Grid Converters For Photovoltaic and Wind Power Systems
Grid Converters For Photovoltaic and Wind Power Systems
Grid Converters For Photovoltaic and Wind Power Systems
and Wind Power Systems
by R. Teodorescu, M. Liserre and P. Rodriguez
ISBN: 978‐0‐470‐05751‐3
Copyright Wiley 2011
Chapter 8
Chapter 8
Grid Synchronization in Three‐Phase
P
Power Converters
C t
Outline
• IIntroduction
t d ti
• Three‐phase voltage vector under grid faults
• Types of voltage sags
Types of voltage sags
• Synchronous Reference Frame PLL (SRF‐PLL)
• Decoupled Double Synchronous Reference
Decoupled Double Synchronous Reference
Frame PLL (DDSRF‐PLL)
• Double Second Order Generalized Integrator
g
FLL (DSOGI‐FLL)
• Conclusions
2
Introduction
• One of the most important aspects to consider in the control of power
converters connected to electrical grids is the proper synchronization with
the three‐phase utility voltages
• The three‐phase voltage should be understood as a vector consisting of three
voltage components
• The module and the rotation speed of the three‐phase grid voltage vector
keep constant when balanced sinusoidal waveforms are present in the three
phases of the system –with equal amplitude, frequency and relative phase
shifting
• The three‐phase grid voltage vector consists of two sequence components
(positive and negative) during grid faults, which result in oscillations in the
module and the rotation speed
p
• The real‐time detection of the sequence‐components of the voltage vector
during grid faults is an essential issue in the control of distributed generation
systems
3
Three‐phase voltage vector under grid faults
• Positive‐sequence voltage vector at the fundamental
frequency interacting with an either positive
positive‐ or
v 1 v 1 negative‐sequence nth order component :
v
1
v
t cos(t ) cos(nt )
1
t v abc v abc
1
v nabc V 1 cos(t 23 ) V n cos(nt 23 )
v
cos(t 23 ) cos(nt 23 )
v 1 v 1
v 1 v 1 • Voltage vector on the d‐q reference frame:
q 3 1 1 3 cos n 1 t
v dq V Vn
b 2 0 2 sin n 1 t
v
v 1 d • Amplitude and phase‐angle:
v 5 t 3 1 2
a
v v2 v2
2 V V
n 2
2V 1V n cos n 1 t
v 5
1 v V n sin n 1 t
v tan t tan 1
1 1
v V V n
cos n 1 t
c 4
Three‐phase voltage vector under grid faults
• Example:
2
1.5
Positive‐ and negative‐
[pu]
sequence
q components
p are 1
p.u.)
0.75
|v| [pu]
Symmetrical components
Va' 1 2 Va ' Va ' Vb '
1 1
V0( a ') Va ' 1 2
Vb ' Va ' Vb '
0 3 3
Va ' 1 1
1 Vc ' Va ' 2Vb '
I a' 2 I
1 a ' jI b '
1 1
I 0(( a ')) I a' 1 2 I b ' jI b '
0 3 3
I 1 1 1 I 0
a ' c '
6
Unbalanced grid voltages during a grid fault
• Line‐to‐line fault
ZS PCC ZF
positive-
