Professional Documents
Culture Documents
A Generic EMT-type Model For Wind Parks With Permanent Magnet Synchronous Generator Full Size Converter Wind Turbines
A Generic EMT-type Model For Wind Parks With Permanent Magnet Synchronous Generator Full Size Converter Wind Turbines
net/publication/334432250
A Generic EMT-type Model for Wind Parks with Permanent Magnet Synchronous
Generator Full Size Converter Wind Turbines
CITATIONS READS
0 15
11 authors, including:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Aboutaleb Haddadi on 12 July 2019.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 2
The phasor domain models of WPs are suitable for transient II. VARIABLE SPEED WIND TURBINES
stability programs (TSPs) and might be adequate for basic
A. Wind Turbine Aerodynamics
power system design studies, such as stability analysis. On the
other hand, detailed electromagnetic transient (EMT) models The wind turbine extracts kinetic energy from the swept area
are typically used for equipment design and the assessment of of the blades. The mechanical power extracted from wind is
fast electromagnetic transients in arbitrary network topologies given by [21]:
with nonlinearities. They are circuit based models and offer the Pt = (1 2 ) A 3C p ( , ) (1)
highest accuracy for a wide range of perturbations, including where is the air density (approximately 1.225 kg/m3), A is
transient stability.
the swept area of the rotor (m2), is upwind free wind speed
Generic EMT-type simulation models for FSC and DFIG
based WPs have been proposed in [9] and [10], respectively. (m/s) and C p is the power coefficient. The C p is usually
However, the WP and WT control structure in both [9] and [10] provided as a set of curves ( C p curves) relating C p to tip-
are not realistic as the outer loops of WT control (the active
power and the voltage/reactive power controls) are moved to speed-ratio with the blade pitch angle as a parameter, as
the point of interconnection as WPC. The EMT-type simulation shown in Fig. 1 [22]. The tip-speed-ratio is defined as
models of DFIG based WPs in [11]-[13] have a WP voltage/ = (t R ) (2)
reactive power control scheme as recommended in [8]. On the where t is the WT rotational speed (rad/s) and R is the blade
other hand, the EMT-type simulation models in [9]-[13] are
alike the generic phasor domain models, i.e. they disregard the radius (m).
possible decoupled sequence control (DSC) scheme in WTs. At a specific wind speed and pitch angle, there is a unique
Ideally, the GSC with traditional coupled sequence control WT rotational speed that achieves the maximum power
(CSC) scheme is not expected to inject any negative sequence coefficient CP max , hence the maximum mechanical power as
currents to the grid during unbalanced loading conditions or shown in Fig. 1.
faults. In practice, it injects a very small amount due to the The mathematical model of WT aerodynamics is shown in
phase shift in low pass measuring filters [14]. Unlike its output Fig. 2. In this modeling approach, the C p curves of the WT are
currents, the GSC terminal voltages contain negative sequence
fitted with high order polynomials on and as
component during unbalanced loading conditions or faults, and
n n
C p ( , ) = ij i j
this causes second harmonic oscillations in the GSC active
(3)
power output as well as the dc bus capacitor voltage. These i =1 j =1
second harmonic oscillations can be eliminated by adopting a
DSC scheme [15]. The DSC scheme in GSC should give the
priority to the positive sequence reactive currents designated by CPmax Curve
the FRT requirement and then use the available converter
capacity for the second harmonic oscillation mitigation [16].
