TELKOM
NIKA
, Vol. 11, No. 7, July 201
3, pp. 3479 ~
3490
e-ISSN: 2087
-278X
3479
Re
cei
v
ed
Jan
uary 23, 201
3
;
Revi
sed Ma
rch 1
3
, 2013;
Acce
pted Ma
rch 2
6
, 2013
Assessment of Wind Power Potential at Hawksbay,
Karachi Sindh, Pakistan
Shahna
w
a
z
Farhan
Khah
ro*1, Amir Mahmood Soo
m
ro
2
, Kav
i
ta Tabba
ssum
3
, Lei Dong
4
,
Liao Xiaozh
ong
5
1,2,
4.5
School of Automatio
n
, Beiji
ng Institute
of
T
e
chnol
og
y,
Beiji
ng 10
00
8
1
, PR Chin
a
2
Departme
n
t of Electrical En
gi
neer
ing,
Me
hra
n
UET
Jamshoro, Sindh, Paki
stan
3
Information T
e
chno
log
y
C
ent
er, Sindh Agr
i
c
u
lture U
n
ivers
i
ty T
andoj
am, Sindh, Pakista
n
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: shahn
a
w
a
z
fa
rhan@
gma
il.co
m
1
A
b
st
r
a
ct
Pakistan is f
a
cing s
e
rio
u
s e
nergy cr
isis at
prese
n
t. T
h
e
gover
n
m
ent
i
s
ai
mi
ng to
utili
z
e
t
h
e
immense
pote
n
tial
of renew
abl
e en
ergy s
ources l
i
ke:
S
o
lar, W
i
nd, et
c, in ad
ditio
n
to intens
ify the
conve
n
tio
nal
s
ources
of e
n
e
r
gy to
over th
e ac
ute s
hort
age
of
ener
gy
. W
i
nd
ener
gy
is th
e fastes
t-
deve
l
op
in
g en
ergy sourc
e
w
o
rldw
id
e. T
he ai
m of th
is pa
per is to expl
o
r
e and esti
mate the w
i
nd po
w
e
r
potenti
a
l of Ha
w
ksbay Karac
h
i, one of
the
locati
ons in s
o
uthern p
a
rt
of Pakistan. W
i
n
d
spee
d data
(in
meters p
e
r se
cond) fro
m
Ap
ril 200
9 to Ap
ril 201
1
at fou
r
different hei
ghts is me
asu
r
ed. W
i
nd po
w
e
r
dens
ities, freq
uency d
i
stributi
on, and W
e
i
b
u
ll distrib
u
tio
n
o
f
w
i
nd speed a
r
e calcul
ated i
n
this study. Thi
s
study als
o
pr
e
s
ents the
ana
l
ysis an
d
co
mp
ariso
n
of 5
nu
mer
i
cal
metho
d
s to det
ermin
e
the W
e
i
bul
l
scal
e
and
sha
p
e
par
ameters for
th
e av
ail
abl
e w
i
nd
data. T
h
e
estimated
w
i
nd
pow
er to
be
gen
erate
d
thro
ug
h
commercia
l w
i
nd turbi
ne is a
l
so incl
ud
ed. T
he yearly
me
a
n
w
i
nd spee
d at Haw
ksbay, Karach
i is 5.9m/s
and
has p
o
w
e
r density of 1
9
7
W
/
m
2
at 80
m heig
h
t w
i
th hi
gh p
o
w
e
r den
sity durin
g Apr
il to Aug
u
st. T
h
e
estimated
cost
per kW
h
is
U
S
$0.03
45. T
h
e
r
efore th
e site
may b
e
co
nsi
dere
d
su
itabl
e
for w
i
nd tur
b
i
n
e
app
licati
ons.
Ke
y
w
ords
: W
i
nd e
ner
gy, Po
w
e
r density fu
nction, W
e
i
b
u
l
l
di
strib
u
tion,
Rayle
i
g
h
Distri
butio
n, Ca
paci
t
y
factor
Copy
right
©
2013 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
The
economi
c
a
n
d
social
stability of
an
y cou
n
try
stro
ngly de
pen
ds upo
n th
e av
ailability
of energy. Th
e per capita e
nergy con
s
u
m
ption is
an i
ndex used to measure
the pro
s
pe
rity of any
so
ciety. Paki
stan i
s
a
de
veloping
co
u
n
try with
pop
ulation of
al
most 1
77 mil
lion
pe
ople
ha
s
averag
e yearl
y
energy con
s
umptio
n of about 45
0k
Wh per
capita,
whe
r
ea
s the
worl
d’s i
s
ab
out
2730
kWh [1]. Pakistan is b
a
si
cally an en
ergy defic
i
ent
country. Almost 37% peo
ple amon
gst the
popul
ation living in rem
o
te and ru
ral a
r
e
a
s are waitin
g to be con
n
e
c
ted to the na
tional grid.
