Dfig LVRT May 2020

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Electric Power Systems Research 186 (2020) 106403

Contents lists available at ScienceDirect

Electric Power Systems Research


journal homepage: www.elsevier.com/locate/epsr

Modified rotor-side converter control design for improving the LVRT T


capability of a DFIG-based WECS
Muhammad Arif Sharafat Alia, Khawaja Khalid Mehmoodb, Shazia Balocha, Chul-Hwan Kima,

a
Department of Electrical and Computer Engineering, Sungkyunkwan University, Suwon 440-746, Republic of Korea
b
U.S.-Pakistan Centre for Advanced Studies in Energy (USPCAS-E), National University of Sciences and Technology, Islamabad 44000, Pakistan

ARTICLE INFO ABSTRACT

Keywords: The rapid integration of wind-power generation with existing power grids has caused reliability and stability
Wind-energy conversion system concerns owing to the negative impact on the dynamic behaviors of power systems. During a fault, large elec-
Doubly fed induction generator tromotive force is induced in the rotor circuit of a doubly fed induction generator (DFIG) as the circuit is highly
Rotor-side converter control vulnerable to it. Such circumstances increase the importance of the low-voltage ride-through (LVRT) capability
Low-voltage ride-through
of a DFIG to ensure stability of the electric grid during transient conditions. Considering these factors, this study
focuses on the mitigation of rotor overcurrents and DC-link voltage variations by modifying the control structure
of the DFIG converter, thereby enhancing its LVRT capability. Additional voltage terms are injected into the
rotor-voltage references to improve the transient behavior of the DFIG control system. In the proposed design,
transient rotor currents and DC-link voltage variations are effectively suppressed. Because the voltage terms are
introduced outside the current loops, there is no impact on their stability. Furthermore, electromagnetic torque
oscillations during the faults are considerably suppressed. Finally, the validity of the proposed design during
abnormal grid conditions is demonstrated via MATLAB/Simulink. The results confirm the feasibility and ef-
fectiveness of the improved converter control design to enhance the LVRT capability of a DFIG.

1. Introduction favorable solution for WECSs owing to their prominent advantages


compared to currently available systems [2,3,6]. Because DFIGs are
In recent times, the rapid integration of renewable energy sources directly connected to the grid and have reduced-capacity power-elec-
such as wind-power plants with existing electric power grids has raised tronic converters (PECs), they are highly vulnerable to grid dis-
serious concerns over the dynamic performance of the power network turbances, primarily voltage sags.
[1,2]. The negative impacts of large-scale integration of wind-power During a fault, the grid voltage drops abruptly, regardless of the
plants has increased power quality and power-system reliability and distance between the fault and the WT location; this induces a large
stability issues [3]. Therefore, it is necessary to alter the integration electromotive force (EMF) in the rotor circuit of the generator [7]. As a
requirements such that grid-connected wind turbines (WTs) and con- result, the DC-link is likely to experience overvoltages and may cause
ventional power plants perform similarly to support the power network overcurrents in the rotor circuit [8]. The magnitude of the rotor over-
during both normal and abnormal grid operating conditions. The current is higher at the beginning and end of the fault duration. Under
aforementioned circumstances have forced several countries to revise the aforementioned circumstances, the rotor-side converter (RSC) of the
their WT grid codes for grid integration and operation and to impose DFIG may damage and also trip the WT, thereby causing the failure of
low-voltage ride-through (LVRT) as a major requirement [2–4] to en- the LVRT. To avoid such situations, appropriate measures should be
sure power-system stability [5,6]. enforced to meet the LVRT conditions for DFIG-based WTs. Proper
According to the LVRT requirements, wind-energy conversion sys- designed protection schemes are necessary to overcome transient rotor
tems (WECSs) must not only remain coupled to the grid but also deliver overcurrents and to protect the DC-link from overvoltages so that safe
reactive power to support grid voltage during faults. The design and operation of the WECSs can be guaranteed.
deployment of WTs with improved and advanced control systems are Motivated by the above arguments, extensive research has been
underway, thereby making the modern power systems more efficient conducted, and different technical solutions have been established in
and reliable. Doubly fed induction generators (DFIGs) are the most response to the increasing LVRT capabilities of DFIG-based WECSs.


Corresponding author.
E-mail address: [email protected] (C.-H. Kim).

https://doi.org/10.1016/j.epsr.2020.106403
Received 7 October 2019; Received in revised form 29 April 2020; Accepted 30 April 2020
Available online 16 May 2020
0378-7796/ © 2020 Elsevier B.V. All rights reserved.
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

