IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO.

4, DECEMBER 2017 1631

Evolution of Topologies, Modeling,

Control Schemes, and Applications
of Modular Multilevel Converters
Apparao Dekka, Student Member, IEEE, Bin Wu, Fellow, IEEE, Ricardo Lizana Fuentes, Member, IEEE,
Marcelo Perez, Senior Member, IEEE, and Navid R. Zargari, Fellow, IEEE

Abstract— Modular multilevel converter (MMC) is one of the low (dv/dt), superior harmonic performance, uses a low
most promising topologies for medium to high-voltage high- cost and low-voltage semiconductor technology, generates an
power applications. The main features of MMC are modularity, output current with a smaller ripple, requires a smaller size
voltage and power scalability, fault tolerant and transformer-less
operation, and high-quality output waveforms. Over the past of input and output filter (if necessary), high-efficiency, and
few years, several research studies are conducted to address the generates low common-mode voltage (CMV) [6]–[8].
technical challenges associated with the operation and control of The integrated multilevel converters are commercially
the MMC. This paper presents the development of MMC circuit available with a voltage rating of 2.3–4.16 kV [9]. Moreover,
topologies and their mathematical models over the years. Also, these converters require significant modifications to increase
the evolution and technical challenges of the classical and model
predictive control methods are discussed. Finally, the MMC their operating voltage and output voltage levels, which is
applications and their future trends are presented. not cost-effective [10]–[14]. Further, these converters should
be shut down during the faults and device failures, which
Index Terms— Capacitor voltage ripple, circulating currents,
high-power converters, high-voltage direct current (HVdc) leads to a significant loss of production in the industrial
transmission, medium-voltage motor drive, model predictive process [15]–[18]. On the other hand, the multicell converters
control (MPC), modular multilevel converters (MMCs), multi- are constructed with a cascade connection of low-power
level converters, power quality, pulsewidth modulation (PWM), submodules (SMs). These converters are suitable for the
submodule (SM) capacitor voltage control. fault tolerant operation as well. The operating voltage of
I. I NTRODUCTION multicell converters can be increased by adding the SMs
in each phase [19]. The commercially available multicell
T HE high-power voltage source converters (VSCs) have
higher market penetration and more evident develop-
ments compared with the current source and the matrix
converters are shown in Fig. 1. Among them, the cascaded
H-bridge (CHB) and H-bridge neutral-point clamped (HNPC)
converters [1]–[5]. Over the past few years, several VSCs are converters require a phase-shifting transformer with multiple
developed and their classification is shown in Fig. 1. Among secondary windings to generate the isolated dc source for
them, very few topologies are commercially available in the each SM. The phase-shifting transformer increases the
form of customized and standard products. The high-power physical size and cost of the converter system. Hence, their
multilevel converters have significant advantages compared operating voltage is limited to 6–13.8 kV [20], [21].
with the two-level converters. Some of the advantages are Recently, the modular multilevel converter (MMC) has
been developed to address the problems associated with the
Manuscript received February 11, 2017; revised May 15, 2017 and CHB and HNPC topologies. The MMC also belongs to
July 23, 2017; accepted August 10, 2017. Date of publication August 21, the multicell converter family, and it was first introduced
2017; date of current version October 30, 2017. This work was supported in
part by the Fondecyt Iniciación 2016 under Project 11160227 and in part by by Prof. R. Marquardt in 2001 for high-voltage applications
the Advanced Center for Electrical and Electronics Engineering AC3E under such as the high-voltage direct current (HVdc) transmission
Grant CONICYT/FB0008. Recommended for publication by Associate Editor systems [22]–[25]. This topology is further extended to other
Jon A. Suul. (Corresponding author: Apparao Dekka.)
A. Dekka and B. Wu are with the Department of Electrical and Computer applications such as medium-voltage motor drives and power
Engineering, Ryerson University, Toronto, ON M5B 2K3, Canada (e-mail: quality improvement [26]–[28].
dapparao@ieee.org; bwu@ee.ryerson.ca). Each application of MMC requires a control scheme to
R. L. Fuentes is with the Department of Environment and Energy,
Universidad Católica de la Santísima Concepción, Concepción 4090541, Chile meet several objectives such as SM capacitor voltage control,
(e-mail: ricardolizana@ucsc.cl). output current control, minimization of circulating currents,
M. Perez is with the Department of Electronics, Universidad Técnica Fed- and the reduction of SM capacitor voltage ripple. Earlier,
erico Santa Maria, Valparaíso 110-V, Chile (e-mail: marcelo.perez@usm.cl).
N. R. Zargari is with the Medium Voltage R&D Department, the open-loop classical control strategy is presented in [29].
Rockwell Automation, Cambridge, ON N1R 5X1, Canada (e-mail: In this method, the imposed modulation indices are calculated
nrzargari@ra.rockwell.com). from the required input and output voltages. Later, the closed-
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org. loop classical control methods in the stationary-abc and the
Digital Object Identifier 10.1109/JESTPE.2017.2742938 synchronous-dq frames are developed [30]. The performance
2168-6777 © 2017 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
1632 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

Fig. 1. Classification of high-power VSCs.

of classical control method depends on the design and tuning

of proportional integral (PI)-controllers, the type of modulation
scheme, and switching frequency. Recently, the direct model
predictive control (D-MPC) approach is developed to over-
come the drawbacks of the classical control method. With the
D-MPC approach, it is easy to achieve the multiple control
objectives of MMC using a single cost function. However,
the computational complexity of D-MPC approach drastically
increases with the number of SMs in the MMC [31]. On the
other hand, an indirect MPC (I-MPC) with reduced com-
putational complexity is presented in [32]. The performance
of MPC method greatly depends on the mathematical model
of the system. The per-phase-based approximate models and
the three-phase-based exact equivalent models are available
in [33] and [34]. These models are used in the design of the
control schemes and the converter stability analysis [35], [36].
Over the years, a lot of research has been published focus-
ing on various aspects of an MMC including SM config-
urations, mathematical modeling, classical control methods,
modulation schemes, predictive control methods, and applica-
tions [37], [38]. This paper provides a comprehensive review Fig. 2. Configuration of a three-phase MMC.
of recent developments in the MMC and its future trends.

group of SMs and an inductor (L). The arm inductor helps

II. OVERVIEW OF M ODULAR M ULTILEVEL C ONVERTER to limit the inrush current caused by the instantaneous voltage
The generalized circuit configuration of a three-phase MMC difference between the arms. Also, it minimizes the magnitude
is shown in Fig. 2. The dc system of MMC is often referred to of ac circulating currents in an MMC.
as a dc-bus or dc-link, and it is connected across the positive The MMC can be designed with a different number of
and negative bars of the converter leg. The three-phase ac SMs, depending on the application, operating voltage, and
system is connected to the midpoint of each converter leg the rating of insulated-gate bipolar transistor (IGBT) devices.
(a, b, and c). Each leg of MMC is divided into two arms. For example, the medium-voltage motor drives with an
The arms connected to the positive bars are referred to as the operating voltage of 3.3–13.8 kV requires 5–20 SM/arm [39],
upper arms (u), while the arms connected to the negative bars the HVdc transmission systems require 200–400 SM/arm
are referred to as the lower arms (l). Each arm consists of a to reach an operating voltage of ± 320 kV (dc) [40]–[43],
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1633

Fig. 3. Circuit configuration of SMs. (a) HB. (b) FB. (c) CHB. (d) NPC. (e) ANPC. (f) NPP. (g) FC. (h) CD-SM.

and the static synchronous compensators (STATCOMs) high efficiency. The output voltage of HB-SM consists of posi-
employ 15–200 SM/arm to reach an operating voltage tive voltage levels (“0” and “vC1 ”) only and cannot support the
of 13.9–220 kV [44], [45]. bipolar operation and dc fault blocking [48]. An antiparallel
The MMC can be realized using different SMs to meet connection of thyristors across the ac output terminals of the
the application requirements such as dc-side fault current HB-SM was proposed in [49] to limit the dc-side fault current.
blocking, smaller capacitor voltage ripple and circulating cur- The full-bridge SM (FB-SM) is also known as an
rents, and high efficiency [46]. The most popular and widely H-bridge circuit and its configuration is shown in Fig. 3(b).
used SM configurations in MMC are shown in Fig. 3. The The FB-SM requires twice the number of semiconductor
half-bridge SM (HB-SM) is often referred to as a chopper- devices as compared with the HB-SM for the same voltage
cell and its circuit configuration is shown in Fig. 3(a). The rating. However, the control and design complexity are similar
HB-SM is commercially used in the MMC-HVdc transmission to that of the HB-SM. In the FB-SM, two devices carry the
systems [47]. Its simple construction results in a simple control current during the normal operation, resulting in a higher
and design. During the normal operation, only one device will device power loss and low efficiency. The FB-SM generates
be in “ON” state. Hence, the HB-SM has low power loss and three voltage levels of “0,” “vC1 ,” and “−vC1 .” The negative
1634 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

voltage level is used to eliminate the dc-side fault current from The three-level flying capacitor SM (3L-FC-SM) has a
the system [50], [51]. With the identical power rating and the nested cell structure, and it requires four semiconductor
number of SMs, the FB-SM-based MMC requires only 50% devices as shown in Fig. 3(g). The FC-SM can be operated
of total dc-bus voltage (Vd ) to generate the same amount of ac as an HB-SM by applying identical gating signals to the
output voltage in comparison with the HB-SM-based MMC. devices S1 and S2 . This feature allows the researchers to
Hence, each FB-SM capacitor can be designed with a voltage investigate the performance of HB-SM, FC-SM, and hybrid
rating equal to 75% of HB-SM rated voltage [52], [53]. Also, MMC (combination of HB and FC-SM) systems using FC-SM
the FB-SM-based MMC has a smaller capacitor voltage ripple itself [62]. The FC-SM-based MMC has a smaller capacitor
compared with HB-SM-based MMC. With the FB-SM, it is voltage ripple and circulating currents compared with the
also possible to create hybrid MMC configurations such as HB-SM-based MMC [63]. Hence, the FC-SM has a better
alternative arm converter and star-channel MMC [54], [55]. efficiency compared with the HB-SM. The FC-SM generates
The performance and efficiency of the SMs can be improved positive voltage levels only and cannot block the dc-side fault
by replacing the standard SMs with the multilevel SMs. These current.
SMs reduce the converter footprint size but increase the design The clamp-double SM (CD-SM) is a three-level hybrid
complexity. With the multilevel SMs, the available redundancy SM [64]. The CD-SM is realized by connecting two HB-SMs
switching states are quite high in an MMC. These redundancy in series along with two additional diodes and a single IGBT
switching states significantly improve the controllability of device as shown in Fig. 3(h). During the normal operation,
the SM capacitor voltage. Over the past few years, several the device S B is continuously in “ON” state and resulting in a
multilevel SMs are developed and studied for an MMC. cascade connection of two HB-SMs. When a dc fault occurs,
Some of the SM configurations are shown in Fig. 3. The all the devices in CD-SM are turned “OFF,” and it generates
CHB-SM (CH-SM) can be realized by connecting two positive or negative voltage levels at the output, depending on
HB-SMs in series as shown in Fig. 3(c). The resultant circuit the current direction. Hence, the CD-SM can be used in the
generates three voltage levels at the SM ac output terminals. HVdc applications to block the dc-side fault current [65], [66].
The main features are easy of control and less design com- However, the power losses, efficiency, and design complexity
plexity. Further, the CH-SM has a low device power loss and are significantly high due to the additional devices. Several
high efficiency similar to the HB-SM. Currently, the CH-SM hybrid SMs are developed to handle the dc-side faults, to
is commercially used in the MMC-based motor drive improve the efficiency with less device count, and to achieve
systems [56]. the better capacitor voltage controllability [67]–[74]. These
The configuration of three-level NPC SM (3L-NPC-SM) SMs are still in the research and development stage, yet to
is shown in Fig. 3(d). The 3L-NPC-SM is constructed with implement in the commercial products.
four semiconductor devices, two clamping diodes, and two
capacitors. The loss distribution between the devices and III. M ATHEMATICAL M ODELING OF M ODULAR
the neutral-point voltage balance are the major issues in the M ULTILEVEL C ONVERTER
NPC-SM [57]. Due to the neutral-point balancing issue, the In an MMC, the interaction between the arm and line quanti-
operating region of NPC-SM is limited at the higher mod- ties (variables) generates low- and high-frequency components
ulation indices [58]. The NPC-SM has higher device power on the ac and dc side of the SM [75]. Due to this interaction,
losses and low efficiency compared with the HB-SM. From it is very difficult to understand the operation and behavior
the control and design perspective, the NPC-SM is not an of the MMC during the steady-state and dynamic conditions.
attractive solution for an MMC. The simulation studies are widely employed to analyze the
The three-level active NPC SM (3L-ANPC-SM) can be behavior of an MMC. However, the simulation process con-
realized by replacing the clamping diodes in the NPC-SM sumes a lot of time, due to the larger number of SMs per arm.
with the active devices. The configuration of 3L-ANPC-SM is Also, a control scheme is required to handle multiple control
shown in Fig. 3(e). The active devices provide the additional objectives of an MMC. The designing of multiobjective control
redundancy switching states to control the neutral point scheme is a difficult and challenging task. To address the above
current and to ensure the equal loss distribution between the problems, several equivalent mathematical models have been
devices [59], [60]. The ANPC-SM generates positive voltage proposed in the literature.
levels only. Hence, it cannot block the dc-side fault current.
The power losses and efficiency of ANPC-SM is quite similar
to the NPC-SM. Another variation of a three-level SM is a A. Equivalent Circuit Models
neutral-point piloted SM (3L-NPP-SM) and its configuration The simplified equivalent model of a three-phase MMC
is shown in Fig. 3(f). In NPP configuration, the neutral point is shown in Fig. 4(a), in which the SM terminal volt-
is connected to the output terminal through an antiseries con- age in each arm is modeled as a single equivalent voltage
nection of IGBT devices. Also, the devices in each leg should source [76]–[80]. This model is very simple and easy to
be designed with two times the voltage blocking capability analyze and the simulation process becomes faster. With this
of the NPC and ANPC SMs [61]. The NPP-SM generates model, it is also possible to obtain the decoupled current
three positive voltage levels only. This SM is not suitable models. These models are used to control various current
for the bipolar operation and cannot block the dc-side fault components flowing through the arm [81]. The simplified
current. model is valid under the following assumptions: 1) the SM
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1635