As discussed in [17]-[19], lack of negative sequence fault
current contribution from the FSC WT with traditional CSC
may cause misoperation of protection system during certain
unbalanced fault conditions. Although the GSC operating under
DSC injects considerable amount of negative sequence currents
during unbalanced faults, the recent VDE-AR-N 4120
Technical Connection Rules [20] includes a negative sequence
reactive current requirement to further reduce the possibility of
protection system misoperation. Hence, the DSC scheme in
FSC WTs (if it exists) should be accounted in EMT simulation
model. Fig. 1. Wind Power C p curves
This paper presents a generic EMT model for FSC WT based
WPs that can be used for a wide range of WP integration
studies. The collector grid and the FSC WTs are represented n n
with their aggregated models, but the overall control structure C p ( , ) = ij i j
i =1 j =1
of the WP is preserved. The FSC WT and the WP control
system models include the non-linearities, necessary transient t
t R Cp
and protection functions; and allow to simulate the accurate =
transient behavior of WPs subjected to external power system
Pt
Pt = A3Cp (, )
disturbances. 1
The first part of this paper briefly presents the WPs with FSC 2
WTs. The EMT model is presented in the second part. The third
Fig. 2. Wind turbine aerodynamics model
part presents illustrative simulation examples.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 3
When the wind speed is above cut −in and below rated speed X
HV /MV
rated , the pitch angle is kept at zero ( = 0 ) and the power 0 Wind Park
Transformer
reference of the WT generator is produced by the MPPT HV Grid
X X
(maximum power point tracking) function to achieve optimal POI +30
Other MV feeders
operation. The conventional method is to calculate the power
X
reference using a cubic function of the turbine angular speed.
Pref = Kopt t3 (4) Fig. 4. Simplified single-line diagram of a typical wind park
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 4
VPOI
KVpoi WPC are controlled by a two-level controller. The slow outer control
calculates the reference dq-frame currents. The fast inner
+ +
-
(current) control produces the converter ac voltage reference.
VPOI V-control WT Measurements Compute
& Filters Variables MSC MSC
Variables
QPOI Control Command
+ + PI
-
GSC GSC
WPC function Q-control Control Command
PPOI QPOI
u[1]
(1 − u[2] ) 2 selector Pitch
Control
Pitch
u
PFPOI [2] Command
Protection
Chopper ON/OFF
PF-control System
WT Breaker Open/Close
Communication Fig. 7. Simplified diagram, FSC WT control and protection system.
V e − sTcom U
Distribution to WTs
MSC GSC
Transfer voltage reference from WPC to WTCs AC DC
DC AC
Fig. 5. Wind park controller
v'qm v'dm v'dg v'qg
IV. FULL SIZE CONVERTER WIND TURBINES iqm iqg
A FSC WT may or may not have a gearbox. A wide range of idm PI PI PI PI idg Inner
electrical generators such as asynchronous, synchronous and Control
permanent magnet can be used. The WT power is transferred i'qm i'dm i'dg i'qg
through an ac-dc-ac converter system, and the dynamics of the
Outer Idq limiter FRT
electrical generator are isolated from the grid [26]. Control
The considered topology in this paper is shown in Fig. 6. It
uses a permanent magnet synchronous generator (PMSG) and f(T’) 0 PI P
Vdc
the ac-dc-ac converter system consisting of two pulse-width
modulated (PWM) voltage source converters (VSCs): machine- T'
side converter (MSC) and grid-side converter (GSC). The dc maximum power ωg V'dc V' = (1+∆V')
resistive chopper is used for the dc bus overvoltage protection. tracking point
.
The line inductor (choke filter) and ac harmonic filters are used Fig. 8. Schematic diagram of FSC WT control.
at the GSC to improve power quality.
A. Machine Side Converter Control
FSC WT
The q- and d-axis currents of MSC ( iqm and idm in Fig. 8)
are used to control the active and reactive power outputs of the
PMSG, respectively. The q-axis current reference ( f (T ) in
Fig. 8) is given by
= T m
iqm (7)
FSC WT
ac terminal where m is the constant flux generated by the permanent
Fig. 6. Schematic diagram of FSC WT.
magnet and T = K opt t2 is the reference for PMSG
Fig. 7 depicts the simplified diagram of FSC WT control and electromagnetic torque given by the MPPT control.
protection systems. The sampled signals are converted to per = 0 ) to
The d-axis current reference is set to zero ( idm
unit and filtered with low-pass filters with the “Measurements achieve unity power factor.
