International Journal of Power Electronics and Drive Systems (IJPEDS)

Vol. 12, No. 4, December 2021, pp. 2372~2383

ISSN: 2088-8694, DOI: 10.11591/ijpeds.v12.i4.pp2372-2383  2372

Enhanced controller for a four-leg inverter operating in a

stand-alone microgrid with unbalanced loads

Effat Ayoubi1, Mohammad Reza Miveh2, Ali Asghar Ghadimi3, Sajad Bagheri4
1,2,4Department
of Electrical Engineering, Arak Branch, Islamic Azad University, Arak, Iran
2Department of Electrical Engineering, Tafresh University, Tafresh, Iran
3Department of Electrical Engineering, Faculty of Engineering, Arak University, Arak, Iran

Article Info ABSTRACT

Article history: Stand-alone low voltage (LV) microgrids supplying small local loads far from
the utility grid are becoming an increasingly popular alternative to a total
Received Jun 24, 2021 reliance on the centralized utility grid. In most of LV microgrids, three-phase
Revised Oct 4, 2021 four-wire distribution systems are used to supply both single- and three-phase
Accepted Oct 11, 2021 loads. Unequal distribution of loads can result in voltage unbalance problems.
The use of the four-leg inverter is one of the best solutions for providing a
neutral current path and compensating unbalanced load conditions in stand-
Keywords: alone LV microgrids. This paper proposes a fast control technique to
compensate unbalanced voltage conditions for a four-leg inverter operating in a
Control stand-alone LV microgrid. The suggested technique provides the current
Dynamic performances controller’s orthogonal component without introducing any additional
Four-leg inverter dynamics or distortions. The major benefits of the recommended per-phase
Microgrid control technique over conventional orthogonal signal generation (OSG)
Voltage unbalance methods are enhanced steady-state and dynamic performances as well as
independency to the system parameters. Several simulation results are provided
to confirm the superior performance of the suggested methods.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Mohammad Reza Miveh
Department of Electrical Engineering
Arak Branch, Islamic Azad University, Arak, Iran
Email: [email protected]

1. INTRODUCTION
A microgrid contains distributed generators (DGs), power electronic devices, storages, and sensitive
loads that can operate in both grid-tied and islanded (stand-alone) modes [1]. Under islanding operation, at
least one of the DG units must work in the grid-forming mode to provide voltage and frequency stability [2].
This voltage and frequency control can be accomplished by inverter-based DGs, synchronous machines, or
other types of generators [3].
Most three-phase islanded microgrids feed unbalanced and/or nonlinear loads [4]. The existence of
single-phase loads in a three-phase microgrid can cause excessive neutral current, reactive power burden, and
unbalanced voltage conditions. However, power quality is highly important for sensitive loads in islanded
microgrids. Therefore, it is necessary to provide high power quality voltages in stand-alone microgrids [5],
[6].
The provision of the neutral wire has become an obligation for three-phase microgrids supplying
unbalanced loads. There exist multiple ways to provide a neutral wire in microgrids. Soto et al. [7], used a
three-phase conventional (three-leg) inverter, connecting its output to a ∆-Y transformer to provide the
neutral wire. This method is expensive compared to other methods and causes excessive copper and magnetic