Fig. 4. Equivalent models of a three-phase MMC. (a) Model with ideal arm voltage sources. (b) Model with nonideal arm voltage sources. (c) AVM.

capacitance is a large enough to model as a constant dc voltage voltage source, whereas the dc side of the SMs is modeled as
source and 2) each arm of MMC should be equipped with a controlled current source [84]. The controlled current source
a larger number of SMs, so that the harmonic components represents the charging and discharging of the SM capacitors.
corresponding to the voltage source can be neglected. The aim of AVM approach is to replicate the average response
Alternatively, each arm of MMC is modeled as a nonlinear of the switching devices, the converter, and the control scheme
capacitor with a time-variant sinusoidal capacitance, while the using simple mathematical equations. The AVM equations are
capacitor voltage is considered as a controlled voltage source very accurate and emulate the fundamental system behavior
as shown in Fig. 4(b). To deal with the capacitor voltage on the ac side of the SM, but it requires a longer simu-
dynamics, the total energy stored in each arm is used to control lation time and is difficult to analyze due to the coupled
the arm capacitor voltage. The summation of upper and lower dynamics. These equations are derived in the abc-frame and
arm energy is regulated corresponding to the load active power are suitable for the time-domain simulations only [85], [86].
demand, and their difference is controlled to maintain the Further, the mathematical models in dq-frame are presented in
voltage balance between the upper and lower arms [82]. [87] and [88]. Using the dq-frame, the arm and line quantities
To simplify the analysis, the average value models (AVMs) can be decoupled, which simplifies the controller design.
are presented in [83]. The AVM of MMC is shown in Fig. 4(c). A state-space approach or matrix representation was pre-
In this model, the ac side of the SMs is modeled as a controlled sented in [89] to simplify the analysis and simulation models.
1636 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

Also, the phasor approach is employed in the mathematical To analyze the dynamic behavior of MMC, the dc cur-
modeling of an MMC [90]. These models are valid in steady- rent (i d ), ac circulating current (i x z ), and ac system cur-
state and use a graphical representation of the converter cur- rent (i ox ) components are considered in the arm current (i yx ).
rents and voltages. The mathematical models are also available From the equivalent model shown in Fig. 4(a), the upper and
in the frequency domain [91], [92]. In this domain, the design lower arm current of phase-a is given by
of the controller is easy and straightforward. In overall, the
simplified models reduce the simulation running time and i ua = 0.5 i oa + i d + i az
provide a straightforward analysis, but the effect of switch- i la = 0.5 i oa − i d − i az . (1)
ing devices and modulation schemes cannot be investigated.
On the other hand, the complex models provide accurate The phase-a ac system current is formulated as
results and a better understanding of the converter operation,
and its behavior. Depending on the application requirement, i oa = 2 i a + 2 i cm . (2)
one of the equivalent models is adopted to analyze the system Similarly, the current flowing through the upper and lower
behavior. dc-bus of MMC is given by

B. Mathematical Modeling i u = i ua + i ub + i uc = 3 i cm + 3 i d
i l = i la + i lb + i lc = 3 i cm − 3 i d . (3)
In this section, a detailed mathematical model of arm
currents in the MMC is presented. The mathematical model From Fig. 4(a), the phase-a and the upper and lower arm
has four independent current components associated with the voltages are given by
input (dc current), output (ac output current), ac circulating, ⎫
and common-mode currents. Even though the common-mode d iu d i ua ⎪
Vu = L d + rd i u + v ua + L ⎪
⎪
current model is not used in the HVdc applications, it could dt dt ⎪ ⎪
⎪
⎪
+ ro i oa + v an ⎪
d i oa
be useful for other applications (motor drives). In overall, the + r i ua + L o ⎬
proposed models are simple, and it can be used to analyze dt (4)
d il d i la ⎪
the converter dynamic behavior. Further, the continuous-time −Vl = L d + rd i l − vla + L ⎪
⎪
dt dt ⎪ ⎪
⎪
and discrete-time models are used to design and implement ⎪
+ ro i oa + v an . ⎪
d i oa
+ r i la + L o ⎭
(simulation and experimental studies) the classical and dt
MPC methods.
The MMC consists of N-SMs in series with an arm inductor Subtraction of lower arm voltage in (4) from the upper arm
in each arm. The N-SMs are modeled as a controlled voltage voltage results in
source (v yx ) as shown in Fig. 4(a). The dc system is repre- d (i u − i l )
sented with a split dc voltage source [Vu = Vl = (Vd /2)], Vd = L d + rd (i u − i l ) + v ua
dt
and it is connected to the converter legs through a dc line. d (i ua − i la )
The ac system of MMC is connected to the ac voltage source + vla + L + r (i ua − i la ). (5)
dt
through an ac line. The list of variables and symbols used in
the mathematical model is defined as follows: Similarly, the summation of phase-a and the upper
and lower arm voltages in (4) outcomes the following
x ∈ {a, b, c} phase;
equation:
y ∈ {u, l} arm;
Vd total dc system voltage; d (i u + i l ) d (i ua +i la )
Vu , Vl upper and lower dc-bus voltages; 0 = Ld + rd (i u + i l ) + v ua − vla + L
dt dt
i u , il upper and lower dc-bus currents; d i oa
v ua , v ub , v uc three-phase upper arm voltages; + r (i ua + i la ) + 2L o + 2ro i oa + 2v an . (6)
dt
i ua , i ub , i uc three-phase upper arm currents;
1) Modeling of DC Current Component: The magnitude
vla , vlb , vlc three-phase lower arm voltages;
of dc current component (i d ) depends on the active power
i la , i lb , i lc three-phase lower arm currents;
demanded by load along with the voltage difference between
v an , v bn , v cn three-phase ac system voltages;
the dc-bus voltage and the SM capacitor voltage in each
i oa , i ob , i oc three-phase ac system currents;
converter leg. From (3), the dc current component is given
i az , i bz , i cz three-phase ac circulating currents;
by
ia , ib, ic three-phase ac output currents;
i cm common-mode current; i u − il
id dc current component; id = . (7)
6
L d , rd , inductance and resistance of a dc line;
L, r , inductance and resistance of an arm inductor; Equations (1) and (7) are substituted in (5), which results in
L o , ro , inductance and resistance of an ac line; d id
() p predictive variable; Vd = 6L d + 6rd i d + v ua + vla
dt
()m measured variable, d id d i az
Ts sampling time. + 2L + 2r i d + 2L + 2r i az . (8)
dt dt
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1637

Similarly, the phase-b and -c voltage equations are obtained With the forward Euler approximation, the discrete-time
as follows: model of three-phase ac circulating currents is formulated as
⎫ ⎡
Vd = 6L d
d id
+ 6rd i d + v ub + vlb ⎪
⎪  v ux p p
⎪
⎪ (k) + vlx (k)
dt ⎪
⎪ ixz (k + 1) = z ⎣
⎪
+ 2r i bz ⎪
d id d i bz 6
+2L + 2r i d + 2L ⎬ x=a,b,c
dt dt (9) p p

Vd = 6L d
d id
+ 6rd i d + v uc + vlc ⎪
⎪ vux (k) + vlx (k)
⎪
⎪ − + z ixz
m
(k), (17)
dt ⎪
⎪ 2
⎪
+ 2r i cz . ⎪
d id d i cz
+2L + 2r i d + 2L ⎭
dt dt
where
The dynamic model of dc current component is obtained by
adding (8) and (9) as Ts r Ts
z = , z = 1 − . (18)
⎡ ⎤ L L
d id 1 V  v ux + vlx 3) Modeling of Common-Mode Current: With the assump-
= ⎣ d
− − (3rd + r ) i d ⎦.
dt 3L d + L 2 6 tion that the ac and dc system grounds are connected, then the
x=a,b,c
magnitude of common-mode current (i cm ) is derived from (3)
(10)
as
The continuous-time model of dc current component given i u + il
in (10) is transformed into the discrete-time domain using i cm = . (19)
6
forward Euler approximation and results in
⎡ ⎤ Substitution of (2) and (19) into (6) results in the following
V m (k)  v
p
(k) + v
p
(k) equation:
i d (k + 1) = d ⎣ d ⎦
p ux lx
−
2 6 d i cm d i cm
x=a,b,c 0 = 6L d + 6rd i cm + v ua − vla + 2L
+ d i dm (k) (11) dt dt
d ia d i cm
+ 2r i cm + 2L + 2r i a + 4L o
where superscript “ p” denotes the predicted variable, dt dt
d ia
“m” denotes the measured variable, and Ts represents the + 4ro i cm + 4L o + 4ro i a + 2v an . (20)
sampling time. The discrete-time parameters are given by dt

Ts (3rd + r ) Ts Similarly, the phase-b and -c voltage equations are given by

d = , d = 1 − . (12) ⎫
3L d + L 3L d + L d i cm d i cm ⎪
0 = 6L d + 6rd i cm + v ub − vlb + 2L ⎪
⎪
dt dt ⎪ ⎪
2) Modeling of AC Circulating Current: The ac circulating ⎪
⎪
+2r i cm + 2L
d ib
+ 2r i b + 4L o
d i cm ⎪
⎪
current flows between the converter legs only. The magnitude ⎪
⎪
dt dt ⎪
⎪
of phase-a circulating current is given by d ib ⎪
⎪
+4ro i cm + 4L o + 4ro i b + 2v bn ⎬
dt
i ua − i la d i cm d i cm ⎪ (21)
i az = − id . (13) 0 = 6L d + 6rd i cm + v uc − vlc + 2L ⎪
⎪
2 dt dt ⎪ ⎪
⎪
d ic d i cm ⎪
⎪
+2r i cm + 2L + 2r i c + 4L o ⎪
⎪
The dynamic model of phase-a, ac circulating current, dt dt ⎪
⎪
⎪
⎪
is obtained by solving (8), (10), and (13) as
+4ro i cm + 4L o
d ic
+ 4ro i c + 2v cn . ⎪
⎭
⎡ ⎤ dt
d i az 1  v ux + vlx v ua + vla
= ⎣ − − r i az ⎦. (14) By adding (20) and (21), the dynamic model of common-
dt L 6 2 mode current is given by
x=a,b,c
⎡
From (14), the generalized representation of three-phase ac
circulating currents is given by ⎢
⎢  v −v
⎡ ⎤
d i cm 1 ⎢ lx ux
= ⎢
 dt 3L d + L + 2L o ⎢ 6
d ixz 1 ⎣ v ux + vlx vux + vlx ⎣ x=a,b,c
= − − r ixz ⎦ (15)  
dt L 6 2 Zero Sequence Voltage
x=a,b,c ⎤
where ⎥
⎥
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎥
i az v ua vla − (3rd + r + 2ro ) i cm ⎥. (22)
⎥
ixz = ⎣i bz ⎦ , vux = ⎣v ub ⎦ , vlx = ⎣vlb ⎦ . (16) ⎦
i cz v uc vlc
1638 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

The discrete-time model of common-mode current is

derived from (22) as
⎡ ⎤
⎢ ⎥
⎢  v p (k) − v p (k) ⎥
⎢ lx ux ⎥
p
i cm (k + 1) = cm ⎢ ⎥ + cm i cm
m
(k),
⎢ 6 ⎥
⎣x=a,b,c ⎦
 
Zero Sequence Voltage
(23) Fig. 5. Block diagram of the classical control system.