The co
untry’
s
install
ed e
l
ectri
c
ity gen
erat
ion
cap
a
c
ity at prese
n
t is approximately
19,566M
W, 3
0
% of which
is from hyd
e
l, 67% fr
o
m
fossil fuel
s (32.3% from
natural
ga
s a
n
d
35.3% from
oil in 2008
-0
9) an
d rem
a
ining from
n
u
cle
a
r en
erg
y
[2-4]. The contri
bution
of
hydroel
ect
r
ici
t
y in the total gene
ration
o
f
the co
untry
has
bee
n de
clinin
g gradu
ally from 70
% in
the 1960
s to
33% at pre
s
e
n
t [5]. This trend ha
s tri
g
g
e
red
a ma
ssi
ve raise in el
ectri
c
ity price
as
well a
s
increa
sed ai
r polluti
on.
Country’s poor economy
does not all
o
w
spe
nding billions of
dolla
rs on fossil fuel
import
s
, parti
cula
rly oil. M
o
st of the im
ported
o
il is
u
s
ed fo
r ele
c
tri
c
ity gene
ratio
n
. The
count
ry’s
limited internal gas
and
oil
base,
with their
exis
ting production
rate,
will
get
exhaust
within
2 t
o
3
decade
s. Pa
kistan’
s imme
n
s
e
coal
poten
tial has
not b
een p
o
tentiall
y utilized
due
many rea
s
on
s.
Similarly technolo
g
y ba
rri
ers,
high
cost, and inte
rn
ational
re
stri
ction
s
a
r
e th
e big
ba
rri
ers i
n
developin
g
n
u
cle
a
r e
n
e
r
g
y
[6]. Thus a
n
immedi
at
e sea
r
ch
for co
st-effe
ctive,
sustain
able an
d
environ
ment-f
riendly ene
rgy
sou
r
ce
s is
ne
ce
ss
a
r
y to meet t
he count
ry’s req
u
ire
m
ent
s.
Ren
e
wable
-
e
nergy
re
so
urce
s p
a
rti
c
ula
r
ly wi
nd
ene
rgy te
chn
o
lo
gy is ra
pidly
growi
ng
en
ergy
resou
r
ce thro
ugho
ut the
world
be
cau
s
e
of its
ample
existen
c
e, redu
ced
cost and having
l
o
w
environ
menta
l
dama
ge [7].
The
win
d
p
o
we
r g
ene
rat
i
on cap
a
city of
the worl
d has
re
ached
to
196,63
0MW
by the end of 2010 [8].
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 7, July 2013
: 3479 – 349
0
3480
In ord
e
r to
e
x
plore th
e wi
nd pote
n
tial, Pakista
n
Met
eorol
ogi
cal
Departm
ent (P
MD)
and
Alternative Energy
Develo
pment Boa
r
d
(AEDB)
in
co-o
rdin
ation
with UNDP h
a
s g
a
thered
wind
data in
co
ast
a
l area
s of P
a
ki
stan.
Wind
data
colle
cte
d
at Hawksba
y
, Karachi is
pre
s
ente
d
in t
h
is
pape
r. Asse
ssment
s a
nd
e
v
aluation
s
of
wind
ene
rgy
potential a
r
e
perfo
rmed
by
many
cou
n
tri
e
s
of the world [
9
-15].
The aim of this articl
e is to highlight the
potential of wi
nd re
sou
r
ce a
t
Hawksb
ay, Kara
ch
i
at the coa
s
t o
f
Sindh. Data
colle
cted
at the site
ha
s b
een a
nalyzed
,
estimated
p
o
we
r availa
bl
e in
the win
d
an
d
elect
r
ical po
wer expe
cted
to be
g
ene
rated
via com
m
ercial wind turbine
ha
s
a
l
so
been
cal
c
ulat
ed in orde
r to supp
ort the
evaluation
a
n
d
planni
ng of
future wi
nd e
nergy p
r
oje
c
t
s
in
south
e
rn
regi
on of Pakista
n
.
2.
Wind As
ses
sment an
d Data Analy
s
is
It is essential
to have the familiarity with
wind characteri
stics
like;
speed, duration of
time that win
d
is availa
ble
and di
re
ction in or
der to
investigate th
e wind
ene
rg
y potential for a
certai
n location. Beside
s, the den
sity of air, de
sign of
turbine a
nd t
he turbin
e to
wer
height aff
e
ct
the power
generated from
the wi
nd.