& Filters” block. The variables used by the FSC WT control The MSC inner control loop is designed based on internal
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 5
model control (IMC) method [27]. This method enables where 0 is the natural frequency of the closed loop system
calculation of dq-frame proportional integral (PI) controller and is the damping factor. H Cdc = ( ECdc S wt ) is the static
parameters (gain and integration time constant) using certain
machine parameters and the desired closed-loop bandwidth. moment of inertia, ECdc is the stored energy in dc bus capacitor
This method simplifies the controller design procedure and (in Joules) and S wt is the WT rated power (in VA).
eliminates (or reduces) the need for trial-and-error. The schematic of the GSC connected to the power system is
The PMSG stator voltages are found from shown in Fig. 9. Z = R + j L represents the total impedance
vdm = − Rs idm − Ld ( d idm dt ) + g Lq iqm (8)
between the GSC and external HV system Thevenin source, i.e.
(
vqm = − Rs iqm − Lq d iqm dt + g ( Ld idm + m ) ) (9) Z = Z HV − grid + ZTR −WP + ZCG + ZTR −WT + Z choke (16)
where Rs is the armature resistance, Ld and Lq are the d- and where Z HV − grid is the external HV system Thevenin
q-axis inductances of PMSG. impedance, Z TR −WP is the wind park transformer impedance,
The idm and iqm errors are processed by the PI controller to Z CG is the equivalent MV collector grid series impedance,
give vdm and vqm , respectively. To ensure good tracking, feed- Z TR −WT is the aggregated wind turbine transformer impedance
forward compensating terms g Lq iqm in (8) and and ZChoke is aggregated wind turbine choke filter impedance.
Using IMC, the PI controller parameters are found as Fig. 9. GSC arrangement.
k pd = c Ld , k pq = c Lq , kid = kiq = c Rs (12)
The voltage equation is given by (bold characters are used
where c is the bandwidth. The relationship between the
for vectors and matrices)
bandwidth and the rise time (10%–90%) is c = ln(9) trise .
(
vabc = R i gabc + L d i gabc dt + v gabc ) (17)
B. Grid Side Converter Control The link between GSC output current and voltage can be
The GSC function is maintaining the dc bus voltage Vdc at described by a transfer function
its nominal value and also controlling its positive sequence ac G ( s) = 1 ( R + sL ) (18)
terminal voltage ( Vt + ) when equipped with AVR (see Fig. 8). Using the IMC method, the PI controller parameters of the
The q-axis reference current is calculated by the proportional inner current control loop are found as
outer voltage control. k p = c L , ki = c R (19)
iqg (
= KV V − Vt + ) (13) Similar to MSC, the feed-forward compensating terms
where KV is the voltage regulator gain. The reference for FSC ( Lchokeiqg + vdt ) (
and − Lchoke idg + vqt ) are added to the d-
positive sequence voltage ( V = 1 + V ) is calculated by the and q-axis voltages calculated by the PI regulators,
WPC (see Fig. 5). respectively. Lchoke is the inductance of the aggregated wind
When AQR is used to control the GSC reactive current turbine choke filter, vdt and vqt are the FSC terminal voltages
output, q-axis reference current is calculated by a PI reactive
in dq reference frame. The FSC terminal is illustrated in Fig. 6.
power regulator.
(
= K p + Ki
iqg
Q Q
)
s ( QWT
− QWT ) (14)
C. Fault-Ride-Through Function
During normal operation, the controller gives priority to the
Q Q active currents, i.e.
where Kp and Ki are the reactive power regulator
I dg
idg lim
parameters, QWT is the reactive power output of the FSC WT
(20)
( ) ( )
2 2
is the reference is
(including harmonic filters) and QWT I qg
iqg lim
= I glim − idg
calculated by the WPC. Reader should refer to [8] for details. lim lim
The d-axis reference current is calculated by the proportional where I dg , I qg and I glim are the limits for d-axis, q-axis and
outer dc voltage control. It is a PI controller tuned based on total GSC currents, respectively.
inertia emulation, with The grid code requirements, such as [28], include the WT
k p = 02 ( 2 H Cdc ) , ki = 20 ( 2 H Cdc ) (15) transient response against severe voltage disturbances (see Fig.
