Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
9
, No
.
3
,
J
un
e
201
9
, pp.
1585~1
597
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v9
i
3
.
pp
1585
-
15
97
1585
Journ
al h
om
e
page
:
http:
//
ia
e
s
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
Dynami
c model
of a
DC
-
DC qu
asi
-
Z
-
sou
rce c
onvert
er (q
-
ZS
C)
Muhamm
ad
Ado
1
, Aw
ang
Ju
s
oh
2
,
Abdul
ha
mi
d
Usm
an Mu
t
awak
kil
3
,
Tole S
ut
ikn
o
4
1,2
School
of El
ectrical
Engi
n
ee
rin
g,
Univer
si
ti T
ek
nologi
Ma
lay
sia
,
Malay
si
a
1,3
Ba
y
ero
Univ
er
sit
y
Kano, Nige
r
ia
4
Depa
rtmen
t of
El
e
ct
ri
ca
l
Eng
in
ee
ring
,
Univ
ersitas Ahm
ad
Dahlan,
Indone
si
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
ug
10, 201
8
Re
vised N
ov 20, 2
018
Accepte
d Dec
11, 201
8
Two
quasi
-
Z
-
so
urc
e
DC
-
DC
co
nver
te
rs
(q
-
ZSCs)
with
buck
-
bo
ost
conve
r
te
r
gai
n
w
ere
r
e
ce
nt
l
y
proposed
.
Th
e
conve
rt
ers
h
av
e
adva
nt
age
s
of
cont
inuous
gai
n
cur
ve
,
high
er
ga
in
m
agnitu
de
and
buck
-
boo
st
oper
a
ti
on
at
ef
fic
i
ent
d
u
t
y
rat
io
ran
ge
whe
n
compare
d
wi
t
h
exi
sting
q
-
ZS
Cs.
Acc
ura
te
d
ynamic
m
odel
s
of
the
se
converte
rs
ar
e
ne
eded
for
global
a
n
d
det
a
il
ed
ov
erv
ie
w
b
y
under
standi
ng
t
hei
r
op
erati
on
li
m
it
s
and
eff
e
ct
s
of
components
siz
es.
A
d
y
namic
m
odel
of
on
e
of
the
s
e
conv
ert
ers
is
pr
oposed
her
e
b
y
f
irst
der
iv
ing
the
g
ai
n
equatio
n,
sta
te
equation
s
and
sta
te
spac
e
m
odel
.
A
gen
er
al
i
ze
d
sm
al
l
signal
m
odel
wa
s
al
so
d
eri
v
ed
b
e
fore
loc
a
li
z
ing
it
to
th
is
topo
log
y
.
The
tra
nsfer
func
ti
ons
(
TF)
were
a
ll
der
iv
e
d,
th
e
po
le
s
an
d
ze
ros
anal
y
z
e
d
with
the
boundar
ie
s
for
st
abl
e
op
era
t
ions
p
rese
nte
d
and
disc
uss
ed.
Som
e
o
f
t
he
f
indi
ngs
inc
lud
e
exi
st
ence
of
righ
t
-
hand
pla
ne
(RHP
)
z
er
o
in
the
dut
y
r
atio
to
outpu
t
ca
pa
ci
tor
voltag
e
TF.
Thi
s
is
c
om
m
on
to
the
Z
-
source
and
qu
asi
-
Z
-
source
topol
ogie
s
and
impli
es
con
tr
ol
l
imitations.
Para
siti
c
r
esistance
s
of
th
e
ca
pa
ci
tors
and
in
duct
ors
aff
ect
th
e
na
ture
and
posi
ti
ons
of
th
e
p
o
le
s
and
ze
ros.
It
was
a
lso
foun
d
and
v
eri
f
ie
d
th
at
r
at
her
th
an
s
ym
m
et
ric
components,
use
of
ca
ref
u
lly
se
lected
sm
al
le
r
as
y
m
m
et
ric
components
L1
and
C1
p
roduc
es
l
ess
par
asitic
vol
ta
g
e
drop,
high
er
o
utput
voltage
a
nd
cur
ren
t
unde
r
the
sam
e
condi
ti
ons
,
thus
be
tt
er
eff
icien
c
y
and
p
erf
orm
anc
e
a
t
red
uc
ed
cost
,
si
z
e
and
weigh
t.
Ke
yw
or
d
s
:
Buck
-
bo
os
t c
onve
rter
Dynam
ic
m
od
el
Im
ped
ance s
ou
rce
Q
-
Z
SC
Sm
a
ll
sign
al
m
od
el
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Aw
a
ng J
usoh,
School
of Elec
tric
al
Engineer
ing
,
Un
i
ver
sit
i Te
knol
og
i M
al
ay
sia
,
81310 Jo
hor
B
ahru,
Mal
ay
sia
.
Em
a
il
: awang@u
tm
.
m
y
1.
INTROD
U
CTION
Im
ped
ance
source
co
nverter
s
(Z
SC/
IS
C)
c
ouple
c
onve
rter’
s
m
ai
n
ci
rcui
t
to
it
s
po
we
r
s
ource
[
1]
.
They
pro
vi
de
add
it
io
nal
feat
ur
es
not
obta
ined
i
n
pr
io
r
c
urren
t
fed
or
volt
age
fe
d
c
onver
te
r
s
su
c
h
as
dead
or
ov
e
rlap
tim
e in addit
ion
t
o
t
he
ir adva
ntages
[
2]
.
Applic
abili
ty
of
Z
-
sourc
e
co
nc
ept
to
ac
-
a
c
[3]
–
[5]
,
ac
-
dc
[6]
,
dc
-
ac
[
7]
–
[
12]
an
d
dc
-
dc
[13],
[
14]
,
[
23]
,
[15]
–
[22]
pow
er
c
onve
rsion
gen
e
rated
a
l
ot
of
i
nterest
a
nd
resea
rch
re
su
lt
ing
in
t
he
de
ve
lop
m
ent
of
vari
ant
an
d
new to
po
l
og
ie
s
[
24
]
. Fi
rst a
ppli
cat
ion
of
Z
S
C was t
he
Z
SI
for fuel cel
l a
ppli
cat
ion
[1]
t
he
n dr
i
ves
[4]
.
Re
fer
e
nce
[25]
pro
posed
a
m
od
i
fied
im
pedance
s
ource
c
onve
rter
(ZSC
)
cal
le
d
quasi
-
ZSC
(
q
-
ZSC)
sh
ow
n
in
Fig
ure
1
by
s
wappi
ng
t
he
po
sit
io
ns
of
switc
hes
a
nd
i
nducto
rs
t
o
s
olv
e
pro
blem
s
li
ke
disco
nt
inu
ou
s
input
cu
rr
e
nt,
hi
gh
ca
pacit
or
volt
age
re
quire
m
ent
fo
r
t
he
vo
lt
age
fed
Z
SCs
and
high
in
duc
tor
c
urren
t
re
quirem
ent
for
cu
rr
e
nt
fe
d
ZSCs.
M
os
t
of
ea
rly
ZSC
and
q
-
ZSCs
[4
]
,
[
26]
,
[
35
]
–
[
40]
,
[27]
–
[34]
fo
c
us
e
d
on
inv
erte
r
app
li
cat
io
ns
e
xcep
t
[
5]
on
ac
-
ac
c
onve
r
te
r
an
d
[6]
on
recti
fie
rs.
Re
fer
e
nce
[13]
exten
ded
Z
SC
and
q
-
Z
SC
c
on
ce
pt
to
DC
-
DC
a
pp
li
cat
ion
s
by
pr
opos
i
ng
f
our
non
-
isolat
ed
DC
-
DC
ZSC
a
nd
q
-
Z
SC
to
polo
gi
es
each,
then
[20],
[
22]
pro
po
se
d
is
olate
d
DC
-
DC
ZSCs
after
w
hi
ch
seve
ral
ot
he
r
isolat
ed
a
nd
non
-
isolat
ed
DC
-
DC
conve
rter top
ol
og
ie
s
h
a
ve bee
n pro
po
se
d.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
3
,
June
2019
:
1585
-
1597
1586
The
m
ajo
r
difference
betwee
n
im
ped
ance
s
ource
dc
-
ac
(i
nverter
s)
a
nd
D
C
-
DC
co
nverte
rs
li
es
on
how
the
outp
ut
is
ta
k
en
.