Journal homepage: http://ijpeds.iaescore.com

Int J Pow Elec & Dri Syst ISSN: 2088-8694  2373

losses [3]. Using a transformer will make the circuit bulky due to the low frequency of the output voltage and
current. Furthermore, using a split dc-link capacitor is proposed in the literature [3], [8]. Balancing the dc
voltage on capacitors, when injecting unbalanced currents, can cause problems in the control procedure.
Moreover, capacitors’ capacities change differently over time and this can complicate the problem. Another
method for providing the neutral wire is to use a four-leg inverter [9]. In this type of inverter, two switches
are added to the three-phase conventional inverter which forms the fourth leg [10]. The neutral wire is
connected to the middle of the fourth leg. This method does not have the disadvantages mentioned above, i.e.
it is not expensive, it is less bulky, the efficiency is high, and balancing the capacitors’ voltages is not an
issue anymore [3].
The output voltage of the inverter must have a ripple-free voltage; hence, a low pass filter must be
used to block the switching frequency harmonics and pass the fundamental frequency [11]. There are several
types of power filters that block the switching frequency [12]. The most known ones are the L, LC, and LCL
filters [13]. L and LCL filters are generally used when the purpose is to control the output current. However,
since we aim to control the output voltage, the LC filter is an appropriate choice.
Various types of switching methods are also proposed to control the switches of all four legs [14],
[15]. Three-dimensional space vector modulation (3-D SVM) is one of the most common methods used for
this approach[10]. Some factors that have made this method so popular are low output voltage/current
distortion, high utilization of the dc-link voltage, the ability to minimize switching losses. Despite its
popularity, this method needs complex calculations for choosing the switching vectors. Another technique for
controlling the switches is to use the carrier-based pulse-width modulation (PWM) method which is proposed
in [15]. It was also proved to be equivalent to the 3-D SVM technique, but it is simpler to implement.
For using the PWM method, one needs a carrier signal and a voltage signal. The voltage signal is the
output of the controller which can come in different types. Well-known controllers that are used for
controlling an inverter are hysteresis controller [16], deadbeat (DB) controller [17], finite-control-set model-
predictive control (FCS-MPC) [18], sliding mode control (SMC) [19], proportional-integral (PI) [20] and
proportional resonant (PR) controller [21]. Variable switching frequency and high current ripple are the
major boundaries with hysteresis control. The problem of the chattering phenomenon in discrete
implementation is the main disadvantage of the SMC. Predictive control is difficult and involves a lot of
calculations. On the other hand, PR controllers are normally used for sinusoidal reference voltages yet, even
PR controllers are unable to fully eliminate the steady state error of a sinusoidal control signal. In order to
overcome this problem, one can use PI controllers with dc signals, because PI controllers can completely
track dc reference signals. One way to convert the AC waveforms into dc ones is to use Park transformation
[22]. The output of this transformation, the dq signals are dc values and can be controlled by PI controllers.
However, when the three-phase signals become unbalanced, the dq signals will not be pure dc anymore. A
double-frequency ripple will be added to both d and q axes. Hence, PI controllers cannot properly control the
signals and the steady-state error will not be zero. The PI controller has good performance by tuning the
parameters for first-order and second-order systems; nonetheless, it becomes more difficult for higher order
systems and it is tuned by try and error. The fuzzy-PID controller has shown good performance in these
situation [23].
Several methods have been proposed to eliminate this ripple. One of them is to utilize low-pass
filters, but this method decreases the response time [24]. Another method is to first calculate the symmetrical
components and then use the Park transformation [25]. Since the symmetrical components are balanced, the
output of the abc-dq transformation will be pure dc. The disadvantage of this method is that calculating
symmetrical components is based on implementing a time-domain delay to the signals. This delay is modeled
as an all-pass filter in the frequency-domain. Therefore, it will be difficult to include the symmetrical
components calculators (SCCs) in the small-signal model [26].
In this paper, a fast control approach for a four-leg inverter in the islanded mode under unbalanced
conditions is proposed. On the outer voltage loop, a second-order generalized integrator (SOGI) is used to
create the α and β components from the real signals. However, using this technique in the current controller
introduces undesired delays. In this paper, a fast and robust method is used to create the orthogonal signal for
the current control loop. This method does not impose any delays to the system and also it is not dependent
on the system parameters.
The paper is structured as shown in: Section II is associated with modeling the three-phase four-leg
inverter and its mathematical equations. The third section describes the proposed control technique in this
paper in detail. The simulation results under different scenarios are given in section IV and it is shown that
this technique can provide balanced voltages in highly unbalanced conditions. Finally, the conclusion is
provided.

Enhanced controller for a four-leg inverter operating in a stand-alone … (Effat Ayoubi)

2374  ISSN: 2088-8694

2. THREE-PHASE FOUR-WIRE INVERTER MODEL

2.1. Topology
Configuration of the three-phase four-leg inverter with an output LC filter is shown in Figure 1. As
shown in this Figure, an ideal battery is used to simulate the dc-link voltage which simplifies the analysis.
Moreover, to control the zero-sequence current, an additional fourth leg is connected to the conventional
three-phase inverter. The additional fourth leg increases the number of gating signals and thus the control
complexity. However, the for-leg inverter can handle balanced or unbalanced loads and linear or nonlinear
loads. The inductor used in the fourth leg, as well as the inductors in the other legs, minimizes the switching
frequency ripples.

R L
iLa iOa va
a
+ iLb iOb vb
vC b
- iLc iOc vc
f c

Load
C

in
n

Figure 1. Four-leg inverter with loads

2.2. Mathematical model

The voltage of each phase in Figure 1 is obtained using Kirchhoff’s voltage law to the three-phase
four-leg inverter as shown in (1):

𝑑 𝑑
𝑒𝑎𝑏𝑐_𝑓 = (𝑅 + 𝐿 ) 𝑖𝐿_𝑎𝑏𝑐 + 𝑣𝑎𝑏𝑐_𝑛 − (𝑅 + 𝐿 ) 𝑖𝑛 (1)
𝑑𝑡 𝑑𝑡

where 𝑒𝑎𝑏𝑐_𝑓 = [𝑒𝑎𝑓 , 𝑒𝑏𝑓 , 𝑒𝑐𝑓 ]𝑇 , 𝑣𝑎𝑏𝑐_𝑛 = [𝑣𝑎𝑛 , 𝑣𝑏𝑛 , 𝑣𝑐𝑛 ]𝑇 and 𝑖𝐿_𝑎𝑏𝑐 = [𝑖𝐿𝑎 , 𝑖𝐿𝑏 , 𝑖𝐿𝑐 ]𝑇 represent the input
voltage, output voltage, and inductor current, respectively. 𝐼𝑛 is also defined as the neutral current.
Furthermore, by applying Kirchhoff’s current law in Figure 1, the following equations are obtained:

𝑑
𝑖𝐿_𝑎𝑏𝑐 = 𝐶 𝑣𝑎𝑏𝑐_𝑛 + 𝑖𝑂_𝑎𝑏𝑐 (2)
𝑑𝑡

𝑖𝑛 = 𝑖𝐿𝑎 + 𝑖𝐿𝑏 + 𝑖𝐿𝐶 (3)

where 𝑖𝑂_𝑎𝑏𝑐 = [𝑖𝑂𝑎 , 𝑖𝑂𝑏 , 𝑖𝑂𝑐 ]𝑇 represents the load current.

