Indonesian J
ournal of Ele
c
trical Engin
eering and
Computer Sci
e
nce
Vol. 2, No. 2,
May 2016, pp
. 268 ~ 274
DOI: 10.115
9
1
/ijeecs.v2.i2.pp26
8-2
7
4
268
Re
cei
v
ed O
c
t
ober 1
1
, 201
5; Revi
se
d Febru
a
ry 20, 2
016; Accepte
d
March 9, 20
16
Arithmatic Coding Based Approach for Power System
Parameter Data Compression
Subhra J. Sarkar*
1
, Nab
e
ndu Kr. Sark
ar
2
, Trisha
y
a
n Dutta
3
, Panchalika De
y
4
,
Aindrila Mukherjee
5
1,3-
5
Dept. of EE,
T
e
chno Indi
a Batana
gar, B7-
360 /
Ne
w
,
W
a
rd No. 30, Putkhali, Ma
hesht
al
a,
Kolkata – 7
0
0
1
41, W
e
st Beng
al, India
2
Dept. of EE, H
a
ldi
a
Institute o
f
T
e
chnolo
g
y
, I
.
C.A.R.E Comple
x, H.I.T
Campus, Hatib
e
ri
a,
PO HIT
,
District Midnap
ore (E
), Ha
ldi
a
, W
e
st Benga
l 72
165
7, India
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: subhro
89@
g
m
ail.com1, ns
a
r
kares@re
diffmail.com,
trisha
ya
nduttak
o
l@gm
ail.c
o
m, panch
a
lik
a.de
y9
5@
gmai
l.co
m, aindril
amuk
herj
ee2
4@gm
ail.com
A
b
st
r
a
ct
For stable p
o
w
e
r system o
p
e
r
ation, vari
ous
system
p
a
ra
meters like vo
lta
ge, current, freque
ncy,
active a
nd r
e
a
c
tive pow
er et
c. are mon
i
tor
ed at
a r
e
g
u
la
r basis. T
hes
e
data ar
e e
i
th
er stored
in th
e
database
of the system
or
trans
m
i
tted to t
he
m
o
nito
r
i
ng
station thr
ough SCAD
A. If these
data c
an
be
compress
ed
b
y
suita
b
le
d
a
ta
co
mpr
e
ssio
n
t
e
chn
i
qu
es, the
r
e w
ill
be
red
u
c
ed
memory r
e
quir
e
ment
as
w
e
ll
as low
e
r e
ner
gy cons
u
m
ptio
n for tr
ans
mitti
ng the
data. In this p
aper, a
n
al
gorith
m
ba
sed o
n
Arith
m
eti
c
Codi
ng
is d
e
ve
lop
ed for c
o
mp
ressin
g
an
d d
e
c
ompressi
ng s
u
ch p
a
ra
met
e
r
s
in MAT
L
AB
envir
on
me
nt. T
h
e
compressi
on r
a
tio of the al
go
rithm cl
early i
n
dicates the
effectiveness of th
e alg
o
rith
m.
Ke
y
w
ords
: D
a
ta Co
mpr
e
ss
ion, Arith
m
eti
c
Codi
ng,
Co
mpr
e
sse
d dat
a, Pow
e
r System, Para
me
ter
mo
nitori
ng
Copy
right
©
2016 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
In mod
e
rn
p
o
we
r
system,
a n
u
mb
er
of alter
nators
conne
cted
in
same
o
r
oth
e
r
pla
n
ts
operate in
pa
rallel to
meet
the loa
d
de
mand
of
the system.
G
r
id implies
the system
comp
ri
sing
of conn
ectin
g
alternato
r
s of all power plant
s in
parallel
wh
ich can be
achieve
d
a
fter
synchro
n
izi
n
g
them
with
e
a
ch
othe
r
or
with
the bus
bar. The voltage, frequ
en
cy, phase a
n
g
le
and pha
se se
quen
ce
of
in
coming altern
a
t
or
an
d
b
u
s
b
a
r sho
u
ld be same. Du
e
to
som
e
sud
d
e
n
transi
ents o
r
some fault
s
in the system,
the sy
stem tend
s to beco
m
e unsta
ble resultin
g devia
tion
from stan
da
rd voltage an
d frequ
ency
value. So,
co
ntinuou
s mo
nitoring of th
e system volt
age
and freq
uen
cy is very important. The continuo
us mo
nitoring of th
e system pa
rameters is d
one
by employin
g suitable hi
ghly sophi
sti
c
ated
comm
unication system built in
within the po
wer
system a
nd u
s
ing a
system
called SCA
D
A (Supervi
sory Control an
d
Data Acqui
si
tion) [1-7].
