c
ise
fuzzy
de
cisi
on an
d le
sse
r
noi
se
whil
e
control. In future
studie
s
, mo
re
element of n
u
trient solutio
n
ion upta
k
e
sho
u
ld be i
n
cluded in
simu
lation to ma
ke
the system m
o
re a
c
curate according to
the actu
al nutrient uptake of the plant.
Referen
ces
[1]
T
r
ejo
T
é
llez
L
I, Gómez-Merino FC.
Editors
. Nutrie
nt Soluti
ons for H
y
dr
op
onic S
y
stems,
H
y
dro
pon
ic
s
– A Standar
d Method
olo
g
y
for Plant Bio
l
o
g
i
c
al Res
earch
e
s
. Rijeka: Cro
a
t
ia. 2012.
[2]
Steiner AA. A
Univers
a
l M
e
th
od for Pre
par
in
g Nutr
i
ent So
lu
tions of a
Cert
ain
Desir
ed C
o
mpositi
on.
Plant an
d Soi
l
. 196
1; 15(2): 13
4-15
4.
[3]
Nemal
i
KS, Van Iersel MW
. Light Intensi
t
y
an
d Fertiliz
er Conc
entrati
on: I. Estimating Optima
l
Fertilizer Conc
entratio
n
from Wate
r-Use Effi
cienc
y
of Wa
x
Bego
nia.
Hort Scienc
e
. 20
04;
39(6):
128
7-
129
2.
[4]
Sonn
evel
d C, Voogt W
.
Plant Nutrition of
Green
ho
use Cro
p
s. Ne
w
York: Sprin
ger.20
09.
[5]
Samarak
oon
UC, W
eeras
in
ghe
PA, W
eer
akkod
y
AP.
Effect of Electric
al C
o
n
ductivit
y
[EC] of th
e
Nutrie
nt Soluti
on on Nutri
e
n
t
Uptake, Grow
t
h
a
nd
Yie
l
d
of Leaf Lettuce (Lactuca
s
a
tiva L.)
i
n
Stationar
y C
u
lt
ure.
T
r
opica
l Agricult
ural R
e
s
earch
. 20
06; 1
8
(1): 13-2
1
.
[6]
Gruda
N. Do
S
o
ill
es C
u
ltur
e S
y
stems
hav
e a
n
Influ
enc
e o
n
Product Qu
alit
y
of Ve
geta
b
le
s?
Jo
urna
l o
f
Appl
ied B
o
tany
and F
ood Qu
a
lity
. 2009; 8
2
(2
): 141-14
7.
[7]
Del
y
a, Buti. D
e
sig
n
of Ebb
and F
l
o
w
Auto
mati
c H
y
dro
p
o
n
ic S
y
stem for
Chil
li Pe
pper
Cultivati
o
n
.
Jurna
l
T
e
knik
Pertani
an.
20
1
5
; 3(3).
[8]
Suprij
ad
i, Nur
a
ini
N, Yus
u
f M
.
Sistem ko
nt
rol n
u
trisi
hi
dro
pon
ik d
e
n
gan
meng
gun
aka
n
logik
a
fu
z
z
y
.
Jurna
l
Otomas
i
, Kontrol, dan I
n
strumentas
i
. 200
9; 1(1):31-
35.
[9]
Domin
g
u
e
s DS
, Camara CAP
Ni
xdorf SL, T
a
kahas
hi HW
. Automated s
y
ste
m
devel
ope
d to control p
H
and
conc
entrat
i
on
of n
u
trie
nt sol
u
tion
ev
alu
a
ted in h
y
dr
op
onic
lettuce
pr
oducti
on.
C
o
mputers an
d
Electron
ics in
Agricult
ure
. 20
12; 84: 53
–61
[10]
Ming D, Bi
ng-j
i
e W
,
Hail
ing
S, Yong-q
i
T
.
St
ud
y
of an I
m
prove
d
F
u
zz
y D
i
rect T
o
rque Co
ntrol o
f
Inductio
n
Moto
r.
TEL
K
OMNIK
A
. 2013; 11(
4): 691-6
98.
[11]
Pra
y
itno A, Ind
r
a
w
ati V,
Utom
o G.
T
r
ajectory T
r
acking of
AR. Drone Qu
adrotor Us
in
g F
u
zz
y
L
ogi
c
Contro
ller.
TELKOMNIKA
. 2014; 12(4): 8
19-
828.
[12]
Prathistha
ya,
ME,
Li
ya
ntono,
Sol
ahu
di
n M.
Ranc
an
g Ba
n
gun
Siste
m
K
e
nda
li Pr
oporti
o
nal I
n
tegra
l
-
Deritativ
e
pa
d
a
Pen
g
e
nda
lia
n Kons
entrasi
Laruta
n
Nutris
i
Hidro
po
nik
. P
r
ocee
din
g
s of
the Nati
ona
l
Semin
a
r FT
IP
UNPAD – PER
T
E
T
A
– HIPI 2014. Jatin
a
n
g
o
r
. 2014: 37
6-38
4
[13]
W
eerakko
d
y
W
AP, W
a
kui K,
Nuka
ya
A. Pl
a
n
t nutri
ent u
p
ta
ke in
recirc
ulati
on cu
lture
of to
mato u
nde
r
gro
w
th sta
ge
base
d
el
ectric
al co
nductiv
i
t
y
adj
ustment.
Journ
a
l of N
a
ti
ona
l Scie
nce
of Sri La
nka
.
201
1; 39(2):1
3
9
-14
7
.
[14]
Derma
w
a
t
i
. Su
bstitusi har
a minera
l
org
ani
k terhada
p in
o
r
gan
ik untuk p
r
oduks
i
tanam
an pakc
h
o
y
(
Brassica rapa
L
.
) secara hi
dr
opo
nik. T
hesis.
Bogor: Postgr
adu
ate IPB. 2006.
[15]
Kusuma
de
w
i
S
,
Purnomo H.
Aplik
asi L
ogik
a
F
u
zz
y
u
n
tuk P
end
uku
ng Ke
p
u
tusan. Yo
g
y
a
k
arta: Graha
Ilmu. 2004.
[16]
Sieg
enth
a
ler J
.
Modern
H
y
dr
onic
Heati
ng:
F
o
r Resi
denti
a
l an
d Li
ght C
o
mmercia
l Bu
il
din
g
s. Ne
w
York: Ceng
ag
e
Learn
i
ng. 2
0
1
1
:335.
[17]
Prast
y
a, And
i
k
a
. Anal
isis K
e
l
a
yaka
n d
an
O
p
timasi
Usah
a
Budi
da
ya
Ba
yam Mera
h d
a
n
Kangk
un
g
Hidro
po
nik D
e
nga
n Sistem N
F
T
(Nutrient F
ilm T
e
chniqu
e)
di PT
. Jo
y
F
a
r
m
, Depok. U
n
dergr
adu
ate
thesis. Bogor: Institut Pertani
a
n
Bogor. 2
009.
Evaluation Warning : The document was created with Spire.PDF for Python.