F
or
in
ve
rters,
it
is
ta
ke
n
a
cr
os
s
a
swit
ch
w
hile
for
t
he
DC
-
DC
c
onver
te
r
,
they
a
r
e
m
os
tl
y
ta
ken
ac
ross
a
capaci
tor
[
13
]
as
show
n i
n Fi
gure
1 (
b)
an
d
(c)
, a
lt
hough
[
15
]
,
[16],
[18],
[19],
[
21]
–
[23]
too
k t
he
ou
t
pu
t
ac
ro
s
s
a
switc
h
al
beit
with
ad
diti
on
al
com
po
ne
nt
s
in
w
hat
is
cal
le
d
P
W
M
DC
-
DC
im
ped
ance
source
conve
rters.
Re
fer
e
nce
[15]
a
naly
sed
t
he
ste
ady
-
sta
te
pe
rfo
rm
ance
of
su
c
h
c
onve
rters
i
n
co
ntinuo
us
c
onduct
io
n
m
od
e (CCM
).
Re
fer
e
nces
[
29
]
,
[41]
–
[
45]
ap
p
li
ed
sta
te
sp
ac
e
aver
a
ging
[46
]
and
Tay
lo
r’
s
series
ex
pa
ns
io
n
an
d
der
i
ved
the
sm
a
ll
sign
al
analy
sis
to
inv
est
ig
at
e
the
dy
nam
ic
char
ac
te
risti
cs
of
dif
f
eren
t
I
SI
to
polog
ie
s
.
Acc
ur
at
e
s
m
al
l
sign
al
m
od
el
i
s
need
e
d
to
obta
in
a
gl
ob
al
a
nd
detai
le
d
ov
erv
ie
w
of
syst
e
m
dynam
ic
s
by
unde
rstan
di
ng
syst
e
m
lim
it
s
and
c
ompone
nts
siz
es
[
44
]
.
It
is
ba
sed
on
the
as
su
m
ption
of
pe
rtu
rbat
ion
s
ar
ound
ste
ady
-
sta
te
operati
ng
po
i
nt
[
47]
.
Sm
al
l
sign
al
pe
rtu
rb
at
io
ns
(
.
.
̃
(
)
,
̃
(
)
,
̃
(
)
)
are
ap
plied
t
o
t
he
ste
ady
sta
te
du
ty
rati
o
(D)
a
nd
input
va
riables
(e.
g.
V
g
a
nd
I
g)
t
o
obta
in
th
e
sm
a
ll
sign
al
m
od
el
.
These
per
t
urbati
ons
c
auses
the
dyna
m
ic
sta
t
e
var
ia
bles
(
.
.
,
,
)
to
va
ry (by
̃
,
̃
,
̃
̃
resp
ect
i
vely
).
Use
of
sm
al
l
si
gn
al
m
od
el
s
to
ob
ta
in
dynam
i
c
m
od
el
s
f
or
co
ntr
oller
desig
n
m
akes
them
ve
ry
im
po
rtant.
They
a
re
al
s
o
us
e
d
to
obta
in
the
tra
ns
fe
r
f
unct
ions
betwee
n
sta
te
var
ia
ble
an
d
syst
em
i
nput
by
ass
umi
ng
ot
her
syst
e
m
inp
uts t
o be ze
ro
[
41
]
,
[44],
[48]
.
In
te
re
sti
ng
ly
,
t
he
existi
ng
publica
ti
on
s
on
dy
nam
ic
m
od
el
s
of
IS
Cs
[41],
[44],
[
45]
,
[
48
]
–
[
51]
fo
c
us
e
d
on in
ver
te
rs
. T
his is m
ai
nly du
e to t
he fact
hi
gh
li
ghte
d by
[
15
]
that m
ajorit
y of
the
li
te
ratur
e
on ISC
s fo
cuses o
n
the
in
ver
te
r
m
od
e
of
operati
on
al
th
ough
[52
]
work
e
d
on
P
WM
DC
-
DC
c
onve
rter.
DC
-
DC
ZSC/
q
-
Z
S
Cs
are
not
ver
y
popula
r
due
to
c
omm
on
def
ic
ie
ncies
li
ke
la
ck
of
buc
k
-
b
oost
ca
pab
il
it
y
at
the
eff
ic
ie
nt
duty
rati
o
ra
ng
e
of
[0.35
t
o
0.6
5]
[53]
,
disco
ntin
uous
gai
n
c
urv
e
an
d
hi
gh
e
r
c
om
po
ne
nts
c
ount
as
c
om
pared
with
the
t
ra
diti
on
al
bu
c
k
-
bo
os
t co
nverter
(
B
BC
)
.
Howe
ver, m
or
e findin
gs
are m
aking
IS
Cs
overc
om
e these ch
al
le
nges su
c
h
as
[
54]
, [55
]
wh
e
re th
e
gain
and
c
on
ti
nuou
s
gain
cu
r
ve
of
BB
C
wer
e
achieve
d
us
i
ng
non
-
i
so
la
te
d
q
-
ZSC
to
polo
gies.
Th
ese
t
opol
og
ie
s
pro
du
ce
d
high
er
m
agn
it
ud
e
ou
t
pu
t
volt
age
s
an
d
c
urre
nts
than
the
c
orr
esp
onding
buc
k
-
boos
t
c
onve
r
te
rs
th
us
giv
in
g
t
hem
p
ot
entia
l adv
a
nta
ges.
In
t
his
pa
pe
r,
t
he
c
on
ce
pt
of
dynam
ic
m
od
el
li
ng
is
exte
nded
to
t
he
DC
-
DC
q
-
ZSC.
Th
is
Extensi
on
is
i
m
po
rtant
bec
ause
their
a
ppli
cabil
it
y
is
increasin
g
whil
e
there
are
no
or
ve
ry
f
ew
existi
ng
dy
nam
ic
m
od
el
s o
f
t
he
m
.
The
m
od
el
li
ng
be
ga
n
by
first
consi
der
i
ng
a
n
ideal
ci
rcu
it
t
o
de
rive
the
ide
al
gain
e
quat
io
n.
Ne
xt,
no
n
-
sy
m
m
e
tric
,
r
ea
l
com
po
ne
nts
wer
e
c
on
si
der
e
d
rathe
r
t
ha
n
t
he
sim
ple
sy
m
m
et
ric
or
i
deal
q
-
ZSC.
T
he
us
e
of
non
-
sy
m
m
e
tric
com
po
nen
ts
al
lo
ws
ide
ntifyi
ng
the
ind
i
vidual
eff
ect
of
eac
h
com
po
ne
nt
w
hi
le
no
n
-
ideal
c
om
po
ne
nts
al
low
a
naly
zi
ng the
ef
fects
of the
par
asi
ti
c re
sist
anc
es
of
t
he
co
m
po
ne
nts.
As
com
m
on
to
ci
rcu
it
s
t
hat
c
ha
ng
e
ov
e
r
swit
chin
g
cy
cl
e,
sta
te
sp
ace
a
ver
a
gi
ng
[
46]
wa
s
us
ed
t
o
desc
ribe
the
ci
rcu
it
.
Stat
e
sp
ace
ave
ra
gin
g
requires g
e
ner
at
in
g
set
s
of
equ
at
io
ns
, w
it
h
each r
ep
rese
nting
a
s
witc
hi
ng
sta
te
[47]
an
d
t
hen a
ver
a
ge
d ov
e
r
t
he
s
witc
hing
pe
rio
d.
IS
Cs ca
n be c
ontr
olled
with
or w
it
hout s
hoot
-
thr
ough
[
34
]
or
op
e
n
sta
te
.
T
his c
onver
te
r was c
ontrolle
d
without
us
in
g
sh
oot
-
t
hro
ugh
or
ope
n
sta
te
s
in
order
to
ena
ble
fai
r
c
om
par
iso
ns
with
the
tradit
io
nal
buck
-
boos
t
conve
rter
wh
i
ch
is
ope
rated
us
in
g
on
ly
t
wo
switc
hi
ng
sta
te
s
(w
it
h
de
ad
-
ti
m
e)
since
they
ha
ve
i
den
ti
cal
gain
e
quat
io
n.
Find
i
ngs
f
ro
m
this
dy
nam
ic
m
od
el
li
ng
s
how
t
hat
the
pa
rasit
ic
resist
ances
of
the
ca
pacit
or
s
and
in
du
ct
or
s
are
am
ong
t
he
m
ajo
r
factors
t
hat
determ
ine
m
os
t
of
the
pol
es
a
nd
ze
ro
s
a
nd
ci
rc
uit
e
ff
ic
ie
ncy
as
detai
l
ed
i
n
the
discuss
i
on sect
ion
.