SCADA sy
ste
m
is used for supe
rvisio
n
and c
ontrol
of remote field
device
s
. The
desi
r
ed
power
syste
m
pa
ramete
rs a
r
e
colle
cted at
re
m
o
te
end
u
s
ing
SCADA
usin
g some
suitable
meters, sen
s
ors,
process
equipm
ent et
c. Th
ese
coll
ected
data
re
ading
s
are
th
en transmit to
the
control
cent
re with
suita
b
l
e wi
red
co
mmuni
cation
ch
annel
in
cl
uding
po
we
r line, teleg
r
a
phic
cabl
e, etc. or throug
h or
wireless commu
nicati
o
n
. SCADA has a
wid
e
appli
c
ation i
n
the indu
stry.
It is extre
m
el
y popul
ar in
electri
c
ity util
ities a
s
it
en
able
s
the
re
mote o
peration &
co
ntrol
of
sub
s
tation
s
and ge
ne
rati
ng station
s
.
Processin
g
o
f
the colle
cted data mig
h
t requi
re
some
openi
ng a
nd
clo
s
ing
of ci
rcuit b
r
ea
ke
rs or valve
s
p
a
r
ticula
rly whe
n
sy
stem p
a
rameters ex
ce
ed
the pred
efine
d
threshold
s
. In SCADA archite
c
ture,
th
e role of Rem
o
te terminal u
n
its (RTUs) a
r
e
to conne
ct sensors in the syst
em, tran
smit the acq
u
ired d
a
ta to the supervi
sory system a
nd
receiving in
struction
s
fro
m
sup
e
rvisory sy
stem. O
n
the othe
r hand, Pro
g
r
amma
ble lo
gic
controlle
rs
(P
LCs) have se
nso
r
s c
onn
ected to it b
u
t d
o
not
have
a
n
y inbuilt tel
e
metry ha
rd
wa
re.
Thus
PLCs can repl
ace RTUs
due
to it
s economy, v
e
rsatility, fl
exibility and configurability. T
h
e
comm
uni
cati
on system b
e
twee
n
field device
s
and
control cent
re
might be wire
d
o
r
wire
less.
Control centre com
p
ri
se
s of Huma
n-Machi
ne Inte
rface
(HMI),
Data Hi
sto
r
ian, co
ntrol/d
a
ta
acq
u
isitio
n server, comm
unication ro
u
t
er etc.
HMI
pre
s
ent
s the pro
c
e
s
sed
data to hu
man
operator
for monitori
ng an
d
interactio
n. Data Hi
st
oria
n ha
s va
riou
s data, eve
n
ts
and
alarms in
a
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 2, May 2016 : 268 –
274
269
databa
se
whi
c
h i
s
HMI
acce
ssi
ble. Bot
h
HMI an
d
Historia
n
are
the
client
s fo
r data
a
c
qui
sition
serve
r
which allows them to acce
ss a
n
y data fr
om fiel
d device
s
u
s
i
ng suita
b
le protocol [8].
Figure 1. SCADA Arc
h
itecture
In
[9
], c
o
mp
ar
is
on
o
f
p
e
r
f
or
ma
nc
e o
f
var
i
o
u
s
lo
ss
les
s
a
n
d
los
s
y
po
w
e
r
s
y
s
t
e
m
r
e
la
te
d
data comp
re
ssion i
s
avail
a
ble. The
r
e a
r
e few
wo
rks f
o
r
comp
re
ssing po
we
r
system data u
s
i
n
g
wavelet transform [10-13]. Power system data co
mpression and its implement
a
tion will be
of
greate
r
si
gnifi
can
c
e while approa
chin
g
towa
rd
s
sm
ar
t
grid [14, 15].
In the prop
o
s
ed m
e
thod,
an
approa
ch to
redu
ce th
e n
u
m
ber of bit
s
o
f
the po
we
r
system mo
nito
ring i
n
form
ation d
e
velop
e
d
in
MATLAB environm
ent and
tested offlin
e. At t
he encoding e
nd, p
o
we
r sy
stem para
m
eters sa
y
voltage, freq
uen
cy and lo
ad po
we
r factor a
r
e
com
p
re
ssed to f
o
rm a
cha
r
a
c
ter
strin
g
u
s
ing
arithmeti
c
co
ding ba
se
d a
ppro
a
ch. At the de
codi
ng
end, the co
mpre
ssed
ch
ara
c
ter
strin
g
is
decomp
r
e
s
se
d to obtain th
e necessa
ry
para
m
eters.
As there i
s
a
redu
ction i
n
the num
ber of
bits
to be tran
sm
itted or sto
r
a
ge, there
mu
st be
a
su
bseque
nt red
u
ction in ene
rg
y requi
red fo
r
transmissio
n and memo
ry
requi
reme
nt. Beside
s tha
t, there is an inherent e
n
cryptio
n
in the
prop
osed alg
o
rithm which provide
s
data
security.