where
Ts (3rd + r + 2ro ) Ts as follows:
cm = , cm = 1 − . ⎡
3L d + L + 2L o 3L d + L + 2L o
(24) ⎢ p
⎢ v (k) − vp (k)
p ⎢ ux
ix (k + 1) = o ⎢ lx
4) Modeling of AC Output Current: The ac output current is ⎢ 2
equally distributed among the upper and lower arms. From (1), ⎣
the magnitude of phase-a, ac output current, is given by
⎤
i ua + i la ⎥
ia = − i cm . (25) ⎥
2  v p (k) − v ux
p
(k) p ⎥
lx
− − vxn ⎥ + o ixm (k)
The dynamic model of phase-a is obtained by solving (20), 6 ⎥
x=a,b,c ⎦
(22), and (25) as  
⎡ Zero Sequence Voltage
(29)
⎢
⎢v − v  vlx − v ux
d ia 1 ⎢ la ua where
= ⎢ −
dt L + 2L o ⎢ 2 6 Ts (r + 2ro ) Ts
⎣ x=a,b,c o = , o = 1 − . (30)
  L + 2L o L + 2L o
Zero Sequence Voltage
⎤ The above mathematical model is developed for the MMC
⎥ with dc ground, but these equations are also applicable to the
⎥ MMC without dc ground. In an MMC without dc ground,
⎥
−v an − (r + 2ro ) i a ⎥. (26) there is no path for the common-mode current. Hence, the
⎥
⎦ dynamic model corresponding to the common-mode current
can be neglected. There is no change in the ac output current,
ac circulating current, and dc current component models.
From (26), the generalized dynamic model of three-phase Equation (27) shows that the ac output current model consists
ac output current is given by of a zero sequence voltage that is used to minimize the ripple
⎡ in the SM capacitor voltage and dc-bus current in an MMC
⎢ during the balanced and unbalanced operating conditions.
⎢v − v  vlx − v ux
d ix 1 ⎢ lx ux
= ⎢ −
dt L + 2L o ⎢ 2 6 IV. C LASSICAL C ONTROL M ETHODS FOR MMC
⎣ x=a,b,c
  The digital control schemes enable a safe, reliable, and
Zero Sequence Voltage
⎤ efficient operation of an MMC. These control schemes provide
a superior dynamic and steady state performance. The control
⎥
⎥ of MMC is quite challenging, and it involves multiple control
⎥
− vxn − (r + 2ro ) ix ⎥ (27) objectives, as shown in Fig. 5. The MMC control objectives
⎥
⎦ are divided into two categories of primary and secondary. The
control of output current and SM capacitor voltages are the
primary objectives, associated with the operation of MMC.
where On the other hand, the minimization of circulating current
⎡ ⎤ ⎡ ⎤ and SM capacitor voltage ripple belongs to the secondary
ia v an
ix = ⎣i b ⎦ , vxn = ⎣v bn ⎦. (28) objectives. These objectives can be achieved using either
ic v cn classical control methods or MPC methods. In this section,
the overview of classical control methods and their limitations
The continuous-time model given in (27) is transformed into are presented. In classical control methods, each objective is
the discrete-time domain using forward Euler approximation achieved using an independent PI regulator. Also, a simple
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1639

interleaving angle, and the results show that the carrier

interleaving improves the total harmonic distortion (THD)
of the output voltage at the expense of higher ripple and
root mean square (RMS) value of the arm current [97].
In PSC-PWM scheme, the effective switching frequency of
the converter is equal to “N × f c ” (without interleaving) and
“2N × f c ” (with interleaving), where N represents the number
of SMs per arm and f c represents the carrier frequency.
These modulation schemes are used in the MMC-fed motor
drive applications to achieve the smaller torque ripple.
The LSC-PWM is further categorized into phase-disposition
(PD-PWM), phase-opposition disposition (POD-PWM), and
Fig. 6. PWM schemes for an MMC.
alternate POD (APOD-PWM) schemes. Among them,
the PD-PWM generates an output voltage with a lowest
pulsewidth modulator is required to generate the gating signals
harmonic distortion compared to the POD-PWM and
for the SMs.
APOD-PWM schemes [98]. The control of MMC with
PD-PWM is presented in [99], and its performance is
A. Pulsewidth Modulation Schemes compared with the PSC-PWM scheme [100]. The PD-PWM
The pulsewidth modulation (PWM) is widely used to con- generates an output line-to-line voltage with a lowest harmonic
trol the ac output voltage of a power converter. The targeted distortion, and it maintains the equal switching frequency
(reference) ac output voltage is generated by controlling the between the switching devices [101]. Also, the optimal
duty cycle of switching devices. The PWM schemes are PD-PWM scheme with interleaving angle improves the output
designed to reduce the output voltage harmonic distortion and voltage harmonic spectrum, and it naturally balances the SM
to increase the magnitude of the output voltage at a given capacitor voltages [102], [103]. But, the PD-PWM scheme
switching frequency. Also, other control objectives such as produces a larger ripple in the SM capacitor voltage, which
capacitor voltage balance, CMV reduction, minimization of increases the magnitude of circulating currents compared
device switching frequency, reduction of power losses, and with the PSC-PWM scheme [104].
the minimization of output current ripple can be achieved with In overall, the carrier modulation schemes are simple and
the modulation scheme [93]. Based on the device switching easy to implement in the digital control platforms. But, the
frequency and the application, the PWM schemes of an MMC change in the number of SMs affects the carrier arrangement
are categorized into high switching frequency, low switching in the modulation scheme. Hence, the carrier PWM modulator
frequency, and fundamental switching frequency modulation should be redesigned by the converter configuration, which
schemes as shown in Fig. 6. further affects the flexibility in extending the existing MMC
The carrier modulation scheme is often referred to as a system to the higher voltage levels [105]. Also, the carrier
sine-triangular modulation, which is well established for the modulation schemes have higher switching losses compared
two-level power converters. In carrier modulation, the mod- to the low and fundamental switching frequency modulation
ulation signals are compared with a single triangular carrier schemes.
signal to generate the gating signals for the switching devices The sampled average modulation (SAM) and space vector
in a two-level converter. The carrier modulation schemes modulation (SVM) schemes are placed in the low switching
are extended to the multilevel converters, where multiple frequency modulation (100 Hz < f sw < 2000 Hz) category.
triangular carrier signals are compared with phase modulation The sampled average approach directly controls the phase
signals to generate the gating signals. This approach is voltages of an MMC, and it generates the line-to-line voltage
referred to as a multicarrier modulation scheme [94]. implicitly [106]. Alternatively, the SVM directly controls the
Depending on the carrier arrangement, the multicarrier line-to-line voltages of an MMC, and it generates the phase
modulation schemes are categorized into phase-shifted voltages implicitly [62], [107]. The three-phase equivalent of
carrier (PSC-PWM) and level-shifted carrier (LSC-PWM). SAM is quite similar to the SVM, except that the distribution
These modulation schemes are placed in the high switching of zero vectors at the starting and end of the switching
frequency ( f sw > 2000 Hz, where f sw represents the device sequence is not equal. Hence, the sampled average approach
switching frequency) modulation category. The natural generates an output voltage with a highest harmonic distortion
balancing of SM capacitors voltage in an MMC can be compared with the SVM approach [108]. The SAM approach
achieved with a higher carrier-frequency-based PSC-PWM is easy to extend for an MMC with any number of SMs without
scheme [95]. The high switching frequency operation causes any modifications in the implementation algorithm. On the
higher switching losses, which is not desirable for the other hand, the SVM approach is difficult to implement for an
high-power applications. Therefore, an additional voltage MMC with a large number of SMs (N > 20).
balancing algorithm is used with a lower carrier-frequency The fundamental switching frequency modulation schemes
(100–500 Hz)-based PSC-PWM scheme, provided that each are very popular for the high-power converters such as
arm is equipped with a large number of SMs [96]. The CHB and MMC [109]. The selective harmonic elimina-
PSC-PWM is further investigated with and without tion (SHE) and staircase modulation schemes belong to this
1640 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

category [110], [111]. In staircase or nearest level modula-

tion (NLM), the power converters are switched at a fundamen-
tal frequency to reduce the switching losses [112]. However,
the harmonic distortion of the output voltage and current
waveforms is significantly high. The harmonic distortion can
be reduced by increasing the number of SMs in an MMC
(i.e., by increasing the number of output voltage levels).
The NLM scheme is widely used in the MMC-based HVdc
applications, where the MMC is built with 200–400 SMs
Fig. 7. Leg voltage control.
in each arm and each SM is modulated at the fundamental
frequency of 50 or 60 Hz. These systems generate an output
voltage with very low THD of <1%. The output waveform
quality can be further improved by designing the modula-
tion schemes with an interleaving angle, which generates an
output voltage with a higher number of voltage levels in
comparison to without interleaving angle. In [113], a modified
NLM method is presented, which generates an output voltage
levels equivalent to the carrier PWM with the interleaving
angle. Further, the fundamental frequency modulation schemes
require a higher sampling frequency to generate an output
voltage corresponding to its reference [114]. Alternatively, the
increment in the output voltage levels can be achieved by Fig. 8. Block diagram of distributed control approach.
increasing the number of SMs in each arm, which is the case
for an MMC-based HVdc system [115]. well established and implemented in the stationary-abc, the
The SHE scheme is also operated at a fundamental fre- stationary-αβ, and the synchronous-dq reference frames [85].
quency [116]. The implementation of SHE scheme involves
the calculation switching angles. The number of switching
angles drastically increases with the number of voltage levels. C. Submodule Capacitor Voltage Control Methods
Also, it is difficult to obtain the solution for such a large The operating voltage of MMC can be increased by adding
number of switching angles [117]. Therefore, the SHE scheme the SMs in each arm. These SMs have floating capacitors, and
is not an ideal solution for an MMC. Other hand, the staircase their voltage should be controlled at its nominal value. The
modulation approach is easy to implement, and it does not SM capacitor voltage control is a fundamental requirement to
involve any complicated calculations like SHE scheme. achieve a high-performance current control, to improve output
In [118], several fundamental switching frequency modula- power quality, and to provide stability and controllability to
tion schemes are presented for an MMC. In these schemes, the the MMC. The capacitor voltage control is divided into three
optimized pulse-pattern or switching angles are calculated in parts: leg voltage control, SM capacitor voltage balance, and
real-time and are applied to the SMs based on their capacitor arm capacitor voltage control [123]. The leg voltage control
voltage and the direction of the arm current. The fundamental regulates the average capacitors voltage of each leg to a
switching frequency modulation schemes are suitable for an value of Vd . This control has an outer voltage and an inner
MMC with a larger number of SMs in each arm [118]. There current control loop as shown in Fig. 7. The outer voltage
are no other modulation schemes in the literature, which can loop regulates the average capacitors voltage of each leg at
be operated at the low switching frequency and suitable for a constant value of vCt∗ , which is equal to Vd . The voltage
an MMC with a smaller number of SMs. The development of control loop minimizes the error between the reference and
low switching frequency modulation schemes for an MMC is actual average voltages, and it gives the dc-bus current com-
another interesting research topic. mand (i d∗ ). At steady-state, the magnitude of the dc-bus current
is proportional to the active power demanded by the load plus
power losses. The dc-bus current command is controlled using
B. Output Current Control Methods an inner current loop. The current loop minimizes the error
The MMC is used commercially in various applications between the reference and actual dc-bus current, and it gives
such as HVdc transmission systems, medium-voltage motor a compensating signal of v dct ∗ [76]. This compensating signal

drives, and STATCOMs [119]. Each application uses a differ- is added to the main modulation signals.
ent technique to control their output currents. For example, The voltage balance between SMs in the arm is achieved
the voltage oriented control (VOC) is used in the HVdc by the capacitor voltage balance method. The capacitor volt-
applications [119], while the field oriented control (FOC) age balancing can be performed at either control stage or
and direct torque control schemes are applied to motor drive modulation stage. Based on the capacitor voltage balancing,
systems [120]. In the case of STATCOM, the output cur- the capacitor voltage control methods are categorized into
rents are controlled to meet the reactive power or power distributed and centralized control approaches. The block
factor requirement [121], [122]. These control techniques are diagram of distributed control approach is shown in Fig. 8.
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1641

Fig. 9. Voltage balance control.

Fig. 11. Voltage balance strategy. (a) Flowchart of balancing strategy.

(b) Comparison logic.

to their nominal voltage. The normalized capacitor voltage is

Fig. 10. Block diagram of centralized control approach.
compared with other capacitors voltage as shown in Fig. 11(b).
The output of each comparison is added together, which results
In this approach, each SM capacitor voltage is closely main- in a virtual index number. These virtual index numbers are
tained at the reference value using a closed loop control. The rearranged in either ascending or descending order based
closed loop control is often referred to as a voltage balancing on the direction of the arm current. When the arm current
control as shown in Fig. 9. The balance control generates a is positive, the SMs with the lowest voltage is chosen to
compensation signal that is added to the modulation signal of be inserted to let them charge and avoid any overcharging
each SM. The distributed control approaches can be imple- of those capacitors with the highest voltages and vice-versa
mented with the PSC-PWM schemes only [124]–[126]. The [110], [111], [127], [128]. The resultant index numbers are
distributed control approach has better controllability, provided compared with the reference index number, which gives an
that the SM switching frequency should be large enough. “ON” and “OFF” state of the SM. These states are applied to
The block diagram of centralized control approach is shown the SMs in each PWM period [58], [133].
in Fig. 10. In this approach, the voltage balance between The direction of the arm current is widely used in the
the SMs in the arm is achieved using a voltage balance implementation of voltage balance methods. The three-phase
strategy. The balancing strategy is implemented after the MMC has total six arms and requires six current sensors to
modulation stage using logical functions. This method selects measure the arm current. The system cost can be reduced by
a certain number of SMs out of N-SMs in an arm to generate eliminating or reducing the number of arm current sensors.
the required voltage level. Therefore, the balance strategy is In [134], the charging and discharging of SM capacitors is ana-
often referred to as a SM selection method. The centralized lyzed based on the pulsewidth duration of gating signals. The
control approach can be implemented with the PSC-PWM, optimum gating signal is applied to the SM, which requires
LSC-PWM, NLM, SHE, SAM, and SVM schemes [62], an extreme charge transfer. This approach eliminates the arm
[108], [110], [111], [127]–[129]. There are several centralized current sensors and improves the reliability of the system
capacitor voltage balancing methods that are presented in the by reducing the number of components that can potentially
literature. fail [134]. These approaches can be implemented with the
The simplest way to balance the SM capacitor voltage is fundamental switching frequency modulation schemes for an
by the periodical rotation of gating signal pattern based on MMC with a smaller number of SMs per arm only. The design
the phase voltage switching state redundancy [130]. However, of generalized voltage balance algorithm without measuring
this method is difficult to apply for MMC with a large number arm currents and suitable to implement with any modulation
of SMs per arm. Other hand, the capacitor voltage balance scheme should be addressed.
can be achieved by adjusting the duration of PWM gating It is also possible to achieve a voltage balance among SMs
pulses in the real-time [131], [132]. These approaches can be by controlling the SMs energy rather than their voltages.
implemented with the PSC-PWM scheme only. An open-loop energy estimation approach was proposed
Alternatively, the sorting and relative comparison logic- in [29], which is simple and suitable to shape the low-
based voltage balancing approaches are developed to control frequency capacitor voltage ripple [135]. Also, the energy
the MMC. The flowchart of relative comparison logic-based balance methods effectively control the SM capacitor voltage
voltage balancing approach is shown in Fig. 11(a). In this at their nominal value even under the SM fails [136].
approach, the SM capacitor voltage is normalized with respect The voltage and energy-based balancing approach show
1642 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

