Received: 13 December 2017 Revised: 17 April 2018 Accepted: 25 May 2018

DOI: 10.1002/etep.2636

RESEARCH ARTICLE

Grid‐connected converter control during unbalanced grid

conditions based on delay signal cancellation

Bane Popadic1 | Boris Dumnic1 | Dragan Milicevic1 | Vladimir Katic1 | Damir Sljivac2

1
Faculty of Technical Sciences,
Summary
Department for Power, Electronic and
Telecommunication Engineering, Considering the recent trend in energy sector transformation towards high share
University of Novi Sad, Novi Sad, Serbia of renewable energy sources, it has become very hard to imagine modern power
2
Faculty of Electrical Engineering, system without the integration of power electronic devices. A grid‐connected
Computer Science and Information
Technology, Department of Power converter will be on the forefront of future energy trading, while simultaneously
Engineering, University of Osijek, Osijek, striving to offer good dynamic behaviour and operation in full accordance with
Croatia
the relevant grid requirements. The control algorithm of the grid‐connected con-
Correspondence verter has to be capable of achieving the stable steady state operation even during
Bane Popadic, Faculty of Technical the most severe faults in the system. More importantly, the power quality of the
Sciences, Department for Power,
Electronic and Telecommunication
injected currents (and thus the power) needs to be kept at the maximum possible
Engineering, University of Novi Sad, Novi level. This paper presents the control strategy for the grid‐connected converter
Sad, Serbia. that offers the possibility of symmetrical grid current injection at the point of
Email: [email protected]
common coupling even during unbalanced grid conditions. Proposed control
Funding information strategy uses delay signal cancellation in the negative sequence synchronous
Ministry of Education and Science, Grant/
rotating reference frame for the mitigation of the respective current components.
Award Number: 042004
The negative influence of asymmetrical grid voltages, present at the point of
common coupling as a result of unbalanced grid loads or faults, will be shown
within the paper. The key features of the improved control method are outlined,
with a special reference to basic theoretical background. The proposed method is
experimentally verified using sophisticated research and development station for
control of grid‐connected converter.

KEYWORDS
control techniques, grid‐connected converter, renewable energy sources, symmetrical current
injection, unbalanced grid faults

List of Symbols: P, active power; Ug, amplitude of the grid voltage; Tδ, current control loop sample time; id, current value in the d‐axis; iq, current value in
the q‐axis; ξ, damping ratio; Kd, direct gain; Ls, equivalent grid inductance; Rs, equivalent grid resistance; ωb g , estimated grid voltage angular velocity; b
θg ,
estimated grid voltage vector angle; Td, grid time constant; Tg, grid voltage period; θg, grid voltage vector angle; p, instantaneous active power; q,
instantaneous reactive power; ωn, natural angular velocity; I +n (I‐n), nth harmonic positive (negative) sequence current amplitude; ϑ+n(ϑ‐n), nth
harmonic positive (negative) sequence current phase shift; U +n (U‐n), nth harmonic positive (negative) sequence voltage amplitude; φ+n(φ−n), nth
harmonic positive (negative) sequence voltage phase shift; φ0n, nth harmonic zero sequence voltage phase shift; T i, PI controller integral time
constant; KP, PI controller proportional gain; Tr, PI controller response time; Ts, PI controller settling time; GPI, PI controller transfer function; ε, PLL
tracking error; u bpαβðnÞ , positive (negative) sequence voltage in stationary reference frame; Q, reactive power; δ, signal overshoot; y, signal settling error
band; uαβ ðt Þ, voltage in stationary reference frame; ud, voltage value in the d‐axis; uq, voltage value in the q‐axis
Abbreviations: DSC, delay signal cancellation; FRT, fault ride through; GCC, grid‐connected converter; PCC, point of common coupling; PI,
proportional‐integral; PLL, phase‐locked loop; RES, renewable energy source; SFPLL, synchronous rotating phase‐locked loop; SOGI, second order
general integrator; SPWM, sinusoidal pulse‐width modulation; VSC, voltage source converter

Int Trans Electr Energ Syst. 2018;e2636. wileyonlinelibrary.com/journal/etep © 2018 John Wiley & Sons, Ltd. 1 of 18
https://doi.org/10.1002/etep.2636
2 of 18 POPADIC ET AL.