2. Data
Com
pression
Data comp
re
ssi
on is the
pro
c
e
ss of re
duci
ng the n
u
mbe
r
of bits (or bytes) in
orde
r to
have e
a
si
er
stora
g
e
or transmi
ssion
o
f
the d
a
ta
. Data comp
re
ssion
can
al
so
lead to
wa
rd
s the
data en
crypti
on thereby p
r
oviding th
e
necess
a
ry data s
e
c
u
rity. In
other
words, it is al
so
be
defined a
s
the recombin
ation the bits (or bytes)
for makin
g
the smaller an
d compa
c
t form of
data by eli
m
inating the
id
entical
data
bits o
r
c
ontin
uou
sly re
cu
rring data. At t
he poi
n
t of d
a
ta
resto
r
atio
n, some de
com
p
ression meth
o
d
must be de
veloped for d
e
co
ding the
compresse
d
d
a
ta
so that the o
r
iginal
data can be
extra
c
t
ed. For the
a
pplication
s
d
ealing b
u
lk in
formation, da
ta
comp
re
ssion
is extremely importa
nt as significa
nt
saving in storage
is ob
serve
d
while stori
ng th
e
data. Besid
e
s that, there i
s
a re
ductio
n
in
time
req
u
ired
for data t
r
an
sfer
and
ene
rgy requi
rem
e
nt
for data tran
smissi
on [1, 18-23].
Figure 2. Block
Diag
ram o
f
Data Comp
ression Syste
m
Data comp
re
ssi
on mig
h
t be eithe
r
lossy or lo
ssl
ess data
comp
ression te
chni
que
s. In
lossy
data co
mpre
ssion, some
d
a
ta whi
c
h are
not
of much
impo
rta
n
ce
are elimi
nated from th
e
HMI
Data
Historian
Contr
o
l
Server
Co
m
m
unication
Router
s
M
odem
RTU /
PLC
M
odem
RTU /
PLC
Co
m
m
unication Mediu
m
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Arithm
atic Co
ding Based A
ppro
a
ch for Powe
r System
Param
e
ter Data …
(Su
bhra J. Sarkar)
270
file to be com
p
re
ssed an
d thus the
r
e is a
redu
ction
of file size. But it is obvious th
at the decod
e
d
file will not be an exact replica of
the actual file. Applications
of this compression includes
comp
re
ssing
multimedia fil
e
s like audi
o, video, image
s etc. e
s
pe
ci
ally for appli
c
ations
su
ch a
s
strea
m
ing
me
dia a
nd inte
rnet telep
hony
. Tran
sf
o
r
m
coding, Ka
rh
u
nen- Lo
eve T
r
an
sform
(KL
T
)
codi
ng,
wavel
e
t ba
sed
co
di
ng et
c. are fe
w of th
e avail
able lo
ssy da
ta com
p
ressi
on te
chni
que
s.
On the othe
r
hand, the
r
e is no loss of da
ta in
lossle
ss
data co
mpression, an
d all the inform
atio
n
can
be
re
con
s
tru
c
ted fo
rm
the comp
re
ssed
data
and
can
be
re
stored. So, this compressio
n can
be u
s
e
d
fo
r
redu
cing
the
size
of imp
o
rta
n
t data
(Z
IP f
ile), me
dical i
m
age
s
etc. T
he a
pproa
ch
of
compression is simil
a
r in most
of the
lossless
co
mpre
ssion te
chni
que
s. Initially a statisti
cal
model fo
r the
input data i
s
gene
rated
wh
ich i
s
then
utilized to
map i
nput data to
b
i
t sequ
en
ce
s i
n
su
ch a way that the most
proba
ble (f
reque
ntly
encountered
) dat
a will pro
d
u
c
e sho
r
ter o
u
tput
than improba
ble data in the
su
bseque
nt steps. Sh
anno
n-F
ano
algorith
m
, Hu
ffman algorit
hm,
Arithmetic
Co
ding et
c. are
the few
com
m
only k
nown
lossle
ss
dat
a co
mpressio
n tech
niqu
es
[1,
16- 21].
Basic a
r
ithme
t
ic
codin
g
i
s
a lo
ssl
ess dat
a comp
re
ssi
o
n
techniq
u
e
whe
r
e
a
data
stri
ng i
s
encode
d in form of a fractional value lying in betwe
en 0 and 1. In this metho
d
, base
d
on the
probability of the content of s
ource m
e
ssage the interval is
narrowed successively [18, 19].
Con
s
id
erin
g some
source
messag
e co
mpri
sing of
symbol
set
S comp
ri
sing of
ch
ara
c
ters {
a
,
b,
c,
d,
s,
k} wit
h
t
heir re
spe
c
t
i
v
e
prob
abil
i
ty of occurre
n
ce {
0
.3, 0.