These characteristic will
be
discussed
in the
subsequent
se
ct
ion
s
.
2.1 Wind
Speed
Char
acteriza
t
ion
The
spee
d of
wind va
rie
s
with hei
ght. Several fun
c
t
i
ons
ca
n be
use
d
to express this
variation. Th
e po
wer
exp
onent fun
c
tio
n
is
comm
on
ly used fu
nct
i
on an
d is
gi
ven in follo
wi
ng
equatio
n;
()
r
r
Z
Vz
V
Z
(1)
whe
r
e
V(z
)
i
s
spe
ed of wind at heig
h
t
Z
agl (above
grou
nd level)
,
V
r
is speed of win
d
at
referen
c
e hei
ght
Z
r
agl an
d
β
is an exp
onent, relie
s
on su
rface ro
ughn
ess len
g
t
h. The lengt
h o
f
surfa
c
e rou
g
h
ness
an
d
β
for variou
s type
s of terrai
n
s a
r
e given in literatu
r
e [16].
2.2
Statis
tical Di
stribu
tion fo
r Wind Da
ta
Wind
spe
ed chang
es with t
i
me at particu
lar lo
cation th
erefo
r
e it is nece
s
sary to carryo
u
t
empiri
cal det
ermin
a
tion
of wind sp
eed
d
i
stributio
n.
Weibull a
nd
Ra
yleigh di
strib
u
tions are two of
the more
often use
d
fun
c
tions a
m
on
g a
numbe
r of a
v
ailable empi
rical fu
nctio
n
s
to fit wind d
a
ta
over pe
riod of
time at particular lo
cation.
2.3.1
Weibull Distr
i
bution Func
tion
The
Weibull
distrib
u
tion
function
offers th
e be
st ag
reem
en
t with a va
riety of
experim
ental
data analy
z
ed [17-20]. It has
widel
y
been u
s
e
d
for the a
s
se
ssment of wi
nd
potential fo
r
different regi
ons i
n
ma
ny cou
n
trie
s [5,
13, 21
-23]. T
he Weibull
probability de
n
s
ity
function (P
DF
) of the wind
spe
ed can be
det
ermin
ed from the follo
wing e
quatio
n [24-26]:
1
()
e
x
p
kk
kV
V
fV
CC
C
(2)
whe
r
e
C
&
k
are
Weib
ull
scale (having
no unit) &
S
hape
(same
unit as of
wi
nd speed; m/
s)
para
m
eters resp
ectively.
The co
rrespo
nding
cumul
a
tive distributio
n function
FV
ca
n be writin
g a
s
follow:
1e
x
p
k
V
FV
C
(3)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
Asse
ssm
ent of wind po
we
r potent
ial at Ha
wksb
ay, Kara
chi Sind
h, Pakista
n
(Sh
ahna
wa
z FK)
3481
2.3.2
Ra
y
l
eigh Distribution Fu
nction
The Raylei
gh
distributio
n is also used e
x
tensively for fitting wind speed d
a
ta. When the
value of sh
ape pa
ramet
e
r
k=
2
the
Weibull di
stribution be
comes
Raylei
gh Di
stributi
on.
Therefore, th
e Rayleig
h
d
i
stributio
n is
said
to b
e
a
one-parame
t
er Wei
bull d
i
stributio
n. The
prob
ability de
nsity and re
spective cu
mu
lative dist
ribu
tion function
s for Rayleigh
distrib
u
tion can
be written a
s
;
2
2
()
e
x
p
24
avg
a
vg
VV
fV
VV
(4)
2
1e
x
p
4
av
g
V
FV
V
(5)
whe
r
e
av
g
V
is average
wind
sp
eed which ca
n be found from the followi
ng equ
ation
s
;
1
1
N
avg
i
i
VV
N
(6)
The Wei
bull distrib
u
tion with
two pa
ra
meters
i
s
g
e
nerally more
versatile wh
ile
the
Rayleig
h
di
st
ribution,
havi
ng o
n
e
pa
ra
meter, i
s
sim
p
ler to u
s
e.
Even someti
mes the
Rayl
eigh
distrib
u
tion p
r
ovide
s
a b
e
tter fitting than the
Wei
bull di
stributi
on for fitting
the mea
s
u
r
ed
prob
ability de
nsity distrib
u
tion [19, 27].
I have
been observed tha
t
the Rayleigh distrib
u
tion
is
biased towards lo
w win
d
speed
s.