Such solutions are primarily categorized into two groups: (1) mod- overcurrents and DC-link overvoltages during faults.
ification of DFIG converter controls and (2) hardware applications Through the proposed control scheme, unlike in the crowbar
support [3–7,9–15]. The aim of improving the control strategy is to method, the faults can ride through without any additional hardware
alleviate the overcurrents and overvoltages of DFIGs and RSCs under requirements, thereby offering a cost-effective solution. A simple im-
grid faults and to enable fast decay of the stator natural flux. Lower cost plementation of the proposed scheme in improving the transient per-
and simple implementation are some of the merits of introducing fault formance of the DFIG demonstrates its effectiveness. In support of the
ride-through (FRT) strategies based on the improved control. Further- above arguments, numerical simulations are used to validate the pro-
more, the DFIG remains under control at all times, and it is easier to posed design to enhance the LVRT capability of a DFIG-based WECS.
switch back to normal operation as compared to hardware-based so- With the support of the proposed control scheme, both the transient
lutions. However, because of reduced capacities of the RSCs, the im- rotor overcurrents and DC-link voltage transients can be limited suc-
proved-control-based FRT strategies are limited owing to their con- cessfully within their established performance bounds. Furthermore,
trollable range. the electromagnetic torque oscillations can be suppressed effectively.
Several modified control strategies have been presented in literature Finally, performance comparison analyses are carried out to confirm if
[2,16–33], of which there are three well-known LVRT strategies based the proposed design is suitable for wind-energy applications.
on improved control: demagnetizing-current control, flux-linkage- The primary objectives of this study are as follows:
tracking control, and feedforward transient-compensation control. In
order to reduce the fault rotor current, demagnetizing methods [19–23] • To develop a simple, efficient, and economical solution to enhance
were proposed to suppress the transients of the induced EMFs in the the DFIG performance (in terms of the compliance of LVRT capacity)
rotor circuits by controlling the RSC outputs. However, these methods by improving the RSC control structure, as it is responsible for
require estimations of the control parameters and are sensitive to generator operation and control.
generator parameters. An improved DC-link voltage control method for • To respond quickly during grid faults to limit the transient rotor
DFIG-based WTs was proposed in [24]. The authors verified the re- overcurrents and DC-link voltage variations within their secured
duction in the magnitude of DC-link voltage variation during grid fault; ranges and to minimize the torque oscillations.
however, the influence of the fault on the rotor current was not ad-
dressed. In [25], the authors proposed an optimal current-proportion The contributions and novelty of this work are summarized as fol-
scheme and a phase-locked loop bandwidth selection method to im- lows:
prove the small-signal stability of the DFIG; however, the proposed
strategy did not examine system behaviors during asymmetrical grid • A transient reconfiguration solution for the DFIG in enhancing its
faults. LVRT capability based on the modified rotor-voltage-reference
To limit the rotor current and expedite the damping of the stator scheme is proposed and evaluated.
natural flux, a flux-linkage-tracking control method was proposed in • The proposed RSC structure focuses on containing the transient
[26]. In this scheme, the rotor-flux linkage is altered to trail a reduced rotor overcurrents and DC-link overvoltages effectively during grid
fraction of the changing stator-flux linkage. However, this scheme re- voltage disturbances and keeping them within their safe operating
quires flux detection and stator-voltage decomposition. Furthermore, it limits, thereby improving the transient performance of the DFIG.
is sensitive to machine-parameter deviations. The authors of [27] pro- Furthermore, the oscillations of the electromagnetic torque for both
posed an improved flux magnitude and angle-control strategy to en- symmetrical and asymmetrical grid faults are effectively suppressed.
hance the LVRT capability by limiting the DFIG power-angle jumps • Additionally, the proposed method has the following advantages:
during grid faults. In [28,29], multiple synchronous reference frame a It is simple to implement, as it does not require any observer,
current controllers were suggested for independent tracking of the po- sequence decomposition algorithm, and additional hardware.
sitive-, transient, and negative-sequence current references. However, b The proposed design does not affect the current-loop stability
the sequence decomposition introduces delays and errors, thus reducing because the additional voltage terms are introduced outside the
the dynamic response and stability of the system. The authors of current loops.
[30,31] suggested a current mechanism based on symmetrical se- c It is insensitive to variations of the DFIG parameters and the DFIG
quences to improve the FRT capability. normal control remains active at all times.
To guarantee safe operation of the DFIG, feedforward control d It does not require any dedicated grid-voltage-dip-detection
[32,33] during LVRT attempts to overwhelm the transient rotor over- method; therefore, the response is fast and suitable for LVRT
currents introduced by the grid-voltage dips. The authors of [32] pro- control.
posed a stator-current-feedforward-control scheme for the RSC that
focused on suppressing the fault currents in the stator and rotor The remainder of this paper is structured as follows. Section 2
windings during grid faults. However, this scheme requires a higher presents a summary of the simulated DFIG-based WECS. A brief
RSC voltage as compared to the maximum rotor voltage under serious mathematical model of the WT, the dynamic structure of the DFIG, and
current-control grid faults. In [33], the authors suggested additional the WT control system are reviewed in Section 3. Section 4 describes
feedforward scheme of rotor-current references for RSC to improve the detailed implementations and analyses of the modified RSC and grid-
system transient performance. Despite achieving good results, the side converter (GSC) control designs to enhance the LVRT capability of
control performance of the proposed method was highly susceptible to a DFIG during severe grid voltage disturbances. Following this, nu-
the deviation of the generator parameters. merical simulation results for a test system are provided and discussed
In the present study, on identifying the limitations and drawbacks of in Section 5 to verify the validity of the proposed scheme embedded
the existing LVRT solutions by altering the RSC control based on the into the RSC structure. Finally, Section 6 presents the concluding re-
above analysis, the authors have attempted to fill the gaps in the ex- marks.
isting literature. To solve the aforementioned problems, a modified
rotor-voltage-reference-based scheme for RSC control to boost the LVRT 2. Configuration of the Simulated DFIG-Based WECS
capability of a DFIG-based WECS is proposed. Under the proposed ap-
proach, the RSC control structure is revised, and new rotor-voltage The simulated grid-connected DFIG-based WECS is presented by an
references are introduced by considering the DC-link voltage variations equivalent aggregated model and is shown in Fig. 1. It contains a WT,
and AC voltage imbalances in the power system. The proposed design is DFIG, RSC, DC-link, GSC, coupling transformer, and control system.
intended to enable effective mitigation of the transient rotor Detailed mathematical modeling and analyses of the simulated test

2
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

In (5), ωm denotes the mechanical speed of the rotor.