2, 0.1, 0.2, 0.1, 0.1}. Based
on
this inform
ation, the rang
e
for a particul
a
r symb
ol
ca
n be obtain
e
d
as given in table 1 [1, 19, 20].
Basic A
r
ithm
etic Co
ding
Algorithm for enco
d
ing
a
nd de
codi
ng
a particular symbol stri
n
g
is
discu
s
sed in the su
bsequ
e
n
t steps [1, 1
6
].
Table 1. Probability distribu
tion table for symbol set S
Sl. No.
Character
Probabilit
y
Cumulative
probabilit
y
Range
r_l
o
w r_hi
1 a
0.3
0.3
0
0.3
2 b
0.2
0.5
0.3
0.5
3 c
0.1
0.6
0.5
0.6
4 d
0.2
0.8
0.6
0.8
5 s
0.1
0.9
0.8
0.9
6 k
0.1
1
0.9
1
2.1. Algorith
m
at Encoding End
Input: Symb
ol string (S)
Outpu
t:
Bina
ry string bin
of floa
t num
ber (num
)
STEP 1: Calculate the leng
th (l) of S.
STEP 2: Initia
lize varia
b
le
s min = 0, max = 1an
d r = 1.
STEP 3: Set i
=
1.
STEP 4: Repeat step
s 5 -
9 until i
l+1.
STEP 5: x=
i
th
characte
r of S.
STEP 6: Corresp
ond
s to x, obtain r_lo
w
and r_
hi.
STEP 7: Update min = min
+ r * r_lo
w a
nd max = min
+ r * r_hi.
STEP 8: r = max – min.
STEP 9: i =
i
+
1.
STEP 10: End of the loop.
STEP 11: Obtain a numbe
r num (min
<n
um< max
)
wit
h
minimum bi
nary stri
ng le
ngth.
STEP 12: bin = Binary equ
ivalent of num
STEP 13: End.
Con
s
id
er, for
a symbol
stri
ng {b, a, c, k}
having 4 ch
a
r
acte
rs, whi
c
h is to be en
coded by
the arithmeti
c
codi
ng al
gorithm.
The
execution results for the f
our
iterations are illustrated in
table 2. T
he
output of th
e
algorith
m
will
be the
bin
a
ry
string
(.) 01
0
1011
corre
s
p
ond
s to th
e fl
oat
numbe
r, n
u
m
=
0.335
937
5
lying b
e
twe
e
n
0.33
54
and
0.336 having
minimum bin
a
ry string
len
g
th
(7 bits
in this
c
a
s
e
) [1, 19].
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 2, May 2016 : 268 –
274
271
Table 2. Itera
tion Steps in Arithmetic Co
ding for st
ring
‘back’
Iteration No.
Character
(x)
min
max
r
1
b
0.3 0.5 0.2
2 a
0.3
0.36
0.06
3 c
0.33
0.336
0.006
4
k
0.3354
0.336
0.0006
2.2. Algorith
m
at Deco
ding End
Input:
Binar
y
string (bin
); Ac
tual string length (l)
Outpu
t:
Ac
tu
al string (ar
r
)
STEP 1: Calculate float nu
mber (num
) correspon
ding
to the binary string.
STEP 2: Define a null arra
y arr.
STEP 3: Set i
=
1.
STEP 4: Repeat step
s unti
l
i
(l+1).
STEP 5: Determine the ran
ge within
whi
c
h num lie
s(correspon
ding
to some
cha
r
acter x).
STEP 6: arr (i
) = x
.
STEP 7: lo =
r_lo
w (x
), hi=
r_hi (x
) a
nd r
= hi -lo.
STEP 8: num = (num – lo
)
/ r.
STEP 9: i =
i
+
1.
STEP 10: End of the loop.
STEP 11: Array arr is the e
n
co
ded
string
.
STEP 12: End.
For bin
a
ry st
ring bein
g
en
coded bi
n:= 0
1010
1 wh
ere
l = 4, the proce
s
s of exe
c
ution of
the algorith
m
to extract the encode
d
strin
g
is as give
n in table 3.
Table 3. Itera
tion Steps to obtain the en
cod
ed stri
ng
Iterati
on No.
(i
)
Val
ue (num)
arr(i
) =
x
l
o
hi
r
1 0.3359375
b
0.3
0.5
0.2
2 0.1796875
a
0
0.3
0.3
3 0.59895833
3
c
0.5
0.6
0.1
4 0.98958333
3
k
0.9
1.0
0.1
3. Proposed
Arithme
t
ic Coding base
d
Algorithm
In the MA
TLAB based al
go
rithm, power
system
pa
ra
meters (say voltage, frequ
e
n
cy and
power fa
ctor) is com
p
ressed to form a
cha
r
a
c
ter
string at the en
codi
ng en
d.