Ia' Ia' Fault
sequence VSa Va Va '
subcircuit
ZS ZF
negative-
Ia'
sequence Va Va '
subcircuit
0 0
ZS ZF
zero- 0
0 Ia' 0 0
sequence Va Va '
subcircuit
Va 1 D 1
1
V 0( pcc ) Va VSa 1 D
0.5
0 2 0
Va
vabc [pu]]
0
1 -0.5
Va 1 1 1 Va
2 1 3
Vabc ( pcc ) Vb 1 Va VSa D -1
2 1 0 2 2
Vc Va
1 3 D -1.5
15
2 2 0 0.02 0.04 0.06 0.08 0.1
t [s]
8
Types of voltage sags
1.5
Va 1
1 1 3
Vb 2 j 2 DVSa,
1 3
0.5
Vc 2 j 2
Vc Va
vabc [pu]
Sag A 0 Va 1
Va 0 DVSa S ,
Three‐phase fault Vb -0.5
0 0
Va
-1
Z F
D
-1.5 Z S Z F
0 0 02
0.02 0.04
0 04 0 06
0.06 0 08
0.08 01
0.1
t [s]
1.5
V
a
D
Vc 1 1
2 j
3
Vb 2 VSa ,
1 3
0.5
Vc 2 j 2
Va
vabbc [pu]
Sag B Va 3 2 D
1
0
1
Single‐phase -0.5
Va 3 1 D VSa ,
0
to ground fault
Vb -1
Va 31 1 D
Z F Z F Z F0
-1.5 D
0 0.02 0.04
t [s]
0.06 0.08 0.1 3Z S Z F Z F Z F0
9
Types of voltage sags
1.5
V
a
1
1 1
V
b 2 j 3
D VSa ,
2
Vc Vc 1 j 3 D
0.5 2 2
Va
1
vabc [pu]
Sag C 0 Va 2 1 D
1
Phase‐to‐phase
V
2
a 1 D V ,
Sa
Vb -0.5
Va 0
0
fault
-1
Z Z
D F F
-1.5
0 0 02
0.02 0.04
0 04 0 06
0.06 0 08
0.08 01
0.1 2Z S Z F Z F
t [s]
1.5
V 1
a
1 1
Vb 2 D j
3
2 D VSa ,
1
Vc 0.5
Vc 2 D j
3
2 D
Va
vabbc [pu]
Sag E 0 Va
1
3 1 2 D
Vb 1
Single‐phase -0.5
Va 3 1 D V
Sa ,
0
to ground fault -1
Va 13 1 D
ZF
-1.5 D
0 0.02 0.04
t [s]
0.06 0.08 0.1 ZS ZF
10
Propagation of voltage sags
ZF
ZS
Vca ( p ) Vc ( p ) Vab ( p ) Vb ( s ) Va ( s )
Va ( p )
Vb ( p )
Vc ( s )
11
Propagation of voltage sags
1.5
Vc 1
Va D
1
0.5 Vb 2 D j 3
2 VSa ,
Vc
1 D j 3
Va 2 2
vabc [pu]
Sag D 0
1
Va 2 1 D
Propagation of -0.5 1
a sag type C
Va 2 1 D VSa
Vb -1 Va0 0
-1.5
0 0 02
0.02 0.04
0 04 0 06
0.06 0 08
0.08 01
0.1
t [s]
1.5
1 Va D
Vc 1
V ,
2 D j
2 D
0.5 Vb 12 Sa
Vc 1
Va
2 D j 2 D
vabbc [pu]
Sag E 0
12
1
Va 3 1 2 D
Propagation of -0.5
1
a sag type E Vb
Va 3 1 D
-1 Va0 0
-1.5
0 0.02 0.04 0.06 0.08 0.1
t [s]
12
Synchronous Reference Frame PLL (SRF‐PLL)
q 1
v 1
va vd vd V 1
vb T vq 1
v
d 1
vc t
s
t 1
v 1
va
d
v 2 cos(( ) cos(( 23 ) cos(( 23 )
v T vb , T sin( ) sin( 2 ) sin( 2 )
q 3 3
vc 3
13
SRF‐PLL under unbalanced and distorted grid
1.5 1.5 1.5
v 1 v 1 v 5 v 1 v 1
1 1 1
orted grid fault
0 0 0
ult
-1 -1 -1
• Unbalanced grid fau
• Balaanced grid fault
6 6 6
• Balaanced and disto
5 5 5
4 4 4
3 3 3
2 2 2
1
t 1
t 1
t
0 0 0
1.5 1.5 1.5
vd V 1 vd V 1 vd V 1
1 1 1
vabc [p.u.]
t, [rad]
[p.u.]
vdq [p.u.]
0 0.5 0
vq 0
abc
3
v
-0.5 -0.5
2
0
-1 1 -1
• Limitation of the PLL bandwidth pplus filteringg is not the most effective solution to
extract the positive‐sequence component from unbalanced three‐phase voltages
1.5 7 1.5 1.5
1
v v 1
t v v 1
6
1 vd V 1 1
1
5
0.5 0.5
vabc [p.u.]
t, [rad]
vabc [p.u.]
vdq [p.u.]