10). To comply with this requirement, an FRT function is
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 6
traditionally added to the WTC. The FRT function is activated positive and negative sequences. The DSC objective is typically
+
when the voltage |1 − VMV | exceeds the pre-defined value elimination of second harmonic pulsations in the active power
+
output of FSC WT during unbalanced loading conditions or
VFRT −ON and deactivated when |1 − VMV | reduces below the faults. The other objective can be injecting desired level
pre-defined value VFRT −OFF after a pre-specified release time negative sequence reactive currents during unbalanced faults
t FRT . In the presented generic model, VFRT −ON = 0.1 pu, [20].
1) Traditional DSC (DSC1)
+
VFRT −OFF = 0.075 pu and t FRT = 250 ms. VMV is the WT The instantaneous active and reactive powers during
transformer MV terminal positive sequence voltage and it is unbalanced grid conditions can be written as [15]
estimated in “Compute Variables” block (see Fig. 7) using the p = P0 + PC 2 cos(2t ) + PS 2 cos(2t )
(21)
WT transformer parameters and measured FSC WT ac terminal q = Q0 + QC 2 cos(2t ) + QS 2 cos(2t )
voltages and currents. During FRT operation, the GSC
where P0 and Q0 are the average values of the instantaneous
controller gives priority to reactive current by reversing the d-
and q-axis current limits given in (20). The limits for d-axis, q- active and reactive powers respectively, whereas PC 2 , PS 2 ,
axis and total GSC currents might be also different during FRT QC 2 and QS 2 represent the magnitudes of the second harmonic
operation. Reader should refer to [16] for details. oscillating terms in these instantaneous powers.
Due to AVR usage, the voltage control is continuous even With DSC usage, four of the six power magnitudes in (21)
inside the dead-band region shown in Fig. 10. On the other can be controlled for a given grid voltage conditions. As the
hand, the reactive current output is limited with the available oscillating terms ( PC 2 and PS 2 ) in active power cause
reserve on GSC as the priority is given to the active currents
oscillations in dc bus voltage Vdc , the GSC current references (
(see (20)).
+ + − −
When AQR is used to for controlling reactive current output idg , iqg , idg and iqg ) are calculated to cancel out these terms
of GSC, it is switched to AVR during FRT operation to achieve (i.e. PC 2 = PS 2 = 0 ).
the desired reactive current injection from the GSC. The voltage
reference of the AVR is set to the pre-disturbance voltage value The GSC DSC implementation in [16] keeps the outer
and AQR input is blocked. Reader should refer [23] for details. control and Idq limiter shown in Fig. 8, to calculate idg ,
, iqg
Depending on the manufacturer and grid code requirements, lim
I dg lim
and I qg . These values are used to calculate the GSC
it is possible to have different control schemes. To imitate the
+ + − −
accurate fault behavior of the WT, more detailed limitation current references idg , iqg , idg and iqg for the DSC. As the
functions can be used (when available), such as reactive power positive sequence reactive current injection during faults is
or reactive current as function of voltage table as described in defined by the grid code (see Fig. 10), the GSC current
IEC 61400-27. reference calculation in [15] is modified as
i + 1 0 0 0
−1
qg
iqg
i + vqg
+ + − −
v v vdg
dg = dg qg
P0 (22)
− vqg
− −
vdg +
vqg +
vdg P
i
qg C2
− −vdg
− − + + P
v v − v S2
idg qg dg qg
when (i+
qg
−
+ iqg )
I qg
lim
, the q-axis current references are
revised as below
+
iqg + lim
= iqg I
qg (i +
qg
−
+ iqg
)
(23)
−
iqg − lim
= iqg I (i + −
+ iqg )
qg qg
Fig. 10. Wind turbine reactive output current during voltage disturbances [28]. + −
where iqg and iqg are the revised reference values for GSC
q-axis positive and negative sequence currents, respectively.
D. GSC Decoupled Sequence Control The revised GSC d-axis positive and negative sequence
DSC scheme enables controlling converter output currents + −
current references idg and idg can be obtained with the same
independently as active and reactive components for both
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 7
lim
approach using I dg . It should be emphasized here that, during
+
(
when iqg + iqg I qg .