Fig. 13. Circulating current control in stationary-abc frame.

signal. The dc signals are easily controlled using a simple

PI regulator [145], [146].
Fig. 12. Circulating current control in the synchronous-dq frame.
The resonant regulators are employed to control the
circulating current in the stationary-abc frame as shown in
the identical performance in the case of HB-based
Fig. 13. In this approach, the resonant regulators are designed
MMC [137], [138]. However, there is no definite answer in
to eliminate the specific dominant harmonic frequency
the case of multilevel submodule-based MMC. The design and
component such as second- and fourth-order harmonic
study of energy balance approach for multilevel submodule-
components from the circulating currents [147]–[149].
based MMC are one of the challenging and interesting
In [150], a parallel combination of repetitive and PI regulator
research topics. Finally, the voltage or energy balance
is used to eliminate multiple harmonics in the circulating
among the arms can be achieved by modifying the current
currents. The parallel combination imposes a limitation on
component in different arms. Particularly, the circulating
the PI regulator performance, and it complicates the design
current component is used to achieve this objective. The
of the repetitive controller. To overcome these problems, a
reference value of circulating current controller is generated
cascade structure of repetitive controller and a PI regulator
using a differential voltage controller (arm capacitor voltage
is presented in [151]. The repetitive controller improves
control) [139]. The circulating current control and voltage
the bandwidth and dynamic performance of the circulating
balance control are sufficient to control the total leg voltage,
current control and the THD of the output voltage.
voltage balancing between the arms, and voltage balancing
During unbalanced operating conditions, the circulating
between the SMs. Therefore, the leg voltage control is
current consists of positive and zero sequence components
eliminated in certain studies under assumption that the
besides the negative sequence components at twice the
dc and ac side powers of MMC are naturally balanced.
line frequency [152]. The positive and negative sequence
components flowthrough the converter legs. On the contrary,
D. Circulating Current Control Methods the zero sequence circulating current flows through the dc-bus
The circulating current flows among the converter legs. and the converter legs, leading to a ripple in the dc-bus power.
These currents originate from the voltage difference between These ripples appear in the dc-bus current, considering that
the upper and lower arm of the converter legs. This current the dc-bus voltage is stiff [152]. The dc-bus current ripple
mainly consists of negative sequence component at twice the increases the SM capacitor voltage ripple and the magnitude
fundamental frequency during the balanced operation [140]. of arm current, which has a cascade effect on the device
The circulating current does not have any impact on the voltage stress and the conduction losses.
ac system voltages and currents. However, the improper con- In [152], the proportional integral resonant regulator is
trol of circulating currents increase the peak/RMS value of employed to control the positive, negative, and zero sequence
the arm current, which consequently increases the rating of circulating currents. In [153], the proportional resonant (PR)
devices, device power losses, and the SM capacitor voltage regulator based on the instantaneous power theory is designed
ripple [141]. By proper sizing of the arm inductors, the to suppress the zero sequence circulating currents. Thereby, the
magnitude of circulating currents can be suppressed to an power fluctuations on the dc side are minimized. In [154], the
extent [142]. However, a closed-loop control technique is suppression of the circulating currents in the stationary-αβ0
necessary to eliminate or minimize the circulating current in frame is discussed. In this approach, the high-pass filter is
the converter legs. used to extract the double line frequency component from the
In [143] and [144], the open-loop methods are presented circulating currents, which affects the dynamic performance
for the circulating current control and harmonic suppression. of the controller [154]. A nonideal PR regulator is presented
These methods are very sensitive to the parameter variation. to adapt the grid frequency variation and to minimize the cir-
The synchronous-dq reference frame-based control approach culating current during unbalanced conditions [155]. In [156],
is widely used for the complete elimination of circulating the circulating currents are controlled using a PR regulator,
current as shown in Fig. 12. In this approach, the circu- where the PR regulator is designed to control the fundamental
lating currents in the stationary-abc frame are transformed and double line frequency components.
into the synchronous-dq frame rotating at twice the funda- Alternatively, the circulating currents are separated into
mental frequency. In the synchronous-dq frame, the double the positive, negative, and zero sequence components in
line frequency circulating current component becomes dc the synchronous-dq0 frame and are controlled using a
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1643

simple PI regulator [157]. In [158], the decoupled double

synchronous-dq0 frame-based current control is presented
to control the sum and the differential energy of each
converter leg. The stationary-abc frame-based control method
is presented in [159], where the combination of the PI
regulator and vector PI regulator is used in the current
control. Other hand, the dc-bus current or SM capacitor
voltage ripple is directly controlled using a PR regulator.
Therefore, the magnitude of zero sequence circulating current Fig. 14. Classification of MPC schemes for an MMC.
is minimized [160]. This controller generates a zero sequence
voltage command, which is added to the reference modulation further [172]–[174]. The CMV and circulating current can be
signals. The performance of classical control methods is injected at the same frequency or different frequencies. Among
greatly limited by the parameters of the current controller, them, the CMV at three-times the output frequency and the
switching frequency, and the type of PWM scheme. circulating current at twice the output frequency are highly
effective in minimizing the voltage ripple [174]. However,
E. Minimization of Capacitor Voltage Ripple these methods require complex analysis and offline optimiza-
tion techniques to generate the reference CMV and circulating
Submodule capacitor voltage ripple is one of the major current. In [30] and [33], a simplified decoupled current model
issues in an MMC-based motor drive system under constant- is presented for an online generation of reference CMV and
torque and low-speed operation [161], [162]. Low-frequency circulating current. The decoupled model is simple and highly
voltage ripple, particularly second and fourth-order harmonic effective with a ripple reduction capability of 60%. These
components, can be observed in the SM capacitor voltage. The methods increase the RMS value of the arm current and the
voltage ripple generates the circulating current on the arm side magnitude of the motor CMV, which further increases the
corresponding to the ripple frequency [163]. The circulating voltage stress on the motor winding and bearings.
current increases the peak value of arm current, thereby Alternatively, the converter configuration can be modified to
increasing the voltage stress on the devices and introducing operate at low-frequency region without increasing the capac-
additional conduction losses. The SM capacitor voltage ripple itor voltage ripple. Several new topologies such as passive
can be reduced at the system level by selecting the proper size cross-connected, active cross-connected arm-based MMCs,
of SM capacitor (capacitance), which increases the cost and and modified MMC are presented in [175]–[177]. The cross-
size of the converter [164], [165]. connected arm topologies consist of a cross-connected branch
The classical control methods are designed to minimize the between the upper and lower arms. This cross-connected
low-frequency voltage ripple components using two additional branch absorbs the power fluctuations in the upper and lower
degrees of freedom, i.e., CMV on the ac-side and a circulating arms during the low-frequency operation [176], [177].
current between the converter legs [166]–[168]. In [166],
a zero-sequence voltage at twice the output frequency is
injected to compensate the double line frequency components V. M ODEL P REDICTIVE C ONTROL
in the capacitor voltage. In [167], an injection of the second M ETHODS FOR AN MMC
harmonic circulating current component is presented to shape In recent years, the MPC has emerged as a powerful tool to
the capacitor voltage ripple. The second and fourth-order control the power converter systems. The MPC approach has
harmonic circulating current components are also used to a simple design and is intuitive to include multiple constraints
minimize the voltage ripple in an MMC [168]. in a cost function. Other interesting features of MPC are a fast
The voltage ripple can be minimized using either proper dynamic response, smaller steady-state error in comparison to
PWM scheme or by adding the CMV to the modulation the classical control methods, ease of handling a control delay,
signals [169], [170]. In [169], a discontinuous-PWM scheme robustness against the system parameter variation [178]–[180].
is presented for the voltage ripple minimization in an MMC. The MPC approach eliminates the use of PI regulators and
This approach includes the CMV in the modulation stage pulsewidth modulators, which significantly improves the
itself. In [170], an improved PSC-PWM is presented to dynamic performance and controllability [181]–[183]. These
minimize the SM capacitor voltage ripple. This scheme features attracted a lot of researchers who rigorously inves-
improves the THD of the output voltage compared with the tigated the MPC approach for various multilevel converters
conventional PSC-PWM scheme. Alternatively, several CMV such as NPC [184], ANPC [185], FC [186], CHB [187], [188],
injection methods such as sinusoidal, third-order harmonic, matrix converters [189], and hybrid converter topologies [190].
and square-wave CMVs are studied in [171]. Among them, the The MPC approach is further extended to an MMC and
square-wave CMV injection method significantly minimizes is suitable to achieve multiple control objectives such as
the voltage ripple without increasing the peak value of the output current regulation, SM capacitor voltage control, and
arm current (50% smaller than that of other methods) and it circulating current reduction. To control the MMC, several
has low conduction losses. optimized per-phase and three-phase-based MPC schemes
A combination of high-frequency CMV and the cir- are presented in the literature. The classification of MPC
culating current is used to reduce the voltage ripple schemes for an MMC is shown in Fig. 14. Among them,
1644 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

Fig. 15. Control block diagram of FCS-MPC. Fig. 17. Control block diagram of dual-stage MPC.

between groups and SMs, respectively [195]. In the predictive

algorithm, the cost function optimization process is limited
to the nearest two switching states rather than all possible
combinations. Hence, the dynamic response is significantly
improved, and the number of switching states are reduced to
(N + 1) [196].
The MPC scheme can be implemented in multiple stages
Fig. 16. Control block diagram of indirect MPC.
using more than one cost function to achieve the control
objectives of an MMC. Fig. 17 shows the control block
diagram of a dual-stage MPC, in which two cost functions
the finite control set MPC (FCS-MPC) is a three-phase-based
are employed to achieve the control objectives [197]. The
approach in which the control objectives of three phases are
objectives of output current and circulating current are eval-
included in a single cost function as shown in Fig. 15. The
N uated in stage-I of MPC to obtain the optimized voltage
cost function is evaluated for 3C2N switching states, and
vector. The switching states corresponding to the optimized
the optimized switching state is selected and applied to the
voltage vector are used in stage-II to achieve the SM capacitor
converter, where N is the number of SMs per arm and C is
voltage control. By doing so, the control complexity can be
the number of combinations [191]. The number of available
minimized without affecting the dynamic response. It is also
switching states drastically increases with the number of SMs.
possible to control every objective of an MMC using an
Hence, the FCS-MPC is difficult to implement in real-time
independent predictive algorithm [198]. However, a further
for an MMC-based HVdc system. To address this issue, the
deeper analysis is required to study the dependence of each
per-phase-based MPC approach, often referred to as a direct
predictive algorithm and its impact on the converter dynamic
MPC (D-MPC), is employed to control the MMC in the
response.
HVdc applications. In this approach, the control objectives of
The performance of MPC scheme greatly depends on the
each phase are included in a cost function and are optimized
mathematical model of the system. Several mathematical
for 22N switching states using an independent predictive
models based on the per-phase and three-phase modeling
algorithm [192]. The optimization process minimizes the
philosophy are presented in the literature. The per-phase
error between the reference and measured output currents.
models are simple and consume less time to simulate, but
With the D-MPC, the converter operates with a variable
not suitable to control the ripple in capacitor voltage and
switching frequency and generates output currents with a
dc-bus current during the balanced and unbalanced operating
higher harmonic distortion. Alternatively, the model predictive
conditions [199]. On the contrary, the three-phase-based math-
direct current control is presented, in which the error between
ematical models consist of zero sequence voltage or CMV
the reference and measured output currents is bounded to a
term inherently [200]. By introducing this voltage component,
certain value. These boundaries set the limits on the THD
the ripple in SM capacitor voltage, dc-bus current, and output
and the converter switching frequency [193]. This approach is
current is significantly minimized [201]. Also, the arm voltage
suitable to implement with a longer predictive horizon as well.
balancing can be achieved without controlling the circulating
To reduce the computational burden, several modified MPC
currents.
strategies have been reported in the literature. The indirect
MPC is one of the simplified approaches that combine the VI. A PPLICATIONS OF MMC
classical voltage balancing approach with the predictive algo-
rithm as shown in Fig. 16 [135]. The predictive algorithm is MMCs are commercialized in the form of standard and
designed to control the output and circulating currents, and customized products for high-power applications. The MMC is
the classical voltage balancing approach takes care of the SM used in several industrial applications such as power transmis-
voltage balancing task. Therefore, the evaluation of switching sion systems, motor drives, and power quality improvement,
states is reduced to (N + 1)2 , but the dynamic response as shown in Fig. 18. The in-depth details of these applications
becomes sluggish [194]. The sorting algorithm in a voltage are discussed in this section.
balancing approach is replaced with a group sorting algorithm,
in which the SMs of each arm are evenly divided into several A. Power Transmission Systems
groups. The computational complexity is effectively reduced The HVdc transmission systems become more viable and
through the implementation of a two-level optimized MPC economical to transmit the power over a long distance with low
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1645

Fig. 19. MMC-based HVdc system.