3.2. Dynamic Modeling of a DFIG

The dynamic behavior of the DFIG is modeled and analyzed using


Park transformation in the form of space vectors. In this context, the
stator- and rotor-voltage equations are described [36] in a dq syn-
chronous reference frame (SRF) and expressed as (6). The SRF is helpful
for simplified tuning of the proportional-integral (PI) controllers.
d ds d qs
vds = Rs ids + dt s qs , vqs = Rs iqs + dt
+ s ds
Fig. 1. Configuration of the simulated grid-connected DFIG-based WECS. d dr d qr
vdr = Rr idr + dt sl qr , vqr = Rr iqr + dt
+ sl dr

= Ls ids + Lm idr , = Ls iqs + Lm iqr


system are presented in subsequent sections. ds qs

The current trend of setting up large-scale WTs using innovative dr = Lr idr + Lm ids , qr = Lr iqr + Lm iqs
concepts for both generators and power electronics as well as their Ls = Lls + Lm , Lr = Llr + Lm , sl = s r (6)
associated control structures is showing higher significance in obtaining
In (6), the subscripts s and r represent the stator- and rotor-side
realistic efficiency and quality restraints of WECSs and catering to the
quantities, respectively; the subscripts d and q represent the d- and q-
stochastic nature of the operating conditions. With its simplified
axis components of the quantities in the SRF; R and L denote the re-
structure, the simulated test system (Fig. 1) is advantageous for a better
sistance and inductance, respectively; v, i, and Ψ are the voltage, cur-
understanding of the employed control mechanism. Furthermore, it
rent, and flux linkage quantities, respectively; Lm is the magnetizing
more easily permits the deep investigation of the proposed modifica-
inductance; Lls and Llr are the stator- and rotor-leakage inductances,
tions in the RSC control structure and their impacts on the entire
respectively; and ωs, ωr, and ωsl are the synchronous, rotor, and slip
system.
angular speeds of the DFIG, respectively.
As stated previously, the stator windings of a DFIG are directly
Stator voltage-oriented control (SVOC) is employed to the DFIG-
linked to the electric grid via a coupling transformer. On the contrary,
based model given in (6) in which the dq-axes are aligned with the
the rotor is indirectly connected through back-to-back (BTB) PECs,
stator voltage. To simplify the analysis, the stator winding resistance
which are coupled using a DC-link. A WECS receives the kinetic energy
(Rs) is ignored. Then, the relationships between the stator and rotor
(KE) of the wind, converts it into electrical energy, and finally transfers
voltages, rotor currents, and stator-flux linkages are obtained. From (6),
it to the grid. The WT control system, comprising a maximum power
the stator currents can be derived as (7):
point tracking (MPPT) controller and a pitch angle controller, is also
implemented. ds Lm idr qs Lm iqr
ids = , iqs =
Ls Ls (7)
3. Mathematical Modeling of a DFIG-Based WECS
Substituting (7) into (6), the rotor fluxes can be expressed as (8):
This section presents a brief review of the mathematical modeling of Lm
= + Lr idr Lm2
a WECS, which contains a WT, a DFIG, and a WT control system.
dr Ls ds
Lm
, =1
Detailed implementation and analyses of the proposed control scheme = + Lr iqr Ls Lr
qr Ls qs (8)
for an RSC are described in the subsequent section.
In (8), σ represents the leakage coefficient of the DFIG. Substituting
3.1. WT Model (8) into (6), a relationship between the rotor voltages and the stator-
flux linkages can be established; it is expressed as (9):
As described in Section 2, a WT is employed to capture the KE from
wind and to transform it into mechanical power (Pm), which is then (
vdr = Rr + Lr dt idr
d
) sl Lr iqr +
Lm
Ls ( d ds
dt sl qs )
directly used by the generator. The KE of the wind (Pw) can be ex-
pressed as (1) [4]:
vqr = (R r +
d
Lr dt )i qr + sl Lr idr +
Lm
Ls ( d qs
dt
+ sl ds ) (9)
Pw = 0.5 r 2Vw3 (1) Equation (6) can be further simplified by ignoring the transient
In (1), ρ is the air density; r is the WT rotor radius; and Vw is the process of the stator-flux linkage, in which context, the stator dq vol-
wind speed. The output mechanical power (Pm) on the shaft of a WT is tages become (10):
given by (2) [34]: vds s qs , vqs s ds (10)
Pm = Pw Cp ( , ) = 0.5 r 2Cp ( , ) Vw3 (2) Substituting (10) into (9), the rotor voltages can be simplified to
(11):
In (2), Cp represents the WT power coefficient; β is the blade pitch
angle; and λ is the tip-speed ratio. Cp value can be calculated using
(3–5) [35]: (
vdr = Rr + Lr dt idr
d
) sl Lr iqr +
Lm sl
v
Ls s ds

= (R )i
d Lm sl
vqr + Lr dt + Lr idr + v
116 21
r qr sl Ls s qs (11)
Cp ( , ) = 0.5176 0.4 5 e
(3) From (11), it can be inferred that the rotor currents are regulated by
1 the rotor voltages and are held for both steady-state and dynamic
1 0.035 conditions. The electromagnetic torque (Tem) expression in the dq re-
=
+ 0.08 3 + 1
(4) ference frame is derived as (12) [37]:
mr Lm
= Tem = 1.5p ( qs idr ds i qr )
Vw (5) Ls (12)

3
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

In (12), p represents the number of pole pairs. The rotor electrical Consequently, the stator q-axis voltage component (vqs) becomes zero.
speed of the generator (ωr) is represented by (13): The resultant stator dq-axes voltage components become (18):
r =p m (13) vds Vs, vqs 0 (18)
The active and reactive powers on the stator- and rotor-sides in the Based on (10) and (18), the stator-flux linkages in the dq reference
dq reference frame can be expressed as (14): frame can be expressed as (19):
Ps = 1.5(vds ids + vqs iqs ), Qs = 1.5(vqs ids vds iqs ) vds
0,
Pr = 1.5(vdr idr + vqr iqr ), Qr = 1.5(vqr idr vdr iqr )
ds qs
(19)
(14) s