1 | INTRODUCTION

Reflecting on several economic and severe energy crises recently, humanity has faced the most important issue going
forward—offering a stable energy supply independent of the fossil fuel limited availability. Still, while the energy
remains the main source of humanity concerns, reconsidering energy technologies also offered very interesting
solutions. To facilitate the change of the perspective, the community has focused on achieving highest possible share
of energy produced using naturally renewable flow of primary energy in the environment. The former, complemented
with the requirement for highest possible combined energy efficiency, define main principles of future energy sector
transformation and sustainable development.
Clearly, renewable energy sources (RES) are by their very nature in full accordance with aforementioned principles.1
Therefore, the rising level of penetration for the RES worldwide is to be expected. The share of RES in the final energy
consumption is swiftly rising, while the leading markets have already surpassed the 2020 goals,2,3 as evident in Figure 1.
Most advances and the highest penetration in the RES sector stems from electrical energy sources, leading the future
power system to adopt a more decentralized approach.5
Previously inconceivable, it is currently very hard to imagine a modern renewable energy conversions system without
an integrated power electronics converter. This is particularly emphasized by the recent development of the technology
and the appropriate decrease of the component prices, making a 2‐level voltage source converter in full back‐to‐back topol-
ogy take the pivotal role as the interconnecting element between the RES system and the grid. This will assert a dominant
position for the grid‐connected converter (GCC) in future energy markets, with the responsibility of achieving necessary
interconnection and power quality requirements. Furthermore, this position will become even more significant when
considering the applications other than RES. Development of the electrical transport and future energy storage sectors will
inevitably offer very interesting solutions almost certainly including the GCC as the interconnecting element.6-9
Therefore, it is not hard to recognize the importance of the proper GCC control, aimed at fulfilling the interconnec-
tion requirements (including power quality). Furthermore, the control algorithm is required to uphold these require-
ments even for the most severe voltage conditions during unbalanced grid faults, offering the fault ride through
(FRT) capabilities required by the relevant grid code.10 Moreover, many distribution networks can frequently have
the voltage asymmetry in the “normal” operating regimes due to unbalanced loads.11-13 This may prove to be a difficult
task for standard GCC control techniques, because they are usually not well equipped to deal with asymmetrical grid
voltages at the point of common coupling (PCC). As presented in Jia et al,14 the control strategy may be power charac-
teristic oriented or voltage support oriented. The 2 direction further separate into many different strategies proposed on
the basis of instantaneous power theory. Within those different control strategies presented in other studies,15-21 the
balanced current control has always attracted special attention. In Ma et al,15 the control strategy for the elimination
of the negative sequence current is explained using well‐known instantaneous power theory, while the cross‐coupling
between positive and negative sequence has not been discussed. Yazdani and Iravani20 use generalized sinusoidal
pulse‐width modulation and voltage source converter sequence subsystem to balance the current or mitigate Direct

FIGURE 1 Renewable energy share in

gross final energy consumption with the
2020 target4
POPADIC ET AL. 3 of 18

Current (DC)‐bus voltage ripples, avoiding classical controllers and having the dynamic and steady state response sus-
ceptible to the parameter variation. Disturbance observer, with relatively complex second order general integrator‐based
estimator decoupling, is proposed in Ozsoy et al19 and tested using simulation. Both Liu et al17 and Jiang et al18 use sta-
tionary reference frame for the control of the currents, and therefore, the decoupling between positive and negative
sequence current is not discussed. Revelo and Silva21 use double decoupled synchronous reference frame phase‐locked
loop (PLL) for the synchronization and the notch filtering in both positive and negative current control loops in order to
regulate the negative sequence currents. In other studies,22-25 mostly harmonic distortion mitigation and synchroniza-
tion under distorted and asymmetrical voltages is discussed.
However, most common control techniques for the GCC are currently based on synchronous rotating reference
frame, utilizing the dynamic performance offered by the classical proportional‐integral (PI) controller. In that regard,
this paper aims at presenting the improved control strategy in the synchronous rotating reference frame for the GCC
operating under voltage asymmetry. The main feature of the proposed control strategy will be the injection of the
symmetrical currents, while operating under asymmetrical voltages at the PCC. This offers increased controllability
of the system, while improving system stability and the overall power quality. Mainly, the proposed control strategy will
ensure the adequate and stable current control, which is the basis of the modern FRT requirements, for any converter
operating under asymmetrical voltages. Additionally, the unbalanced conditions should not be amplified by the
converter with the uncontrolled injection of asymmetrical current at the PCC. Furthermore, expected decrease in both
active and reactive power oscillations will contribute to the stability of the system. Increased controllability and a
beneficial influence on system stability will offer the possibility for the long‐term (steady state) converter operation
under previously unimaginable conditions, increasing power system availability. In that regard, considering the
complete scope of the term (with major indicators being continuity of the service and waveform shape of transient
currents), overall power quality of the system should be increased.
The proposed control strategy is experimentally verified through an extensive series of tests, using custom made
advanced laboratory prototype for control of the inverter‐based distributed energy sources.

2 | G C C CO N T R O L

Most straight‐forward approach in the control of the GCC is based on the implementation of the well‐known principles
from the instantaneous power theory.26 Assuming the common voltage‐oriented control in synchronous reference
frame, active and reactive power control can be achieved in open‐ or closed‐loop control structure.27 When adequate
synchronization algorithm is applied, active and reactive power values can be obtained using the following equations:

3  3
P¼ ud id þ uq iq ¼ ud id ; (1)
2 2

3  3
Q¼ uq id −ud iq ¼ − ud iq : (2)
2 2

Therefore, the control algorithm can calculate the reference values for the d‐ and q‐axis current directly using

1  
i*d ¼ ⋅ u2d ⋅P−u2q ⋅Q ; (3)
u2d þ uq
2

1  
i*q ¼ ⋅ u 2
⋅P þ u 2
⋅Q : (4)
u2d þ u2q q d

To compensate for the power flow from the primary energy source, a DC‐link controller will be included in the
active power control path, modifying the reference. Sometimes, direct power control or virtual flux‐based control are
used as an alternative approach to the active and reactive power (PQ) control‐based structure.28,29 Figure 2 outlines
the basic control structure for the GCC in arbitrary application.
Obviously, the central part of the GCC control is occupied by the current control mechanism in the synchronous
reference frame.30 Considering extraordinary performance of the PI‐based controller in the synchronous reference
frame, the dominant position of the technique for the current application was never an issue. The transfer function
4 of 18 POPADIC ET AL.

FIGURE 2 Control algorithm structure for the grid‐connected converter. Abbreviations: FRT, fault ride through; PCC, point of common
coupling; PI, proportional‐integral

of the controller, presented in the Equation 5, implies the implementation simplicity, while offering high dynamic
performance.
 