At the decodi
ng
end, the po
wer
system
para
m
eter
s a
r
e extra
c
ted from the co
mp
re
ssed st
ring by arithmetic
decomp
r
e
ssi
on based alg
o
rithm.
The d
a
ta symbol
set D compri
ses of the characters bet
we
en 0-
9.If the probability of each character i
s
assumed
to be equal, the probability distri
bution table for
D is given in t
able 4.
Table 4. Probability distributi
on table for data symbol
set D
Sl. No.
Character
Probabilit
y
Cumulative
probabilit
y
Range
r_l
o
w r_hi
1 0
0.1
0.1 0
0.1
2 1
0.2
0.1
0.2
3 2
0.3
0.2
0.3
4 3
0.4
0.3
0.4
5 4
0.5
0.4
0.5
6 5
0.6
0.5
0.6
7 6
0.7
0.6
0.7
8 7
0.8
0.7
0.8
9 8
0.9
0.8
0.9
10 9
1
0.9
1
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Arithm
atic Co
ding Based A
ppro
a
ch for Powe
r System
Param
e
ter Data …
(Su
bhra J. Sarkar)
272
3.1. Algorith
m
at Encoding End
Input: Sys
t
e
m
voltage (v
) in kV; System fr
eque
nc
y (f) in Hz; L
o
ad powe
r
factor
(pf)
Outpu
t:
Cha
r
acter arra
y (char)
STEP 1: Convert v, f and pf decimal st
ri
ng vstr, fstr, p
f
str.
STEP 2: pfv:= Co
ncatenat
e (pfstr, fstr, vstr).
STEP 3: arith:= Binary arra
y correspon
d
s
to the arith
m
etic coding
of pfv.
STEP 4: num:= 12- L
ength
(pfv).
STEP 6: binum:= Binary e
quivalent of n
u
m.
STEP 7: nwari:= Co
ncaten
ate (binu
m
, arith).
STEP 8: len:= Array length (nwari
).
STEP 9: num:= len%7.
STEP 10: zero:= Zero arra
y having num
numbe
r of elements.
STEP 11: nwar:=
Con
c
ate
nate (zero, wari).
STEP 12: count:= Array length (n
wa
r) / 7.
STEP 13: spar:= Split nwa
r
in (7 X c
o
unt) array.
STEP 14: Define null array
cha
r
and d
e
ci
of size (1 X 7).
STEP 15: Initialize i:= 1 a
n
d
repe
at step
s 16-18 until i
count.
STEP 16: deci (i):
= De
cim
a
l equivalent
of nwar
(7 X i).
STEP 17: asci:= ASCII cha
r
acte
r corre
s
pond
s to de
ci
(i).
STEP 18: char(i
):= a
s
ci. i:= i+1.
STEP 19: End of the loop.
STEP 20: lnt:
=
ASCII c
h
arac
ter
c
o
rres
p
onds
to len.
STEP 21: com:= Co
ncate
nate (lnt, cha
r
).
STEP 22: End.
3.2. Algorith
m
at Encoding End
Input:
Char
a
c
ter ar
ray (c
har)
Outpu
t: Sy
st
em
voltage
(
v
) in kV; Sys
t
em
freq
uen
cy (f
) in Hz
;
Load pow
er
fac
t
or (p
f)
STEP 1: len:= Array length (co
m
bo
).
STEP 2: actlen:= ASCII value co
rrespon
ds to the ch
aracter
com
bo (1).
STEP 3: numzero:= 7*
(len
-1) – actle
n
.
STEP 4: nwar:= New a
rray
of size (1X(le
n-1
)) formed f
r
om combo af
ter rem
o
ving comb
o(1).
STEP 5: null:= Null a
r
ray. Initialize i:= 1
and re
peat st
eps 6
-
8 until i
(len-1).
STEP 6: p:=
nwa
r(i
).
STEP 7: bin:= (1X7
) bina
ry array co
rrespond
s to bina
ry value of p.
STEP 8: null:= Co
ncatenat
e (null, bin
)
. i:= i+1.
STEP 9: End
of the loop
STEP 10: nwbin:= New bi
n
a
ry array formed from bin
after rem
o
ving numzero nu
mber of zero
s
from the begi
nning.
STEP 11: strsz(1X2
)
:= Array contai
ni
ng
first two bits
of nwbin.
STEP 12: num:= De
cimal
equivalent of
strsz.
STEP 12: modbin:= Mo
dified bina
ry array formed fro
m
nwbin afte
r removing
strsz.