Substituting (19) into (7), the relationship between stator and rotor
While the stator delivers active power to the grid, the power fed into
currents in the dq reference frame is established. It can be represented
or out of the rotor depends on the rotor rotational speed. The rotor
by (20):
feeds power to the grid in super-synchronous mode, whereas the grid
delivers power to the rotor in sub-synchronous mode. The relationship Lm vds Lm
ids = idr , iqs = iqr
between the turbine mechanical and generator electromagnetic torques Ls s Ls Ls (20)
(Tm and Tem) is expressed by (15) [36]:
Substituting (19) into (12) yields the Tem expression:
d r 1 B
= (Tem Tm) Lm
dt J J
r
(15) Tem = 1.5p vds idr
s Ls (21)
Here, J is the system inertia and B represents the friction coefficient.
Substituting (18) into (14), the stator active and reactive powers can
be expressed as (22):
3.3. Maximum Power Point Tracking
Ps = 1.5vds ids , Qs = 1.5vds iqs (22)
Typically, a WT produces mechanical power by following a power
curve based on the wind-speed variation [35]. A WT is activated in two Substituting (20) into (22) gives the modified expressions for stator
operating regimes, namely the partial-load regime and full-load regime, active and reactive powers as (23) in terms of rotor currents.
depending on wind speed. Lm vds vds
An MPPT controller supports the WECS to extract the maximum Ps = 1.5 vds idr , Qs = 1.5 + Lm iqr
Ls Ls s (23)
possible energy from the wind. It provides reference signals [38] for the
RSC that permits the WECS to trail the maximum power point trajectory Substituting (18) into (11), analytical expressions for rotor voltages
at different wind speeds. An MPPT controller only functions in the (11) are converted to (24).
partial-load regime. In this study, the generator torque (Tem-ref) is se-
lected as a reference signal as it develops a quadratic relationship with (
vdr = Rr + Lr dt idr
d
) sl Lr iqr +
Lm sl
v
Ls s ds
the rotational speed of the WT [6,37]. For MPPT, the following re-
= (R )i
d
vqr + Lr dt + Lr idr
lationship (16) [6] is established:
r qr sl
(24)

Tem ref = km r
2
F r D (16) The electromagnetic torque and reactive power are controlled by
the outer open loops. The torque expression (21) is utilized to generate
Here, km denotes the maximum power at base wind speed, while F the rotor d-axis current reference (idr-ref), whereas the rotor q-axis cur-
and D represent the coefficients of the losses. rent reference (iqr-ref) is obtained using the reactive power (second
equation of (23)). The resultant expressions for rotor-current references
4. Proposed Control Scheme are presented as (25):

This section presents an explanation of the proposed modifications idr ref = s Ls


T
1.5pLm vds em ref
to the RSC control design for LVRT capability enhancement of the DFIG- Ls vds
iqr = Q
based WT. As stated earlier, the proposed scheme, which is based on ref 1.5Lm vds s ref s Lm (25)
generating suitable rotor-voltage references, is designed to restrict the
The stator reactive-power reference (Qs-ref) is considered to be zero,
transient rotor overcurrents and to mitigate the DC-link voltage varia-
as the magnetizing current is provided by the rotor [37]. As a result, the
tions during abnormal grid-voltage conditions, thus, improving the
expression (25) can be rewritten as (26):
dynamic performance of the DFIG by its enhanced LVRT capability.
First, in subsection 4.1, the principle and modeling of the RSC are ex- idr = s Ls
T
ref 1.5pLm vds em ref
plained, following which the proposed modifications to the RSC control
vds
structure are described in subsection 4.2. The detailed modeling of the iqr ref =
s Lm (26)
GSC control structure is presented in subsection 4.3. In terms of LVRT
requirements, the WT equipped with DFIG must fulfill the following The decoupling terms (vdr-comp. and vqr-comp.) are added to the outputs
minimum technical criteria (17): (1) the maximum acceptable limits of of the PI current regulators to produce the rotor dq voltage-reference
the stator and rotor currents are 1.5 times their rated values [4,18,26]; components. Decoupling terms are introduced to eliminate the inter-
(2) the allowable range of the DC-link voltage variations is limited ference between the d- and q-axis current controls. The expressions for
to ± 15% of its rated value [4,6,18]. the resultant rotor-voltage references and decoupling terms are given in
(27) and (28), respectively.
ir 1.5Ir , rated, is 1.5Is, rated
sl Lm
0.85Vdc ref vdc 1.15Vdc ref (17) vdr ref = PI (idr ref idr ) + vdr comp . + vds
s Ls
vqr ref = PI (iqr ref iqr ) + vqr comp . (27)
4.1. RSC Control vdr comp . = sl Lr iqr , vqr comp . = sl Lr i dr (28)

In the present study, SVOC is employed to control the RSC; it can be The obtained voltage references are converted into pulse-width
realized by aligning the d-axis of the SRF with the stator-voltage vector. modulation (PWM) switching signals to control the RSC.

4
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

Fig. 2. Block diagram of improved RSC control structure.