2.3
Metho
d
s for
Determining the Weibull Parameter
Weib
ull para
m
eters ca
n b
e
determi
ned
by us
ing diff
erent p
a
ra
me
ter estimatio
n
method.
The Weibull
distrib
u
tion is important for the as
sessm
ent of the wi
nd ene
rgy po
tential and wi
nd
cha
r
a
c
t
e
ri
st
ic
s;
t
h
e
r
ef
o
r
e,
i
t
is
ne
ce
ss
ar
y
t
o
find
it
s para
m
eters
p
r
ope
rly. If its para
m
eters a
r
e
obtaine
d well
, the Weibull
distributio
n not onl
y agree better wit
h
wind spee
d data, but also
rep
r
e
s
ent the
wind po
we
r potential mu
ch more a
c
curately [28].
Several te
ch
nique
s
are
d
e
scrib
ed i
n
th
e text to d
e
te
rmine
the
We
ibull p
a
ra
met
e
rs to fit
Wei
bull
distrib
u
tion to
the mea
s
u
r
e
d
data
at a certain l
o
cation. Some of t
hese meth
od
s, whi
c
h
provide
easy, effe
ctive an
d
accu
rate meth
od f
o
r
determinat
ion of
Weibu
ll pa
ramete
rs, are
de
scri
b
e
d
belo
w
;
2.3.1 Graphic
a
l
Method
:
This metho
d
is impl
eme
n
ted u
s
in
g the
co
ncept
of l
east
sq
uares to fit straigh
t
line t
o
wind
data, where
the time
-se
r
ie
s d
a
ta
must b
e
sort
ed into
bins.
In this meth
o
d
, the shap
e
“
k
”
and scale
“
C
” paramete
r
can b
e
d
e
termi
ned
by re
-a
rrangin
g
a
nd ta
king
natu
r
al l
og of
cum
u
lat
i
ve
Weib
ull distri
bution fun
c
tio
n
, given in eq
uation (3
), wh
ich yield
s
;
ln
l
n
1
l
n
l
n
FV
k
V
k
C
(7)
This is
simila
r to linear equ
ation
y
ax
b
in whi
c
h;
ln
ln
1
yF
V
,
ln
x
V
,
ak
and
ln
bk
C
(8)
Her
e
x
and
y
are calcula
t
ed throu
gh t
he mea
s
u
r
ed
wind
spe
e
d
data. The
sl
ope
a
and
t
he
intercept
b
can be
determined th
rou
g
h
stan
dard le
ast
squa
re
re
gre
ssi
on m
e
thod. Finally,
the
scale an
d sh
ape pa
ram
e
ters
can b
e
ca
lculate
d
as;
ka
and
b
k
Ce
(9)
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 7, July 2013
: 3479 – 349
0
3482
2.3.2 Empirical
Method
:
In this
method the
Weibull sc
ale ‘
k
’ and
sh
ape
‘
C
’ p
a
r
a
me
ter
c
a
n b
e
de
te
r
m
ined
us
in
g
averag
e win
d
spee
d and
standa
rd deviat
i
on as follo
ws [29-30];
1.086
av
g
k
V
(10
)
1
1
av
g
V
C
k
(11
)
whe
r
e
is stan
dard d
e
viatio
n whi
c
h can b
e
found from
the followin
g
equatio
ns;
1
2
2
1
1
1
N
ia
v
g
i
VV
N
(12
)
2.3.3
Maximum Likelihood Estimation Method:
This
method
is
comm
onl
y used
to d
e
te
rmin
e the
Wei
bull p
a
rameters. Th
e shap
e
para
m
eter
ca
n be estimate
d by followi
ng
equation iteratively [18, 2
9
-30];
1
11
1
ln
ln
NN
k
ii
ii
N
k
i
i
VV
V
i
k
N
V
(13
)
whe
r
e
N
i
s
th
e total nu
mb
er of
win
d
sp
eed m
e
a
s
ure
m
ents
and
i
V
is the
mea
s
u
r
ed
wind
sp
ee
d
value for the
i
th measure
m
ent. After calcul
ating the
value of sh
ape p
a
ram
e
ter
k
, the
sc
ale
para
m
eter
C
can b
e
determined u
s
ing f
o
llowin
g
equ
a
t
ion;
1
1
1
N
k
k
i
i
CV
N
(14
)
2.3.4
Modified Ma
ximum Li
kelihood Estima
tion Meth
od:
The m
odified
maximum
li
kelih
ood
met
hod
ca
n b
e
con
s
id
ere
d
if
win
d
spe
e
d
data i
s
available in f
r
eque
ncy di
stribution format
. The Wei
bull
para
m
eters
are
cal
c
ulate
d
as foll
ows [
18,
29-3
1
];
1
11
1
ln
(
)
ln
(
)
(0
)
()
NN
k
ii
i
i
ii
N
k
ii
i
V
V
fV
V
i
fV
k
fV
Vf
V
(15
)
whe
r
e
i
V
is the
wind
sp
eed
central to bi
n
i
.