STEP 11: str:= Ch
ara
c
te
r string co
rrespo
nds to t
he a
r
ithmetic de
co
d
i
ng of modbin
using a
c
tlen.
STEP 12: Usi
ng num, split
str to obtain v
,
f and pf.
STEP 13: End.
4. Result a
n
d Analy
s
is
Effectiveness of any compre
ssion al
gorit
hm
can
be determi
ned by the value of
comp
re
ssion
ratio, i.e. ratio of the size of
uncom
p
r
esse
d data
to that of compre
ssed da
ta.
Higher will be the comp
ression ratio, better will be the comp
ressi
on. High com
p
ression ratio also
implies re
du
ced
mem
o
ry
and
ene
rgy
req
u
ire
m
ent
for d
a
ta
storag
e a
nd
d
a
ta tran
smi
s
sion
respe
c
tively. The p
e
rfo
r
ma
nce
of the al
gorithm
i
n
terms
comp
re
ssion ratio
as well as
execution
time requi
red
at enco
d
ing a
nd de
codi
ng
end is give
n in table 5.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 2, May 2016 : 268 –
274
273
Table 5. Co
m
p
re
ssi
on ratio
and executio
n ti
me requi
re
d for variou
s
voltage levels
Sl.
No
Input param
eters
Compression res
u
lt
Voltage
(i
n k
V
)
Freque
nc
y
(i
n Hz
)
pf
No. of
integer
Actual
data
size
(i
n
by
t
e
s)
Compressed
data size (in
by
t
e
s)
Compr
e
ssio
n ratio
Encoding
time
Decodin
g time
1 0.41
49.95
0.965
9
226
98
2.306
0.423
0.139
2 3.31
50.01
0.931
10
226
98
2.306
0.438
0.121
3 11.02
50.02
0.802
11
242
114
2.123
0.428
0.137
4 220.06
50.01
0.875
12
258
114
2.263
0.425
0.129
From ta
ble 5
,
it is clea
r t
hat the com
p
re
ssi
on
rati
o for thi
s
arit
hmetic
codi
n
g
ba
sed
algorith
m
is a
bove 2 for an
y possi
ble in
put. As
the length of enco
ded bin
a
ry array is de
pen
d
ent
on the
conte
n
t of input a
r
ray length b
u
t not on the
l
e
ngth of the in
put array, it is quite
po
ssi
ble
that at similar voltage levels with sli
g
h
t
ly diffe
rent inputs, there is a
variation
of compre
ssed
string l
ength
and thu
s
there is some variation in
the compressio
n ratio. Few
su
ch example
s
a
r
e
given in table
6.
Table 6. Vari
ation of numb
e
r of ch
ara
c
te
rs in
comp
re
ssed
string
with inputs
Sl. No.
Input param
eters
Number of
characters in
compr
e
ssed
string
Voltage
(i
n k
V
)
Freque
nc
y
(i
n Hz
)
Power
factor
1 1.13
49.98
0.675
7
2 1.11
50.01
0.758
6
3
33.12
49.96
0.752
6
4
33.12
50.01
0.798
7
5. Conclusio
n
s
From th
e ab
o
v
e discu
ssi
on
s, it is
clea
r t
hat t
here is a
signifi
cant
re
ductio
n
in
size of the
input data wit
hout loss of any sigle bit. This im
pli
e
s
a redu
ce
d m
e
mory re
qui
rement for sto
r
ing
bulk volu
me
of su
ch inp
u
t data. The
r
e
is an in
he
ren
t
encryptio
n i
n
the algo
rith
m, thereby
can
deal
with dat
a se
cu
rity problem
s successfully. Re
d
u
ce
d amo
unt
of data tra
n
sfer
also re
sults
lowe
r e
nergy requi
rem
ents
for commu
nication pu
rp
ose
s
. Tho
ugh th
e above
re
sul
t
s are obtai
ne
d
by executin
g
the algo
rith
m offline, this alg
o
ri
thm
can al
so b
e
i
m
pleme
n
ted
for onlin
e testing
whe
r
e the en
codi
ng and d
e
co
ding p
r
og
ram
s
are exe
c
uted
simulta
neou
sly. The variation of total
executio
n tim
e
(i
n
se
c.)
wit
h
total
numb
e
r
of
de
ci
mal
n
u
mbe
r
s in
the
input
pa
rame
ters is given
i
n
figure 3.