As shown in (1)-(3) represent the system model in the abc frame. Using the Park transformation, the
dq0 frame model of the system is given by (4) and (5).

𝑑
𝐸𝑑 = 𝑉𝑑 + (𝑅 + 𝐿 ) 𝐼𝑑 − 𝐿𝜔𝐼𝑞
𝑑𝑡
𝑑
𝐸𝑞 = 𝑉𝑞 + (𝑅 + 𝐿 ) 𝐼𝑞 − 𝐿𝜔𝐼𝑑 (4)
𝑑𝑡
𝑑
{ 𝐸0 = 𝑉0 + (𝑅 + 𝐿 ) 𝐼0
𝑑𝑡

𝑑
𝐼𝐿𝑑 = 𝐶 𝑉 + 𝐼𝑂𝑑 − 𝐶𝜔𝑉𝑞
𝑑𝑡 𝑑
𝑑
𝐼𝐿𝑞 = 𝐶 𝑉 + 𝐼𝑂𝑞 + 𝐶𝜔𝑉𝑑 (5)
𝑑𝑡 𝑞
𝑑
{ 𝐼𝐿0 = 𝐶 𝑉 + 𝐼𝑂0
𝑑𝑡 0

where ω is the coordinate angular frequency for the Park transformation. Park transformation matrix is
shown in (6):

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2372 – 2383
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2375

2𝜋 2𝜋
cos(𝜃) 𝑐𝑜𝑠 (𝜃 − ) 𝑐𝑜𝑠 (𝜃 + )
3 3
2 2𝜋 2𝜋
𝑇= sin(𝜃) 𝑠𝑖𝑛 (𝜃 − ) 𝑠𝑖𝑛 (𝜃 + ) (6)
3 3 3
1 1 1
[ 2 2 2 ]

To design the system controller and analyze the stability of the inverter, the per-phase method can
be used [27]. Here the single-phase inverter representation of the four-leg inverter is used for design and
analysis [28]. The per-phase model for one phase to the neutral point of a four-leg inverter is demonstrated in
Figure 2. According to this figure, IL and VC are inductor current and capacitor voltage, respectively. Input
voltage and the load connected to that phase are represented as an ideal control voltage source and current
source, respectively.
Figure 3 depicts the small-signal block diagram of the inverter with output LC filters. The block
diagram is in the dq frame. Control of signal in the synchronous reference frame is so easy with PI
controllers; because PI controllers have a simple implementation. As it is shown, one pair coupling exists in
this figure while they omitted by feed-forwards that’s design in the Figure 4. So, control of the system is well
done in dq frame without any coupling and d and q components are controlled independently.

+ iL +
einv C vC iO
- -

Figure 2. Per-phase/single-phase representation of an inverter

Vd 1/RLoad
Iod
-

V*d + I*Ld ILd + ICd

Gc2 Gc1 1/(R+Ls) 1/Cs Vd
+
-

ωC

ωC

-
-

V*q + I*Lq + + ICq

Gc2 Gc1 1/(R+Ls) 1/Cs Vq
ILq
-

Ioq
Vq 1/RLoad

Figure 3. Small signal block diagram of the grid-forming inverter with output LC filter

3. THE PROPOSED CONTROL TECHNIQUE

3.1. Per-phase controller
The control block diagram of the three-phase four-leg inverter is shown in Figure 4. As it is shown
in this figure, this controller includes a cascaded voltage control loop and a current control loop to
compensate unbalanced voltage conditions in a stand-alone LV microgrid. Also, VCa and ILa represent the
capacitor voltage and inductor current of the output filter of phase a, respectively. V α and Iα are the same as
VCa and ILa. Vβ is the orthogonal component of Vα. Vβ is attained from the SOGI block [29]. SOGI is one of
the OSG types and its structure is represented in Figure 5. As it is shown in Figure 5, the SOGI consists of
two integrators that result slow dynamic responses (In Figure 5, k is set to 1 in this work). In order to obtain a
better dynamic response, a faster method is needed to compute the orthogonal component. For solving this
problem, the article uses another OSG method that is explained in section B and I β is obtained from this
method. The most important difference between Figure 4 and the block diagram shown in [28] is the method
of obtaining the Iβ that has no dependency on the grid’s parameters. After the OSG blocks, the d and q
components of the capacitor voltage, inductor current, and output current are attained by the park
Enhanced controller for a four-leg inverter operating in a stand-alone … (Effat Ayoubi)
2376  ISSN: 2088-8694

transformation blocks. It is needed to provide a synchronous reference frame in order to use simple
controllers like Fuzzy-PIDs to regulate the reference values. The PI controller gains were designed as done in
[28] and the Fuzzy-PID were designed as done in [23]. Eventually, Eα is equal to the Ea fed into the PWM
block, while Eβ is discarded. As shown in Figure 4, two feed-forward (FF) blocks are used in both the current
controller and the voltage controller loops. The main purpose of these four FF blocks is to compensate the
coupling between the d and q axes. Moreover, sinθ and cosθ values are needed for converting ac voltages and
currents into dc ones. Since the inverter is working in the grid-forming mode, sinθ and cosθ values are known
and they do not need to be measured or calculated.