2.
CIRC
UIT
ANALYS
IS
This
sect
io
n
is
cl
assifi
ed
int
o
two:
gain
de
riv
at
ion
a
nd
sta
te
equ
at
io
ns
de
ri
vation.
Ci
rc
uit
analy
sis
was
done
us
i
ng
ide
al
an
d
real
ci
rc
uits
f
or
the
gai
n
a
nd
sta
te
e
quat
ion
s
de
rivati
on
res
pecti
vely
.
T
he
a
naly
ses
wer
e
do
ne
us
in
g
tw
o
s
witc
hing
m
od
es
with
re
sp
ect
t
o
S
1
w
hile
S
2
is
com
ple
m
entaril
y
switc
hed
w
it
h
res
pect
to
S
1
giv
i
ng
rise to tw
o
ope
rati
on
m
od
es s
how
n
in Fi
gure
2
. T
he du
ty
r
a
ti
o
of
t
he
m
od
es are
′
D
′
a
nd
′1
−
D
′
fo
r
m
od
e
s I
a
nd
II
res
pecti
vely
.
C
1
,
C
2
,
L
1
an
d
L
2
a
re
ca
pac
it
or
s
a
nd
in
duc
tors
with
c
urre
nts
I
C1
,
I
C2
,
I
L1
an
d
I
L2
,
a
nd
pa
rasit
ic
resist
ances
R
1
, R
2
,
r
1
,
an
d
r
2
re
sp
ect
ively
w
hile
V
g
,
I
g
,
R
O
a
nd
I
O
are
i
nput
volt
age,
in
put
current,
l
oad
r
es
ist
ance
and loa
d
c
urre
nt r
es
pecti
vely
.
2.1
.
G
ain
De
ri
vat
ion
Fo
r
sim
plici
ty
,
the
ideal
ci
rc
ui
t
of
Fi
gure
1
(
b) w
as
u
sed
t
o
de
rive
the
to
polog
y
’s
i
deal g
ai
n
eq
uatio
n
by
assum
ing
p
a
ras
it
ic
r
esi
sta
nces
R
1
, R
2
an
d r
1
, r
2
of the ca
pacit
or
s
and i
nduct
or
s
of F
i
gure
2 t
o be
neg
li
gib
l
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
D
yn
amic
mo
de
l of a DC
-
DC
quas
i
-
z
-
s
ource
con
ve
rte
r (q
-
Z
SC)
(
Awan
g
J
usoh
)
1587
Mod
e
1
:
In this
m
od
e as s
how
n
in
Fig
ure
2(b
),
S
1
is O
N w
hile S
2
is O
FF. T
he du
ty
rati
o for this m
od
e is
D.
V
L1
=
V
O
−
V
C1
(1)
V
L2
=
V
g
(2)
Mod
e
II
:
In
th
is
m
od
e,
S
1
is
OF
F
w
hile
S
2
is
O
N
as
s
ho
wn
in
Fi
gure
2(
c
).
T
he
duty
rati
o
for
this
m
od
e
is
D
′
=
1
−
D
.
V
L1
=
V
g
−
V
C1
(3)
V
L2
=
V
O
(4)
Applyi
ng Volt
-
Seco
nd
-
Ba
la
nc
e on L
1
a
nd L
2
yi
el
ds
V
̅
L1
=
D
V
O
+
V
g
−
V
C1
−
D
V
g
=
0
(5)
V
̅
L2
=
D
V
g
−
V
O
(
D
−
1
)
=
0
(6)
Fr
om
(
6),
V
O
=
−
D
1
−
D
V
g
(7)
(
7)
Is
t
he
ideal
ste
ady
-
sta
te
ou
tpu
t
vo
lt
age
f
or
this
co
nverte
r.
It
is
the
sam
e
as
the
ideal
ste
ady
sta
te
outp
ut
vo
lt
ag
e
of
buck
-
boos
t
conve
rter
wh
e
r
e
the
tw
o
s
witc
hes
a
re
switc
he
d
c
om
pli
m
entaril
y
and
D
is
th
e
duty
ra
ti
o
of
S
1
[55]
.
2.2
.
State
eq
ua
ti
on
s
deriv
ati
on
The
no
n
-
i
deal
ci
rcu
it
s
of
Fig
ur
e
2
we
re
us
e
d
t
o
de
rive
the
sta
te
eq
uatio
ns.
T
he
ci
rc
uit’s
two
ope
rati
on
m
od
es
are
pre
sented
i
n
Fi
gure
2(
b)
a
nd
Fi
gure
2(
c
)
a
nd
their
du
ty
r
at
io
s
are
"D"
a
nd
"1
−
D"
for
m
ode
I
a
nd
m
od
e II
res
pec
ti
vely
.
V
̇
C1
,
V
̇
C2
,
I
̇
C1
a
nd
I
̇
L2
are
the
sta
te
v
a
riables
wh
il
e
i
nput
vol
ta
ge (
V
g
), i
nput
curre
nt
(I
g
), a
n
d
ou
t
pu
t
c
urre
nt
(I
O
)
were
ch
ose
n
as
in
puts
w
hile
capaci
tor
vo
lt
age
s
V
C1
a
nd
V
C2
,
in
put
curr
ent
(I
g
)
an
d
outp
ut
vo
lt
age
(V
O
)
a
s
ou
t
pu
ts.
T
hi
s
is
to
ide
ntify
their
s
uitabil
it
y
f
or
c
on
tr
olle
r
desig
n
as
wi
ll
be
re
vealed
by
the
aver
a
ge
d
m
od
e
l.
V
S
C
2
C
1
L
1
S
1
L
2
S
2
(a)
C
2
L
1
S
1
L
2
S
2
R
O
V
g
C
1
(b)
S
1
L
1
S
2
L
2
C
1
(c)
Figure
1.
(a
) G
ener
ic
q
-
ZSC
(
b) D
e
rive
d DC
-
DC
q
-
ZSC
C
1
r
1
V
g
C
2
L
1
R
1
R
2
R
O
r
2
L
2
S
2
S
1
(a)
C
1
r
1
V
g
C
2
L
1
R
1
R
2
R
O
r
2
L
2
I
C
2
I
L
1
I
C
1
I
O
I
L
2
I
g
(b)
C
1
r
1
V
g
C
2
L
1
R
1
R
2
R
O
r
2
L
2
I
O
I
C
2
I
C
1
I
L
1
I
g
(c)
Figure
2.
(a
)
C
on
si
der
e
d
ci
rcui
t
with p
a
rasit
ic
r
esi
sta
nces
(b
)
Ci
rc
uit i
n
m
od
e
I
(c
)
Ci
rc
uit i
n
m
od
e
II
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
3
,
June
2019
:
1585
-
1597
1588
Mod
e I:
I
n
this
m
od
e
as
s
how
n
in Figure
2(b
),
S
1
is ON
w
hi
le
S
2
is
OF
F
.
L
2
is
charge
d
by
the
in
pu
t v
oltage
d
ue
to
the
res
ulti
ng
par
al
le
l
conne
ct
ion
.
T
he
loa
d,
C
1
,
L
1
an
d
C
2
are
al
l
isolat
ed
from
the
inp
ut
vo
lt
age
.
C
1
and
L
1
discha
rg
e
to
get
her to t
he
loa
d wh
il
e the
out
put fil
te
r
C
2
a
bsor
bs
t
he
ac
rip
ples. T
he
m
od
e
eq
uations are
V
̇
C1
=
I
L1
C
1
(8)
V
̇
C2
=
−
I
L1
C
2
−
I
O
C
2
(9)
I
̇
L1
=
−
V
C1
L
1
−
(
R
1
+
r
1
L
1
)
I
L1
+
I
O
R
O
L
1
(10)
I
̇
L2
=
−
I
L2
r
2
L
2
+
V
g
L
2
(11)
Ex
pr
essi
ng in s
ta
te
sp
ace
form
X
̇
i
=
A
i
X
+
B
i
U
w
he
re
i
=
1
for
m
od
e
1
yi
el
ds
X
̇
1
=
[
V
̇
C1
V
̇
C2
I
̇
L1
I
̇
L2
]
=
[
0
0
1
C
1
0
0
0
−
1
C
2
0
−
1
L
1
0
−
(
R
1
+
r
1
)
L
1
0
0
0
0
−
r
2
L
2
]
[
V
C1
V
C2
I
L1
I
L2
]
+
[
0
0
0
0
0
−
1
C
2
0
0
R
O
L
1
1
L
2
0
0
]
[
V
g
I
g
I
O
]
(12)
Fo
r
the
outp
ut,
V
C1
, V
C2
, I
g
an
d V
O
are
consi
der
e
d
a
nd the
outp
ut equati
ons
are
V
C1
=
V
C1
(13)
V
C2
=
V
C2
(14)
I
g
=
−
I
L2
(15)
V
O
=
I
O
R
O
(16)
Ex
pr
essi
ng
t
he
ou
t
pu
t
e
quat
io
ns
in
t
he
sta
te
sp
ace
f
or
=
+
w
her
e
i
ind
ic
at
es
the
m
od
e,
i
=
1
for
m
ode
1
a
nd
i
=
2
for
m
ode 2
.