Currently, the MMC-based VSC-HVdc technology is com-

mercially available for the bulk power transfer. The MMC
provides a substantial increase in the output voltage levels and
output switching frequency, enabling a reduction in the filter
size and power losses [25]. Even though the overall converter
Fig. 18. Applications of MMC. switching frequency is high, but each SM can be switched at a
very low frequency of a few hundred hertz. Other prominent
power losses. The HVdc systems have higher controllability features are modularity and scalability and handle a higher
and stability, improve the power quality, and require less dc-voltage using a low-cost and low-voltage semiconductor
expensive bus-bar structures on the ac-side in comparison with technology [27].
the high-voltage alternating current systems [202]. The HVdc The typical back-to-back configuration of MMC-HVdc
systems are suitable to interconnect multiple asynchronous ac system is shown in Fig. 19. The ac-side of MMC-I and
systems and integrate offshore wind farms to the transmission MMC-II is connected to the grid through a transformer, and
grids using submarine sea cables. their dc systems are interconnected through a dc cable of
The line-commutated current source converters (CSCs) are several hundreds of kilometer long. The MMC provides a
widely used in the traditional HVdc technology. The line- high-quality output voltage and current waveforms, and there
commutated CSCs operate at lagging power factor and require is no need for the harmonic filters in the dc and ac-sides of
external harmonic filters to supply the deficit reactive power MMC-HVdc system [209]. In HVdc applications, the MMC
to the grid [203], [204]. Later, the IGBT-based VSCs are designed with a larger number of SMs of 200–400 SMs/arm
employed in the HVdc technology. The VSC-HVdc technol- to reach an operating voltage of ± 320 kV [27], [47]. In the
ogy transmits active power and can provide reactive power at future, it is expected that the MMC-HVdc system can reach
both sending and receiving end. The VSC technology even an operating voltage of ± 500 to ± 600 kV.
permits the black start and is suitable to form a multiterminal The HB-SM is widely used in the MMC-HVdc systems.
HVdc system [205]. One of the major challenges associated with the HB-SM-based
The first generation VSC-HVdc systems are developed MMC-HVdc systems is the lack of dc-side fault blocking
with PWM-based two-level power converters. The two-level capability. In the case of dc-side faults, the fault current
converter uses series-connected IGBTs to handle a higher flows through the antiparallel diodes of the HB-SMs [49].
blocking voltage; therefore, the dc-bus voltage level of HVdc Therefore, the HB-SM-based MMC does not have any
systems can be increased. In addition, the PWM converters capability to block the dc-side faults. Alternatively, several
switched at high frequencies (1–2 kHz) resulted in higher new SMs with dc-fault blocking capability are studied for
power losses [206]. Over the past few years, several multilevel MMC-HVdc applications [66]. Some of the SMs are FB,
converter topologies are reported in the literature. Among CD, cross-connected, and asymmetric unipolar FB [28], [50],
them, the PWM-based three-level NPC is utilized in the sec- [65], [71], [73]. It is also possible to modify the complete
ond generation HVdc systems [207]. The existing multilevel arm structure to achieve the dc-fault blocking capability
converters significantly improve the output power quality but in the MMC [72], [210]. With the new development, the
intolerant to the dc-side faults. In the case of dc-side faults, fault current can be completely eliminated from the MMC.
the converter acts as an uncontrolled rectifier due to the However, compared to the HB-SM-based MMC, the new
antiparallel diode of the IGBT, even if the IGBTs are blocked. solution sacrifices the cost and power losses. Currently, the
Earlier, the ac-side breakers are used to limit the fault current. new solutions are under research and development stage only,
Also, significant research has been conducted on the dc yet to implement in the commercial products.
circuit breakers such as solid-state, hybrid, and resonant circuit Currently, four types of HB-SM-based MMC-HVdc tech-
breakers for HVdc systems [208]. These circuit breakers are in nologies are commercially available under different trade
the research and development stage only and yet to implement names of Siemens-HVdc Plus, ABB-HVdc Light Gen.4,
in the commercial HVdc projects. Considering that it is more C-EPRI-HVdc Flexible, and GE/(Alstom)-HVdc Maxsine,
advantageous to adapt the fault tolerant power converters for which can reach up to a voltage of ± 320 kV [41]–[43], [47].
the HVdc applications. The list of MMC-based HVdc projects are summarized
1646 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

TABLE I TABLE II
C OMMERCIAL MMC-BASED HVDC S YSTEMS C OMMERCIAL M ULTITERMINAL MMC-BASED HVDC S YSTEMS
(P OINT- TO -P OINT C ONNECTION )

Fig. 21. MMC-HVdc with submarine cable for an off-shore wind farm.

TABLE III
C OMMERCIAL MMC-HVDC-BASED O FFSHORE W IND FARMS

Fig. 20. Multiterminal MMC HVdc system.

in Table I. San Francisco’s Trans Bay project is the first

MMC-based HVdc system installed by Siemens in 2010. This
project handles a power of 400 MW at a voltage rating of
±200 kV through a subsea cable of 85 km long [40]. In this Typically, the offshore wind farms are located in the sea,
project, the HB-SMs are connected in cascade to form an which is far from the utility grid. These offshore wind farms
MMC [211]. are connected to the ac-grid for the subsequent distribution and
Another important application of MMC is the multiterminal consumption of the generated power [217]–[219]. The MMC-
HVdc system as shown in Fig. 20. Earlier, the CSCs are HVdc system is a suitable candidate to transfer the power over
employed to form a multiterminal HVdc system, in which a long distance. The connection of an offshore wind farm with
one terminal is operated in the voltage control mode to set the ac-grid through a submarine cable and the MMC-HVdc
the system voltage, and other terminals are operated in the network is shown in Fig. 21. Several projects relevant to this
current controlled mode [208]. However, the control and application are indicated in Table III. These projects are mainly
implementation of CSC-based multiterminal HVdc systems installed in the European region with a maximum voltage of
are quite difficult. Due to the voltage nature of MMC dc-bus, ±320 kV and a power rating of 900 MW.
it is possible to interconnect multiple HVdc systems to form
an HVdc grid with a high controllability, efficiency, and
B. Medium-Voltage Motor Drives
reliability than the ac grids [212]–[216]. The China Electric
Power Research Institute (C-EPRI) installed an MMC-based The MMC offers voltage scalability by connecting the SMs
five-terminal HVdc system in Zhoushan to distribute the in series, operates with standard motors without output filters,
power among different regions as indicated in Table II. In this and offers a simple construction with less engineering effort,
project, the HVdc Flexible technology is employed to handle unlike conventional multilevel converters [120]. However,
a voltage of ± 200 kV. Another three-terminal HVdc system the SM capacitor voltage ripple is one of the main techni-
(Nanao) is installed in Shantou by China Southern Power cal challenges in MMC-based motor drive systems during
Grid (CSG) to deliver the wind power from the Island of constant-torque, low-speed operation [173]. The voltage ripple
Nanao to the main land [215], [216]. can be reduced at the system level using a larger size of
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1647

TABLE IV
C OMMERCIAL MMC-BASED M OTOR D RIVE S YSTEMS

Fig. 22. MMC-based motor drive with passive rectifier.

Fig. 24. Configuration of MMC-based STATCOM.

Fig. 23. MMC-based motor drive with active rectifier. TABLE V

C OMMERCIAL MMC-BASED S TATCOM

SM capacitors, which increases the cost and physical size of

the converter [164]. On the other hand, the CMV is injected
into the system, which minimizes the capacitor voltage ripple,
but increases the voltage stress on the motor winding and
causes bearing failure [171], [220]. Hence, a transformer is
required to block the common-mode current in an MMC-based
motor drive system.
The typical configuration of the MMC-based motor drive
with passive front-end rectifier is shown in Fig. 22. The motor
drive uses a 12-pulse diode bridge rectifier on the ac-grid
C. Power Quality Improvement
side and an MMC with a cascade HB SM on the motor
side [56]. The 12-pulse rectifier eliminates the lower order har- Another potential application of MMC is to operate as a
monics in the line current and improves power quality on the STATCOM to compensate both harmonics and reactive power
ac-grid side. Also, the transformer on the ac-grid side blocks demanded by the distorted and unbalanced loads. The MMC-
the common-mode current from entering into the system. based STATCOM has modularity and voltage scalability
Other hand, the MMC consists of six SMs in each arm features, and it allows the direct connection (without trans-
and generates a phase-voltage with 13-levels at the motor former) to the medium or high-voltage power system networks
terminals. Hence, the output filters are not required. as shown in Fig. 24. In this application, the HB and FB SMs
For regenerative applications, the motor drive systems are the preferred choices for an MMC [223]. In MMC-based
employ an MMC on both the ac-grid and motor side as STATCOM technology, the arms are connected in delta
shown in Fig. 23. The MMC-I is controlled with a VOC configuration instead of star configuration. The MMC-based
approach, and it generates a sinusoidal current on the ac-grid STATCOM technologies were developed by Siemens to handle
side and improves the grid power factor. The MMC-II is a voltage of 33 kV without a transformer and 220 kV with
controlled with a FOC to meet the motor requirements such a transformer as shown in Table V. These technologies are
as speed/torque and flux. Currently, the MMC-based motor installed in the offshore wind farms, substations, and the arc
drives are commercially available with a maximum voltage furnace industries to supply the reactive power [121], [122].
of 7.2 kV and the power rating of 13.7 MVA by Siemens The unified power quality conditioner (UPQC) is another
as shown in Table IV [39]. The MMC is also used in important technology in the power quality improvement. The
railway traction as a power electronic transformer, in which UPQC system has two stages of compensation with the series
the single-phase structure of MMC is employed to connect and parallel connection of the power converters. The UPQC
the low-frequency and the high-voltage line to a variable- can be realized using a back-to-back connection of MMC as
frequency-based medium-voltage three-phase machine by a shown in Fig. 25. The MMC-I is connected to the ac-grid
medium-frequency transformer [23], [221], [222]. through a transformer. It provides the series compensation,
1648 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

positive and negative bars of the upper and lower arms,

respectively. These power units control the power among the
arms. Also, the middle SM is employed to reduce the required
number of SMs in each arm. Inherently, the hybrid MMC has
a smaller SM capacitor voltage ripple. Hence, a smaller SM
capacitance is required, which reduces the size and cost of the
converter [175]. Therefore, the hybrid MMC is suitable for the
variable speed motor drive applications. However, the hybrid
topology sacrifices the modularity and reliability.
Fig. 26(d) shows an interesting topology referred as a
Fig. 25. Configuration of MMC-based UPQC. FC MMC (FC-MMC). In this configuration, the midpoint of
the upper and lower arms is connected through a capacitor. The
where it handles the voltage sag and voltage swell problems cross-connected capacitor is used to achieve the power balance
of the ac-grid. The MMC-II provides a shunt compensa- between the arms and to minimize the capacitor voltage ripple
tion, which is directly connected to the ac-grid without any without increasing the converter CMV [176]. However, the
transformer. Eliminating the transformer in shunt compen- capacitor carries a large magnitude of current proportional
sation is one of the advantages of MMC [224], [225]. The to the injected high-frequency voltage corresponding to the
shunt compensator handles the reactive power and harmonic reduction in the voltage ripple. Hence, the devices should be
compensation of the distorted load. The high-quality output over-rated. The FC-MMC is available with an even number of
waveforms and the fast current control of MMC help to SMs to create the midpoint connection for the capacitor.
improve the active filtering capability of the UPQC system. The FC in FC-MMC is replaced with a stack of SMs,
Other interesting applications of MMC are the high-voltage which results in the active cross-connected MMC (ac-MMC),
dc–dc conversion to form a dc-micro grid [51], [226]–[228] as shown in Fig. 26(e). The ac-MMC has similar features of
and grid integration of photovoltaic (PV) systems [229], [230], FC-MMC such as smaller capacitor voltage ripple at zero/
wind energy conversion systems [231], and battery energy low-speed operation [177]. The ac-MMC requires higher
storage systems [232]. These new applications are current number of passive and active components compared to the
research focus, and expecting that the MMC plays a key role ac-MMC. Hence, the cost is high. Also, the converter size
in the development of smart grids. and its control complexity are significantly increased in
comparison with the FC-MMC.
The back-to-back configuration of MMC is widely
VII. A DVANCED AND H YBRID C ONFIGURATIONS OF MMC
employed in the interconnection of two ac systems. Also,
In the literature, several advanced and hybrid MMC con- it is possible to connect two ac systems directly using a
figurations are presented to address the operational and tech- matrix MMC as shown in Fig. 27(a). In the matrix MMC, the
nical issues of a conventional MMC. Fig. 26(a) shows the required number of arms is reduced to 9 compared with the
circuit configuration of an alternate arm MMC, in which each 12 arms in the back-to-back conventional MMC configuration.
arm is equipped with a stack of H-bridge SM, a director However, the matrix MMC requires a SM that can produce
switch, and an arm inductor [210]. The stack of H-bridge both the polarities of the output voltage. At the same time,
SM generates a multilevel stepped waveform and limits the the arm inductors can be eliminated in the matrix MMC. But,
dc-side fault current [233]. The director switch is composed the control complexity of matrix MMC is quite high, due
of several IGBT devices, which are connected in series to to the presence of input and output current components in
handle the required operating voltage. The director switch each arm [234]. Another interesting topology is a hexagonal
is used to select the conducting arm such that the upper MMC that requires only six arms to connect two ac systems
arm generates the positive half-cycle of ac voltage, while the Fig. 27(b). The hexagonal MMC has also required an SM
lower arm generates the negative half-cycle (i.e., at any given with a capability of generating both polarities of the output
time one arm conducts). By doing so, the converter losses voltage. In hexagonal MMC, the control actions are limited
can be significantly reduced compared with the conventional compared with the matrix MMC, particularly for unbalanced
HB-SM-based MMC [54]. grid conditions [235].
Fig. 26(b) shows the MMC with a middle SM, in which
the upper and lower arms are connected through a middle VIII. F UTURE T RENDS
SM. The new MMC structure requires less number of SMs, Over the past few years, there was a wide range of devel-
while producing the voltage waveform with a similar number opments in the MMC in terms of topologies, modulation
of voltage levels as that of the conventional MMC [171]. The schemes, control methods, and applications. Still, there are
middle SM helps to achieve the power balance between the several aspects that need to be addressed to expand its potential
arms but increases the complexity of voltage balancing among applications. One of the drawbacks is the expensive and a
the SMs. Also, it is hard to realize the redundancy for a middle larger size of capacitors required to minimize the voltage
SM, which affects the reliability of the converter. ripple during zero and low-frequency operation. To handle
The hybrid configuration of MMC is shown in Fig. 26(c), this problem, it is necessary to design and implement a
in which two three-phase power units are connected to the high-performance control method. By solving this problem,
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1649