1 1 þ T is
GPI ¼ K P ⋅ 1 þ ¼ Kp (5)
Tis Tis

Naturally, the selection of the PI controller parameters can be performed using different tuning methods, while the
main goal is achieving fast dynamic response with least possible overshoot. Using modulus optimum targeted perfor-
mance can be achieved, with response time Tr = 4.7 ⋅ Tδ, settling time Ts = 8.4 ⋅ Tδ, and overshoot less than 4.3%.
The parameters acquired by modulus optimum are calculated by the following equations:

Td
KP ¼ ; (6)
2K d T δ

Kp
Ki ¼ : (7)
Td

The system dominant time constant Td is equivalent to the grid time constant designated as τi = Ls/Rs, the Tδ is
sample time for the current control loop, and the direct gain equals Kd = 1/Rs. Clearly, when the grid parameters cannot
be determined precisely, some manual parameter tuning may be required for the actual system.
The most commonly used mechanism for the synchronization to the represent of the grid voltage vector is based on
the PLL structure. Because the original application is not related to the GCC control, the adopted structure was modi-
fied, and a significant number of consequent PLL‐based techniques was developed.31 The synchronization unit for the
GCC is usually implemented in the synchronous rotating (dq) reference frame, creating synchronous rotating PLL. As
evident in Figure 3, the Park transformation is used in order to adjust the voltage values at the PCC for the use in
synchronous rotating PLL.
Using a PI controller for the loop filter, the reference frame is aligned to the grid voltage component in the d‐axis.
Therefore, when the error signal reaches zero value, the estimated angle will become equal to the traced angle as per the
following set of rules:

FIGURE 3 Synchronous frame phase‐locked loop

POPADIC ET AL. 5 of 18

 
ε ¼ 0−uq ¼ −U g sin θg −b
θg ∧ε ¼ 0 ⇒ b
θg ¼ θ g ; (8)

 
ε≈U g θg −b
θg ; (9)

 
Ki 1
Δωg ¼ Gfil ðsÞεðsÞ ¼ Kp þ εðsÞ∧b
θg ¼ ωb g ðsÞ: (10)
s s

The parameters of the loop filter are tuned according to the general transfer function of the PLL and the desired
damping ratio (ξ), desired overshoot (δ), signal settling error band (y), and natural frequency by

U gK ps þ U gK i 2ξωn s þ ω2n
GPLL ðsÞ ¼ ¼ ; (11)
s2 þ U g K p s þ U g K i s2 þ 2ξωn s þ ω2n

jln δj ln y
ξ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∧t set ¼ − ; (12)
2 ξωn
π 2 þ ðln δÞ

ω2n 2ξ
Ki ¼ ∧ K p ¼ rffiffiffiffiffiffi: (13)
Ug Ug
Ki

3 | IMPROVED CONTROL TECHNIQUE UNDER ASYMMETRICAL VOLTAGE

CO NDITIONS U S I N G DE LAY S I G NA L C A NC E L L A T I O N

The GCC techniques, presented previously, are fairly easy to implement and offer good dynamic characteristics while
operating under rated grid conditions. However, any deviation from the rated grid conditions can have adverse conse-
quences on any part of the control structure. The deviation can range from inadequate power quality characteristics to
asymmetrical voltages at the PCC coupling causing the converter to enter and operate in inadmissible states.
Unlike the previous times, for modern power system, respective grid codes have adopted severe regulations in regard
to the operation of the GCC under arbitrary grid voltage (different from the rated conditions). These rules require the
GCC to offer grid voltage support during faults, ie, offer FRT capabilities. In that regard, it becomes mandatory to
uphold the proper control of the desired values, while negating the negative impact of the voltage asymmetry on the
control structure. Under such operating conditions, the control strategy can offer a wide variety of different optimization
paths. Most commonly, for the control, the choice is set at either controlling the symmetrical current injection or
controlling the desired power output characteristics.
Arbitrary voltage conditions at the PCC can occur during balanced or unbalanced grid faults, when the value of the
voltage can be presented by
0 2 3 2 3 1
cosðnωt þ φþn Þ cosðnωt þ φ−n Þ
 7  7 2   3
B 6 6 cos nωt þ φ0n C
B 6 cos nωt− þ φþn 7
2π 6 cos nωt þ 2π þ φ−n 7 C
∞ B 6 7 6 7 6 7C
u¼ ∑n¼1 BU þn 6 3 7 þ U −n 6 3 7 þ U 0n 4 cosðnωt þ φ0n Þ 5C: (14)
B 6  7 6  7 C
@ 4 4π 5 4 4π −n
5 cosðnωt þ φ0n Þ A
þn
cos nωt− þ φ cos nωt þ þφ
3 3

The 3 components represent nth harmonic component of the positive, negative, and zero sequence components,
respectively, designated by +n, −n, 0n. Arbitrary voltage will produce arbitrary currents, usually containing the same
order of harmonics and inherent asymmetry excluding the zero sequence components. Therefore, the injected current
can be represented as follows:
6 of 18 POPADIC ET AL.