()
i
f
V
the freq
u
ency fo
r wi
nd
spe
ed
ran
g
i
ng withi
n
bin
i
, and
(0
)
fV
is the pro
bability
for wind
spe
ed equ
al to or exceedi
ng zero.
2.3.5
Energ
y
Pattern Factor Me
thod
:
The ene
rgy p
a
ttern facto
r
method is
rel
a
t
ed to the average
d data
of wind sp
ee
d and is
defined by th
e followin
g
eq
uation
s
[29-3
0
, 32];
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
Asse
ssm
ent of wind po
we
r potent
ial at Ha
wksb
ay, Kara
chi Sind
h, Pakista
n
(Sh
ahna
wa
z FK)
3483
3
3
()
()
av
g
pf
av
g
V
E
V
(16
)
2
3.
69
1
pf
k
E
(17
)
whe
r
e
p
f
E
is the energy patte
rn facto
r
.
The scale p
a
rameter
can b
e
estimated u
s
ing e
quatio
n
(11).
2.4
Prediction P
e
rforma
nce
of Weibull Di
stribu
tion M
odel
The co
rrelati
on coeffici
ent
2
R
a
nd ro
ot
m
ean sq
ua
re e
rro
r (
RMSE
) analysi
s
hav
e been
carrie
d o
u
t in
order to d
e
te
rmine
which
one
of
the Weibull parame
t
er cal
c
ulatio
n
metho
d
s
gi
ves
a better re
sult
. These p
a
ra
meters ca
n b
e
cal
c
ulate
d
from the follo
wing e
quatio
ns [29, 33];
22
2
11
2
1
NN
ii
ii
N
i
i
yz
x
z
R
yz
(18
)
1
2
2
1
1
N
ii
i
RM
S
E
y
x
N
(19
)
whe
r
e
i
y
is the
ith
ac
tual data,
i
x
is the
ith
predi
cted
dat
a with
the
Weibull
distri
bu
tion,
z
is the
mean of actu
al data and
N
is the numb
e
r of observa
tions. The hig
h
2
R
and low value of
RMSE
will give the b
e
tter model.
2.5 Wind
Po
w
e
r
Gener
a
tion
The wi
nd po
wer P (W), with air den
sity
ρ
(Kg/m3), a
nd win
d
turbi
ne swept a
r
e
a
T
A
(m
2
)
can b
e
cal
c
ul
ated by [34];
3
1
2
pT
P
CA
V
(20
)
whe
r
e
p
C
is Betz limit and is
equal to 0.59
3
The win
d
po
wer exp
r
e
s
se
d in terms of
area,
inde
pe
ndent of the wind turbine
area, is
kno
w
n a
s
wi
n
d
power de
nsity (
WPD
). It can b
e
obtain
ed from the e
quation (20
)
, as given b
e
lo
w;
3
1
2
p
T
P
WP
D
C
V
A
(21
)
The win
d
ene
rgy (E) extra
c
ted by a wind
turbine, ca
n be determine
d as [24];
0
.
E
TP
V
f
V
d
V
(22
)
whe
r
e T is ti
me peri
od an
d
PV
is the win
d
turbin
e’s p
o
w
e
r cu
rv
e
Substitute e
q
uation
(2
) int
o
(2
2), th
e
wind
ene
rgy
in term
s
of Weib
ull di
stri
bution i
s
obtaine
d, and
is given by;
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 7, July 2013
: 3479 – 349
0
3484
1
ex
p
.
kk
KV
V
E
TP
V
d
V
CC
C
(23
)
The
wind tu
rbine’
s produ
ctivity or some
other
po
wer
gene
ration fa
cility is me
asured
by
an ele
m
ent
kno
w
a
s
ca
pacity facto
r
(
f
C
). It evaluates the
rati
o of the tu
rbine'
s a
c
tual
production to the rated power the turbine ru
nning with entire
capability for the same time
duratio
n [35]. It can be cal
c
ulated a
s
;
e
n
e
r
g
y
p
r
od
uc
e
d
(
W
h/
y
e
a
r
)
%1
0
0
r
a
t
e
d
w
i
nd
e
n
e
r
g
y
p
r
od
uc
e
d
(
W
h/
y
e
a
r
)
f
Wi
n
d
C
(24
)
3. Resul
t
s
and
Discus
s
ion
The
Ha
wksb
ay co
ast, n
e
a
r Ka
ra
chi, i
s
in
southe
rn
part
of Pa
ki
stan. T
he
ge
ogra
phi
c
locatio
n
of
wi
nd
spee
d m
e
asu
r
em
ent
site is
24°
52’
0
2
.025’’
N
an
d
66°
51’ 4
1
.98
3
’’ E.