Y
1
=
[
V
C1
V
C2
I
g
V
O
]
=
[
1
0
0
0
0
1
0
0
0
0
0
−
1
0
0
0
0
]
[
V
C1
V
C2
I
L1
I
L2
]
+
[
0
0
0
0
0
0
0
0
0
0
0
R
O
]
[
V
g
I
g
I
O
]
(18)
Mod
e
II
:
In
t
hi
s
m
od
e,
S
1
is
OF
F
w
hile
S
2
is
ON
a
s
s
how
n
in
Fig
ur
e
2(c
).
Durin
g
this
i
nter
val,
C
1
an
d
L
1
are
charge
d by the
input v
oltage
V
g
due t
o
the
s
eries co
nnect
io
n betwee
n
the
m
w
hile L
1
is i
so
la
te
d from
the supp
ly
.
L
1
disc
harges t
o
the
loa
d wh
il
e the
ou
t
pu
t
f
il
te
r
a
bs
or
bs
the
ri
pp
le
s.
V
̇
C1
=
I
L1
C
1
(19)
V
̇
C2
=
−
I
L2
C
2
−
I
O
C
2
(20)
I
̇
L1
=
−
V
C1
L
1
−
(
R
1
+
r
1
L
1
)
I
L1
+
V
g
L
1
(21)
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
D
yn
amic
mo
de
l of a DC
-
DC
quas
i
-
z
-
s
ource
con
ve
rte
r (q
-
Z
SC)
(
Awan
g
J
usoh
)
1
589
I
̇
L2
=
−
I
L2
r
2
L
2
+
I
O
R
O
L
2
(22)
Ex
pr
essi
ng in s
ta
te
sp
ace
form
X
̇
i
=
A
i
X
+
B
i
U
w
he
re
i
=
2
for
m
od
e
2
yi
el
ds
X
̇
2
=
[
V
̇
C1
V
̇
C2
I
̇
L1
I
̇
L2
]
=
[
0
0
1
C
1
0
0
0
0
−
1
C
2
−
1
L
1
0
−
(
R
1
+
r
1
)
L
1
0
0
0
0
−
r
2
L
2
]
[
V
C1
V
C2
I
L1
I
L2
]
+
[
0
0
0
0
0
−
1
C
2
1
L
1
0
0
0
0
R
O
L
2
]
[
V
g
I
g
I
O
]
(23)
The o
utput eq
ua
ti
on
s
for
m
ode II are:
V
C1
=
V
C1
(2
4)
V
C2
=
V
C2
(2
5)
I
g
=
−
I
L1
(2
6)
V
O
=
I
O
R
O
(2
7)
Ex
pr
essi
ng the
outp
ut in
t
he f
or
m
=
+
w
her
e
i
=
2
f
or m
od
e
2
yi
el
ds
Y
2
=
[
V
C1
V
C2
I
g
V
O
]
=
[
1
0
0
0
0
1
0
0
0
0
−
1
0
0
0
0
0
]
[
V
C1
V
C2
I
L1
I
L2
]
+
[
0
0
0
0
0
0
0
0
0
0
0
R
O
]
[
V
g
I
g
I
O
]
(28)
The
sta
te
e
quat
ion
s
are
t
hen a
ver
a
ge
d
a
nd express
ed
as
X
̇
=
AX
+
BU
(29)
Y
=
EX
+
FU
(30)
Wh
e
re
A
=
∑
A
i
D
i
,
n
i
=
1
B
=
∑
B
i
D
i
n
i
=
1
,
E
=
∑
E
i
D
i
n
i
=
1
,
F
=
∑
F
i
D
i
n
i
=
1
,
n
is
the
nu
m
ber
of
switc
hi
ng
sta
te
s
involve
d,
i
=
s
witc
hed
sta
te
and
D
is
t
he
du
t
y
rati
o
of
the
s
witc
hed
sta
te
.
Fo
r
this
ci
rc
uit,
n
=
2
si
nce
tw
o
switc
hing
sta
te
s
are
invo
lved
(as
in
ty
pi
cal
bu
c
k
-
boost
conver
te
r)
,
D
1
=
D
a
nd
D
2
=
D
′
=
1
−
D
fo
r
m
od
es
I
an
d
I
I
resp
ect
ively
.
The
refor
e
,
A
=
A
1
D
+
A
2
(
1
−
D
)
,
B
=
B
1
D
+
B
2
(
1
−
D
)
,
E
=
E
1
D
+
E
2
(
1
−
D
)
an
d
F
=
F
1
D
+
F
2
(
1
−
D
)
.
X
̇
=
[
V
̇
C1
V
̇
C2
I
̇
L1
I
̇
L2
]
=
[
0
0
1
C
1
0
0
0
−
D
C
2
−
(
1
−
D
)
C
2
−
1
L
1
0
−
(
R
1
+
r
1
)
L
1
0
L
1
0
0
0
−
r
2
L
2
]
[
V
C1
V
C2
I
L1
I
L2
]
+
[
0
0
0
0
0
−
1
C
2
(
1
−
D
)
L
1
0
DR
O
L
1
D
L
2
0
(
1
−
D
)
R
O
L
2
]
[
V
g
I
g
I
O
]
(31)
Y
=
[
V
C1
V
C2
I
g
V
O
]
=
[
1
0
0
0
0
1
0
0
0
0
−
(
1
−
D
)
−
D
0
0
0
0
]
[
V
C1
V
C2
I
L1
I
L2
]
+
[
0
0
0
0
0
0
0
0
0
0
0
R
O
]
[
V
g
I
g
I
O
]
(32)
(31)
And
(
32)
are
t
he
m
od
el
le
d
ave
rag
e
d
st
eady
-
sta
te
e
qu
at
ion
s
of
the
ci
rcu
it
.
T
he c
hoic
e
of
V
O
a
nd
I
O
as
outp
ut
and
i
nput
res
pe
ct
ively
re
s
ulted
in
the
feedf
orward
m
at
rices
in
(
18),
(
28)
a
nd
(
32)
no
nzero.
I
f
V
O
is
not
c
onside
red
as outp
ut,
al
l t
hese
fee
dforwa
rd m
at
rices will
b
e ze
ro.
How
ever, the
choic
e of
Ig as
bo
t
h ou
t
pu
t
an
d
i
nput d
i
dn’t
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
3
,
June
2019
:
1585
-
1597
1590
aff
ect
the
fee
dforwar
d
m
at
rice
s
nor
a
ny
in
pu
t
m
at
ri
x
beca
use
the
syst
em
’s
ste
ady
-
sta
te
re
sp
onse
is
i
nd
e
pende
nt
of the i
nput I
g
bu
t
V
g
a
nd I
O
.
This is im
po
rta
nt in
c
ontr
oller
d
esi
gn.
3.
SMALL
SIGNAL
A
NA
L
Y
SIS
Sm
a
ll
sign
al
pe
rturbati
ons
d
̃
(
t
)
,
v
̃
g
(
t
)
,
i
̃
g
(
t
)
and
i
̃
O
(
t
)
are
a
pp
li
ed
to
the
ste
ady
-
s
t
at
e
du
ty
rati
o
(
D)
a
nd
input
var
ia
bles
(
V
g
,
I
g
,
a
nd
I
O
)
res
pecti
vely
to
obta
in
the
s
m
al
l
sign
al
m
od
el
.
T
hese
pe
rturbati
ons
ca
use
the
dynam
ic
sta
te
var
ia
bles
v
C1
,
v
C2
,
i
L1
a
nd
i
L2
to
va
ry
by
v
̃
C1
,
v
̃
C2
,
i
̃
L1
an
d
i
̃
L2
resp
ect
ivel
y
[
56
]
.