0 2   3 2 31
cos nωt þ ϑþn cosðnωt þ ϑ−n Þ
B 6  7 6   7C
B 6 2π 7 6 cos nωt þ 2π þ ϑ−n 7C
∞ B 6 cos nωt− þ ϑþn 7 6 7C
i¼ ∑n¼1 BI þn 6 3 7 þ I −n 6 3 7C: (15)
B 6  7 6   7C
@ 4 4π 5 4 4π 5A
cos nωt− þ ϑþn cos nωt þ þ ϑ−n
3 3

Considering the instantaneous power theory, active and reactive power can be obtained from arbitrary voltages and
currents as26

∞ 0   1
∑n¼1 U þn I þn cos φþn −ϑþn þ U −n I −n cosðφ−n −ϑ−n Þþ
B ∞ ∞  þn
 C
B ∑m ¼ 1 ∑n¼1 U þm I þn cos ðωm −ωn Þt þ φþm −ϑ þ C
B C
3B C
m≠n
B ∞ ∞ −m −n −m −n C
p ¼ u⋅i ¼ B ∑m ¼ 1 ∑n¼1 U I cosððωm −ωn Þt þ φ −ϑ Þþ C; (16)
2B C
B m≠n C
B ∞ ∞ þm −n þm −n C
@ ∑m¼1 ∑n¼1 −U I cosððωm þ ωn Þt þ φ þ ϑ Þþ A
∞ ∞  
∑m¼1 ∑n¼1 −U −m I þn cos ðωm þ ωn Þt þ φ−m þ ϑþn þ

∞ 0   1
∑n¼1 U þn I þn sin φþn −ϑþn −U −n I −n sinðφ−n −ϑ−n Þþ
B ∞ ∞  þn
 C
B ∑m ¼ 1 ∑n¼1 U þm I þn sin ðωm −ωn Þt þ φþm −ϑ þ C
B C
3B C
m≠n
B ∞ ∞ C
q ¼ ju×ij ¼ B ∑m ¼ 1 ∑n¼1 −U −m I −n sinððωm −ωn Þt þ φ−m −ϑ−n Þþ C: (17)
2B C
B m≠n C
B ∞ ∞ C
@ ∑m¼1 ∑n¼1 −U þm I −n sinððωm þ ωn Þt þ φþm þ ϑ−n Þþ A
∞ ∞  
∑m¼1 ∑n¼1 U −m I þn sin ðωm þ ωn Þt þ φ−m þ ϑþn þ

It is evident that, when synchronous rotating reference frame transformation is applied and only the fundamental
frequency is considered, there are 6 distinct components of instantaneous power, controlled with 4 degrees of freedom,
ie, by positive and negative sequence of the d‐ and q‐axis currents.
Considering most common grid code requirements demand a defined value of reactive current component to be
injected during grid faults, the reference calculation is assumed by the FRT algorithm. Furthermore, the injection of
the negative sequence current during faults can lead to further degradation of the voltage and the inability to control
the desired current values. Consequently, the control algorithm should look to mitigate the negative sequence current
components.
However, the mitigation of the negative sequence current can prove to be a difficult task for the standard control
techniques. The main issue for the present techniques is the ineffectiveness of the Park transformation for the asymmet-
rical voltage or current vectors. Applying the standard Park transformation, with the negative sequence orientation, to
the arbitrary current vector in Equation 15, following expression is obtained:
" #

i−
d cosðϑ− Þ  þ
 cosð2ωtÞ  þ
 − sinð2ωtÞ
i− ¼ ¼I − þ
þ I cos ϑ þ
þ I sin ϑ : (18)
dq
i−
q sinðϑ− Þ sinð2ωtÞ cosð2ωtÞ

Obviously, the current in the synchronous rotating reference frame has a constant component and a component at
twice the fundamental frequency originating from negative sequence component. If we designate the expression 18 as
negative sequence d‐ and q‐axis current, it can be proven that the positive sequence d‐ and q‐axis currents will contain
components at twice the fundamental frequency originating from negative sequence current in the natural frame.
Applying the same reasoning, the exact components can be recognized in the d‐ and q‐axis voltages. These undesired
components will reduce the effectiveness of the PI controller, making the GCC control imprecise and at certain
instances even non‐existent. The synchronization unit will have an inherent error in the angle estimation, introducing
the oscillatory behaviour and further degrading the transformation. The current control will be unable to reach the
designated reference values required by the respective grid code.
It is then easy to imagine why the improvement of the control strategy is of the utmost importance.
POPADIC ET AL. 7 of 18

To offer the flexible control of the positive and negative sequence current components, it is necessary to achieve the
decoupling in either positive or negative sequence orientated transformation. Once the mitigation of the current com-
ponents is achieved in negative sequence reference frame, the positive reference frame will contain the cross‐coupling
component, and the standard control technique will be sufficient to achieve the desired steady state. For implementa-
tion of the proper negative sequence current control, the decoupling between positive and negative sequence
components needs to be achieved in the negative sequence synchronous rotating reference frame.
The paper proposes using delay signal cancellation (DSC) technique to extract the value of the negative sequence
current component in the negative sequence reference frame. The DSC method is commonly used for the separation
of the phase sequences in unbalanced power systems.32 This method is usually defined in the stationary reference frame
complex domain as

uαβ ðtÞ ¼ upαβ ðt Þ þ unαβ ðt Þ ¼ U p ⋅e jðωg tþφp Þ þ U n ⋅e− jðωg tþφp Þ ; (19)

p 1   n 1  
b
u αβ ðtÞ ¼ uαβ ðtÞ þ juαβ t−T g =4 buαβ ðt Þ ¼ uαβ ðt Þ− juαβ t−T g =4 : (20)
2 2

However, implementation of the DSC in the synchronous rotating reference frame is far more practical. Observing
the principle in Figure 4, the simplification used for the DSC implementation in synchronous reference frame is obvious
(where * designates arbitrary sequence reference frame). The parameter Tg represents a grid voltage period, while the
actual delay is implemented by the defined amount of sample delays. Clearly, the inherent delay introduced in the
control structure will influence the parameters of the negative sequence current PI controllers.
Proving that the DSC is capable of extracting the desired value of the current in the arbitrary sequence reference
frame is very simple:

FIGURE 4 Delay signal cancellation in A, a stationary and B, a synchronous reference frame

FIGURE 5 Block diagram of the improved GCC control strategy. Abbreviations: DSC, delay signal cancellation; GCC, grid‐connected
converter; PCC, point of common coupling
8 of 18 POPADIC ET AL.