The wi
nd
data con
s
ide
r
ed in thi
s
pa
per i
s
for
25
months from
Apr 20
09 to
Apr 20
11 m
e
asu
r
ed
every
10
minutes. T
h
e
averag
e ho
u
r
ly, daily and
monthly
win
d
spe
ed h
a
ve b
een
cal
c
ulate
d
at 10m, 3
0
m
,
60m and 8
0
m
heights
.
T
able 1 sh
ows Monthly mean wind
spe
e
d
for all 25 months. It ca
n be
observed
that
monthly
average
win
d
spe
ed at
80m
h
e
i
ght is mo
re
than
5m/s.
Th
e me
an
hou
rl
y,
daily and mo
nthly wind sp
eed value
s
at
different heig
h
ts are
sho
w
n in Figure 1.
Figure 1. Averag
e Win
d
Speed at Hawksbay, Kara
chi
(a) Hourly
(b) Daily
(c) Monthly
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
Asse
ssm
ent of wind po
we
r potent
ial at Ha
wksb
ay, Kara
chi Sind
h, Pakista
n
(Sh
ahna
wa
z FK)
3485
Table 1. Mont
hly mean win
d
Speed (m/
s
) at Ha
wksba
y
Time
(Mon
th
)
W
i
nd Spee
d (
m
/
s
)
80 m
60 m
30 m
10 m
Apr-09
6.215
5.999
5.481
5.041
Ma
y
-
0
9
6.605
6.437
6.018
5.673
Jun-09
6.852
6.681
6.246
5.883
Jul-09 7.643
7.390
6.825
6.325
Aug-09
7.346
7.172
6.778
6.392
Sep-09
6.574
6.331
5.698
5.231
Oct-09
4.521
4.345
3.940
3.567
Nov-09
4.782
4.595
4.046
3.416
Dec-09
4.923
4.682
4.048
3.405
Jan-10
4.447
4.267
3.768
3.213
Feb-10
5.162
4.974
4.471
3.876
Mar-10
5.641
5.464
5.037
4.644
Apr-10
6.739
6.526
5.976
5.520
Ma
y
-
1
0
7.928
7.656
7.038
6.577
Jun-10
7.336
7.081
6.491
6.060
Jul-10 6.897
6.736
6.316
5.958
Aug-10
6.149
5.996
5.521
5.157
Sep-10
5.511
5.310
4.837
4.439
Oct-10
4.474
4.319
3.903
3.479
Nov-10
4.724
4.529
4.001
3.414
Dec-10
5.210
4.896
4.204
3.484
Jan-11
4.926
4.721
4.176
3.582
Feb-11
4.924
4.752
4.247
3.704
Mar-11
5.302
5.086
4.620
4.214
Apr-11
5.821
5.625
5.128
4.740
The yearly mean wi
nd power
de
nsi
t
y obtained at Hawksb
ay
is 197.1
85W/m
2
,
179.30
4W/m
2
, 143.222
W/m
2
and 11
7.309
W/m
2
at 80m, 60m, 30m an
d 10m heig
h
ts
respe
c
tively. Thu
s
the
available
annu
al ene
rgy den
sity woul
d be
1727.3
3
6
k
Wh/m
2
,
1570.6
9
9
k
Wh/m
2
, 1254.6
22kWh/m
2
a
nd
1
027.6
2
4
k
Wh/m
2
of
rotor
are
a
fo
r the fo
ur hei
ghts
r
e
spec
tively.
Monthly wind
powe
r
den
si
ty and avera
ge ene
rg
y de
nsity of the rotor area at the four
height
s is sh
own in Fig
u
re 2 and Figu
re 3 re
sp
e
c
ti
vely. It can be noted that the wind po
wer
den
sity in
su
mmer du
ring
the mo
nths o
f
April to
A
u
g
u
st i
s
fai
r
ly g
ood. T
he ye
a
r
ly average
wind
spe
ed, mean
wind p
o
wer d
ensity and e
n
e
rgy den
sity
at different he
ights are give
n in Table 2.