4.2. Modified RSC Control link voltage varies considerably. In this scenario, the difference be-
tween the reference and measured values of DC-link voltage is used to
The amplitude of the grid voltage is the key factor in deciding the obtain the reference value of DC-link current based on the error signal
system operation approach. It is assumed that during fault condition, by the PI controller. The measured DC-link current is subtracted from
the grid voltage (vg) drops below 90% of its nominal value. When the its reference value, and the error is provided to the PI controller to
grid fault occurs, i.e., vg ≤ 0.9 pu, the proposed control scheme will be obtain vDCC. When the DC-link voltage regains its reference value, the
initiated immediately to ensure the dynamic stability of the DFIG RSC returns to its MPPT function.
system during the LVRT.
The schematic of the improved RSC structure is presented in Fig. 2; 4.2.2. Modified RSC q-Axis Control
the rotor-position angle (θr) is referenced to the stator frame, while the As stated earlier, the DFIG is used to produce the stator active and
angle between the dq frame and the rotor is the slip angle (θsl). It is used reactive powers. During the stator voltage drop and the post-fault vol-
in RSC for abc/dq rotor current transformation and dq/abc transfor- tage recovery in the event of a grid fault, the system demands reactive
mation of rotor-voltage references and is determined by (29). power support from the DFIG. In this context, an additional stator
= (29) voltage regulator (SVR) is incorporated in the original control system of
sl s r
the RSC to contribute to the transient voltage-stability control. The
block diagram of the q-axis control including the SVR is illustrated in
4.2.1. Modified RSC d-Axis Control the lower part of Fig. 2.
To facilitate the transient-voltage stability of a DFIG-based WECS, a Comparing the stator reference and measured voltage values during
DC voltage-coupling controller (DCC) is employed between the RSC and the grid faults, and the error signal is input to a proportional-type
GSC, as highlighted in the upper part of Fig. 2. The output of the DCC regulator to produce vSVR. The output of the SVR is added to vqr-ref (27)
generates the reference value for the rotor d-axis voltage reference to determine the actual value of the rotor q-axis voltage reference (vqr-ref
(vDCC). This value is superimposed with vdr-ref (27), and their sum is the (actual)). The analytical expression for the actual rotor q-axis voltage
actual value of the rotor d-axis voltage reference (vdr-ref(actual)). The DCC reference is expressed by (31):
control system contains dual PI regulators in a cascade structure for the
DC-link voltage and current loops. The analytical expression for actual vSVR = P ( vs ref vs )
rotor d-axis voltage reference is obtained as (30): vqr ref (actual) = vqr ref+ vSVR (31)

idc ref = PI (Vdc ref vdc )


vDCC = PI (idc ref idc ) 4.3. GSC Control
vdr ref (actual) = vdr ref + vDCC (30)
As discussed, the key responsibility of a GSC is to maintain a con-
In the steady-state, the DCC does not activate, as the DC-link voltage stant DC-link voltage for the RSC. It also delivers some reactive power
is at its rated value. The RSC functions to control the DFIG and to im- support for the grid voltage [39]. A cascaded dual-control loop struc-
plement the MPPT of the WECS. However, during grid faults, the DC- ture is applied to perform the above-mentioned tasks, as presented in

5
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

Fig. 3. Block diagram of GSC control structure.

Fig. 3. A GSC control involves fast inner current-control loops and slow vdg ref = PI (idg ref idg ) + vdg comp . + vdg
outer-control loops to regulate the DC-link voltage and reactive power, vqg = PI (iqg iqg ) + vqg
ref ref comp . (37)
respectively. A GSC may deliver or receive active power to or from the
grid depending on the DFIG speed operation. The GSC is synchronized vdg comp . = g Lg i qg , vqg comp . = g Lg idg (38)
with the grid to avoid undesirable transient currents caused by the
phase difference between the GSC and the grid voltage. Here, an SRF
phase-locked loop is employed to serve this function. Equation (32) 5. Simulation Results and Discussion
expresses the dynamic model of a GSC with a grid filter (Lg).
didg The performance of the proposed improvements in RSC control
vdg = Rg idg Lg dt
+ g Lg i qg + vd design for a 1.5 MW DFIG-based WT test system, as depicted in Fig. 1,
vqg = Rg iqg Lg
diqg are examined herein. In support of the theoretical exploration and
g Lg i dg + vq (32)
dt
mathematical modeling of the improved RSC control structure for en-
In (32), the subscript g represents the grid, while vd and vq represent hancing the LVRT capability of the DFIG presented in Section 4, ex-
the GSC output voltage vectors in the dq frame. The d-axis of the dq tensive numerical simulations were carried out in MATLAB/Simulink.
reference frame is oriented with the grid voltage. As a result, the dq The DFIG parameters were borrowed from [40], and the relevant
components of the grid voltage are expressed as (33): parameters are summarized in Appendix A. The results confirm that the
suggested improvements in RSC produce satisfactory performance not
vdg Vg , vqg 0 (33) only in mitigating the transient rotor overcurrents during grid faults but
The grid active and reactive powers can be written as (34): also in effectively suppressing the DC-link variations according to the
criteria given in (17). Moreover, the performance of the proposed
Pg = 1.5(vdg idg + vqg iqg ) = 1.5vdg idg structure is verified for both symmetrical and asymmetrical grid faults.
Qg = 1.5(vqg idg vdg iqg ) = 1.5vdg iqg (34) A constant wind speed of 10 m/s is applied to the test system. The
DC-link voltage is held at 1150 V, and the utility grid voltage frequency
From (34), it is inferred that the grid active and reactive powers are is 50 Hz. The switching frequency is 2 kHz, whereas the sampling fre-
related to idg and iqg, respectively, thus keeping the grid voltage con- quency is taken as 200 kHz in this study. At first, the DFIG operates at
stant. The DC-side equation of the GSC is written as (35): the sub-synchronous mode of speed, and the initial slip is 0.5.
dvdc vdg Moreover, performance comparisons of the proposed control design
Cdc = idc 1.5 idg with two other methods, i.e., the conventional PI control method and
dt vdc (35)
the method presented in [33], are carried out.
To achieve a unity power factor, the grid reactive power reference
(Qg-ref) is taken as zero. Therefore, the grid-side q-axis current reference 5.1. Performance Comparison for Symmetrical Fault
(iqg-ref) becomes zero. By comparing the reference and measured values
of the DC-link voltage, the grid-side d-axis current reference (idg-ref) is To demonstrate the effectiveness of the modified RSC control
obtained according to the error signal by the PI regulator. The grid dq- structure for enhancing the LVRT ability of the DFIG-based WECS,
axes current references can be written as (36): comparative analyses were performed with two other methods to
Qg ref highlight the effects of a symmetrical fault on the behavior of the si-
idg ref = PI (Vdc ref vdc ), iqg ref = =0 mulated system. Figure 4 presents the numerical simulation results of
1.5vdg (36)
the selected transient waveforms for all three methods. A three-phase
The resultant rotor-current references are compared with their fault is applied to the LV-side of the coupling transformer for 300 ms at
corresponding measured values to obtain the error signals and applied the time instant of 6 s. Consequently, the stator voltage dropped below
to the inner current PI regulators. The outputs of the current regulators 10%, as can be seen in Fig. 4(a).
are added to the decoupling terms (vdg-comp. and vqg-comp.) to generate the As shown in Fig. 4(a), the maximum allowable rotor current is
grid-side dq voltage references (37) for the GSC. The decoupling terms successfully controlled within ± 1.5 pu with the proposed design. Thus,
are expressed by (38). there is no possibility of overcurrent for the RSC during the symmetrical