T
he
relat
io
ns
hi
p
betwee
n
a
dynam
ic
v
ariable
x,
it
s stea
dy sta
te
v
al
ue X a
nd
per
t
urbati
on
x
̃
is give
n
a
s
x
=
X
+
x
̃
(33)
Diff
e
re
ntiat
ing (
33) wit
h respe
ct
to
ti
m
e y
ie
l
ds
x
̇
=
X
̇
+
x
̃
̇
(34)
Stea
dy
-
sta
te
va
riables in
(2
9)
are s
ubsti
tuted
with
dynam
ic
v
aria
bles for
s
m
al
l si
gn
al
an
a
ly
sis t
o
yi
el
d
(
35)
x
̇
=
(
A
1
d
+
A
2
(
1
−
d
)
)
x
+
(
B
1
d
+
B
2
(
1
−
d
)
)
u
(35)
Substi
tuti
ng (3
3) an
d (34
)
int
o (35),
neg
le
ct
i
ng pr
oducts
of
two
sm
al
l si
gnal
p
ert
urbati
on
s and
rear
rangi
ng
yi
el
ds
X
̇
+
x
̃
̇
=
AX
+
BU
+
A
x
̃
+
B
u
̃
+
[
(
A
1
−
A
2
)
X
+
(
B
1
−
B
2
)
U
]
d
̃
(36)
(36) Is
the
ge
ne
rali
sed
la
r
ge s
ign
al
sta
te
e
quat
ion
for
a m
odel
. Mat
chin
g
st
eady sta
te
and
per
t
urbati
on te
rm
s
tog
et
he
r
s
how
s
X
̇
=
AX
+
BU
=
0
(37)
x
̃
̇
=
A
x
̃
+
B
u
̃
+
[
(
A
1
−
A
2
)
X
+
(
B
1
−
B
2
)
U
]
d
̃
(38)
(37)
=
0
beca
use
der
i
vative
of
a
c
onsta
nt
(st
eady
sta
te
)
X
̇
=
0
.
(
37)
I
s
t
he
ge
neral
ise
d
st
ea
dy
st
at
e
m
od
el
w
hile
(38) is
gen
e
rali
sed
sm
al
l si
gn
a
l
m
od
el
.
Si
m
plifyi
ng
(3
6) furt
her
yi
el
ds
X
=
−
BU
A
−
1
(39)
Si
m
il
arly
, f
or
t
he
ste
a
dy stat
e
ou
t
pu
t
Y
=
EX
+
FU
,
it
s d
y
nam
ic
sign
al
aft
er s
m
al
l si
gn
al
an
al
ysi
s is g
i
ve
n
as
y
=
Y
+
y
̃
=
EX
+
FU
+
E
x
̃
+
F
u
̃
+
[
(
E
1
−
E
2
)
X
+
(
F
1
−
F
2
)
U
]
d
̃
(40)
(40) Is
the
ge
ne
rali
zed lar
ge
s
ign
al
ou
t
pu
t
equati
on for a m
od
el
. Mat
chi
ng
ste
ady stat
e an
d pert
urbati
on
te
rm
s
tog
et
he
r
s
how
s
Y
=
EX
+
FU
(41)
y
̃
=
E
x
̃
+
F
u
̃
+
[
(
E
1
−
E
2
)
X
+
(
F
1
−
F
2
)
U
]
d
̃
(42)
(41) Is
the
ge
ne
rali
zed s
te
a
dy stat
e outp
ut equati
on whil
e (
42)
is the
ge
ner
a
li
sed
sm
all sign
al
equati
on.
Substi
tuti
ng (3
9) into
(41) yi
e
lds
Y
=
−
EBU
A
−
1
+
FU
(43)
Y
=
(
F
−
EB
A
−
1
)
U
(44)
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
D
yn
amic
mo
de
l of a DC
-
DC
quas
i
-
z
-
s
ource
con
ve
rte
r (q
-
Z
SC)
(
Awan
g
J
usoh
)
1591
Evaluati
ng a
nd
sim
plify
ing
(38) yi
el
ds
(4
5)
t
o (48)
v
̃
̇
C1
=
i
̃
L1
C
1
(45)
v
̃
̇
C2
=
−
D
i
̃
L1
−
(
1
−
D
)
i
̃
L2
−
i
̃
0
+
(
I
L2
−
I
L1
)
d
̃
C
2
(46)
i
̃
̇
L1
=
−
v
̃
C1
−
(
R
1
+
r
1
)
i
̃
L1
+
(
1
−
D
)
v
̃
g
+
(
DR
0
)
i
̃
0
+
(
I
0
R
0
−
V
g
)
d
̃
L
1
(47)
i
̃
̇
L2
=
−
r
2
i
̃
L2
+
D
v
̃
g
+
(
1
−
D
)
R
0
i
̃
0
+
(
V
g
−
I
0
R
0
)
d
̃
L
2
(48)
Takin
g
La
plac
e trans
f
or
m
an
d
sim
plific
at
ion
yi
el
ds
sC
1
v
̃
c1
(
s
)
=
i
̃
L1
(
s
)
(49)
s
C
2
v
̃
c2
(
s
)
=
−
D
i
̃
L1
(
s
)
−
(
1
−
D
)
i
̃
L2
(
s
)
−
i
̃
0
(
s
)
+
(
I
L2
−
I
L1
)
d
̃
(
s
)
(50)
(
s
L
1
+
R
1
+
r
1
)
i
̃
L1
(
s
)
=
−
v
̃
C1
(
s
)
+
(
1
−
D
)
v
̃
g
(
s
)
+
(
DR
0
)
i
̃
0
(
s
)
+
(
I
0
R
0
−
V
g
)
d
̃
(
s
)
(51)
(
sL
2
+
r
2
)
i
̃
L2
(
s
)
=
i
̃
L2
(
s
)
+
D
v
̃
g
(
s
)
+
R
0
(
1
−
D
)
i
̃
0
(
s
)
+
(
V
g
−
I
0
R
0
)
d
̃
(
s
)
(52)
Fu
rt
her sim
plif
ic
at
ion
a
nd sub
sti
tuti
on
s yi
el
ds
v
̃
c1
(
s
)
=
(
1
−
D
)
v
̃
g
(
s
)
+
(
DR
0
)
i
̃
0
(
s
)
+
(
I
0
R
0
−
V
g
)
d
̃
(
s
)
(
s
L
1
+
R
)
(
sC
1
+
1
)
(53)
v
̃
c2
(
s
)
=
−
s
2
(
L
1
+
L
2
)
+
s
(
R
+
r
2
)
+
1
C
1
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
(
s
L
2
+
r
2
)
s
C
2
C
1
D
D
′
v
̃
g
−
s
3
L
1
L
2
+
s
2
(
L
2
(
D
2
R
0
+
R
)
+
L
1
(
R
0
D
′2
+
r
2
)
)
+
s
(
R
0
D
′2
R
+
r
2
(
R
+
D
2
R
0
)
+
L
2
C
1
)
+
R
0
D
′2
+
r
2
C
1
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
(
s
L
2
+
r
2
)
s
C
2
C
1
i
̃
0
(
s
)
+
s
3
L
1
L
2
I
+
s
2
(
L
1
r
2
I
+
L
2
RI
+
DL
2
V
−
D
′
L
1
V
)
+
s
(
R
r
2
I
+
L
2
C
1
I
+
D
r
2
V
+
D
r
2
−
D
′
RV
)
+
r
2
I
−
D
′
V
C
1
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
(
s
L
2
+
r
2
)
s
C
2
C
1
d
̃
(
s
)
(54)
i
̃
L1
(
s
)
=
S
C
1
(
1
−
D
)
v
̃
g
(
s
)
+
S
C
1
DR
0
i
̃
0
(
s
)
+
S
C
1
(
I
0
R
0
−
V
g
)
d
̃
(
s
)
s
2
C
1
L
1
+
sC
1
R
+
1
(55)
i
̃
L2
(
s
)
=
D
v
̃
g
(
s
)
+
R
0
(
1
−
D
)
i
̃
0
(
s
)
+
(
V
g
−
I
0
R
0
)
d
̃
(
s
)
s
L
2
+
r
2
(56)
w
he
re
R
=
R
1
+
r
1
,
V
=
V
g
−
V
0
,
D
′
=
1
−
D
an
d
I
=
I
L2
−
I
L1
.