The Weibull
Scale “
C
” an
d sha
pe “
k
” a
nd the mod
e
l
predi
ction p
e
r
forma
n
ce pa
ramete
rs
R
2
an
d
RMSE
are
dete
r
mi
ned th
rou
gh f
i
ve method
s f
o
r fou
r
diffe
re
nt height
s in
this
study. Th
e
Weib
ull para
m
eters and a
nalysi
s
re
sult
s are m
ention
ed in Table 3.
Figure 2. Monthly average
wind
po
we
r den
sity at Hawksb
ay
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 7, July 2013
: 3479 – 349
0
3486
Figure 3. Monthly energy
den
sity at Hawksb
ay
Table 2. Ann
ual mean
Win
d
spe
ed, win
d
power an
d energy den
sities at Ha
wksbay
Height
(m)
V
av
g.
(m/s)
Power densit
y
(W
/m
2
)
Energ
y
(k
W
h
/m
2
)
80 5.939
197.185
1727.336
60 5.742
179.304
1570.699
30 5.241
143.222
1254.622
10 4.767
117.309
1027.624
Table 3. Statistical a
nalysi
s
for wi
nd dat
a measured a
t
Hawksbay Kara
chi
Heigh
t
Metho
d
k
C
RMSE
R^2
80m
MLM 2.5139
6.5913
0.00004263
0.99999994
MM 2.5321
6.5976
0.00004223
0.99999994
GM
2.2159
5.9833
0.00048150
0.99999199
MMLM 2.5376
6.5942
0.00004280
0.99999994
EPFM 2.5004
6.5998
0.00004236
0.99999994
60M
MLM 2.4939
6.3662
0.00005225
0.99999987
MM 2.5073
6.3713
0.00005224
0.99999987
GM
2.1870
5.7868
0.00049397
0.99998859
MMLM 2.5167
6.3679
0.00005328
0.99999987
EPFM 2.4823
6.3728
0.00005142
0.99999988
30M
MLM 2.3410
5.8010
0.00007261
0.99999971
MM 2.3529
5.8039
0.00007352
0.99999970
GM
2.1609
5.1062
0.00055225
0.99998322
MMLM 2.3647
5.8041
0.00007497
0.99999969
EPFM 2.3436
5.8046
0.00007251
0.99999971
10M
MLM 2.0979
5.2784
0.00016402
0.99999789
MM 2.1065
5.2727
0.00016593
0.99999784
GM
2.0503
4.4907
0.00060251
0.99997149
MMLM 2.1225
5.2794
0.00016826
0.99999778
EPFM 2.1239
5.2733
0.00016903
0.99999776
It is obse
r
v
ed that the
maximum l
i
kelih
ood m
e
thod, empi
rical meth
od,
modified
maximum li
kelihoo
d meth
od a
nd
ene
rg
y pattern
fact
or m
e
thod, to
dete
r
mine
scale
an
d
sha
p
e
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
Asse
ssm
ent of wind po
we
r potent
ial at Ha
wksb
ay, Kara
chi Sind
h, Pakista
n
(Sh
ahna
wa
z FK)
3487
para
m
eters
of Weibull di
stributio
n, gi
ve the good
fitting result
s for wi
nd d
a
ta colle
cted
at
Ha
wksb
ay site. The grap
hi
cal metho
d
didn’t give goo
d fitting result
s with mea
s
u
r
ed wi
nd sp
e
e
d
at four height
s in this stu
d
y.
The Weibull
prob
ability di
stributio
ns
ob
tained fro
m
five method
s
were comp
ared to the
measured wi
nd data to find their su
itabi
lity. Figure 4 depi
cts the We
ib
ull pro
b
ability distribu
tions
of wind spee
ds over frequ
ency di
stributi
on of
wind
sp
eed
s at four d
i
fferent altitudes.
The
com
pari
s
on
bet
wee
n
Wei
bull
an
d Rayleigh
distrib
u
tion
o
f
wind
spee
ds
over
freque
ncy di
stribution at 8
0
m height, fo
r above me
nt
ioned five me
thods, is
sh
o
w
n in Fig
u
re
5
.
Usually at low wind velo
cities the Wei
bull dist
ri
butions fit poorly
, however, it is endurable
as
extremely sm
all power is g
enerat
ed at these win
d
sp
eed
s. It can
be ob
serve
d
that the Weib
ull
distrib
u
tion gi
ve better fitting than Raylei
gh distri
butio
n for wind
sp
eed data at th
e can
d
idate
site
Figure 4. Wei
bull probabilit
y distributions of
wind speeds at four
heights at Hawksbay
To cal
c
ulate t
he yearly ene
rgy pro
d
u
c
e
by a
wind turbine is o
ne of
the nece
s
sa
ry steps
to assess an
y wind energ
y
project. Th
e powe
r
(o
r energy) prod
uce
d
at
a given site by any
particula
r win
d
turbine d
e
p
end
s not onl
y upon t
he existence of good win
d
but
also on its h
ub
height, roto
r area a
nd the
cap
a
city facto
r
(o
r efficien
cy).