4
0 0.5 0
3 vq 0
-0.5
05 -0.5
05
2
0
-1 -1
1
Unbalanced input voltage vector v on the DSRF when the
angular position of the positive reference frame dq+1 matches v
the angular position of the positive‐sequence voltage vector v+1,
that is, ’ = t. t d 1
v 1 1 cos(2 t )
v dq 1 d Tdq 1 v V 1 V 1
vq 1 0 sin(2 t )
t
v dq 1
v 1 cos(2 t )
d Tdq1 v V 1 V 1 1 d 1
vq1
sin(2 t )
0
v 1
The dc values on the dq q+1 and the dqq-1 frames correspond to the amplitude the sinusoidal
p p
signals of v+1 and v-1, while the oscillations at 2 correspond to the coupling between
axes appearing as a consequence of the voltage vectors rotating in opposite direction
16
Decoupled Double Synchronous Reference
Frame PLL (DDSRF‐PLL)
• The decoupling network
Generic input voltage vector v:
v n cos( n t ) m cos( m t )
n m
v v v V
n m
n
V m
v
sin( n t ) sin(m t )
If a perfect synchronization of the PLL is possible, with ’ = t, the voltage vector v can be
expressed on generic dqn and dqm reference frames as follows:
vd n vd n vd n n cos( )
n
m cos(( n m ) t ) m sin(( n m ) t )
v dq n V n
V m
cos( ) sin((n m) t ) V m
sin( )
vq n vq n vq n sin(
) cos((n m) t )
dc terms ac terms
vd m vd m vd m m cos( )
m
n cos(( n m ) t ) n sin(( n m) t )
v dqm V m
V n
cos( ) sin((n m) t ) V n
sin( )
v v v
q q q
m m m sin(
) cos((n m) t )
dc terms ac terms
The amplitude of the ac terms in the dpn axes depends on the dc terms of the signals on the dqm
axes, and vice versa. 17
Decoupled Double Synchronous Reference
Frame PLL (DDSRF‐PLL)
• The decoupling network
vd m vq m
vd n d n dm qm *
d n* vd n
The decoupling cell cancels out the
The decoupling cell cancels out the
oscillations generated by the voltage vqn q n q n* vq*n
vector vm on the dqn axes signals. For
cancelling out the oscillations in the dqm
cancelling out the oscillations in the dq
axes signals, the same structure may be
used but swapping the m and n indexes
in it.
n
DC
m
18
Decoupled Double Synchronous Reference
Frame PLL (DDSRF‐PLL)
vd n
dn vd* n
T n vq n n* vd*n
dq n d
q n DC vq*n
m q n* vq*n
n m qm
' d
v abc v
T
vd m
m dn qn
d vd* m
T m vqm m* vd*m
dq m d
q m DC vq*m
n q m* vq*m
' m
' from
f the
h PLL
A cross‐feedback decoupling network is used to estimate the value of the dc
terms on the positive‐ and negative‐reference frames. The LPF block is a low‐
terms on the positive‐ and negative‐reference frames The LPF block is a low‐
pass filter such as: f
LPF ( s )
s f 19
Analysis of the DDSRF‐PLL
Unbalanced input voltage vector v:
v 1 cos( t ) 1 cos( t )
1 1
1 1
v v v V 1
V 1
v
sin( t ) sin( t )
d signals on the positive‐
d‐q i l h i i and negative‐reference frames:
d i f f
vd 1 1 cos( )
1
1 cos(2 t ) sin(2 t ) cos( 1 )
v dq1 V 1
V
v
q
1
sin(
i ( ) sin(2 (2 t ) sin(
i (2 t ) cos(2 i ( 1 )
vd 1 1 cos( )
1
1 cos(2 t ) sin(2 t ) cos( 1 )
v dq1 V 1
V
vq1 sin(
i ( ) sin(2 (2 t ) sin(
i (2 t ) cos(2 i ( 1 )
dq 1 vq 1
sin( 1
)
20
Analysis of the DDSRF‐PLL
The relationship between the signals on the positive‐ and negative‐reference frame is given by:
The estimated values at the output of the DDSRF can be written as:
p
vd*1
v *
dq 1 v
q 1
* F v dq 1 Tdq 2 v*dq 1
LPF ( s ) 0
F
0 LPF ( s )
vd*1
v *
dq 1
vq 1
* F v dq 1 Tdq 2 v*dq 1
v dq1 v*dq 1
T 1
dq F v*dq1
T 2
dq
v
T 2
dq
T 1
dq v 1 * F v*dq1
dq
v dq 1
21
Structure of the DDSRF‐PLL
vq*1 vq*1
q d 2 q2
ff
vd 1
1 vd* 1 vd*1 V 1
1 1*
T
dq 1
vq1
d
q 1
DC d
1 vq*1
LPF
vq*1
1*
1 q 1 q LPF
d
v abc v
T
vd 1 vd*1
1 vd* 1
T
1 1
d d q
vq1 d 1* LPF
1 1
dq
q 1 v *
vq*1
DC q 1* q 1
LPF
1
22
Response of the DDSRF‐PLL
1.5 7 1.5 1.5
v 1 v 1 t v 1
1 6 vd 1
1
5
1 vd*1
0.5 0.5
[rad]
vabc [p.u.]
[p.u.]
4
[p.u.]
0 0.5 0
t,
abcc
3
dq
+
v
v
-0.5 -0.5
2
0
-1 1 -1
vq1 vq*1
-1.5
1.5 0 -0.5
0.5
-1.5
15
100 125 150 175 200
t [ms]
400 1.5 1.5
1.5
v 1
1 1
300 1 vd 1
05
0.5 05
0.5
vq1 vd*1
[p.u.]