6
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

Fig. 4. Simulation results for symmetrical grid voltage fault: (a) proposed method, (b) conventional PI control method, and (c) method in [33].

grid fault. It is worth noting out that the oscillation of the generator when the conventional PI control method is implemented during the
torque Fig. 4(a)) as per the proposed design is much smaller than those fault, the rotor current exceeds ± 2.0 pu, resulting in the overcurrent of
of the two other methods, as shown in Figs. 4(b)–(c). Moreover, the the RSC. A similar case occurs for stator current, as its magnitude
waveform of the DC-link voltage shows that the transient voltage var- violates the criteria ((17) noted in Section 4. The postfault generator
iations remain within the limitations ( ± 15%). As shown in Fig. 4(b), torque oscillates considerably and exceeds 1.5 pu at the instant of the

7
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

Fig. 5. Simulation results for asymmetrical grid voltage fault: (a) proposed method, (b) conventional PI control method, and (c) method in [33].

fault. In contrast, when the control method presented in [33] is em- observed in comparison to the result achieved by the proposed method.
ployed, the rotor current exceeds ± 1.5 pu and violates the successful These results confirm the reasonable performance of the modified
LVRT criteria (17) with high torque oscillations, as shown in Fig. 4(c). RSC control design in suppressing transient rotor overcurrents and al-
Furthermore, the DC-link voltage is always restricted within the range leviating DC-link voltage variations during severe symmetrical grid
of ± 15% of its rated value. However, significant oscillations are faults while retaining them within their permissible ranges. Therefore,

8
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

considerably during asymmetrical fault, which can be seen in Fig. 6(b).


Finally, the total harmonic distortion (THD) results of the selected
waveforms for both the symmetrical and asymmetrical faults are pre-
sented. For this purpose, a THD block from the Simulink library was
used, and the results were obtained by multiplying the output of the
THD block by a factor of 100, as shown in Fig. 7. Figures 8 and 9 show
the THD results of the selected waveforms for symmetrical and asym-
metrical grid faults, respectively.

Fig. 6. Simulation results of grid-voltage-angle estimation: (a) symmetrical


6. Conclusions
fault and (b) asymmetrical fault.

In this study, a modified RSC control structure was designed and


presented for addressing the LVRT issue of DFIG-based WTs during grid
faults. The feasibility and validity of the proposed design for severe grid
voltage faults were confirmed via numerical simulations. A comparative
Fig. 7. Block diagram showing the procedure of obtaining the THD.
study was performed with two other methods to highlight the perfor-
mance of the proposed scheme. The following conclusions are drawn:
the potential of applying the proposed control design in the RSC
structure to control DFIG transient performance is confirmed, and the
LVRT requirements are fulfilled.
• A simple implementation of the proposed design based on gen-
erating suitable rotor-voltage references was achieved to meet the
LVRT requirements of a grid-connected DFIG, thus enabling this
5.2. Performance Comparison for Asymmetrical Fault scheme to attain system stability and reliability.

The dynamic behavior of the simulated system after incorporating


• The comparative analyses revealed that the proposed design de-
monstrated good transient performance for limiting the transient
the proposed control design was examined and compared with two rotor currents (< ± 1.5 pu of the rated current) and DC-link over-
other methods by applying an asymmetrical grid fault. A step single- voltages (< ± 15% of the rated DC voltage). Thus, these confirmed
phase to ground fault was applied, resulting in a voltage drop of over the suitability of the proposed design for wind-energy applications.
90% under the same operating conditions used in the case of the
symmetrical fault. For comparison purposes, Figs. 5(a)–(c) present the
• The electromagnetic torque oscillations under symmetrical and
asymmetrical grid faults were significantly suppressed with the
simulation results of these selected parameters. The results are com- proposed design.
parable to those of the case of the symmetrical fault. However, com-
pared with the symmetrical fault scenario, the dynamics of the rotor dq
• The proposed approach was easy to implement without any addi-
tional hardware requirements to deal with the grid fault, thereby
currents, generator torque, DC-link voltage, and stator active and re- making it cost-effective.
active powers contained more transients, as shown in Figs. 5(a)–(c).
Given the achieved simulation results, it is readily concluded that not Future works in this area are expected to focus on applying the
only the transient rotor currents but also the DC-link voltage variations proposed design to a real system and redesigning the proposed scheme
are restricted well within their prescribed performance bounds (17) via during voltage recovery.
the proposed improvements in the RSC control structure. As a result,
the LVRT capability of the DFIG is enhanced.
Credit Author Statement
The comparative analyses indicated that the proposed control
scheme is capable of effectively suppressing both the fault rotor current
Muhammad Arif Sharafat Ali: Conceptualization, Methodology,
and the torque oscillations of the generator and are considered its
Software, Writing- Original draft preparation. Khawaja Khalid
greatest advantages as compared to two other control methods.
Mehmood: Visualization, Investigation. Shazia Baloch: Writing-
To highlight the advantage of the SRF-PLL, the grid-voltage angle is
Reviewing and Editing. Chul-Hwan Kim: Supervision, Resources,
estimated by (39) and the results obtained for symmetrical and asym-
Writing- Reviewing and Editing
metrical grid faults are shown in Fig. 6.
v s
s est . = tan 1
Acknowledgement
v s (39)
In case of symmetrical grid fault, the variations of the grid-voltage This work was supported by the National Research Foundation of
angle are significant at the instants of fault occurrence and clearance Korea (NRF) grant funded by the Korea government (MSIP) (No.
only, as shown in Fig. 6(a). However, the grid-voltage angle varies 2018R1A2A1A05078680).