0     1
I − cosðϑ− Þ þ I þ cos ϑþ cosð2ωt Þ−I þ sin ϑþ sinð2ωt Þþ
iDSC−
d ¼@  þ   π  þ  þ   A
π  ¼
− − þ
þI cosðϑ Þ þ I cos ϑ cos 2 ωt− −I cos ϑ sin 2 ωt−
 þ 2   2
! : (21)
2⋅I cosðϑ Þ þ I cos ϑ cosð2ωt Þ−I cos ϑþ sinð2ωt Þþ
− − þ þ
¼     ¼
þI þ cos ϑþ cos2ωt cosð−π Þ−I þ cos ϑþ sin2ωt cosð−π Þ
¼ 2⋅I − cosðϑ− Þ

FIGURE 6 Advanced laboratory

prototype for testing distributed energy
resources: ① dSPACE processor board, ②
grid side converter, ③ grid emulator, ④
and ⑤ switching and protection
distribution cabinets, ⑥ adapter block, ⑦
electrical machines, ⑧ torque controlled
drives, and ⑨ DC/DC converter

FIGURE 7 Structural organization of the control algorithm for the real‐time implementation
POPADIC ET AL. 9 of 18

Same procedure can be applied to all the current sequences in every rotating reference frame, making it possible
for the current control to achieve a stable steady state response with the desired references. Because the mitigation of
the negative sequence current can be performed in the respective synchronous reference frame and the DSC can
influence the parameters, similar decoupling in the positive sequence reference frame is unnecessary and should
be avoided.
Using the described DSC method, the standard control strategy for the GCC will be improved by the introduction of
the negative sequence current mitigation in the synchronous reference frame. The block diagram of the improved con-
trol strategy is presented in Figure 5. The positive and negative sequences are controlled independently, and the DSC
algorithm is added only in negative sequence controller. The output of the negative sequence controller is added to
the output of the positive sequence controller in the stationary reference frame.

FIGURE 8 Experimental verification for 3‐phase voltage dip—Standard control technique: A, d‐ and q‐axis currents in the positive
sequence frame; B, d‐ and q‐axis currents in the negative sequence frame; C, natural frame currents; D, grid‐connected converter active
and reactive power; and E, voltage at the point of common coupling
10 of 18 POPADIC ET AL.

4 | EXPER IM ENTAL VERIFIC ATION O F THE PR OPOS ED C ONTROL

STRATEGY

The experimental verification of the proposed control strategy is performed using advanced laboratory prototype for
testing distributed energy resources, developed at the Faculty of Technical Sciences in Novi Sad. The state‐of‐the‐art
hardware, presented in Figure 6, represents a scaled model of the inverter‐based distributed generator and allows for
extensive testing of proposed control strategies. Consisting of modular and highly versatile dSPACE processor board,
the system is capable of achieving extraordinary performance even for the most challenging applications in testing
different RES power technologies (ie, wind, wave, and solar energy conversion systems; combined heat and power
systems; electricity‐based storage systems).

6
3
0
-3
6
-6
23.28 23.3 23.32 23.34 23.36
3

-3
-6
17.41 17.43 17.45 17.47 17.49

1000

800

600

400
17.3 17.4 17.5 17.6

FIGURE 9 Experimental verification for 1‐phase voltage dip—Standard control technique: A, d‐ and q‐axis currents in the positive
sequence frame; B, d‐ and q‐axis currents in the negative sequence frame; C, natural frame currents; D, grid‐connected converter active
and reactive power; and E, voltage at the point of common coupling
POPADIC ET AL. 11 of 18

One of the most important elements for the verification of the proposed control strategy during different grid voltage
conditions at the PCC is the grid emulator. The grid emulator offers the possibility to set different voltage value with
arbitrary harmonic content. In addition, different faults including voltage variation (3 N–0‐480 V), frequency variation
(40‐400 Hz), harmonic sequence (up to 15th harmonic), and a combination of those may be provided for the GCC
control testing.
The control algorithm for the GCC is organized according to the proposed theoretical considerations and for the
real‐time implementation divided in several subsystems, as presented in Figure 7. The measurements and system pro-
tection are performed at 8 kHz sampling frequency, which is triggered by the interrupt at the midpoint of the PWM
period. The current control loop, including the negative sequence current control is executed 2 times slower, operating
at 4 kHz. The data acquisition for visualization, as the slowest operation, is set at 2 kHz sampling frequency.

FIGURE 10 Experimental verification for 2‐phase voltage dip—Standard control technique: A, d‐ and q‐axis currents in the positive
sequence frame; B, d‐ and q‐axis currents in the negative sequence frame; C, natural frame currents; D, grid‐connected converter active
and reactive power; and E, voltage at the point of common coupling
12 of 18 POPADIC ET AL.