+ − lim
)
faults, the priority is to provide iqg specified by the grid code. The positive sequence active current reference is generated
+ −
The remaining reserve in GSC is used for eliminating PC 2 and by the dc bus voltage regulator ( idg ) and idg
= idg = 0 as there
PS 2 . Hence, its performance reduces with decreasing electrical is no active power exchange on negative sequence.
distance between the WP and the unbalanced fault location. During normal operation, the controller gives priority to the
+ + − − positive sequence active current (see (20)). On the other hand,
As idg , iqg , idg and iqg are controlled, the DSC contains
four PI regulators and requires sequence extraction for GSC during FRT operation, the GSC controller gives priority to
currents and voltages. Numerous methods have been proposed reactive current by reversing the d- and q-axis current limits
for sequence extraction in the literature. The implementation in given in (20).
[16] uses the sequence decoupling method [30] shown in Fig. As there is no dead-band region in VDE-AR-N 4120, AVR
11. In this method, a combination of a low-pass filter (LPF) and usage is essential to control GSC reactive current output.
double-line frequency Park transformation ( P −2 and P+2 ) is
used to produce the oscillating signal, which is then subtracted. Additional reactive current required
The blocks C and P represent the Clarke and Park 1.0
ΔiB1 , ΔiB2
transformation matrices, and the superscripts ±1 and ±2
correspond to direct and inverse transformation at line 0.5
frequency and double-line frequency, respectively.
Δu1 , Δu2
+
i dq
+
P
+1
-
Σ -0.3 -0.1 0.1 0.3
-2
LPF P
iabc i αβ Voltage drop or increase, respectively
C -0.5 Representation in related quantities:
+2 Δu1 = voltage change in the + sequence system
LPF P - - Δu2 = voltage change in the - sequence system
i dq
P
-1
+ Σ 2 k 6 ΔiB1 = current variation in the + sequence system
-1.0 ΔiB2 = current variation in the - sequence system
Fig. 11. Sequence extraction using decoupling method.
Fig. 12. Wind turbine reactive output current during voltage disturbances [20].
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 8
The FSC WT control offers both DSC options (DSC1 and reader should refer to [16], [29], [37] for wind park and 120 kV
DSC2) in addition to the traditional CSC option when AVR is test system details.
selected to control the GSC reactive currents. DSC2 option is
not available for AQR selection. S1 S3 S5 + p
The generic model has two versions: detailed model (DM) va
+
and average value model (AVM). In the DM, the WT vb Vdc g1
RLC +
converters are represented based on the circuit of Fig. 13.a in vc
which the IGBT/diode is modeled by an ideal switch and S2 S4 S6
- n
nonlinear resistors (shown in Fig. 13.b) to mimic the actual
behavior accurately. Simulation of such switching circuits with (a) (b)
variable topology requires many time consuming mathematical Fig. 13. (a) Two level VSC circuit. (b) IGBT valve model.
operations and the high frequency PWM signals force small
simulation time-step usage. These computational inefficiencies
+
can be eliminated by using AVM, which replicates the average
+
va vb vc Vdc
response of converters through simplified functions and
controlled sources [32]. AVMs are used for wind generation -
technologies [33],[34]. AVM of FSC is obtained by replacing
ac voltage function dc voltage function
the DMs of converters with voltage-controlled sources on the
ac side and current-controlled sources on the dc side (see Fig. Fig. 14. AVM diagram for the two level VSC.
14) [35].
The sampled signals are converted to per unit and filtered at
“Measurements & Filters” block. The sampling frequencies are
set by the user in addition to the PWM frequency (12.5 kHz and
2.5 kHz, respectively, for both MSC and GSC in the presented
generic model) and sampling function is deactivated when
AVM is used for FSC converters. In the presented generic
model, second order Bessel type low pass filters are used. The
cut-off frequencies of the filters are set to 2.5 kHz for both MSC
and GSC. However, the order (up to 8th order), the type (Bessel
and Butterworth) and the cut-off frequencies of the low pass
filters can be modified through the device mask (device data
input function). The measuring filter parameters may have
significant impact on WT behavior in some phenomenon such
Fig. 15. 120 kV test system.
as subsynchronous control interaction [12].