Fig. 26. Advanced and hybrid configurations of MMC. (a) Alternate arm MMC. (b) Middle-SM MMC. (c) Hybrid MMC. (d) FC MMC. (e) Active
cross-connected MMC.

the research on the MMC can be opened to the new applica- of battery energy systems. Another interesting application is
tions such as medium/high-voltage dc–dc converters to form the variable-speed wind energy conversion systems. Currently,
the dc-grids, medium-frequency traction applications as a 8-MW wind turbines are available in the market, and some
power electronic transformer, and medium-voltage variable- manufacturers announced a wind turbine of 10–15 MW in
speed drive with constant torque applications. their future projects. Considering this change, the MMC can
The modular characterization of MMC can be used in the be a good fit for these applications, and it can be used in the
PV plants to increase the energy storage so that the plants future wind energy projects.
can supply the power to consumers during the night time. Another important issue is the dc-side fault blocking capa-
This modular feature is also useful for the grid integration bility in the HVdc systems. Most of the MMC-HVdc systems
1650 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

Fig. 27. Modular ac–ac converters. (a) Matrix MMC. (b) Hexagonal MMC.

are designed with HB-SM and cannot block the dc-side fault Also, the state-of-the-art and emerging technologies in MMCs
current. The FB-SM and CD-SM can eliminate the dc-side are discussed. Finally, the list of commercial applications
fault current, but these SMs are not used in the commercial based on the MMC and their technical details are provided.
products due to their cost, efficiency, and the design com-
plexity. This problem opened up the new research on HVdc R EFERENCES
circuit breakers to handle the internal and external faults. In the [1] J. Rodríguez et al., “Multilevel converters: An enabling technology for
future, it is possible to interconnect both the ac and dc systems high-power applications,” Proc. IEEE, vol. 97, no. 11, pp. 1786–1817,
to form a super grid using the MMC technology. Nov. 2009.
[2] S. Kouro, J. Rodríguez, B. Wu, S. Bernet, and M. Perez,
Other interesting research areas are the design and imple- “Powering the future of industry: High-power adjustable speed drive
mentation of capacitor voltage control methods with a lesser topologies,” IEEE Ind. Appl. Mag., vol. 18, no. 4, pp. 26–39,
number of voltage sensors, which reduces the system cost Jul. 2012.
[3] H. Abu-Rub, J. Holtz, J. Rodríguez, and G. Baoming, “Medium-voltage
and improves the reliability. The multiphase MMCs for motor multilevel converters: State of the art, challenges, and requirements
drives are another interesting topic. The multiphase system in industrial applications,” IEEE Trans. Ind. Electron., vol. 57, no. 8,
increases the converter power handling capacity, at the same pp. 2581–2596, Aug. 2010.
[4] P. W. Wheeler, J. Rodríguez, J. C. Clare, L. Empringham, and
time the drive performance is greatly improved (small torque A. Weinstein, “Matrix converters: A technology review,” IEEE Trans.
and flux ripples). However, these systems require new modu- Ind. Electron., vol. 49, no. 2, pp. 276–288, Apr. 2002.
lation and control schemes to extend their applications. [5] L. Empringham, J. W. Kolar, J. Rodríguez, P. W. Wheeler, and
J. C. Clare, “Technological issues and industrial application of matrix
IX. C ONCLUSION converters: A review,” IEEE Trans. Ind. Electron., vol. 60, no. 10,
pp. 4260–4271, Oct. 2013.
The attractive features of the MMC played a key role in the [6] J. Rodríguez, S. Bernet, B. Wu, J. O. Pontt, and S. Kouro, “Multi-
development of new HVdc transmission systems, medium- level voltage-source-converter topologies for industrial medium-voltage
drives,” IEEE Trans. Ind. Electron., vol. 54, no. 6, pp. 2930–2945,
voltage motor drives, and power quality improvement Dec. 2007.
technologies. These technologies are commercialized by [7] S. Kouro et al., “Recent advances and industrial applications of
various leading industrial manufacturers such as GE, Alstom, multilevel converters,” IEEE Trans. Ind. Electron., vol. 57, no. 8,
pp. 2553–2580, Aug. 2010.
ABB, Siemens, and C-EPRI. Depending on the application, [8] L. G. Franquelo, J. Rodríguez, J. I. Leon, S. Kouro, R. Portillo, and
the MMC has several technical issues such as circulating M. A. M. Prats, “The age of multilevel converters arrives,” IEEE Ind.
currents, capacitor voltage ripple, and dc-bus faults. Also, Electron. Mag., vol. 2, no. 2, pp. 28–39, Jun. 2008.
[9] B. Wu and M. Narimani, High-Power Converters and AC Drives.
a complex control system is required to meet the several Hoboken, NJ, USA: Wiley, 2017.
control objectives of an MMC. Over the past few years, [10] J. Rodríguez, S. Bernet, P. K. Steimer, and I. E. Lizama, “A survey on
numerous studies are conducted to understand the behavior of neutral-point-clamped inverters,” IEEE Trans. Ind. Electron., vol. 57,
the MMC, resulting in new topologies, mathematical models, no. 7, pp. 2219–2230, Jul. 2010.
[11] H. D. T. Mouton, “Natural balancing of three-level neutral-point-
and control schemes. This paper presents a review of the recent clamped PWM inverters,” IEEE Trans. Ind. Electron., vol. 49, no. 5,
developments in the MMC in terms of the SM configurations, pp. 1017–1025, Oct. 2002.
mathematical models, pulsewidth modulation schemes, clas- [12] T. Brückner, S. Bernet, and H. Guldner, “The active NPC converter and
its loss-balancing control,” IEEE Trans. Ind. Electron., vol. 52, no. 3,
sical control schemes, and high-performance MPC methods. pp. 855–868, Jun. 2005.
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1651

[13] M. Narimani, B. Wu, and N. R. Zargari, “A novel five-level voltage [35] Q. Song, W. Liu, X. Li, H. Rao, S. Xu, and L. Li, “A steady-state
source inverter with sinusoidal pulse width modulator for medium- analysis method for a modular multilevel converter,” IEEE Trans.
voltage applications,” IEEE Trans. Power Electron., vol. 31, no. 3, Power Electron., vol. 28, no. 8, pp. 3702–3713, Aug. 2013.
pp. 1959–1967, Mar. 2016. [36] A. Antonopoulos, L. Angquist, L. Harnefors, K. Ilves, and H.-P. Nee,
[14] S. S. Fazel, S. Bernet, D. Krug, and K. Jalili, “Design and comparison “Global asymptotic stability of modular multilevel converters,” IEEE
of 4-kV neutral-point-clamped, flying-capacitor, and series-connected Trans. Ind. Electron., vol. 61, no. 2, pp. 603–612, Feb. 2014.
H-bridge multilevel converters,” IEEE Trans. Ind. Appl., vol. 43, no. 4, [37] M. A. Perez, S. Bernet, J. Rodríguez, S. Kouro, and R. Lizana, “Circuit
pp. 1032–1040, Jul. 2007. topologies, modeling, control schemes, and applications of modular
[15] C. Wang and Y. Li, “Analysis and calculation of zero-sequence volt- multilevel converters,” IEEE Trans. Power Electron., vol. 30, no. 1,
age considering neutral-point potential balancing in three-level NPC pp. 4–17, Jan. 2015.
converters,” IEEE Trans. Ind. Electron., vol. 57, no. 7, pp. 2262–2271, [38] S. Debnath, J. Qin, B. Bahrani, M. Saeedifard, and P. Barbosa, “Oper-
Jul. 2010. ation, control, and applications of the modular multilevel converter:
[16] J.-S. Lai and F. Z. Peng, “Multilevel converters—A new breed A review,” IEEE Trans. Power Electron., vol. 30, no. 1, pp. 37–53,
of power converters,” IEEE Trans. Ind. Appl., vol. 32, no. 3, Jan. 2015.
pp. 509–517, May 1996. [39] SIEMENS. Sinamics sm120. Accessed: 2013. [Online]. Avail-
[17] B. P. McGrath and D. G. Holmes, “Natural capacitor voltage balancing able: http://www.industry.siemens.com/drives/global/en/converter/mv-
for a flying capacitor converter induction motor drive,” IEEE Trans. drives/sinamics-sm120-cm/
Power Electron., vol. 24, no. 6, pp. 1554–1561, Jun. 2009. [40] S. Teeuwsen, “Modeling the trans bay cable project as voltage-sourced
[18] M. Narimani, B. Wu, Z. Cheng, and N. R. Zargari, “A new nested converter with modular multilevel converter design,” in Proc. IEEE
neutral point-clamped (NNPC) converter for medium-voltage (MV) Power Energy Soc. Gen. Meet., Jul. 2011, pp. 1–8.
power conversion,” IEEE Trans. Power Electron., vol. 29, no. 12, [41] ALSTOM. HVDC MaxSine. Accessed: Aug. 2013. [Online]. Available:
pp. 6375–6382, Dec. 2014. https://www.alstom.com/grid/products-and-services/engineered-
[19] M. Malinowski, K. Gopakumar, J. Rodríguez, and M. A. Pérez, energy-solutions/HVDC-transmission-systems/HVDC_Maxsine/
“A survey on cascaded multilevel inverters,” IEEE Trans. Ind. [42] ABB. HVDC Light Gen 4. Accessed: Aug. 2013. [Online]. Available:
Electron., vol. 57, no. 7, pp. 2197–2206, Jul. 2010. https://www.abb.com/systems/HVDC/
[20] I. Etxeberria-Otadui, A. Lopez-de-Heredia, J. San-Sebastian, [43] EPRI. HVDC Flexible. Accessed: 2013. [Online]. Available: https://
H. Gaztanaga, U. Viscarret, and M. Caballero, “Analysis of a www.epri/sgcc.com.cn/prgc/english/Product_Solution/HVDC_Flexible/
H-NPC topology for an AC traction front-end converter,” in Proc.
[44] M. Pereira, D. Retzmann, J. Lottes, M. Wiesinger, and G. Wong, “SVC
13th Power Electron. Motion Control Conf. (EPE-PEMC), Sep. 2008,
PLUS: An MMC STATCOM for network and grid access applications,”
pp. 1555–1561.
in Proc. IEEE Trondheim PowerTech, Jun. 2011, pp. 1–5.
[21] J. Rodríguez, P. W. Hammond, J. Pontt, R. Musalem, P. Lezana, and
[45] SIEMENS. The Efficient Way. Accessed: Aug. 2013. [Online]. Avail-
M. J. Escobar, “Operation of a medium-voltage drive under faulty
conditions,” IEEE Trans. Ind. Electron., vol. 52, no. 4, pp. 1080–1085, able: http://www.energy.siemens.com/ru/pool/hq/power-transmission/
Aug. 2005. FACTS/SVC_PLUS_The%20efficient%20Way.pdf
[22] M. Glinka and R. Marquardt, “A new AC/AC multilevel converter [46] G. Konstantinou, J. Zhang, S. Ceballos, J. Pou, and V. G. Agelidis,
family,” IEEE Trans. Ind. Electron., vol. 52, no. 3, pp. 662–669, “Comparison and evaluation of sub-module configurations in modular
Jun. 2005. multilevel converters,” in Proc. IEEE 11th Int. Conf. Power Electron.
[23] M. Glinka and R. Marquardt, “A new AC/AC-multilevel converter Drive Syst., Jun. 2015, pp. 958–963.
family applied to a single-phase converter,” in Proc. 5th Int. Conf. [47] SIEMENS. The Smart Way HVDC Plus-One Step Ahead.
Power Electron. Drive Syst. (PEDS), vol. 1. Nov. 2003, pp. 16–23. Accessed: Aug. 2013. [Online]. Available: https://www.
[24] H. Akagi, “Classification, terminology, and application of the modular energy.siemens.com/us/en/power-transmission/HVDC/HVDCplus/
multilevel cascade converter (MMCC),” IEEE Trans. Power Electron., [48] A. Nami, J. Liang, F. Dijkhuizen, and G. Demetriades, “Modular
vol. 26, no. 11, pp. 3119–3130, Nov. 2011. multilevel converters for HVDC applications: Review on converter
[25] H. J. Knaak, “Modular multilevel converters and HVDC/FACTS: cells and functionalities,” IEEE Trans. Power Electron., vol. 30, no. 1,
A success story,” in Proc. 14th Eur. Conf. Power Electron. Appl. (EPE), pp. 18–36, Jan. 2015.
Aug. 2011, pp. 1–6. [49] X. Li, Q. Song, W. Liu, H. Rao, S. Xu, and L. Li, “Protection of
[26] A. Lesnicar and R. Marquardt, “An innovative modular multilevel nonpermanent faults on DC overhead lines in MMC-based HVDC
converter topology suitable for a wide power range,” in Proc. IEEE systems,” IEEE Trans. Power Del., vol. 28, no. 1, pp. 483–490,
Power Tech Conf., vol. 3. Bologna, Italy, Jun. 2003, pp. 23–26. Jan. 2013.
[27] S. Allebrod, R. Hamerski, and R. Marquardt, “New transformerless, [50] R. Zeng, L. Xu, L. Yao, and D. J. Morrow, “Precharging and DC
scalable modular multilevel converters for HVDC-transmission,” in fault ride-through of hybrid MMC-based HVDC systems,” IEEE Trans.
Proc. IEEE Power Electron. Specialists Conf. (PESC), Jun. 2008, Power Del., vol. 30, no. 3, pp. 1298–1306, Jun. 2015.
pp. 174–179. [51] G. J. Kish, M. Ranjram, and P. W. Lehn, “A modular multilevel DC/DC
[28] R. Marquardt, “Modular multilevel converter: An universal concept converter with fault blocking capability for HVDC interconnects,”
for HVDC-networks and extended DC-bus-applications,” in Proc. Int. IEEE Trans. Power Electron., vol. 30, no. 1, pp. 148–162, Jan. 2015.
Power Electron. Conf. (IPEC), Sapporo, Japan, Jun. 2010, pp. 502–507. [52] L. He, K. Zhang, J. Xiong, S. Fan, and Y. Xue, “Low-frequency
[29] L. Angquist, A. Antonopoulos, D. Siemaszko, K. Ilves, M. Vasiladiotis, ripple suppression for medium-voltage drives using modular multilevel
and H.-P. Nee, “Open-loop control of modular multilevel converters converter with full-bridge submodules,” IEEE Trans. Emerg. Sel. Topics
using estimation of stored energy,” IEEE Trans. Ind. Appl., vol. 47, Power Electron., vol. 4, no. 2, pp. 657–667, Jun. 2016.
no. 6, pp. 2516–2524, Nov. 2011. [53] G. P. Adam and I. E. Davidson, “Robust and generic control of
[30] R. Lizana, M. Perez, S. Bernet, J. Espinoza, and J. Rodríguez, “Control full-bridge modular multilevel converter high-voltage DC transmission
of arm capacitor voltages in modular multilevel converters,” IEEE systems,” IEEE Trans. Power Del., vol. 30, no. 6, pp. 2468–2476,
Trans. Power Electron., vol. 31, no. 2, pp. 1774–1784, Feb. 2016. Dec. 2015.
[31] J. Qin and M. Saeedifard, “Predictive control of a three-phase DC–AC [54] M. M. C. Merlin and T. C. Green, “Cell capacitor sizing in multilevel
modular multilevel converter,” in Proc. IEEE Energy Convers. Congr. converters: Cases of the modular multilevel converter and alternate arm
Expo. (ECCE), Sep. 2012, pp. 3500–3505. converter,” IET Power Electron., vol. 8, no. 3, pp. 350–360, 2015.
[32] F. Zhang and G. Joos, “A predictive nearest level control of modular [55] S. Du, B. Wu, and N. Zargari, “A star-channel modular multilevel con-
multilevel converter,” in Proc. IEEE Appl. Power Electron. Conf. verter for zero/low fundamental frequency operation without injecting
Expo. (APEC), Mar. 2015, pp. 2846–2851. common-mode voltage,” IEEE Trans. Power Electron., to be published,
[33] R. Lizana, M. A. Perez, D. Arancibia, J. R. Espinoza, and J. Rodríguez, doi: 10.1109/TPEL.2017.2701905.
“Decoupled current model and control of modular multilevel con- [56] BENSHAW. M2l 3000 Series Medium Voltage Motor Drive.
verters,” IEEE Trans. Ind. Electron., vol. 62, no. 9, pp. 5382–5392, Accessed: 2013. [Online]. Available: http://www.benshaw.com/
Sep. 2015. uploadedFiles/Literature/Benshaw_M2L_MVFD_2.3-6.6kV.pdf
[34] A. Dekka, B. Wu, V. Yaramasu, and N. R. Zargari, “Model predictive [57] N. Celanovic and D. Boroyevich, “A comprehensive study of neutral-
control with common-mode voltage injection for modular multilevel point voltage balancing problem in three-level neutral-point-clamped
converter,” IEEE Trans. Power Electron., vol. 32, no. 3, pp. 1767–1778, voltage source PWM inverters,” IEEE Trans. Power Electron., vol. 15,
Mar. 2017. no. 2, pp. 242–249, Mar. 2000.
1652 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