As proposed by the theoretical consideration in Sections 2 and 3, the classical control strategy is more than capable
of handling symmetrical voltage variation at the PCC. This is demonstrated by Figure 8, where the behaviour of the
GCC is investigated under 10% 3‐phase voltage dip. This type of voltage variation, by all relevant grid codes, requires
the GCC to stay indefinitely connected to the distribution network, supplying the power mostly according to maximum
power point tracking algorithm.
During the experiment, the current reference for GCC was set to 2 A in the positive sequence, while the q‐axis
current reference is kept at 0. After the voltage dip has occurred, the reference value of the positive sequence current
was changed to 3 A. Evidently, the control strategy has managed to achieve the desired steady state after voltage

FIGURE 11 Experimental verification for 1‐phase voltage dip—Proposed control technique: A, d‐ and q‐axis currents in the positive
sequence frame; B, d‐ and q‐axis currents in the negative sequence frame; C, natural frame currents; D, grid‐connected converter active
and reactive power; and E, voltage at the point of common coupling
POPADIC ET AL. 13 of 18

variation that happened at 30 s. Negative sequence currents had zero value, and therefore, the injected currents were
symmetrical, while active and reactive power follow the respective references of the positive sequence currents.
Nevertheless, it remains to be seen how the classical GCC control technique behaves under 1‐ or 2‐phase 10%
voltage dip at the PCC. These voltage conditions, which are present in unbalanced networks, still require the GCC to
stay indefinitely connected to the distribution network and provide the desired values of positive sequence currents.
Figures 9 and 10 represent the behaviour of the classical GCC control strategy under 1 and 2‐phase voltage dip,
respectively. The positive sequence current references were kept the same as for the 3‐phase voltage dip (2 and 3 A).
The standard control strategy was not able to keep the desired steady state upon voltage dip occurrence because the
positive sequence reference had significant oscillations, while the negative sequence current appeared. Additionally,
significant oscillations of active and reactive powers can be noted for both cases.

FIGURE 12 Experimental verification for 2‐phase voltage dip—Proposed control technique: A, d‐ and q‐axis currents in the positive
sequence frame; B, d‐ and q‐axis currents in the negative sequence frame; C, natural frame currents; D, grid‐connected converter active
and reactive power; and E, voltage at the point of common coupling
14 of 18 POPADIC ET AL.

After 1‐phase voltage dip, negative sequence d‐ and q‐axis currents were both 1.2 A, while they increased to 1.3 and
1.8 A, respectively, after positive sequence reference was changed. For a 3 A positive sequence current reference, natural
frame currents were highly asymmetrical with 2‐phase amplitudes reaching 5 A. The oscillations of ±100 W, originating
after voltage dip, can be recognized in the active power for both positive sequence current reference, while the oscilla-
tions of the reactive power were ±75 VAr with 75 VAr mean value.
Similarly, the 2‐phase voltage dip caused higher natural frame currents asymmetry with amplitude reaching 5.6 A,
for the higher current reference. Additionally, significant introduction of the negative sequence q‐axis current can be
recognized at 2.3 A, with −0.5 A of negative sequence d‐axis component. This led to similar oscillation of positive
sequence currents, while active and reactive power had the difference of ±100 W and ±80 VAr, respectively. Consider-
ing the requirement for the indefinite operation of the GCC under 10% voltage dip, presented operation would be
considered inadequate and the GCC could be regarded as uncontrolled. Therefore, improvements of the standard
control techniques are necessary in order to allow the appropriate operation of the GCC in unbalanced grids. To test
the proposed control strategy, the same set of experiments was performed after the additional negative sequence current
cancellation was added to the control structure.
Figures 11 and 12 demonstrate the response of the improved control strategy for the cases of 1‐phase and 2‐phase
voltage dip, respectively.
Clearly, the improved control strategy had no problems operating under asymmetrical voltages, successfully regulat-
ing the positive sequence currents to the desired references. The current oscillations, present in the positive sequence for
the standard control strategy, are clearly mitigated. Furthermore, negative sequence currents are completely mitigated,
which is clearly obvious from the symmetry of the natural frame current. The dynamic response of the proposed control
strategy is very high (considering that voltage transition is not instantaneous), achieving steady state in less than 40 ms,
for both the reference change and the disturbance. The current overshoot is within the expected range of 20%. In respect
to the positive sequence reference, the injected current amplitude is equal to 2 A before and after the voltage dip and 3 A
after the reference change. The active and reactive power oscillations under unrated voltage conditions for the improved

(A) 8 (B) 1500

4 1000

0 500

-4 0

-8 -500
17.5 17.6 17.7 17.8 17.5 17.6 17.7 17.8

(C) 8

-4

-8
17.5 17.55 17.6 17.65 17.7 17.75 17.8

FIGURE 13 Experimental verification for 1 phase voltage dip—Standard control technique: A, d‐ and q‐axis currents in the negative
sequence frame; B, grid‐connected converter active and reactive power; and C, natural frame currents
POPADIC ET AL. 15 of 18

control strategy are noticeably lower, thus avoiding the possibility to lead the power system to instability and increasing
the reliability. The remaining oscillations are a consequence of the negative sequence voltage at the PCC, because the
current is located in the positive sequence.
To further verify the proposed theory, standard and improved control techniques are tested under 1‐phase 30% volt-
age dip starting at 17.6 s. The response of the standard control structure is presented in Figure 13, where the d‐ and q‐
axis currents in the negative sequence reference frame, active power, reactive power, and natural reference frame
currents are presented. Clearly, prior to the voltage dip event, the negative sequence currents hold zero value, while
the currents in natural frame are symmetrical at a predefined value of 2 A. Reactive power components were not
injected in the power system, while the active power was 750 W. When the voltage dip has occurred, the significant
increase in the injected current values can be noticed. Even if such an increase is expected at the moment of the fault,
the fact that the currents reach new steady state with the asymmetrical currents 3 times the value prior to the fault
should be addressed carefully.
The achieved steady state, different from the intended, undoubtedly proves that the GCC has relinquished the
control, operating in the quasi‐stable operating point. The inadequacy of the control technique can be proven once more
by observing the injected active and reactive power, where the significant oscillations can be noticed. These oscillations
may lead the power system to instability, especially where the penetration of the inverter‐based renewable energy
systems is high. The natural frame currents and the oscillations of the injected power are a consequence of the unnat-
urally high negative sequence current components in both d‐ and q‐axis, with 2.7 and −6 A, respectively. Therefore, as
proposed in this paper, with the mitigation of the negative sequence currents, it is expected that the remaining control
will be able to achieve the desired steady state (controlled state). The slight oscillations, within the negative sequence
currents originate from the positive sequence current components.
After proving the ineffectiveness of the classical control strategy, the proposed negative sequence current mitiga-
tion with the DSC has been implemented. The improved control strategy is tested against the same conditions at the
PCC, with the same d‐axis current reference prior to the fault. Figure 14 presented the behaviour of the GCC