Iq*

Id*
tan 1  q* 
 Id 
 
 I* 

Iq*

Id*
 cos()

I d*  I q*
2
Ea ILa Vca IOa

α a a
×
×

αβ αβ αβ

Eα ILa=Iα Vα Vβ IOα IOβ

Iβ

αβ cos  αβ αβ cos  αβ
dq sin  dq dq sin  dq

Ed Eq Vq Vd Vq Vd IOd IOq

Id IOq
FF FF
Vq Vd
+ - Iq* + -
Fuzzy_PID PI Vq*

- -
Fuzzy_PID PI Vd*
+ Id* +
Iq IOd
FF FF
Vd Vq

Figure 4. Block diagram of the proposed per-phase controller

-

v v'
+ k + ×
- ×

qv’
×
× 

SOGI

Figure 5. SOGI block diagram

If the inverter were supposed to connect to a grid and inject current, it would be necessary to
measure θ using a phase locked loop (PLL) and calculate sinθ and cosθ values. Therefore, in this situation,
the small-signal model of the PLL should also be added to the system which causes delays. By using this
method each phase is controlled independently with any extra dynamics into the current control loop and
finally, the unbalanced load voltages are compensated in the best way.

3.2. The proposed OSG technique

The OSG used in this paper generates an orthogonal current (I β). The idea of this method has been
proposed in [27], but it is used in another topology of the inverter and another topology of the grid. As it is
shown, one of the fastest alternatives for the SOGI is the OSG depicted in Figure 6.
This technique is applied in the four-leg inverter for computing the β axis of the load current. Using
this method, the reference values of the load currents, Iα and Iβ, are obtained from (7) and (8), respectively [27].

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2372 – 2383
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2377

𝐼𝛼∗ = 𝐵sin(𝜑 + 𝛾) (7)

𝐼𝛽∗ = 𝐵cos(𝜑 + 𝛾) (8)

where, B, γ and φ are as shown in (4)-(5),

𝐵 = √𝐼𝑑∗2 + 𝐼𝑞∗2 (9)

𝐼∗
𝛾 = 𝑡𝑎𝑛−1 ( 𝑞∗ ) (10)
𝐼𝑑

𝑡
𝜑 = ∫ 𝜔(𝜏)𝑑𝜏 (11)
0

By using (7)-(11), the beta component of the reference current is generated as a function of 𝐼𝑑∗ and
𝐼𝑞∗ .
The advantage of this method is its independence on the grid’s parameters such as the inductance or
resistance of the output filter. Also, this method does not add any distortion and dynamics to the system and
generates the orthogonal component very fast, while the SOGI has an inherent delay in generating the beta
component. Eventually, the output current response is fast and smooth and the load voltage is constant under
balanced and unbalanced loads.

Id *
I d*  I q*
2 2

*
Iq

×
θ +  cos() × Iˆ
+
I d*  I* 
tan 1  q* 
I 
I q*  d

Figure 6. Block diagram of the used OSG to obtain Iβ

3.3. PWM Technique

Fuzzy-PID controllers in the current control create the reference signals. Then these signals are
transformed back to the abc frame. In the next step, the abc signals are imported to the PWM block to create
the switching signals and produced the three-phase reference voltages of the inverter. The carrier-based
PWM method has been used in this research for the generation of the inverter output voltages [14]. This
method has been applied because of its strong performance and simple implementation. The block diagram of
carrier-based PWM is presented in Figure 7.

eao
eaf + + a+
+ -
ebo
ebf + + b+
+ -
eco
ecf + + c+
+ -
Offset Voltage efo
+ f+
Calculation -

+Vc/2

-Vc/2
Carrier

Figure 7. Block diagram of carrier-based PWM

Enhanced controller for a four-leg inverter operating in a stand-alone … (Effat Ayoubi)

2378  ISSN: 2088-8694

eaf, ebf, and ecf are inputs of the α component obtained from the four-leg inverter. They can be represented as
(12).