, [rad/s]
[p.u.]
[p.u.]
0 200 0.5 0
beta
vq*1
abc
-
dq
-
v
v
v
-0.5
-0.5
-1
1 100 0
-1
-1.5
-1.5 -1 -0.5 0 0.5 1 1.5 0 -0.5 -1.5
valpha [p.u.] 100 125 150 175 200
t [ms]
23
Response of the DDSRF‐PLL
Sag (dip) type C ( V 0,6 20º F 0,)9 10º
Sag (dip) type D ( V 0,6 20º F 0,9) 10º
24
Double Second Order Generalized Integrator
FLL (DSOGI‐FLL)
• Instantaneous symmetrical components
va 1 a a 2 va
1
b T v abc
v abc ; vb a 2
b 1 a vb
3
vc a a2 1 vc
va 1 a2 a va
1
v abc v abc v abc v 0abc v abc T v abc ; vb a 1 a 2 vb
3 2
vc a a 1 vc
va0 1 1 1 va
v 0 T v ; v 0 1 1 1 1 vb
abc 0 abc b
3
vc0 1 1 1 vc
1 1 1 q
v T T T v T v ; T
2 q 1 j
qe 2
1 1 1 q
v T T T v T v ; T
2 q 1
25
Structure of the DSOGI
v
v v ( ) k v ( ) v v v abc
k
1
2
T1
v
v abc v
T SOGI-QSG( )
v ( ) qv 1
2
v v( ) k v ( ) v
k
1
2
SOGI-QSG( )
v( ) qv
1
v
2
v v abc
T
1
v
PNSC
26
Response of the DSOGI
• DSOGI transfer function
j)
Complex frequency (s=+j
Complex frequency (s
1 D ( s ) Q ( s ) 1 k s
v T v v s v
2 Q ( s ) D ( s )
2 s 2 k s 2
( )
Harmonic frequency (s=j
f
v 1 D ( j ) Q ( j ) v 1 k j v
v 2 Q ( j ) D ( j ) v 2 jk
2 2
j v
20
v ( j ) jv ( j )
k 2
P ( j )
0
v 1 k v |P(j)| , ||N(j)| (dB)
2
v 2 k j v
2
-20
N ( j )
P ( j ) v
v -40
N ( j ) v
v -60 -2 -1 0 1 2
10 10 10 10 10
|/|
27
Structure of the FLL for the DSOGI
• The
h DSOGI uses a single
l FLL in whichh h the h frequency
f error signals
l generated
d by
b the
h
QSGs of the and signals have been combined by calculating a average error
signal.
f ( ) 1
f f ( ) qv qv
2 2
v ( ) k
f ( )
(v ) 2 (v ) 2 ff
qv
v( )
1
2
f ( )
qv
28
Response of the DSOGI
1.5 7 1.5 1.5
v 1 v 1 t v 1
1 6
1
5
1
v 1 v 1 v
1
0.5 0.5
[p.u.]
[rad]
vabc [p.u.]
vabcc [p.u.]
4
0 0.5 0
v , v
+
t,
3
+
+
-0.5 -0.5
2
0
-1 1 -1
-1.5
1.5 0 -0 5
-0.5 -1.5
15
100 125 150 175 200
t [ms]
1.5 400 1.5 1.5
v 1
1 1
300 1
05
0.5 05
0.5
v
[p.u.]
v 1 v 1 1
v , v [p.u.]
, [rad/s]
vabc [p.u.]
0
200 0.5 0
beta
-
v
-0.5
-
-0.5
-1
1 100 0
-1
-1.5
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5
valpha [p.u.] 0 -0.5 100 125 150 175 200
t [ms]
29
Response of the DSOGI
Voltage
sag
Frequency
variation
Voltage
distortion
0.35Hz/div
30
Conclusion
• Grid synchronization of three‐phase power converter is an
essential issue to control the delivered active/reactive
/ power and
p
the grid support services
• Conventional SRF‐PLL is not the most suitable technique for
synchronizing with unbalanced grid voltages during grid faults
• Specific synchronization techniques should be used to estimate
the instantaneous positive and negative sequence components of
the unbalanced grid voltage during faults
• The
h DDSRF‐PLL makes k possiblebl a good d synchronization
h d
during
unbalanced conditions by decoupling axis signals on the positive‐
and negative‐reference frames
• The DSOGI‐FLL is a very effective solution, based on adaptive
filtering, for grid monitoring and synchronization under generic
grid operating conditions
31