(A) (B)
8 1500

4 1000

0 500

-4 0

-8 -500
15.25 15.35 15.45 15.55 15.25 15.35 15.45 15.55

(C) 8

-4

-8
15.25 15.3 15.35 15.4 15.45 15.5 15.55

FIGURE 14 Experimental verification for 1‐phase voltage dip—Delay signal cancellation‐based control technique: A, d‐ and q‐axis
currents in the negative sequence frame; B, grid‐connected converter active and reactive power; and C, natural frame currents
16 of 18 POPADIC ET AL.

operated using the proposed control strategy. Obviously, the current in the negative sequence reference frame have
been mitigated in less than 0.1 s after the fault has started (15.35 s). The inrush current at the moment of the fault
is still close to 3 times the current prior to the fault, difference being the achieved steady state where we observe the
current equal to the desired 2 A. After the control is established, the currents in the natural reference frame are
clearly symmetrical.
With the successful mitigation of the negative sequence currents, the oscillation of the injected active and reac-
tive power is significantly lower. It is evident that the proposed control strategy is capable of achieving the effi-
cient control of the GCC under unbalanced grid conditions. With the improved control strategy, the GCC can
preserve complete functionality during asymmetrical grid voltages, fulfilling the respective grid code requirements
and consequently enabling the further increase of RES integration level. Furthermore, prior to the occurrence of
the grid fault, there is no adverse effect of the control strategy on the GCC behaviour. Therefore, the proposed
technique can replace standard algorithm, not requiring the separate algorithm for the fault detection or the
switch on time.

5 | CONCLUSION

Recent advances in the power electronics technology resulted in the significant increase of the converters intercon-
nected with the power system. The number of the high‐power GCCs (mainly inverters) reached a new high recently,
heavily supported by RES penetration in the leading markets worldwide. However, the influence of the power electronic
converters on power system operation can be disregarded no more. Especially in large quantities, the GCCs can have
adverse impact on power quality including power availability and network stability. On the other hand, inadequate grid
conditions, especially regarding voltage conditions at the PCC, can adversely influence the GCC to have undesired
current components injection.
After acknowledging the known problems of the standard GCC control techniques under asymmetrical voltages, this
paper presented the possibility for the improvement of the control strategy. To improve the controllability of the GCC
during unbalanced grid faults (with asymmetrical voltages at the PCC), the paper proposed the use of DSC method in
the negative sequence reference frame for the mitigation of the negative sequence currents. As proved by the experi-
mental verification this, computationally nondemanding algorithm, was sufficient to allow the GCC to preserve control-
lability and enter the desired steady state in regard to the current reference. This is particularly important for the GCC
that aim at offering FRT capabilities during unbalanced faults. The decoupled control is achieved for both active and
reactive power, as well as for the positive and negative sequence currents.
Future research is aimed at improvements that include the remaining elements of the standard control technique,
especially PLL unit that would still be affected by asymmetrical voltages, even after mitigation of the negative sequence
currents. The broader improvements of the GCC control techniques will enable further increase of the RES integration
level in the power system. The final goal would, therefore, be the improvement of the power system availability,
stability, and ultimately reliability.

A C K N O WL E D G E M E N T
This paper is a result of the scientific project no. III 042004 of Integrated and Interdisciplinary Research entitled “Smart
Electricity Distribution Grids Based on Distribution Management System and Distributed Generation,” funded by
Republic of Serbia, Ministry of Education and Science.

ORCID
Bane Popadic http://orcid.org/0000-0002-2353-2312

R EF E RE N C E S
1. Twidell J, Weir T. Renewable Energy Resources. 2nd ed. London: Routledge; 2005.