𝑒𝑎𝑜 = 𝑒𝑎𝑓 + 𝑒𝑓𝑜

{𝑒𝑏𝑜 = 𝑒𝑏𝑓 + 𝑒𝑓𝑜 (12)
𝑒𝑐𝑜 = 𝑒𝑐𝑓 + 𝑒𝑓0

where, ’o’ stands for the fictive midpoint of the dc-link and efo is the offset voltage that is calculated as shown
in (13),

𝑒𝑚𝑎𝑥 𝑒𝑚𝑖𝑛 𝑒𝑚𝑖𝑛 +𝑒𝑚𝑖𝑛

𝑒𝑓𝑜 = 𝑚𝑖𝑑 (− ,− ,− ) (13)
2 2 2

where efo corresponds to the medium value of the mentioned variables. Also, emax and emin correspond to the
maximum and minimum instantaneous values of eaf, ebf, and ecf respectively. They are also represented in (14)
and (15),

𝑒𝑚𝑎𝑥 = 𝑚𝑎𝑥(𝑒𝑎𝑓 , 𝑒𝑏𝑓 , 𝑒𝑐𝑓 ) (14)

𝑒𝑚𝑖𝑛 = 𝑚𝑖𝑛(𝑒𝑎𝑓 , 𝑒𝑏𝑓 , 𝑒𝑐𝑓 ) (15)

And their constraints are represented as (16),

−𝑉𝑑𝑐 ≤ 𝑒𝑎𝑓 , 𝑒𝑏𝑓 , 𝑒𝑐𝑓 ≤ 𝑉𝑑𝑐 (16)

Also, the ON-times of the upper switches of the respective legs can be achieved from (17).
𝑇𝑠 𝑒𝑎𝑜
𝑇𝑎 = + 𝑇
2 𝑉𝑑𝑐 𝑠
𝑇𝑠 𝑒𝑐𝑜
𝑇𝑏 = + 𝑇
2 𝑉𝑑𝑐
𝑇𝑠 𝑒𝑐𝑜 (17)
𝑇𝑐 = + 𝑇
2 𝑉𝑑𝑐
𝑇𝑠 𝑒𝑑𝑜
{ 𝑇𝑓 = 2
+
𝑉𝑑𝑐
𝑇

where Ts is the period of the triangular carrier signal.

4. SIMULATION AND RESULTS

The average model of the three-phase four-leg inverter shown in Figure 1 has been simulated in
MATLAB/Simulink software. Then, the performance of the transient and steady-state response of the
suggested control method is tested and evaluated when different loads are connected to the stand-alone grid.
At last, the performance of the offered technique is compared with the conventional method [26] and fictive
axis emulation (FAE) method [28]. System parameters under study are indicated in Table 1.

Table 1. System parameters under study

Variable description Value
DC link voltage 460 [V]
Load voltage 180 [V]
Filter capacitor 25 [μF]
Filter inductor 3 [mH]
The internal resistance of filter inductor 0.1 [Ω]
Grid frequency 60 [Hz]
Switching scheme carrier-based PWM
FF1 gain 2∗π*60*25*10-6
FF2 gain 2*π*60*3*10-3
Vd* 220 [V]
Reference voltages
Vq* 0 [V]
Prop. (Kp) 1
PI voltage controller gains
Integ. (Ki) 1

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2372 – 2383
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2379

Three different scenarios for loading conditions are considered. At first, the four-leg inverter
supplies a resistive balanced load (Ra = Rb = Rc = 10.75 [Ω]). Next, at 0.3 [s], an inductive load is added
between phases B and C (R=30 [Ω] and L=15 [mH]), and finally, at 0.5s, a resistive load is added on phase C
(R=15 [Ω]). Figure 8 illustrates the output waveform of the load voltages and current in the abc frame for all
three methods. Figures 8 (a)-(c) is also depicted the unbalanced load condition that is happened in 0.3 [s],
when an inductive load is added between phases B and C. Also, the results of the proposed technique are
compared with the FAE method [28] and conventional technique [26]. It is observed that the waveforms of
all control methods remain balanced in steady-state at all the various load conditions that are considered in
this study. However, as seen, the transient response of the proposed method is faster than the two other
conventional methods. In addition, the proposed method is compensated for two phase unbalances with
smaller disturbance.
Figure 9 also depicts the output voltage and current waveforms of the inverter for unbalanced load
conditions. In this test, a resistive load is added on phase C (R=15Ω) at 0.5 [s]. Figures 9 (a) depicts the
voltages and currents of the conventional method, respectively. Figures 9 (b) also show the voltages and
currents of the FAE method, Figures 9 (c) illustrates voltages and currents of the suggested method,
respectively. As can be seen, the transient response of the proposed method is so faster than the FAE and
conventional methods, and the load voltage is compensated under highly unbalanced conditions.

Time (s) Time (s) Time (s)

(a) (b) (c)

Figure 8. Output voltage and current waveforms of the inverter for unbalanced load condition at t=0.3[s];
(a) conventional technique voltage and current, respectively; (b) FAE technique voltage and current,
respectively; and (c) the proposed technique voltage and current, respectively

Time (s) Time (s) Time (s)

(a) (b) (c)

Figure 9. Output voltage and current waveforms of the inverter for unbalanced load condition at t=0.5[s]; (a)
conventional technique voltage and current, respectively; (b) FAE technique voltage and current,
respectively; and (c) the proposed technique voltage and current, respectively

Enhanced controller for a four-leg inverter operating in a stand-alone … (Effat Ayoubi)