Fig. 8. THD results of the selected waveforms for symmetrical grid voltage fault.

9
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

Fig. 9. THD results of the selected waveforms for asymmetrical grid voltage fault.

Appendix A

Table A1.

Table A1
Parameters of 1.5 MW DFIG system.
Parameter Value Unit

Rated mechanical power 1.5 MW


Rated stator line-to-line voltage 690 V (rms)
Rated frequency 50 Hz
Pole pair number 2 Nos.
Stator winding resistance 2.65 mΩ
Stator leakage inductance 0.1687 mH
Rotor winding resistance 2.63 mΩ
Rotor leakage inductance 0.1337 mH
Magnetizing inductance 5.4749 mH
Rotational inertia 3 s
DC-link voltage reference 1150 V

References Sustain. Energy 4 (4) (2013) 1071–1081.


[15] M. Rahimi, Analytical assessment of the impact of stator-series passive resistive
hardware (SSPRH) on transient response and fault current contribution in DFIG
[1] A. Habib, R. Abbassi, A.J. Aristizábal, A. Abbassi, Forecasting model for wind power based wind turbines, Electric Power Syst. Res. 177 (2019) 105959.
integrating least squares support vector machine, singular spectrum analysis, deep [16] R. Zhu, Z. Chen, Closure to Discussion on Virtual Damping Flux-Based LVRT Control
belief network, and locality‐sensitive hashing, Wind Energy 23(2020) 235–257. for DFIG-Based Wind Turbine, IEEE Trans. Energy Convers. 31 (1) (2016) 408–409.
[2] Y.M. Alsmadi, et al., Detailed Investigation and Performance Improvement of the [17] M.J. Morshed, A. Fekih, A new fault ride-through control for DFIG-based wind
Dynamic Behavior of Grid-Connected DFIG-Based Wind Turbines Under LVRT energy systems, Electric Power Syst. Res. 146 (2017) 258–269.
Conditions, IEEE Trans. Ind. Appl. 54 (5) (2018) 4795–4812. [18] D. Xie, Z. Xu, L. Yang, J. Østergaard, Y. Xue, K.P. Wong, A Comprehensive LVRT
[3] M. Firouzi, G.B. Gharehpetian, LVRT Performance Enhancement of DFIG-Based Control Strategy for DFIG Wind Turbines With Enhanced Reactive Power Support,
Wind Farms by Capacitive Bridge-Type Fault Current Limiter, IEEE Trans. Sustain. IEEE Trans. Power Syst. 28 (3) (2013) 3302–3310.
Energy 9 (3) (2018) 1118–1125. [19] D. Xiang, L. Ran, P.J. Tavner, S. Yang, Control of a doubly fed induction generator
[4] Y. Shen, D. Ke, Y. Sun, D.S. Kirschen, W. Qiao, X. Deng, Advanced Auxiliary Control in a wind turbine during grid fault ride-through, IEEE Trans. Energy Convers. 21 (3)
of an Energy Storage Device for Transient Voltage Support of a Doubly Fed (2006) 652–662.
Induction Generator, IEEE Trans. Sustain. Energy 7 (1) (2016) 63–76. [20] S. Hu, X. Lin, Y. Kang, X. Zou, An Improved Low-Voltage Ride-Through Control
[5] G. Rashid, M.H. Ali, Nonlinear Control-Based Modified BFCL for LVRT Capacity Strategy of Doubly Fed Induction Generator During Grid Faults, IEEE Trans. Power
Enhancement of DFIG-Based Wind Farm, IEEE Trans. Energy Convers. 32 (1) (2017) Electron 26 (12) (2011) 3653–3665.
284–295. [21] L. Zhou, J. Liu, S. Zhou, Improved Demagnetization Control of a Doubly-Fed
[6] M.A.S. Ali, K.K. Mehmood, J.S. Kim, C.H. Kim, ESD-based Crowbar for Mitigating Induction Generator Under Balanced Grid Fault, IEEE Trans. Power Electron 30 (12)
DC-link Variations in a DFIG-based WECS, Proc. Int. Conf. Power Syst. Transients, (2015) 6695–6705.
Perpignan, France, 2019, pp. 1–6. [22] Q. Huang, X. Zou, D. Zhu, Y. Kang, Scaled Current Tracking Control for Doubly Fed
[7] A.M.A. Haidar, K.M. Muttaqi, M.T. Hagh, A Coordinated Control Approach for DC Induction Generator to Ride-Through Serious Grid Faults, IEEE Trans. Power
link and Rotor Crowbars to Improve Fault Ride-Through of DFIG-Based Wind Electron 31 (3) (2016) 2150–2165.
Turbine, IEEE Trans. Ind. Appl. 53 (4) (2017) 4073–4086. [23] D. Zhu, X. Zou, L. Deng, Q. Huang, S. Zhou, Y. Kang, Inductance-Emulating Control
[8] W. Chen, F. Blaabjerg, N. Zhu, M. Chen, D. Xu, Doubly Fed Induction Generator for DFIG-Based Wind Turbine to Ride-Through Grid Faults, IEEE Trans. Power
Wind Turbine Systems Subject to Recurring Symmetrical Grid Faults, IEEE Trans. Electron 32 (11) (2017) 8514–8525.
Power Electron. 31 (2) (2016) 1143–1160. [24] J. Yao, H. Li, Y. Liao, Z. Chen, An Improved Control Strategy of Limiting the DC-
[9] Y.K. Gounder, D. Nanjundappan, V. Boominathan, Enhancement of transient sta- Link Voltage Fluctuation for a Doubly Fed Induction Wind Generator, IEEE Trans.
bility of distribution system with SCIG and DFIG based wind farms using STATCOM, Power Electron 23 (3) (2008) 1205–1213.
IET Renew. Power Gener. 10 (8) (2016) 1171–1180. [25] R. Liu, J. Yao, X. Wang, P. Sun, J. Pei, J. Hu, Dynamic Stability Analysis and
[10] C. Wessels, F. Gebhardt, F.W. Fuchs, Fault Ride-Through of a DFIG Wind Turbine Improved LVRT Schemes of DFIG-Based Wind Turbines During a Symmetrical Fault
Using a Dynamic Voltage Restorer During Symmetrical and Asymmetrical Grid in a Weak Grid, IEEE Trans. Power Electron 35 (1) (2020) 303–318.
Faults, IEEE Trans. Power Electron 26 (3) (2011) 807–815. [26] S. Xiao, G. Yang, H. Zhou, H. Geng, An LVRT Control Strategy Based on Flux
[11] J. Yao, et al., Enhanced Control of a DFIG-Based Wind-Power Generation System Linkage Tracking for DFIG-Based WECS, IEEE Trans. Ind. Electron 60 (7) (2013)
With Series Grid-Side Converter Under Unbalanced Grid Voltage Conditions, IEEE 2820–2832.
Trans. Power Electron 28 (7) (2013) 3167–3181. [27] X. Li, X. Zhang, Z. Lin, Y. Niu, An Improved Flux Magnitude and Angle Control With
[12] M. Firouzi, G.B. Gharehpetian, B. Mozafari, Power-Flow Control and Short-Circuit LVRT Capability for DFIGs, IEEE Trans. Power Syst. 33 (4) (2018) 3845–3853.
Current Limitation of Wind Farms Using Unified Interphase Power Controller, IEEE [28] M. Reyes, P. Rodriguez, S. Vazquez, A. Luna, R. Teodorescu, J.M. Carrasco,
Trans. Power Deliv. 32 (1) (2017) 62–71. Enhanced Decoupled Double Synchronous Reference Frame Current Controller for
[13] G. Rashid, M.H. Ali, Transient Stability Enhancement of Doubly Fed Induction Unbalanced Grid-Voltage Conditions, IEEE Trans. Power Electron 27 (9) (2012)
Machine-Based Wind Generator by Bridge-Type Fault Current Limiter, IEEE Trans. 3934–3943.
Energy Convers. 30 (3) (2015) 939–947. [29] H. Geng, C. Liu, G. Yang, LVRT Capability of DFIG-Based WECS Under
[14] S. Alaraifi, A. Moawwad, M.S. El Moursi, V. Khadkikar, Voltage Booster Schemes for Asymmetrical Grid Fault Condition, IEEE Trans. Ind. Electron 60 (6) (2013)
Fault Ride-Through Enhancement of Variable Speed Wind Turbines, IEEE Trans. 2495–2509.