The
sm
al
l
sign
al
equ
at
io
ns
of
the
sta
te
s
v
̃
c1
(
s
)
an
d
v
̃
c2
(
s
)
as
show
n
i
n
(
53)
a
nd
(54
)
ar
e
no
t
i
den
ti
cal
,
li
kew
ise
i
̃
L1
(
s
)
a
nd
i
̃
L2
(
s
)
as
s
how
n
i
n
(55
)
an
d
(
56)
are
al
so
non
ide
ntica
l.
A
n
ex
planati
on
to
this
no
n
-
i
den
ti
cal
it
y
is
due
t
o
t
he
a
sy
m
m
e
try
of
t
his
t
opology.
This
a
sym
m
et
r
y
is
ex
plaine
d
by
the
di
ff
e
re
nce
i
n
t
he
gai
n
c
urve
s
ob
ta
ine
d
w
hen
ta
kin
g
the
outp
ut
acr
os
s
C
1
a
s
done
i
n
[13]
an
d
wh
e
n
ta
ken
a
cro
ss
C
2
as
done
in
this
prese
nt
at
ion
.
Th
e
gain
of t
he
two va
riant t
opol
og
ie
s
sho
ws
that f
or a
ny g
i
ven ope
rati
on
a
l par
am
et
ers,
V
C1
≠
V
C2
.
The
m
od
el
s
pr
ese
nted
in
[
41]
,
[44]
hav
e
the
a
bove
-
m
entione
d
sta
te
s
t
o
be
ide
ntica
l
be
cause
i
nv
e
rter
s
we
re
c
onside
red
an
d
no
t
DC
-
DC
c
onve
rter
th
us
th
e
to
po
l
o
gies
a
re
e
ntirel
y
different.
Howe
ve
r,
the
pole
s
of
i
̃
L1
(
s
)
a
nd
i
̃
L2
(
s
)
are
con
ta
ine
d
i
n
t
he
poles
of
V
̃
C2
(
s
)
th
us (
55)
a
nd (5
6)
cou
l
d be
re
-
wr
i
tt
en
as
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
3
,
June
2019
:
1585
-
1597
1592
i
̃
L1
(
s
)
=
S
C
1
(
1
−
D
)
v
̃
g
(
s
)
+
S
C
1
DR
0
i
̃
0
(
s
)
+
S
C
1
(
I
0
R
0
−
V
g
)
d
̃
(
s
)
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
(
s
L
2
+
r
2
)
s
C
2
C
1
(
s
L
2
+
r
2
)
s
C
2
C
1
(55a)
i
̃
L2
(
s
)
=
D
v
̃
g
(
s
)
+
R
0
(
1
−
D
)
i
̃
0
(
s
)
+
(
V
g
−
I
0
R
0
)
d
̃
(
s
)
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
(
s
L
2
+
r
2
)
s
C
2
C
1
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
s
C
2
C
1
(56
b)
3.1
.
Tr
an
s
fer
f
unc
tio
ns
The
sm
all
signa
l
m
od
el
s
pres
ented
in
(
53)
to
(56
)
wer
e
use
d
t
o
obta
in
t
he
tra
ns
fe
r
func
ti
on
s
(
G
inp
u
t
̃
s
t
at
e
̃
)
betwee
n
sta
te
var
ia
ble
a
nd
sy
stem
input.
Thi
s
was
done
by
consi
der
i
ng
on
e
syst
em
inp
ut
at
a
ti
m
e
and
a
ssu
m
ing
oth
e
r
syst
em
inp
uts
to be ze
ro
[41],
[44],
[48]
.
̃
̃
1
=
(
1
−
)
(
1
+
)
(
1
+
1
)
(57)
G
i
0
v
̃
C1
=
DR
0
(
s
L
1
+
)
(
sC
1
+
1
)
(58)
G
d
̃
v
̃
C1
=
(
I
0
R
0
−
V
g
)
(
s
L
1
+
)
(
sC
1
+
1
)
(59)
G
v
̃
g
v
̃
C2
=
−
[
s
2
(
L
1
+
L
2
)
+
s
(
R
+
r
2
)
+
1
C
1
]
D
D
′
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
(
s
L
2
+
r
2
)
s
C
2
C
1
(60)
G
i
̃
0
v
̃
C2
=
−
s
3
+
s
2
(
(
L
2
(
D
2
R
0
+
R
)
+
L
1
(
R
0
D
′2
+
r
2
)
L
1
L
2
)
+
s
(
R
0
C
1
D
′2
R
+
r
2
C
1
(
R
+
D
2
R
0
)
+
L
2
L
1
L
2
C
1
)
+
R
0
D
′2
+
r
2
L
1
L
2
C
1
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
(
s
L
2
+
r
2
)
s
C
2
C
1
(61)
G
d
̃
v
̃
C2
=
s
3
L
1
L
2
I
+
s
2
(
L
1
r
2
I
+
L
2
RI
+
DL
2
V
−
D
′
L
1
V
)
+
s
(
R
r
2
I
+
L
2
C
1
I
+
D
r
2
V
+
D
r
2
−
D
′
RV
)
+
r
2
I
−
D
′
V
C
1
(
s
2
C
1
L
1
+
sC
1
R
+
1
)
(
s
L
2
+
r
2
)
s
C
2
C
1
(62)
G
v
̃
g
i
̃
L1
=
S
C
1
(
1
−
D
)
s
2
C
1
L
1
+
sC
1
R
+
1
(63)
G
i
̃
0
i
̃
L1
=
S
C
1
DR
0
s
2
C
1
L
1
+
sC
1
R
+
1
(64)
G
d
̃
̃
L1
=
S
C
1
(
I
0
R
0
−
V
g
)
d
̃
s
2
C
1
L
1
+
sC
1
R
+
1
(65)
G
v
̃
g
̃
L2
=
D
s
L
2
+
r
2
(66)
G
i
̃
0
i
̃
L2
=
R
0
(
1
−
D
)
s
L
2
+
r
2
(67)
G
d
̃
i
̃
L2
=
(
V
g
−
I
0
R
0
)
d
̃
(
s
)
s
L
2
+
r
2
(68)
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
D
yn
amic
mo
de
l of a DC
-
DC
quas
i
-
z
-
s
ource
con
ve
rte
r (q
-
Z
SC)
(
Awan
g
J
usoh
)
1593
4.
A
NA
L
YS
I
S
The
pole
s
an
d
zero
s
of
the
tra
ns
fe
r
f
un
ct
i
on
s
are
discusse
d
i
n
t
his
sect
ion.
Ro
ots
of
f
unct
ion
s
not
gr
eat
e
r
than
de
gr
ee
2
a
re
f
ully
disc
us
s
ed
wh
il
e
t
ho
se
of
de
gr
e
e
3
a
nd
4
are
just
i
ntrodu
ce
d
due
t
o
th
e
com
plexity
inv
ol
ve
d.
Po
le
-
ze
ro
m
aps
hav
e
bee
n
use
d
to
a
naly
se
dy
nam
ic
m
od
el
s
of
dc
-
ac
I
SCs
[41],
[
44
]
,
[45],
[
48
]
,
[49],
[51]
,
analy
ti
cal
m
et
h
od is
us
e
d here
for analy
ses
due t
o
the
asym
m
et
ry o
f
t
his to
po
l
og
y
w
hich
re
su
lt
ed
i
n (62
) havi
ng
so
m
any v
aria
bl
es.
a.
Starti
ng
with
t
he
fi
rst
tran
sfe
r
f
unct
ion
G
V
̃
g
V
̃
C1
,
to
ge
ther
with
G
i
̃
0
V
̃
C1
a
nd
G
d
̃
V
̃
C1
,
they
ha
ve
tw
o
pole
s
al
l
neg
at
ive
locat
ed
at
s
=
−
R
1
+
r
1
L
1
a
nd
s
=
−
1
C
1
. T
hey all
h
a
ve
no ze
ro.
b.
G
V
̃
g
V
̃
C2
tog
et
her
with
G
i
̃
0
V
̃
C2
a
nd
G
d
̃
V
̃
C2
,
they
hav
e
f
our
pole
s
al
l
non
-
po
sit
iv
e
lo
cat
ed
at
0
,
−
r
2
L
2
an
d
−
2
L
1
±
√
(
R
L
1
)
2
−
(
4
C
1
L
1
)
2
. T
he
pole
−
R
2
L
1
+
√
(
L
1
)
2
−
(
4
C
1
L
1
)
2
is
al
so
non
posit
iv
e
beca
us
e
(
L
1
)
2
−
4
(
1
C
1
L
1
)
4
≯
(
R
2
L
1
)
2
⟹
−
1
C
1
L
1
≯
0
s
ince
C
1
a
nd
L
1
are
al
l
posit
ive.