2380  ISSN: 2088-8694

The reference and actual waveforms of voltage in phases A, B, and C for the grid-forming inverter
in the q and d components are depicted in Figure 10 and Figure 11, respectively. Figure 10 (a)-(c)
demonstrated the obtained q voltage components of phase A, B and C via the FAE method and proposed
method, respectively and Figures 11 (a)-(c) depicted obtained d voltage components of phase A, B and C via
the FAE method and proposed method, respectively. By comparing the proposed method’s operation with the
FAE method one can better observe in this article’s scheme is the better speed of transient response rather
than the FAE to following load changes. Deviations in the voltage waveforms are also shorter and smaller.
The reference and actual waveforms of current in phases A, B, and C for the grid forming inverter in
the q and d components are depicted in Figure 12 and Figure 13, respectively. Figure 12 (a)-(c) demonstrated
the obtained q current components of phase A, B and C via the FAE method and proposed method,
respectively and Figure 13 (a)-(c) depicted obtained d current components of phase A, B and C via the FAE
method and proposed method, respectively. As can be seen, the reference current following in the proposed
method is so faster than the FAE and its deviation is so shorter and smaller. Fast transient response is the
result of the proposed method.

Time (s) Time (s) Time (s)

(a) (b) (c)

Figure 10. Comparison between the q components of the reference voltage waveform (V *q) with actual
voltage waveform (Vq) in three-phase. (a) phase A with FAE technique and proposed technique, respectively;
(b) phase B with FAE technique and proposed technique, respectively; and (c) phase C with FAE technique
and proposed technique, respectively

Time (s) Time (s) Time (s)

(a) (b) (c)

Figure 11. Comparison between the d components of the reference voltage waveform (V *d) with actual
voltage waveform (Vd) in three-phase. (a) phase A with FAE technique and proposed technique, respectively;
(b) phase B with FAE technique and proposed technique, respectively; and (c) phase C with FAE technique
and proposed technique, respectively

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2372 – 2383
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2381

Time (s) Time (s) Time (s)

(a) (b) (c)

Figure 12. Comparison between the q components of the reference current waveform (I *q) with actual current
waveform (Iq) in three-phase. (a) phase A with FAE technique and proposed technique, respectively; (b)
phase B with FAE technique and proposed technique, respectively; and (c) phase C with FAE technique and
proposed technique, respectively

Table 2 shows the total harmonic distortion (THD) for all three load combinations and the Phase
Voltage Unbalance Rate (PVUR) that are tested in steady-state. They are all evaluated for the proposed
technique, the FAE, and the conventional OSG methods. Based on IEEE Std 141, PVUR is [30]:

𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑣𝑜𝑙𝑡𝑎𝑔𝑒 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑝ℎ𝑎𝑠𝑒 𝑣𝑜𝑙𝑡𝑎𝑔𝑒

%𝑃𝑉𝑈𝑅 = × 100 (18)
𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑝ℎ𝑎𝑠𝑒 𝑣𝑜𝑙𝑡𝑎𝑔𝑒

According to IEEE’s recommendation, the value of PVUR should be less than 2% for the system’s
distribution. As shown in Table 2, the THD and PVUR of the voltages are quite low when the proposed
control method is used, for example, PVUR in the unbalanced#1 scenario improved 98 and 99 percent
relative to the FAE and conventional methods, respectively. Also, the THD of the phase B voltage in the
same scenario improved 82% and 92% with respect to the FAE and conventional methods, in sequence.

Time (s) Time (s) Time (s)

(a) (b) (c)

Figure 13. Comparison between the d components of the reference current waveform (I *d) with actual current
waveform (Id) in three-phase. (a) phase A with FAE technique and proposed technique, respectively; (b)
phase B with FAE technique and proposed technique, respectively; and (c) phase C with FAE technique and
proposed technique, respectively

Enhanced controller for a four-leg inverter operating in a stand-alone … (Effat Ayoubi)

2382  ISSN: 2088-8694

Table 2. Simulation results

Conventional method FAE method Proposed method
THD(%) THD(%) THD(%)
Load type PVUR PVUR PVUR
Phase Phase Phase Phase Phase Phase Phase
(%) PhaseB (%) PhaseA (%)
A B C A C B C
Balanced 0.94 1.24 1.13 0.104 0.50 0.52 0.52 0.040 0.05 0.04 0.07 0.001
Unbalanced#1 1.32 2.41 2.52 0.652 0.51 1.83 1.83 0.100 0.03 1.59 1.59 0.003
Unbalanced#2 3.48 3.05 3.45 0.611 1.16 1.23 2.83 0.312 0.02 0.22 1.92 0.005

5. CONCLUSION
An improved per-phase control strategy for a grid-forming four-leg inverter using a fast orthogonal
signal generation (OSG) technique operating under unbalanced conditions is suggested in this paper. The use
of the new OSG technique allows the four-leg inverter to provide balanced output voltages under unbalance
load conditions without introducing any extra dynamics into the current control loop and distortions into the
generated orthogonal signal. The suggested controller has presented outstanding voltage control performance,
such as fast transient response and tiny steady-state error under unbalanced load conditions. In addition, its
operation does not depend on the system parameters. The superior performance of the suggested technique is
shown using simulation studies under various scenarios.