[58] E. Solas, G. Abad, J. Barrena, S. Aurtenetxea, A. Carcar, and L. Zajac, [79] S. Rohner, J. Weber, and S. Bernet, “Continuous model of modular
“Modular multilevel converter with different submodule concepts— multilevel converter with experimental verification,” in Proc. IEEE
Part I: Capacitor voltage balancing method,” IEEE Trans. Ind. Energy Convers. Congr. Expo., Sep. 2011, pp. 4021–4028.
Electron., vol. 60, no. 10, pp. 4525–4535, Oct. 2013. [80] H. Saad et al., “Dynamic averaged and simplified models for MMC-
[59] T. Brückner and S. Bemet, “Loss balancing in three-level voltage source based HVDC transmission systems,” IEEE Trans. Power Del., vol. 28,
inverters applying active NPC switches,” in Proc. IEEE 32nd Annu. no. 3, pp. 1723–1730, Jul. 2013.
Power Electron. Specialists Conf., vol. 2. Jun. 2001, pp. 1135–1140. [81] M. A. Pérez, R. L. Fuentes, and J. Rodríguez, “Decoupled current
[60] P. Barbosa, P. Steimer, J. Steinke, L. Meysenc, M. Winkelnkemper, and control of modular multilevel converter for HVDC applications,” in
N. Celanovic, “Active neutral-point-clamped multilevel converters,” in Proc. IEEE Int. Symp. Ind. Electron., May 2012, pp. 1979–1984.
Proc. IEEE 36th Power Electron. Specialists Conf. (PESC), Jun. 2005, [82] A. Antonopoulos, L. Ängquist, and H.-P. Nee, “On dynamics and
pp. 2296–2301. voltage control of the modular multilevel converter,” in Proc. 13th Eur.
[61] M. Schweizer and J. W. Kolar, “Design and implementation of a highly Conf. Power Electron. Appl., Sep. 2009, pp. 1–10.
efficient three-level T-type converter for low-voltage applications,” [83] J. Xu, A. M. Gole, and C. Zhao, “The use of averaged-value model
IEEE Trans. Power Electron., vol. 28, no. 2, pp. 899–907, Feb. 2013. of modular multilevel converter in DC grid,” IEEE Trans. Power Del.,
[62] A. Dekka, B. Wu, N. R. Zargari, and R. L. Fuentes, “A space-vector vol. 30, no. 2, pp. 519–528, Apr. 2015.
PWM-based voltage-balancing approach with reduced current sensors [84] D. C. Ludois and G. Venkataramanan, “Simplified terminal behav-
for modular multilevel converter,” IEEE Trans. Ind. Electron., vol. 63, ioral model for a modular multilevel converter,” IEEE Trans. Power
no. 5, pp. 2734–2745, May 2016. Electron., vol. 29, no. 4, pp. 1622–1631, Apr. 2014.
[63] A. Dekka, B. Wu, and N. R. Zargari, “Dynamic voltage bal- [85] G. Bergna et al., “A generalized power control approach in ABC frame
ancing algorithm for modular multilevel converter with three- for modular multilevel converter HVDC links based on mathematical
level flying capacitor submodules,” in Proc. Int. Power Electron. optimization,” IEEE Trans. Power Del., vol. 29, no. 1, pp. 386–394,
Conf. (IPEC-Hiroshima-ECCE-ASIA), May 2014, pp. 3468–3475. Feb. 2014.
[64] Y. Xue, Z. Xu, and G. Tang, “Self-start control with grouping sequen- [86] J. Xu, C. Zhao, W. Liu, and C. Guo, “Accelerated model of modular
tially precharge for the C-MMC-based HVDC system,” IEEE Trans. multilevel converters in PSCAD/EMTDC,” IEEE Trans. Power Del.,
Power Del., vol. 29, no. 1, pp. 187–198, Feb. 2014. vol. 28, no. 1, pp. 129–136, Jan. 2013.
[65] Y. Xue and Z. Xu, “On the bipolar MMC-HVDC topology suitable for [87] A. Jamshidifar and D. Jovcic, “Small-signal dynamic DQ model of
bulk power overhead line transmission: Configuration, control, and DC modular multilevel converter for system studies,” IEEE Trans. Power
fault analysis,” IEEE Trans. Power Del., vol. 29, no. 6, pp. 2420–2429, Del., vol. 31, no. 1, pp. 191–199, Feb. 2016.
Dec. 2014. [88] V. Najmi, M. N. Nazir, and R. Burgos, “A new modeling approach
[66] J. Zhang and C. Zhao, “The research of SM topology with DC fault for modular multilevel converter (MMC) in D-Q frame,” in Proc.
tolerance in MMC-HVDC,” IEEE Trans. Power Del., vol. 30, no. 3, IEEE Appl. Power Electron. Conf. Expo. (APEC), Mar. 2015,
pp. 1561–1568, Jun. 2015. pp. 2710–2717.
[67] R. Lizana, A. Dekka, S. Rivera, and B. Wu, “Modular multilevel [89] J. Wang, R. Burgos, and D. Boroyevich, “Switching-cycle state-space
converter based on 5-level submodule with DC fault blocking capa- modeling and control of the modular multilevel converter,” IEEE Trans.
bility,” in Proc. 42nd Annu. Conf. IEEE Ind. Electron. Soc. (IECON), Emerg. Sel. Topics Power Electron., vol. 2, no. 4, pp. 1159–1170,
Oct. 2016, pp. 3482–3487. Dec. 2014.
[68] A. Nami, L. Wang, F. Dijkhuizen, and A. Shukla, “Five level cross [90] D. Jovcic and A. A. Jamshidifar, “Phasor model of modular multilevel
connected cell for cascaded converters,” in Proc. 15th Eur. Conf. Power converter with circulating current suppression control,” IEEE Trans.
Electron. Appl. (EPE), Sep. 2013, pp. 1–9. Power Del., vol. 30, no. 4, pp. 1889–1897, Aug. 2015.
[69] J. Qin, M. Saeedifard, A. Rockhill, and R. Zhou, “Hybrid design [91] H. Nademi, A. Das, and L. Norum, “An analytical frequency-domain
of modular multilevel converters for HVDC systems based on var- modeling of a modular multilevel converter,” in Proc. 3rd Power
ious submodule circuits,” IEEE Trans. Power Del., vol. 30, no. 1, Electron. Drive Syst. Technol. (PEDSTC), Feb. 2012, pp. 86–91.
pp. 385–394, Feb. 2015. [92] S. Norrga, L. Ängquist, K. Ilves, L. Harnefors, and H.-P. Nee,
[70] Y. Zhang, G. P. Adam, T. C. Lim, S. J. Finney, and B. W. Williams, “Frequency-domain modeling of modular multilevel converters,” in
“Hybrid multilevel converter: Capacitor voltage balancing limits and its Proc. 38th Annu. Conf. IEEE Ind. Electron. Soc. (IECON), Oct. 2012,
extension,” IEEE Trans Ind. Informat., vol. 9, no. 4, pp. 2063–2073, pp. 4967–4972.
Nov. 2013. [93] J. I. Leon, S. Kouro, L. G. Franquelo, J. Rodríguez, and B. Wu,
[71] X. Li, W. Liu, Q. Song, H. Rao, and S. Xu, “An enhanced MMC “The essential role and the continuous evolution of modulation tech-
topology with DC fault ride-through capability,” in Proc. 39th Annu. niques for voltage-source inverters in the past, present, and future power
Conf. IEEE Ind. Electron. Soc. (IECON), Nov. 2013, pp. 6182–6188. electronics,” IEEE Trans. Ind. Electron., vol. 63, no. 5, pp. 2688–2701,
[72] R. Oliveira and A. Yazdani, “A modular multilevel converter with DC May 2016.
fault handling capability and enhanced efficiency for HVDC system [94] B. P. McGrath and D. G. Holmes, “Multicarrier PWM strategies
applications,” IEEE Trans. Power Electron., vol. 32, no. 1, pp. 11–22, for multilevel inverters,” IEEE Trans. Ind. Electron., vol. 49, no. 4,
Jan. 2017. pp. 858–867, Aug. 2002.
[73] H. Kim, J. Kang, S. Kim, C. K. Kim, and K. Hur, “DC fault protection [95] W. van der Merwe, “Natural balancing of the 2-cell modular multilevel
for modular multilevel converter HVDC using asymmetrical unipolar converter,” IEEE Trans. Ind. Appl., vol. 50, no. 6, pp. 4028–4035,
full-bridge submodule,” in Proc. 9th Int. Conf. Power Electron. ECCE Nov. 2014.
Asia (ICPE-ECCE Asia), Jun. 2015, pp. 1083–1089. [96] G. S. Konstantinou and V. G. Agelidis, “Performance evaluation of
[74] K. Ilves, F. Taffner, S. Norrga, A. Antonopoulos, L. Harnefors, and half-bridge cascaded multilevel converters operated with multicarrier
H.-P. Nee, “A submodule implementation for parallel connection sinusoidal PWM techniques,” in Proc. 4th IEEE Conf. Ind. Electron.
of capacitors in modular multilevel converters,” IEEE Trans. Power Appl., May 2009, pp. 3399–3404.
Electron., vol. 30, no. 7, pp. 3518–3527, Jul. 2015. [97] Z. Li, P. Wang, H. Zhu, Z. Chu, and Y. Li, “An improved pulse
[75] K. Ilves, A. Antonopoulos, S. Norrga, and H.-P. Nee, “Steady-state width modulation method for chopper-cell-based modular multi-
analysis of interaction between harmonic components of arm and level converters,” IEEE Trans. Power Electron., vol. 27, no. 8,
line quantities of modular multilevel converters,” IEEE Trans. Power pp. 3472–3481, Aug. 2012.
Electron., vol. 27, no. 1, pp. 57–68, Jan. 2012. [98] R. Darus, J. Pou, G. Konstantinou, S. Ceballos, and V. G. Agelidis,
[76] E. Solas, G. Abad, J. A. Barrena, A. Cárear, and S. Aurtenetxea, “Circulating current control and evaluation of carrier dispositions in
“Modelling, simulation and control of modular multilevel converter,” modular multilevel converters,” in Proc. IEEE ECCE Asia Downun-
in Proc. 14th Int. Power Electron. Motion Control Conf. (EPE-PEMC), der (ECCE Asia), Jun. 2013, pp. 332–338.
Sep. 2010, pp. T2-90–T2-96. [99] J. Mei, K. Shen, B. Xiao, L. M. Tolbert, and J. Zheng, “A new
[77] M. A. Pérez and J. Rodríguez, “Generalized modeling and simulation selective loop bias mapping phase disposition PWM with dynamic
of a modular multilevel converter,” in Proc. IEEE Int. Symp. Ind. voltage balance capability for modular multilevel converter,” IEEE
Electron., Jun. 2011, pp. 1863–1868. Trans. Ind. Electron., vol. 61, no. 2, pp. 798–807, Feb. 2014.
[78] S. Norrga, L. Ängquist, K. Ilves, L. Harnefors, and H.-P. Nee, “Decou- [100] X. Shi, Z. Wang, L. M. Tolbert, and F. Wang, “A comparison of phase
pled steady-state model of the modular multilevel converter with half- disposition and phase shift PWM strategies for modular multilevel
bridge cells,” in Proc. 6th IET Int. Conf. Power Electron., Mach. converters,” in Proc. IEEE Energy Convers. Congr. Expo. (ECCE),
Drives (PEMD), Mar. 2012, pp. 1–6. Sep. 2013, pp. 4089–4096.
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1653