10
M.A.S. Ali, et al. Electric Power Systems Research 186 (2020) 106403

[30] R. Abbassi, S. Saidi, S. Marrouchi, S. Chebbi, Novel flexible algorithm for the op- [35] M.A.S. Ali, K.K. Mehmood, C.H. Kim, Full operational regimes for SPMSG-based
eration of renewable source grid interface VSCs under unbalanced voltage sags, WECS using generation of active current references, Int. J. Electr. Power Energy
IEEE 15th Int. Conf. Environment and Electrical Engineering (EEEIC), Rome, Italy, Syst. 112 (2019) 428–441.
2015, pp. 1414–1419. [36] F. Blaabjerg, D. Xu, W. Chen, N. Zhu, Advanced control of doubly fed induction
[31] R. Abbassi, S. Marrouchi, S. Saidi, A. Abbassi, S. Chebbi, Optimal Energy generator for wind power systems, Wiley, 2018.
Management Strategy and Novel Control Approach for DPGSs Under Unbalanced [37] G. Abad, J. Lopez, M. Rodriguez, L. Marroyo, G. Iwanski, Doubly fed induction
Grid Faults, J. Circuits Syst. Comput. 28 (4) (2019). machine: Modeling and control for wind energy generation, Wiley-IEEE, 2011.
[32] F. Lima, A. Luna, P. Rodriguez, E. Watanabe, F. Blaabjerg, Rotor Voltage Dynamics [38] R. Abbassi, S. Saïdi, M. Hammami, S. Chebbi, Analysis of Renewable Energy Power
in the Doubly Fed Induction Generator During Grid Faults, IEEE Trans. Power Systems: Reliability and Flexibility during Unbalanced Network Fault, Handbook of
Electron 25 (1) (2010) 118–130. Research on Advanced Intelligent Control Engineering and Automation, USA:IGI
[33] D. Zhu, X. Zou, S. Zhou, W. Dong, Y. Kang, J. Hu, Feedforward Current References Global, 2015, pp. 651–686.
Control for DFIG-Based Wind Turbine to Improve Transient Control Performance [39] M.A.S. Ali, K.K. Mehmood, S. Baloch, C.H. Kim, Wind-Speed Estimation and
During Grid Faults, IEEE Trans. Energy Convers 33 (2) (2018) 670–681. Sensorless Control for SPMSG-Based WECS Using LMI-Based SMC, IEEE Access 8
[34] J. Sim, I. Song, Y. Lee, H. Lee, Y Choi, Improvements in active power control (2020) 26524–26535.
methods for a wind farm based on modified wind turbine control in Danish grid [40] B. Wu, Y. Lang, N. Zargari, S. Kouro, Power conversion and control of wind energy
codes, J. Electr. Eng. Technol. 13 (4) (2018) 1438–1449. systems, Wiley-IEEE, 2011.

11

You might also like