P
r
ov
i
ded
(
L
1
)
2
≥
4
(
1
C
1
L
1
)
,
or
sim
ply
L
1
≤
C
1
4
R
2
,
it
has
al
l
real
non
-
po
sit
ive
po
le
s
.
It
has
two
ze
ros
at
−
+
r
2
2
(
L
1
+
L
2
)
±
√
(
R
+
r
2
2
(
L
1
+
L
2
)
)
2
−
(
1
C
1
(
L
1
+
L
2
)
)
al
l
the
zer
os
are
neg
at
ive
beca
us
e
(
+
r
2
2
(
L
1
+
L
2
)
)
2
−
(
1
C
1
(
L
1
+
L
2
)
)
≱
(
R
+
r
2
2
(
L
1
+
L
2
)
)
2
since
−
1
C
1
(
L
1
+
L
2
)
≱
0
due
t
o
the
fact
that
both
C
1
,
L
1
a
nd
L
2
a
re
al
l
po
sit
ive
.
c.
G
i
̃
O
v
̃
C2
has t
hr
ee
zer
os
an
d t
hei
r
loca
ti
on
s
ca
n
be a
na
ly
sed base
d
on
the
gi
ven
ope
rati
ng
co
ndit
ion
s
beca
us
e
th
e
po
ly
nom
ia
l
being
of
de
gr
ee
three
an
d
with
so
m
any
pa
ra
m
et
ers
m
akes
it
diff
ic
ult
to
present
a
gen
e
r
al
ise
d
analy
sis.
d.
G
d
̃
v
̃
C2
has
th
ree
zer
os
and
a
re
gi
ve
n
by
the
roots
of
the
poly
nom
ia
l
s
3
L
1
L
2
I
+
s
2
(
L
1
r
2
I
+
L
2
RI
+
DL
2
V
−
D
′
L
1
V
)
+
s
(
R
r
2
I
+
L
2
C
1
I
+
D
r
2
V
+
D
r
2
−
D
′
RV
)
+
r
2
I
−
D
′
V
C
1
.
An
al
ysi
ng
th
e
be
ha
viour
of
al
l
the
possible
r
oots
of
t
his
cu
bic
poly
no
m
ia
l
analy
ti
cal
l
y
is
co
m
plex
a
nd
in
vo
l
ves
s
o
m
uch
m
at
hem
atics
beyond
th
e
sco
pe
of
t
his
pap
e
r
beca
us
e
I
a
nd
V
a
re
var
ia
bles
who
se
val
ues
va
ry
f
or
dif
fere
nt
op
e
rati
ng
po
i
nt
s.
This
is
e
vid
en
t
as
[41],
[
44
]
al
so
analy
se
d
the
ir
qu
a
drat
ic
G
̃
v
̃
C1
by
consi
der
i
ng
t
he
par
am
et
ers
of
a
giv
e
n
ci
rc
uit
under
giv
e
n
conditi
ons. H
oweve
r, l
i
m
it
ed
cases wil
l be
c
on
si
der
e
d suc
h as
I
L1
=
I
L2
. If
I
L1
=
I
L2
,
the
po
ly
no
m
ia
l red
uces
to
d
e
gr
ee
tw
o
as
s
2
+
s
(
D
r
2
+
D
r
2
V
−
D
′
R
DL
2
−
D
′
L
1
)
−
D
′
(
DL
2
−
D
′
L
1
)
C
1
. It
s ro
ots ar
e
giv
e
n by
−
D
r
2
+
D
r
2
V
−
D
′
R
2
(
DL
2
−
D
′
L
1
)
±
√
(
D
r
2
+
D
r
2
V
−
D
′
R
2
(
DL
2
−
D
′
L
1
)
)
2
−
D
′
(
DL
2
−
D
′
L
1
)
C
1
.
I
L1
=
I
L2
an
d
V
g
=
V
O
,
the e
qu
at
i
on r
e
duces to
s
D
r
2
thu
s
the
zer
o
e
xi
st at
o
rigi
n (s
= 0).
As
s
how
n
by
t
hese
tw
o
cases
,
the
nature
of
the
zer
os
var
ie
s
for
diff
e
re
nt
po
i
nts.
An
im
po
rta
nt
point
to
no
t
e
is
that
rig
ht
-
ha
nd
plane
(R
HP)
ze
ro
m
ay
exist
outsi
de
the
conditi
ons
of
c
ase
I
I.
T
he
exi
ste
nce
of
t
his
RHP
zero
was
al
s
o
noti
ced
in
ZSI
a
nd
q
-
ZS
I
w
hic
h
im
plies
con
tr
ol
lim
it
a
ti
on
s
a
nd
hi
gh
gain
in
sta
bili
ty
[4
1],
[
44
]
,
[45],
[48],
[49]
there
by
destab
il
iz
ing
the
feedback
lo
op.
e.
G
v
̃
g
i
̃
L1
,
G
i
̃
O
i
̃
L1
a
nd
G
d
̃
i
̃
L1
hav
e
al
l
ne
gative
p
oles
locat
e
d
at
−
2
L
1
±
√
(
R
L
1
)
2
−
(
4
C
1
L
1
)
2
.
The
po
le
s
are
al
l
neg
at
iv
e
becau
s
e
(
L
1
)
2
−
(
4
C
1
L
1
)
≱
(
R
L
1
)
2
since
−
4
C
1
L
1
≱
0
.
The
pol
es
are
real pro
vid
e
d
L
1
≤
C
1
(
2
)
2
.
They
al
l
hav
e
si
ng
le
z
e
ro
and is locat
e
d at
s
=
0
.
f.
G
v
̃
g
i
̃
L2
,
G
i
̃
O
i
̃
L2
a
nd
G
d
̃
i
̃
L2
ha
ve
a
sin
gle
po
le
a
nd
no zer
o.
The
pole is l
ocated at
s
=
−
r
2
L
2
.
Fr
om
the
ab
ov
e
analy
sis,
it
ca
n
be
de
duced
t
hat
the
tra
nsfer
f
unct
ion
s
G
v
̃
g
v
̃
C1
,
G
i
̃
O
v
̃
C1
a
nd
G
d
̃
v
̃
C1
de
ri
ved
from
the
st
ate
v
̃
c1
(
s
)
and
G
v
̃
g
i
̃
L2
,
G
i
̃
O
i
̃
L2
a
nd
G
d
̃
i
̃
L2
de
rive
d
f
ro
m
the
sta
te
i
̃
L2
(
s
)
are
ge
ner
al
ly
sta
ble
re
gardless
of
par
am
et
er
values
.
All
their
pole
s
are
ne
gat
ive
-
r
eal
and
ha
ve
no
zer
os
.
Sm
a
ll
er
L
1
an
d
C
1
increase
the
s
ta
bili
ty
of
the
trans
fer
functi
ons
G
v
̃
g
v
̃
C1
,
G
i
̃
O
v
̃
C1
a
nd
G
d
̃
v
̃
C1
by
push
in
g
their
po
l
es
a
way
from
ori
gin
.
Also,
sm
al
le
r
L
2
will
in
crease
the
syst
e
m
sta
bili
ty
du
e
t
o
G
v
̃
g
i
̃
L2
,
G
i
̃
O
i
̃
L2
a
nd
G
d
̃
i
̃
L2
by
pu
s
hing
t
heir
po
le
s
further
a
way
from
the
or
i
gin.