The MSC and GSC overcurrent protections use the root mean Several simulations are performed for different fault types
square (rms) values of the current values. When the current at and locations using the simulation models (M1 to M4)
any phase exceeds the used defined limit, it blocks the presented in Table I. However, only the 250 ms double line to
overloaded converter temporarily. The user defined converter ground (DLG) fault at BUS4 scenario is presented below due to
pickup current and reset time are set to 2 pu and 50ms in the space limitations. A long duration fault is applied for testing
presented generic model. purpose. The simulation time-step is 10 µs (a typical value in
The low voltage and overvoltage relays use rms voltages on DM usage) and total simulation time is 2 seconds.
each phase at FSC WT ac terminals and send a trip signal to the
FSC circuit breaker when any of the phase rms voltage violates Table I: Simulation Models
the limits defined as a function of time by the user. The voltage- Model M1 M2 M3 M4
time characteristics of the low voltage and overvoltage relays GSC Control CSC DSC1 DSC2 DSC1
Converter Model DM DM DM AVM
are set based on the technical requirements of Hydro-Quebec
for the integration of wind generation [36].
As shown in Fig. 16, the simulated unbalanced fault results
The reader should refer to [16] for the modeling,
into second harmonic pulsations in the active power output of
implementation and utilization details of the generic model
presented in this paper. M1. These second harmonic pulsations ( PC 2 , PS 2 ) are
eliminated in M2 at the expense of a reduction in the active
VI. SIMULATIONS power output of FSC WT ( P0 ), as seen in Fig. 17. M3 achieves
The single-line diagram of the 120 kV, 60 Hz test system is injection of desired negative sequence reactive current at the
shown in Fig. 15. The WP includes 45 WTs rated at 1.5 MW. expense of a reduction in in the active power output of FSC WT
The WP is operating at full load (under nominal wind speed) (see Fig. 17) and an increase in second harmonic oscillations in
and under Q-control function of WPC with QPOI = 0 . The the GSC active power output (Fig. 16). On the other hand, the
reactive power output of the FSC WT is similar in M1, M2 and
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 9
P0 (pu)
DSC objectives cannot be achieved in both M2 and M3 when
0.8 M1
the required GSC current output exceeds its rating. It should be
M2
noted that, when the electrical distance between the WP and 0.7 M3
unbalanced fault decreases, larger GSC currents are required to
0.6
achieve the DSC objectives in both M2 and M3. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
The negative and positive sequence fault currents ( I n and t (s)
0.4
I p ) of the WP in M1, M2 and M3 are illustrated in Fig. 18. The
0.3 M1
small negative sequence current injection in M1 is due to phase M2
Q0 (pu)
0.2 M3
shift in low pass measuring filters [14]. M2 injects a
0.1
considerable amount of negative sequence current to achieve
mitigation of second harmonic power oscillations, but still quite 0
low compared to M3. It should be noted that, this difference -0.1
strongly depends on the unbalanced fault type, its electrical 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
t (s)
distance to the WP and GSC rating. It becomes less noticeable Fig. 17. P0 and Q0 of aggregated FSC WT in M1, M2 and M3.
especially for the electrical distant faults such as an unbalanced
fault at BUS6. 0.6
As shown in Fig. 19 - Fig. 21, AVM usage instead of DM
0.55
provides acceptable accuracy even for 50 µs time step usage
Ip (kA) M1
while providing a significant computational gain as illustrated 0.5 M2
in Table II. M4* in Fig. 19 - Fig. 21 is the M4 solution with 50 M3
µs time step. In this simulation, the computational gain over 0.45
DM is more than 5 when AVM is used with 50 µs time step.
However, a higher computational gain can be expected while 0.4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
simulating a large scale power system. t (s)
0.3
0.1
0.1
PC2 (pu)
0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
t (s)
M1
-0.1 M2 Fig. 18. I p and I n of WP in M1, M2 and M3.