REFERENCES
[1] N. Khosravi, H. R. Abdolmohammadi, S. Bagheri, and M. R. Miveh, “Improvement the harmonic conditions of the
AC/DC microgrids with the presence of filter compensation modules,” Renewable and Sustainable Energy Reviews
vol. 143, 110898, 2021, doi: 10.1016/j.rser.2021.110898.
[2] R. S. Mongrain and R. Ayyanar, “Control of nonideal grid-forming inverter in islanded microgrid with hierarchical
control structure under unbalanced conditions,” International Journal of Electrical Power & Energy Systems, vol.
119, p. 105890, 2020, doi: 10.1016/j.ijepes.2020.105890.
[3] M. R. Miveh, M. F. Rahmat, A. A. Ghadimi, and M. W. Mustafa, “Control techniques for three-phase four-leg
voltage source inverters in autonomous microgrids: A review,” Renewable and Sustainable Energy Reviews, vol.
54, pp. 1592-1610, 2016, doi: 10.1016/j.rser.2015.10.079.
[4] D. I. Brandao, F. E. G. Mendes, R. V. Ferreira, S. M. Silva and I. A. Pires, “Active and Reactive Power Injection
Strategies for Three-Phase Four-Wire Inverters During Symmetrical/Asymmetrical Voltage Sags,” IEEE Trans.
Industry Applications, vol. 55, no. 3, pp. 2347-2355, May-June 2019, doi: 10.1109/TIA.2019.2893135.
[5] P. P. Vergara, J. C. López, M. J. Rider, H. R. Shaker, L. C. da Silva, and B. N. Jørgensen, A stochastic
programming model for the optimal operation of unbalanced three-phase islanded microgrids, International
Journal of Electrical Power & Energy Systems, vol. 115, p. 105446, 2020, doi: 10.1016/j.ijepes.2019.105446.
[6] M. R. Miveh, M. F. Rahmat, A. A. Ghadimi, and M. W. Mustafa, “Power quality improvement in autonomous
microgrids using multi-functional voltage source inverters: a comprehensive review,” Journal of Power
Electronics, vol. 15, no. 4, pp. 1054-1065, 2015, doi: 10.6113/JPE.2015.15.4.1054.
[7] D. Soto, C. Edrington, S. Balathandayuthapani, and S. Ryster, “Voltage balancing of islanded microgrids using a
time-domain technique,” Elec. power sys. Res., vol. 84, no. 1, pp. 214-223, 2012, doi: 10.1016/j.epsr.2011.11.022.
[8] M. Pichan and H. Rastegar, “Improved Integral Sliding Mode Control Scheme for Inverter-Based Four-Wire
Autonomous Power Supply Units,” IETE Journal of Research, pp. 1-12, 2019, doi:
10.1080/03772063.2019.1644972.
[9] A. Saim, A. Houari, M. A. Ahmed, A. Djerioui, M. Machmoum and J. M. Guerrero, “Adaptive Reference
Trajectory for Power Quality Enhancement in Three-Phase Four-Wire Standalone Power Supply Systems With
Nonlinear and Unbalanced Loads,” in IEEE Journal of Emerging and Selected Topics in Power Electronics, vol. 8,
no. 2, pp. 1593-1603, June 2020, doi: 10.1109/JESTPE.2020.2966923.
[10] R. Zhang, V. H. Prasad, D. Boroyevich and F. C. Lee, “Three-dimensional space vector modulation for four-leg
voltage-source converters,” in IEEE Transactions on Power Electronics, vol. 17, no. 3, pp. 314-326, May 2002,
doi: 10.1109/TPEL.2002.1004239.
[11] S. Pettersson, M. Salo and H. Tuusa, “Applying an LCL-filter to a four-wire active power filter,” 2006 37th IEEE
Power Electronics Specialists Conference, 2006, pp. 1-7, doi: 10.1109/pesc.2006.1711972.
[12] W. Cao, K. Liu, Y. Ji, Y. Wang, and J. Zhao, “Design of a four-branch lcl-type grid-connecting interface for a
three-phase, four-leg active power filter,” Energies, vol. 8, no. 3, pp. 1606-1627, 2015, doi: 10.3390/en8031606.
[13] F. Liu, X. Zha, Y. Zhou and S. Duan, “Design and research on parameter of LCL filter in three-phase grid-
connected inverter,” 2009 IEEE 6th International Power Electronics and Motion Control Conference, 2009, pp.
2174-2177, doi: 10.1109/IPEMC.2009.5157762.
[14] J.-H. Kim, and S.-K. Sul, “A carrier-based PWM method for three-phase four-leg voltage source converters,” IEEE
Trans. Power Elec., vol. 19, no. 1, pp. 66-75, Jan. 2004, doi: 10.1109/TPEL.2003.820559.
[15] J. Kim, S. Sul, and P. N. Enjeti, “A Carrier-Based PWM Method With Optimal Switching Sequence for a
Multilevel Four-Leg Voltage-Source Inverter,” in IEEE Transactions on Industry Applications, vol. 44, no. 4, pp.
1239-1248, July-aug. 2008, doi: 10.1109/TIA.2008.926201.