S
m
al
le
r
values
of
L
2
rather
tha
n
la
r
ge
r
values
of
r
2
are
pr
e
ferre
d
be
cause
r
2
bei
ng
a
parasi
ti
c
resi
sta
nce
will
incr
ease
non
-
i
deali
ty
su
c
h
as p
a
rasit
ic
volt
age dr
op there
by r
e
duci
ng ef
f
ic
ie
ncy.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
3
,
June
2019
:
1585
-
1597
1594
The
t
ran
s
fe
r
functi
ons
G
v
̃
g
i
̃
L1
,
G
i
̃
O
i
̃
L1
a
nd
G
d
̃
i
̃
L1
de
rive
d
from
the
sta
te
i
̃
L1
(
s
)
hav
e
m
arg
ina
l
gain
sta
bili
ty
du
e
to
zer
o
a
t
the
ori
gin
w
hich i
m
plies
con
tr
ol
li
m
it
a
ti
on
[
44]
. Alt
ho
ugh al
l
their
pole
s
ar
e
al
l
neg
at
ive,
os
ci
ll
at
ion
s
m
a
y
occur
if
L
1
>
C
1
(
2
)
2
du
e
t
o
th
e
e
xistence
of
a
c
om
plex
co
njug
at
e
pair,
el
se,
t
he
po
le
s
are
ne
gative
a
nd
re
al
with
a
sm
a
ll
er
value o
f
L
1
push
i
ng th
e
m
f
ur
t
her awa
y from
the o
ri
gi
n.
It
is
now
cl
ear
that
the
tra
ns
fe
r
functi
ons
G
v
̃
g
v
̃
C2
,
G
i
̃
O
v
̃
C2
a
nd
G
d
̃
v
̃
C2
deri
ved
f
ro
m
v
̃
c2
(
s
)
are
the
m
os
t
cru
ci
al
beca
us
e
they
ind
ic
at
e
m
arg
inal
sta
bil
it
y
du
e
to
the
existe
nce
of
a
po
le
at
or
igi
n
and
os
ci
ll
at
ion
m
ay
occu
r
if
L
1
>
C
1
4
2
du
e
to
the
e
xistence
of c
om
plex
c
onju
gate
pole
p
ai
r
. T
he
zer
os
of
G
v
̃
g
v
̃
C2
are all
ne
gative.
F
ro
m
all
the a
bove
analy
sis,
it
sh
ows
that
the
pos
sibil
it
y
of
posit
ive
r
oo
ts
only
exists
in
the
ze
ro
s
of
G
i
̃
O
v
̃
C2
a
nd
G
d
̃
v
̃
C2
wh
ic
h
sign
i
fies
con
t
ro
l
lim
it
ati
on
a
nd
high
ga
in
i
ns
ta
bili
ty
a
nd
al
s
o
exists
in
t
he
ZS
I
a
nd
q
-
Z
SI.
This
s
hows
t
hat
t
he
Z
SI
,
q
-
Z
SI
and this
DC
-
D
C q
-
ZSC a
re
non
-
m
ini
m
u
m
p
hase s
yst
em
s
[48]
.
5.
VERIF
IC
ATI
ON
To
ve
rify
the
se
fin
dings
,
oper
at
ion
s
of
t
wo
c
onve
rters w
e
re
com
par
e
d
by
s
i
m
ulati
ng
their
pe
rfor
m
ance
on
in
put
volt
ag
e
V
g
=
12
V
,
du
ty
rati
o
D
=
0
.
63
a
nd
7
Ω
load
us
i
ng
MATLAB
S
I
MULI
NK.
O
n
on
e
side
was
a
conve
rter
ba
se
d
on
ar
bitra
ry
sy
m
m
e
tric
com
po
nen
ts
as
1
=
2
=
400
,
1
=
2
=
500
,
1
=
2
=
0
.
03
Ω
,
1
=
2
=
0
.
47
Ω
wh
il
e
on
t
he
oth
e
r
was
an
oth
e
r
c
onve
rter
with
car
efu
ll
y
sel
ect
ed
asy
m
m
et
ric
com
po
ne
nts
ba
sed
on
the
opti
m
iz
at
ion
equ
at
ion
s
de
rive
d
in
(62
)
by
only
m
od
ify
ing
the
op
ti
m
iz
ation
c
apacit
or
and
in
du
ct
or
t
o
C
1
=
80
μ
F
an
d
L
1
=
4
μ
H
as
show
n
i
n
Table
1.
T
he
new
sm
aller
va
lues
of
C
1
a
nd
L
1
push
es
the poles
of
G
V
̃
g
V
̃
C1
,
G
d
̃
V
̃
C1
a
nd
G
i
̃
O
V
̃
C1
further away
on t
he
le
ft
ha
nd p
l
ane
(LHP)
.
Althou
gh
the
va
lues o
f
R
1
an
d
r
1
a
re p
r
oport
ion
al
t
o
C
1
a
nd R
1
resp
e
ct
ively
,
an
d
each
can
inf
lue
nce
t
he
po
sit
io
n,
the
c
hoic
e
of
sm
al
le
r
L
1
a
nd
C
1
are
prefe
rr
e
d
due
to
the
inef
fic
ie
nc
y
associat
ed
wi
th
par
a
sit
ic
resist
ances
and
oth
er
c
on
st
raints
s
uc
h as
weig
ht
a
nd
siz
e
ass
ociat
ed
with
la
r
ger
ca
pac
it
or
s
a
nd
in
duc
tors.
T
he
ne
w
values
of
C
1
an
d
L
1
al
s
o ens
ur
es
that
th
e
pole
s
of
the
TFs
of
v
̃
C2
a
nd
i
̃
L1
are
r
ea
l
an
d
no
n
-
posit
iv
e
i
ns
te
ad
o
f
the
c
om
plex
po
le
that
e
xist
ed
from
C
1
=
400
μ
F
an
d
L
1
=
500
μ
H
.
T
he
r
esp
on
se
of
the
tw
o
ci
rc
uits
w
it
h
res
pect
t
o
outp
ut
vo
lt
age
(
V
O
)
,
ou
t
pu
t
c
urre
nt
(I
O
)
an
d
i
nput
current
(I
g
)
a
re
presente
d
in
F
igure
3.
Fig
ur
e
4
(a)
sho
ws
t
he
ideal
gain
c
ur
ve
of t
he
c
onve
rter.
Their
operati
ons
wer
e
al
s
o
c
om
par
ed
us
in
g
ideal
com
po
ne
nts
by
neg
le
c
ti
ng
the
pa
rasit
ic
resist
ances
R
1
,
R
2
,
r
1
a
nd
r
2
for
both
t
he
optim
iz
ed
and
s
ymm
et
ric
ci
rcu
it
s
in
order
t
o
com
par
e
their
ou
t
pu
t
vo
lt
ag
e
s
with
the
ideal
ste
ad
y
sta
te
ou
tp
ut
volt
age
of
(7)
a
nd
ide
ntify
the
e
ff
ect
s
of
t
he
pa
rasit
ic
resist
an
ces
as
sho
wn
i
n
Fig
ure
4(b) an
d (c)
.
Tab
le
1.
Param
et
er v
al
ues use
d for
sim
ulati
on
Para
m
eter
Valu
e
Sy
m
m
et
ric
Op
ti
m
ized
V
g
(
V
)
12
12
D
0
.63
0
.63
f
(
KHz
)
100
100
C
1
(
μ
F
)
400
80
C
2
(
μ
F
)
400
400
L
1
(
μ
H
)
500
4
L
2
(
μ
H
)
500
500
R
1
(
Ω
)
0
.03
0
.03
R
2
(
Ω
)
0
.03
0
.03
r
1
(
Ω
)
0
.47
0
.47
r
2
(
Ω
)
0
.47
0
.47
Load
(
Ω
)
7
7
6.
RE
SULTS
AN
D DISC
US
SION
Re
su
lt
s
of
Fi
gure
3
c
onfirm
the
validit
y
of
th
ese
equ
at
i
on
s
be
cause
the
ou
t
put
volt
a
ge
an
d
ou
t
pu
t
c
urren
t
of
the
optim
ized
ci
rc
uit
are
15
.25
V
a
nd
2.1
8
A
a
gainst
13.15
V
a
nd
1.8
7
A
ob
ta
i
ne
d
without
opti
m
iz
at
ion
resp
ect
ively
.
T
his
is
becau
se
the
op
ti
m
iz
a
ti
on
ca
pacit
or
C
1
and
in
du
ct
or
L
1
wer
e
sel
ect
ed
base
d
on
the
e
qu
at
io
ns
der
i
ved
from
thi
s
m
od
el
as
discu
ssed
in
the
A
naly
sis
an
d
Ve
rificat
ion
sect
i
on
s
r
at
her
t
han
sym
m
e
try
.
This
i
ncr
ease
r
epr
ese
nts
a
m
a
gn
it
ude
inc
reas
e
of
16.
35
%
a
nd
16.58%
f
or
the
ou
t
pu
t
volt
age
a
nd
ou
t
pu
t
cu
rr
e
nt
Evaluation Warning : The document was created with Spire.PDF for Python.