M3
-0.2
1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
t (s) 0.9
P0 (pu)
0.6
0.8 M2
0.4 M4
M1 0.7 M4*
PS2 (pu)
M2
0.2 M3 0.6
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
t (s)
0 0.4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.3 M2
t (s) M4
Q0 (pu)
Fig. 16. PC 2 and PS 2 of aggregated FSC WT in M1, M2 and M3. 0.2 M4*
0.1
-0.1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
t (s)
Fig. 19. P0 and Q0 of aggregated FSC WT in M2 and M4.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 10
M4*
current injection is much larger compared the DSC1. Hence, it
0.45 can reduce the possibility of protection system misoperation
further. However, the desired negative sequence current
injection is achieved at the expense of very large second
0.4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 harmonic pulsations in active power output.
t (s)
The presented generic EMT model also offers two converter
modeling options: detailed model (DM) and average value
0.1 M2
M4
model (AVM). Simulation results demonstrated that, AVM
0.08
M4* usage instead of DM provides acceptable accuracy even for
0.06
In (kA)
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JPETS.2019.2928013, IEEE Power
and Energy Technology Systems Journal
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 11
Power Systems," IEEE Trans. Power Syst., vol. 28, no. 3, pp. 3096-3104, power systems,” Electric Power Systems Research, vol. 77, Issue 11, pp.
Aug. 2013. 1514-1520, Sep. 2007.
[10] D. N. Hussein, M. Matar and R. Iravani, "A Wideband Equivalent Model [32] S. R. Sanders, J. M. Noworolski, X. Z. Liu, and G. C. Verghese,
of Type-3 Wind Power Plants for EMT Studies," IEEE Trans. Power Del., “Generalized averaging method for power conversion circuits,” IEEE
vol. 31, no. 5, pp. 2322-2331, Oct. 2016. Trans. Power Electron., vol. 6, no. 2, pp. 251–259, Apr. 1991.
[11] O. Goksu, M. Altin, J. Fortmann and P. E. Sorensen, "Field Validation of [33] J. Morren, S. W. H. de Haan, P. Bauer, J. Pierik, and J. Bozelie,
IEC 61400-27-1 Wind Generation Type 3 Model With Plant Power Factor “Comparison of complete and reduced models of a wind turbine with
Controller," IEEE Trans. Energy Convers., vol. 31, no. 3, pp. 1170-1178, Doubly-fed Induction Generator,” Proc. 10th Eur. Conf. Power Electron.
Sept. 2016. Appl., Toulouse, France, Sep. 2003, pp. 1–10D.
[12] U. Karaagac, J. Mahseredjian, S. Jensen, R. Gagnon, M. Fecteau, I. Kocar, [34] J. G. Slootweg, H. Polinder, and W. L. Kling, “Representing wind turbine
"Safe Operation of DFIG based Wind Parks in Series Compensated electrical generating systems in fundamental frequency simulations,”
Systems," IEEE Trans. Power Del., vol. 33, no. 2, pp. 709 - 718, April IEEE Trans. Energy Convers., vol. 18, no. 4, pp. 516-524, Dec. 2003.
2018. [35] J. Peralta, H. Saad, U. Karaagac, J. Mahseredjian, S. Dennetière and X.
[13] M. Ghafouri, U. Karaagac, H. Karimi, S. Jensen, J. Mahseredjian, S. O. Legrand, “Dynamic Modeling of MMC-based MTDC Systems for The
Faried, "An LQR Controller for Damping of Subsynchronous Interaction Integration of Offshore Wind Generation,” CIGRE Canada conference on
in DFIG-Based Wind Farms," IEEE Trans. Power Syst., vol. 32, no. 6, pp. Power Systems, September 2012, Montreal.
4934-4942, Nov. 2017. [36] Transmission Provider Technical Requirements for the Connection of
[14] Impact of Renewables on System Protection: Wind/PV Short-Circuit Power Plants to the Hydro-Quebec Transmission System, Hydro Quebec
Phasor Model Library and Guidelines for System Protection Studies. Transenergie, 2009.
EPRI, Palo Alto, CA: 2016, 3002008367. [37] T. Kauffmann; U. Karaagac; I. Kocar; S. Jensen; J. Mahseredjian; E.