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2372 – 2383
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2383

[16] C. N. Ho, V. S. P. Cheung, and H. S. Chung, “Constant-Frequency Hysteresis Current Control of Grid-Connected
VSI Without Bandwidth Control,” in IEEE Transactions on Power Electronics, vol. 24, no. 11, pp. 2484-2495,
Nov. 2009, doi: 10.1109/TPEL.2009.2031804.
[17] M. Pichan, and H. Rastegar, “A New Hybrid Controller for Standalone Photovoltaic Power System with
Unbalanced Loads,” International Journal of Photoenergy, vol. 2020, 2020, doi: 10.1155/2020/5373914.
[18] S. Bayhan, M. Trabelsi, H. Abu-Rub and M. Malinowski, “Finite-Control-Set Model-Predictive Control for a
Quasi-Z-Source Four-Leg Inverter Under Unbalanced Load Condition,” in IEEE Transactions on Industrial
Electronics, vol. 64, no. 4, pp. 2560-2569, April 2017, doi: 10.1109/TIE.2016.2632062.
[19] A. Houari, A. Djerioui, A. Saim, M. Ait-Ahmed, and M. Machmoum, “Improved control strategy for power quality
enhancement in standalone systems based on four-leg voltage source inverters,” IET Power Electronics, vol. 11, no.
3, pp. 515-523, 2017, doi: 10.1049/iet-pel.2017.0124.
[20] W. Sinsukthavorn, E. Ortjohann, A. Mohd, N. Hamsic and D. Morton, “Control Strategy for Three-/Four-Wire-
Inverter-Based Distributed Generation,” in IEEE Transactions on Industrial Electronics, vol. 59, no. 10, pp. 3890-
3899, Oct. 2012, doi: 10.1109/TIE.2012.2188871.
[21] H. Nazifi, and A. Radan, “Current control assisted and non-ideal Proportional-Resonant voltage controller for four-
leg three-phase inverters with time-variant loads,” 4th Annual International Power Electronics, Drive Systems and
Technologies Conference, 2013, pp. 355-360, doi: 10.1109/PEDSTC.2013.6506732.
[22] C. Cai, L. Wang, and G. Yin, “A three-phase active power filter based on park transformation,” 2009 4th
International Conference on Computer Science & Education, 2009, pp. 1221-1224.
[23] Z.-Y. Zhao, M. Tomizuka, and S. Isaka, “Fuzzy gain scheduling of PID controllers,” in IEEE Transactions on
Systems, Man, and Cybernetics, vol. 23, no. 5, pp. 1392-1398, Sept.-Oct. 1993, doi: 10.1109/21.260670.
[24] Y. Suh, V. Tijeras and T. A. Lipo, “A control method in dq synchronous frame for PWM boost rectifier under
generalized unbalanced operating conditions,” 2002 IEEE 33rd Annual IEEE Power Electronics Specialists
Conference. Proceedings (Cat. No.02CH37289), 2002, pp. 1425-1430 vol.3, doi: 10.1109/PSEC.2002.1022376.
[25] R. Lliuyacc, J. M. Mauricio, A. Gomez-Exposito, M. Savaghebi, and J. M. Guerrero, “Grid-forming VSC control in
four-wire systems with unbalanced nonlinear loads,” Electric Power Systems Research, vol. 152, pp. 249-256,
2017, doi: 10.1016/j.epsr.2017.07.010.
[26] I. Vechiu, O. Curea, and H. Camblong, “Transient Operation of a Four-Leg Inverter for Autonomous Applications
With Unbalanced Load,” in IEEE Transactions on Power Electronics, vol. 25, no. 2, pp. 399-407, Feb. 2010, doi:
10.1109/TPEL.2009.2025275.
[27] M. Ebrahimi, S. A. Khajehoddin and M. Karimi-Ghartemani, “Fast and Robust Single-Phase $DQ$ Current
Controller for Smart Inverter Applications,” in IEEE Transactions on Power Electronics, vol. 31, no. 5, pp. 3968-
3976, May 2016, doi: 10.1109/TPEL.2015.2474696.
[28] N. A. Ninad and L. Lopes, “Per-phase vector control strategy for a four-leg voltage source inverter operating with
highly unbalanced loads in stand-alone hybrid systems,” International Journal of Electrical Power & Energy
Systems, vol. 55, pp. 449-459, 2014/02/01/ 2014, doi: 10.1016/j.ijepes.2013.09.019.
[29] P. Rodríguez, A. Luna, I. Candela, R. Mujal, R. Teodorescu and F. Blaabjerg, “Multiresonant Frequency-Locked
Loop for Grid Synchronization of Power Converters Under Distorted Grid Conditions,” in IEEE Transactions on
Industrial Electronics, vol. 58, no. 1, pp. 127-138, Jan. 2011, doi: 10.1109/TIE.2010.2042420.
[30] “IEEE Recommended Practice for Monitoring Electric Power Quality,” in IEEE Std 1159-2019 (Revision of IEEE
Std 1159-2009) , vol., no., pp.1-98, 13 Aug. 2019, doi: 10.1109/IEEESTD.2019.8796486.

Enhanced controller for a four-leg inverter operating in a stand-alone … (Effat Ayoubi)