Figure 3. Vari
ation of total execut
io
n time with numb
e
r
of integers
0.545
0.55
0.555
0.56
0.565
0.57
9
1
01
11
2
Total
execution
time
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Arithm
atic Co
ding Based A
ppro
a
ch for Powe
r System
Param
e
ter Data …
(Su
bhra J. Sarkar)
274
It is clear fro
m
figure 3 th
at the execut
i
on time is in
depe
ndent of
the numbe
r
of th
e
interge
r
s in the input p
a
rameters. The
algorithm i
n
pre
s
ent form can
deal
with thre
e po
we
r
system
pa
ra
meters du
e t
o
software li
mitations. Bu
t if the limitation can
de ta
ken
ca
re
of, it is
possibl
e to compress mo
re numb
e
r
of para
m
eter
s t
o
improve th
e syste
m
pe
rforman
c
e. T
h
ere
will be i
m
provement in the compre
ssion ratio further,
if probability distri
bution can
be prepared
by con
s
ulting
the real tim
e
data value
s
. This
algo
ri
thm can
also
be implem
e
n
ted with oth
e
r
los
s
le
ss d
a
t
a
comp
re
ssi
on
t
e
chniq
u
e
s
t
o
comp
are a
nd improve the perfo
rma
n
c
e, if possibl
e
.
The work
ca
n also be ext
ende
d for
co
mpre
ssing
l
a
rge multi-dime
nsio
nal data
array co
ntaini
ng
different po
wer sy
stem pa
rameters moni
tored at different time insta
n
ce
s.
Referen
ces
[1]
SJ Sarkar, B Das, T
Dutta,
P Dey
,
A. Muk
herjee,
An Alte
rnative Vo
ltag
e
and F
r
eq
uenc
y Monitori
n
g
Sche
me for S
C
ADA bas
ed
Co
mmun
icati
o
n in Pow
e
r Sy
stem us
in
g Dat
a
Co
mpr
e
ssi
on
, International
Confer
ence
an
d W
o
rkshop o
n
Computi
ng a
n
d
Commu
ni
cati
on (IEMCON), Vanco
u
ver. 20
15: 1- 7.
[2]
JB.
Gupta.
Power System
Analysis
. 1
st
Edition, S. K. Kataria
& Sons, Ne
w
Delhi. 2
0
1
1
.
[3]
Leo
nard
L
Grig
sb
y
.
Pow
e
r Sy
stem Stab
ility
& Control
. 3
rd
E
d
itio
n, CRC Pr
ess, Boca Rato
n. 2012.
[4]
AE F
i
tzgerald,
Charl
e
s Kin
g
sl
e
y
, Step
he
n. D Umans.
Electri
c
Machin
ery
. 6
th
Edition, Ne
w
York. 2003.
[5]
CL
Wadh
w
a
. Electrical P
o
w
e
r Systems
. 6
th
ed., Ne
w
A
ge In
ternatio
nal P
u
b
lishers, N
e
w
D
e
lhi. 2
0
1
4
.
[6]
John J Grai
ng
er, W
illiam D S
t
evenso
n
Jr.
Power System
A
nalysis
. International Edit
io
n
,
McGraw
Hi
ll
Inc., Singap
ore
.
1994.
[7]
BM W
eed
y, BJ
Cor
y
,
N J
enki
n
s, JB Ekan
a
y
ake, G Strbac.
Electric Power
System
. 5
th
Ed
i
t
ion, Wile
y,
West Sussex
.
2012.
[8]
Nicol
eta ARG
H
IRA, Dani
ela
HOSSU, Ioan
a F
Ă
G
Ă
R
ĂŞ
AN, Sergiu Stelian
ILIESCU, D
ani
el R
ă
zvan
COST
IANU. Moder
n SCA
D
A
Phil
osop
h
y
in
Po
w
e
r
S
y
stem
Operati
o
n
–
A
Surve
y
.
U.P.B. Sci. Bull.
,
Series C
. 20
11
; 73(2).
[9]
Sarasij
D
a
s, M.S.
T
hesis (
E
ngi
neer
in
g). Po
w
e
r S
y
st
em
Data
Com
p
re
ssion f
o
r Arch
iving. I
ndi
a
n
Institute of Science. 20
07.
[10]
Shan
g
L, Jae
g
e
r J, Kre
b
s R.
Efficiency
Analysis
of data
compre
ssion of power system
transient
s
usin
g w
a
velet transfor
m
. Po
w
e
r T
e
ch Conference Proc
ee
di
ngs, IEEE Bolo
gna. 20
03; 4.
[11]
Gabrie
l
G
ăş
p
ă
resc,
P
o
wer Quality Data Compress
ion
, Av
ail
a
ble. [Onl
i
n
e
]
http://cdn.intec
hopen.com/pdfs/44193/InT
e
chPo
w
e
r_qualit
y
_
dat
a_com
p
ression.pdf.
[12]
Jia
x
in
N
i
ng,
Jia
nhu
i W
a
n
g
, W
enzh
o
n
g
Gao,
Con
g
L
i
u. A
W
a
vel
e
t b
a
sed
d
a
ta com
p
ressi
o
n
tech
ni
qu
e
for smart grid.