[15] R. Teodorescu, M. Liserre, P. Rodriguez, Grid Converters for Farantatos, "An Accurate Type III Wind Turbine Generator Short Circuit
Photovoltaic and Wind Power Systems, IEEE-Wiley, 2011. Model for Protection Applications," IEEE Trans. Power Del. , vol. 32, no.
[16] U. Karaagac, J. Mahseredjian, H. Gras, H. Saad, J. Peralta and L. D. 6, pp. 2370-2379, Dec. 2017.
Bellomo, “Simulation Models for Wind Parks with Variable Speed Wind
Turbines in EMTP-RV”, research report, Polytechnique Montréal, March
2017.
[17] I. Erlich, T. Neumann, F. Shewaega, P. Schegner and J. Meyer, "Wind
Turbine Negative Sequence Current Control and its Effect on Power
System Protection," Proc. 2013 IEEE PES Gen. Meet., Vancouver, BC,
July 21-25, 2013.
[18] M. Nagpal and C. Henville, “Impact of Power-Electronic Sources on
Transmission Line Ground Fault Protection”, IEEE Trans. Power Del.,
vol. 33, no. 1, pp. 62–70, February 2018.
[19] “Impact of Inverter Based Generation on Bulk Power System Dynamics
and Short-Circuit Performance”, PES-TR68, IEEE/NERC Task Force on
Short-Circuit and System Performance Impact of Inverter Based
Generation, July 2018.
[20] Technische Regeln für den Anschluss von Kundenanlagen an das
Hochspannungsnetz und deren Betrieb (TAR Hochspannung), VDE-AR-
N 4120 Anwendungsregel: 2018-11.
[21] O. Anaya-Lara, N. Jenkins, J. Ekanayake, P. Cartwright, and M. Hughes,
Wind Energy Generation: Modelling and Control, Wiley, 2009, John
Wiley & Sons, Ltd.
[22] N. W. Miller, W. W. Price, and J. J. Sanchez-Gasca, “Dynamic modeling
of GE 1.5 and 3.6 wind turbine-generators,” GE-Power System Energy
Consulting, General Electric International, Inc., Schenectady, NY, USA,
Oct. 2003.
[23] Wind Turbines - Part 27-1: Electrical Simulation Models - Wind
Turbines, IEC Standard 61400-27-1, 2015.
[24] G. Abad, J. Lopez, M. A. Rodriguez, L. Marroyo, G. Iwanski, Doubly Fed
Induction Machine: Modeling and Control for Wind Energy Generation,
2011, Wiley.
[25] M. Singh and S. Santoso, “Dynamic Models for Wind Turbines and Wind
Power Plants,” tech. rep., University of Texas at Austin and National
Renewable Energy, Oct. 2011.
[26] V. Akhmatov, A. H. Nielsen, J. K. Pedersen, O. Nymann, "Variable-speed
wind turbines with multi-pole synchronous permanent magnet generators.
Part I: Modelling in dynamic simulation tools", Wind Eng., vol. 27, no. 6,
pp. 531-548, Dec. 2003.
[27] L. Harnefors and H. Nee, "Model-based current control of ac machines
using the internal model control", IEEE Trans. Ind. Appl., pp. 133-141,
Jan/Feb. 1998.
[28] “Grid code - high and extra high voltage,” E.ON Netz GmbH, Bayreuth,
Germany, April 2006.
[29] Impact of Renewables on System Protection: Short-Circuit Phasor
Models of Renewables and Impact of Renewables on Power Swing
Detection and Distance Protection,. EPRI, Palo Alto, CA: 2016.
3002005765.
[30] P. Rodriguez, J. Pou, J. Bergas, J. I. Candela, R. P. Burgos and D.
Boroyevich, "Decoupled Double Synchronous Reference Frame PLL for
Power Converters Control," IEEE Trans. Power Electron., vol. 22, no. 2,
pp. 584-592, March 2007.
[31] J. Mahseredjian, S. Dennetière, L. Dubé, B. Khodabakhchian and L.
Gérin-Lajoie: “On a new approach for the simulation of transients in
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
View publication stats