In this articl
e
the annu
al power g
ene
rati
on potenti
a
l of Nordex
N82/1
500 a
n
d
Suzlo
n
S66 wind tu
rbine
s
are
cal
c
ulate
d
as a
n
example.
Th
e detail of pa
ramete
rs
of Nordex N8
2/1500
[36] and Suzl
on S66 [37] is given in Tabl
e 4.
Annual
outpu
t energy for
Ha
wksb
ay, Kara
chi i
s
fou
nd to b
e
5.4
1
GWh
with theoretica
l
cap
a
city facto
r
of 0.411 thr
ough
Nordex N82/1
500 at
80m. Annual
output ene
rgy
expected to
be
gene
rated th
ough Su
zlo
n
S66 at the same
80m h
e
i
ght is 3.5
1
G
Wh
with th
eoreti
c
al
cap
a
city
factor of 0.3
20. The
an
n
ual o
u
tput e
nergy
expe
cted to
be
ge
nerate
d
th
ro
ugh th
e two
wind
turbine
s
, use
d
in this study
, at height 80m and 60m i
s
given in Tabl
e 5.
The
study a
bout a
s
sessment of wi
nd
ene
rgy pot
e
n
tial at Hawksbay
sho
w
s
that it is
suitabl
e site for win
d
po
we
r pro
d
u
c
tion p
r
oje
c
ts, like; small stand
-al
one sy
stems
and/or la
rge
off-
grid
or gi
rd
conne
cted
wi
n
d
farm
s i
n
Si
ndh Pa
ki
st
an.
It is
wo
rth m
entionin
g
h
e
re that a
d
e
tailed
eco
nomi
c
ev
aluation
mu
st be
done
bef
ore i
n
sta
lli
ng
a large
win
d
energy p
r
oje
c
t. Study sh
o
w
s
that there i
s
stron
g
wind
e
s
pe
cially i
n
summer
at
Ha
wksb
ay, wh
e
n
ele
c
tri
c
ity d
e
mand
am
plified
due to
cooli
n
g load
s. Th
u
s
, it coul
d be
benefi
c
ia
l to
install
wind
energy proje
c
ts to m
eet the
con
s
um
er’
s
requireme
nts.
(a) At 80m
(b) At 60m
(c) At 30m
(d) At 10m
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 7, July 2013
: 3479 – 349
0
3488
Figure 5. Co
mpari
s
o
n
of Weib
ull & Ra
yleigh dist
ri
bu
tion with mea
s
ured wi
nd d
a
ta at 80m he
ight
usin
g (a
) ML
M; (b) EM; (c) GM; (d) MM
LM; and (e
) EPFM
Table 4. Main
characte
ri
stics of wi
nd turbine
s
Turbi
ne m
ode
l
Nordex
N82/1500 [36]
Suz
l
on
S66 [37]
Rated po
wer (k
W)
1500
1250
Rotor diamete
r
(
m
)
82
66
Hub height (m
)
80
54,63,72
Cut-in
w
i
nd spee
d (m/s)
3.5
3
Rated
w
i
nd spee
d (m/s)
12
14
Cut-out
w
i
nd
speed (m/s)
25
22
Table 5. Esti
mated ann
ual
energy ge
ne
rated thr
oug
h
wind turbi
n
e
s
and
co
st/kWh at Ha
wksbay,
Karac
h
i
Turbine mod
e
l
Nordex
N82/150
0
Suzlon S66
Height (m)
80m
60m 80m 60m
Power Gene
rate
d
(kW)
617.19
561.23
399.83
363.58
Energ
y
P
r
oduce
d
(MWh)
5406.59
4916.37
3502.55
3184.96
Capacit
y
Facto
r
0.411
0.374
0.320
0.291
Cost/kWh (US$)
0.0345
0.0408
0.0443
0.0525
4. Cos
t
An
aly
s
is
The e
s
timatio
n
of the cost
s of kWh of e
ner
gy p
r
o
d
u
c
ed by the tu
rbine
s
ha
s b
e
en do
ne
with following as
sumptions:
(a) T
he lifetime of turbine ‘
t’
was a
s
sume
d to be 20 years.
(a) (b)
(c) (d)
(e)
Evaluation Warning : The document was created with Spire.PDF for Python.