2. Global Wind Energy Council—GWEC. Global wind report 2015, 2015.

POPADIC ET AL. 17 of 18

3. REN21. Global status report—renewables 2016. REN21, 2015. [Online]. Available: http://www.ren21.net/status‐of‐renewables/global‐
status‐report/. [Accessed: 22‐Feb‐2017].
4. Eurostat. Key Figures on Europe: 2017 Edition. Luxembourg: Publications Office of the European Communities; 2017.
5. Anees AS. Grid integration of renewable energy sources: challenges, issues and possible solutions. In 2012 IEEE 5th India International
Conference on Power Electronics (IICPE), 2012, pp. 1–6.
6. Amditis A, Brusaglino G, Spessa E. Optimizing vehicle to grid electric energy system with new technologies. 2015, pp. 353–359.
7. Kim B‐G, Ren S, van der Schaar M, Lee J‐W. Bidirectional energy trading for residential load scheduling and electric vehicles. 2013,
pp. 595–599.
8. Saponara S, Moras R, Roncella R, Saletti R, Benedetti D. Performance measurements of energy storage systems and control strategies in
real‐world e‐bikes. 2016, pp. 1–6.
9. Trabelsi M, Ben‐Brahim L, Gastli A, Abu‐Rub H. Enhanced low‐voltage ride‐through capability of flying capacitors inverter using model
predictive control. Int Trans Electr Energy Syst. 2017;27(12):e2430.
10. Diaz M, Cardenas R. Analysis of synchronous and stationary reference frame control strategies to fulfill LVRT requirements in wind
energy conversion systems. 2014, pp. 1–8.
11. Khushalani S, Schulz N. Unbalanced distribution power flow with distributed generation. 2006, pp. 301–306.
12. Nasr‐Azadani E, Canizares CA, Olivares DE, Bhattacharya K. Stability analysis of unbalanced distribution systems with synchronous
machine and DFIG based distributed generators. IEEE Trans Smart Grid. 2014;5(5):2326‐2338.
13. Sharma I, Bozchalui MC, Sharma R. Smart operation of unbalanced distribution systems with PVs and energy storage. 2013, pp. 1–6.
14. Jia J, Yang G, Nielsen AH. A review on grid‐connected converter control for short circuit power provision under grid unbalanced faults.
IEEE Trans Power Deliv. 2017;1‐1.
15. Ma K, Liserre M, Blaabjerg F. Power controllability of three‐phase converter with unbalanced AC source. 2013, pp. 342–350.
16. Ji Y, Sun P, Wu Y, Du X, Tai H‐M, Gu S. Power oscillation analysis and control of three‐phase grid‐connected voltage source converters
under unbalanced grid faults. IET Power Electron. 2016;9(11):2162‐2173.
17. Liu Y, Li N, Fu Y, Wang J, Ji Y. Stationary‐frame‐based generalized control diagram for PWM AC‐DC front‐end converters with unbal-
anced grid voltage in renewable energy systems. 2015, pp. 678–683.
18. Jiang W, Ma W, Wang J, Wang L, Gao Y. Deadbeat control based on current predictive calibration for grid‐connected converter under
unbalanced grid voltage. IEEE Trans Ind Electron. 2017;64(7):5479‐5491.
19. Ozsoy E, Padmanaban S, Mihet‐Popa L, et al. Control strategy for a grid‐connected inverter under unbalanced network conditions—a
disturbance observer‐based decoupled current approach. Energies. 2017;10(7):1067.
20. Yazdani A, Iravani R. A unified dynamic model and control for the voltage‐sourced converter under unbalanced grid conditions. IEEE
Trans Power Deliv. 2006;21(3):1620‐1629.
21. Revelo S, Silva CA. Current reference strategy with explicit negative sequence component for voltage equalization contribution during
asymmetric fault ride through: current strategy with negative sequence component. Int Trans Electr Energy Syst. 2015;25(12):3449‐3471.
22. Lai N, Kim K‐H. An improved current control strategy for a grid‐connected inverter under distorted grid conditions. Energies.
2016;9(3):190.
23. Xiao P, Corzine KA, Venayagamoorthy GK. Multiple reference frame‐based control of three‐phase PWM boost rectifiers under
unbalanced and distorted input conditions. IEEE Trans Power Electron. 2008;23(4):2006‐2017.
24. Galecki A. Control system of the grid‐connected converter based on a state current regulator with oscillatory terms. PRZEGL□D
ELEKTROTECHNICZNY. 2015;1(1):67‐71.
25. Bobrowska‐Rafal M, Rafal K, Jasinski M, Kazmierkowski M. Grid synchronization and symmetrical components extraction with PLL
algorithm for grid connected power electronic converters—a review. Bull Pol Acad Sci Tech Sci. 2011;59(4).
26. Akagi H, Watanabe EH, Aredes M. Instantaneous Power Theory and Applications to Power Conditioning. 1st ed. Hoboken, NJ:
Wiley‐IEEE Press; 2007.
27. Teodorescu R, Liserre M, Rodriguez P. Grid Converters for Photovoltaic and Wind Power Systems. 1st ed. Piscataway N.J: Wiley; 2011.
28. Malinowski M, Kazmierkowski MP, Hansen S, Blaabjerg F, Marques GD. Virtual‐flux‐based direct power control of three‐phase PWM
rectifiers. IEEE Trans Ind Appl. 2001;37(4):1019‐1027.
29. Liu X, Wang D, Peng Z. Predictive direct power control for three‐phase grid‐connected converters with online parameter identification:
predictive DPC for grid‐connected converters. Int Trans Electr Energy Syst. 2017;27(1):e2240.
30. Hojabri M, Ahmad AZ, Toudeshki A, Soheilirad M. An overview on current control techniques for grid connected renew energy.
Presented at the 2nd International Conference on Power and Energy Systems (ICPES 2012), Pune, India, 2012.
31. Guo X‐Q, Wu W‐Y, Gu H‐R. Phase locked loop and synchronization methods for grid‐interfaced converters: a review. Przegląd
Elektrotechniczny Electr Rev. 2011;R 87(4):182‐187.
18 of 18 POPADIC ET AL.

32. Svensson J, Bongiorno M, Sannino A. Practical implementation of delayed signal cancellation method for phase‐sequence separation.
IEEE Trans Power Deliv. 2007;22(1):18‐26.

How to cite this article: Popadic B, Dumnic B, Milicevic D, Katic V, Sljivac D. Grid‐connected converter
control during unbalanced grid conditions based on delay signal cancellation. Int Trans Electr Energ Syst. 2018;
e2636. https://doi.org/10.1002/etep.2636