IEEE Transactions on S
m
art
Grid
. 2011; 2(
1
)
: 212-21
8.
[13]
G. Panda, PK Dash, AK Prad
han, SK Me
her
. Data
Compr
e
ssion of Po
w
e
r
Qualit
y Event
s Using th
e
Slantl
e
t T
r
ansform.
IEEE Tra
n
sactions on P
o
wer Deliv
ery
. 200
2; 17(2): 66
2-66
7.
[14]
K Nithi
yan
anth
an, V Ramach
andr
an. Effective Data Comp
ressio
n
Mode
l for On-line Po
w
e
r S
y
st
e
m
Appl
icatio
ns.
Internatio
nal J
o
u
r
nal of Electric
al Ener
gy
. 201
4; 2(2).
[15]
Michel
P T
c
heou, Lis
a
n
d
ro
L
o
visol
o
, Mois
e
s
V Rib
eiro, E
duar
do AB
da
Silva, Marc
o A
M
Rodri
g
u
e
s,
João
MT
Romano, P
a
u
l
o S
R
Din
iz. T
he Co
mpressio
n
of E
l
ectric Si
gn
al
W
a
veforms for
Smart Grids
:
State of the Art and F
u
ture T
r
ends.
IEEE Transactions on S
m
art Grid
. 20
1
4
; 5(1): 291-3
0
1
.
[16]
T
r
ibeni Pras
ad
Ban
e
rje
e
, Am
it Kon
a
r,
Ajith
Abra
ham. CA
M bas
ed
Hig
h
-
spee
d C
o
mpr
e
ssed
Da
t
a
Commun
i
cati
o
n
S
y
stem
Dev
e
lo
p
m
ent us
in
g F
P
GA, [Onlin
e] Avai
lab
l
e
:
http://
w
w
w
.
ac
a
demi
a
.edu/
237
408
2/CAM_
ba
sed_
Hig
hsp
e
e
d_C
ompress
e
d
_
Data
_Comm
u
nicati
on_S
y
stem_Dev
elo
p
m
ent_us
ing
_
F
P
GA.
[17]
Difference betw
een Lossless
& Lossy
Data Compression
, [Online] Available:
http:// w
w
w
.
rf
w
i
reless-
w
o
rld.c
o
m/T
e
rminol
og
y/l
o
ssl
ess-data-c
o
mp
ression-vs-loss
y
-
data-com
pression.html.
[18]
T
heor
y
of Data
Compressi
on.
[Online]. Avail
a
bl
e: http://
w
w
w
.
data-compres
s
ion .com/theo
r
y
.shtml.
[19]
Z
e
-Nian
Li, Ma
rk S Dre
w
, Jia
n
g
chu
an L
i
u.
F
unda
mentals of
Multi
m
ed
ia
. 2
nd
Edition, Spri
ng
er. 2014.
[20]
Clau
d
i
o
Iomb
o, MS.
T
hesi
s
, Predictive
Da
ta C
o
mpre
ssion
usin
g A
daptiv
e Arithm
etic Co
din
g
.
Lou
isia
na Stat
e Univ
ersit
y
&
Agricult
ural a
n
d
Mecha
n
ica
l
Coll
eg
e. Augus
t 2007.
[21]
Z
S
R Koditu
w
a
kku, US Amar
asin
ghe. C
o
m
paris
ons of L
o
ssless Dat
a
C
o
mpressi
on A
l
gorithms for
T
e
x
t
Data.
Indian Jour
nal
of Computer
Sci
e
n
c
e and En
gi
ne
erin
g
. 1(4): 406
-425.
[22]
Rupi
nd
er Si
ng
h Brar, Bikr
a
m
jeet Si
ng
h.
A Surve
y
on
Different C
o
m
p
ressi
on T
e
ch
niq
ues & B
i
t
Reduction A
l
gorithms for Compre
ssion of
T
e
x
t
/ Lossles
s Data.
Intern
ation
a
l Jo
urn
a
l
of Advanc
e
d
Rese
arch in C
o
mputer Sci
e
n
c
e & Softw
are
Engi
neer
in
g
. 2013; 3(3): 5
79-
582.
[23]
Cha
ng, Ch
ia-
H
ui Wan
g
, Chi
n
g-Chi
a
Hs
ieh.
Da
ta C
o
mpres
s
ion for E
ner
g
y
Efficient Com
m
unic
a
tion
o
n
Ubiq
uito
us Sen
s
or Net
w
ork.
T
a
mk
an
g Journ
a
l of Scienc
e a
nd Eng
i
n
eeri
n
g
. 2011; 14(
3): 345-3
54.
Evaluation Warning : The document was created with Spire.PDF for Python.