[101] R. Darus, G. Konstantinou, J. Pou, S. Ceballos, and V. G. Agelidis, [122] M. Hagiwara, R. Maeda, and H. Akagi, “Negative-sequence reactive-
“Comparison of phase-shifted and level-shifted PWM in the power control by a PWM STATCOM based on a modular multilevel
modular multilevel converter,” in Proc. Int. Power Electron. cascade converter (MMCC-SDBC),” IEEE Trans. Ind. Appl., vol. 48,
Conf. (IPEC-Hiroshima-ECCE ASIA), May 2014, pp. 3764–3770. no. 2, pp. 720–729, Mar./Apr. 2012.
[102] B. P. McGrath, C. A. Teixeira, and D. G. Holmes, “Optimised phase [123] G. P. Adam, O. Anaya-Lara, G. M. Burt, D. Telford, B. W. Williams,
disposition (PD) modulation of a modular multilevel converter using and J. R. McDonald, “Modular multilevel inverter: Pulse width modu-
a state machine decoder,” in Proc. IEEE Energy Convers. Congr. lation and capacitor balancing technique,” IET Power Electron., vol. 3,
Expo. (ECCE), Sep. 2015, pp. 6368–6375. no. 5, pp. 702–715, Sep. 2010.
[103] B. McGrath, C. Teixeira, and G. Holmes, “Optimized phase disposi- [124] M. Hagiwara and H. Akagi, “Control and experiment of pulsewidth-
tion (PD) modulation of a modular multilevel converter,” IEEE Trans. modulated modular multilevel converters,” IEEE Trans. Power
Ind. Appl., to be published, doi: 10.1109/TIA.2017.2697953. Electron., vol. 24, no. 7, pp. 1737–1746, Jul. 2009.
[104] A. Hassanpoor, S. Norrga, H.-P. Nee, and L. Ängquist, “Evaluation of [125] S. Du, J. Liu, and T. Liu, “Modulation and closed-loop-based DC
different carrier-based PWM methods for modular multilevel converters capacitor voltage control for MMC with fundamental switching fre-
for HVDC application,” in Proc. 38th Annu. Conf. IEEE Ind. Electron. quency,” IEEE Trans. Power Electron., vol. 30, no. 1, pp. 327–338,
Soc. (IECON), Oct. 2012, pp. 388–393. Jan. 2015.
[105] D. Siemaszko, A. Antonopoulos, K. Ilves, M. Vasiladiotis, L. Angquist, [126] S. Du and J. Liu, “A study on DC voltage control for chopper-
and H.-P. Nee, “Evaluation of control and modulation methods for mod- cell-based modular multilevel converters in D-STATCOM application,”
ular multilevel converters,” in Proc. Int. Power Electron. Conf. (IPEC), IEEE Trans. Power Del., vol. 28, no. 4, pp. 2030–2038, Oct. 2013.
Jun. 2010, pp. 746–753. [127] J. Mei, B. Xiao, K. Shen, L. M. Tolbert, and J. Y. Zheng, “Modular
[106] A. Dekka, B. Wu, and N. R. Zargari, “A novel modulation scheme multilevel inverter with new modulation method and its application to
and voltage balancing algorithm for modular multilevel converter,” in photovoltaic grid-connected generator,” IEEE Trans. Power Electron.,
Proc. IEEE Appl. Power Electron. Conf. Expo. (APEC), Mar. 2015, vol. 28, no. 11, pp. 5063–5073, Nov. 2013.
pp. 1227–1233. [128] D. Siemaszko, “Fast sorting method for balancing capacitor voltages in
[107] Y. Deng, Y. Wang, K. H. Teo, and R. G. Harley, “A simplified space modular multilevel converters,” IEEE Trans. Power Electron., vol. 30,
vector modulation scheme for multilevel converters,” IEEE Trans. no. 1, pp. 463–470, Jan. 2015.
Power Electron., vol. 31, no. 3, pp. 1873–1886, Mar. 2016. [129] A. Dekka, B. Wu, N. R. Zargari, and R. L. Fuentes, “Dynamic
[108] A. Dekka, B. Wu, and N. R. Zargari, “A novel modulation scheme and voltage balancing algorithm for modular multilevel converter: A unique
voltage balancing algorithm for modular multilevel converter,” IEEE solution,” IEEE Trans. Power Electron., vol. 31, no. 2, pp. 952–963,
Trans. Ind. Appl., vol. 52, no. 1, pp. 432–443, Jan. 2016. Feb. 2016.
[109] Y. Deng and R. Harley, “Space-vector versus nearest-level pulse width [130] Y. Zhang, G. P. Adam, T. C. Lim, S. J. Finney, and B. W. Williams,
modulation for multilevel converters,” IEEE Trans. Power Electron., “Analysis of modular multilevel converter capacitor voltage balancing
vol. 30, no. 6, pp. 2962–2974, Jun. 2015. based on phase voltage redundant states,” IET Power Electron., vol. 5,
[110] G. Konstantinou, M. Ciobotaru, and V. G. Agelidis, “Space-vector no. 6, pp. 726–738, Jul. 2012.
versus nearest-level pulse width modulation for multilevel converters,” [131] F. Deng and Z. Chen, “A control method for voltage balancing in
IET Power Electron., vol. 6, no. 1, pp. 96–107, Jan. 2013. modular multilevel converters,” IEEE Trans. Power Electron., vol. 29,
[111] P. M. Meshram and V. B. Borghate, “A simplified nearest level no. 1, pp. 66–76, Jan. 2014.
control (NLC) voltage balancing method for modular multilevel [132] A. Gholizad and M. Farsadi, “A novel state-of-charge balancing method
converter (MMC),” IEEE Trans. Power Electron., vol. 30, no. 1, using improved staircase modulation of multilevel inverters,” IEEE
pp. 450–462, Jan. 2015. Trans. Ind. Electron., vol. 63, no. 10, pp. 6107–6114, Oct. 2016.
[112] G. Konstantinou, J. Pou, S. Ceballos, R. Darus, and V. G. Agelidis, [133] R. Lizana, A. Dekka, B. Wu, and M. Perez, “Control of HVDC
“Switching frequency analysis of staircase-modulated modular multi- transmission system based on MMC with three-level flying capacitor
level converters and equivalent PWM techniques,” IEEE Trans. Power submodule,” in Proc. IEEE Int. Conf. Ind. Technol. (ICIT), Mar. 2015,
Del., vol. 31, no. 1, pp. 28–36, Feb. 2016. pp. 3046–3051.
[113] P. Hu and D. Jiang, “A level-increased nearest level modulation method [134] F. Deng and Z. Chen, “Voltage-balancing method for modular multi-
for modular multilevel converters,” IEEE Trans. Power Electron., level converters switched at grid frequency,” IEEE Trans. Ind. Electron.,
vol. 30, no. 4, pp. 1836–1842, Apr. 2015. vol. 62, no. 5, pp. 2835–2847, May 2015.
[114] Q. Tu and Z. Xu, “Impact of sampling frequency on harmonic distortion [135] K. Ilves, L. Harnefors, S. Norrga, and H.-P. Nee, “Predictive sorting
for modular multilevel converter,” IEEE Trans. Power Del., vol. 26, algorithm for modular multilevel converters minimizing the spread
no. 1, pp. 298–306, Jan. 2011. in the submodule capacitor voltages,” IEEE Trans. Power Electron.,
[115] R. Darus, J. Pou, G. Konstantinou, S. Ceballos, R. Picas, and vol. 30, no. 1, pp. 440–449, Jan. 2015.
V. G. Agelidis, “A modified voltage balancing sorting algorithm for [136] P. Hu, D. Jiang, Y. Zhou, Y. Liang, J. Guo, and Z. Lin, “Energy-
the modular multilevel converter: Evaluation for staircase and phase- balancing control strategy for modular multilevel converters under
disposition PWM,” IEEE Trans. Power Electron., vol. 30, no. 8, submodule fault conditions,” IEEE Trans. Power Electron., vol. 29,
pp. 4119–4127, Aug. 2015. no. 9, pp. 5021–5030, Sep. 2014.
[116] A. Perez-Basante, S. Ceballos, G. Konstantinou, M. Liserre, J. Pou, [137] A. Zama, S. A. Mansour, D. Frey, A. Benchaib, S. Bacha, and
and I. M. de Alegria, “Circulating current control for modular mul- B. Luscan, “A comparative assessment of different balancing control
tilevel converter based on selective harmonic elimination with ultra- algorithms for modular multilevel converter (MMC),” in Proc. 18th Eur.
low switching frequency,” in Proc. 18th Eur. Conf. Power Electron. Conf. Power Electron. Appl. (EPE ECCE Eur.), Sep. 2016, pp. 1–10.
Appl. (EPE ECCE Eur.), Sep. 2016, pp. 1–10. [138] A. E. Leon and S. J. Amodeo, “Energy balancing improvement of
[117] M. Moranchel, E. J. Bueno, F. J. Rodríguez, and I. Sanz, “Selective modular multilevel converters under unbalanced grid conditions,” IEEE
harmonic elimination modulation for medium voltage modular multi- Trans. Power Electron., vol. 32, no. 8, pp. 6628–6637, Aug. 2017.
level converter,” in Proc. IEEE 7th Int. Symp. Power Electron. Distrib. [139] S. Fan, K. Zhang, J. Xiong, and Y. Xue, “An improved control system
Generat. Syst. (PEDG), Jun. 2016, pp. 1–6. for modular multilevel converters with new modulation strategy and
[118] K. Ilves, A. Antonopoulos, S. Norrga, and H.-P. Nee, “A new modula- voltage balancing control,” IEEE Trans. Power Electron., vol. 30, no. 1,
tion method for the modular multilevel converter allowing fundamental pp. 358–371, Jan. 2015.
switching frequency,” IEEE Trans. Power Electron., vol. 27, no. 8, [140] S. Teeuwsen, “Simplified dynamic model of a voltage-sourced con-
pp. 3482–3494, Aug. 2012. verter with modular multilevel converter design,” in Proc. IEEE/PES
[119] M. Saeedifard and R. Iravani, “Dynamic performance of a modular Power Syst. Conf. Expo. (PSCE), Mar. 2009, pp. 1–6.
multilevel back-to-back HVDC system,” IEEE Trans. Power Del., [141] R. Darus, J. Pou, G. Konstantinou, S. Ceballos, and V. G. Agelidis,
vol. 25, no. 4, pp. 2903–2912, Oct. 2010. “Controllers for eliminating the AC components in the circulating
[120] M. Hagiwara, K. Nishimura, and H. Akagi, “A medium-voltage motor current of modular multilevel converters,” IET Power Electron., vol. 9,
drive with a modular multilevel PWM inverter,” IEEE Trans. Power no. 1, pp. 1–8, 2016.
Electron., vol. 25, no. 7, pp. 1786–1799, Jul. 2010. [142] Y. Li and F. Wang, “Arm inductance selection principle for modular
[121] H. P. Mohammadi and M. T. Bina, “A transformerless medium-voltage multilevel converters with circulating current suppressing control,” in
STATCOM topology based on extended modular multilevel converters,” Proc. 28th Annu. IEEE Appl. Power Electron. Conf. Expo. (APEC),
IEEE Trans. Power Electron., vol. 26, no. 5, pp. 1534–1545, May 2011. Mar. 2013, pp. 1321–1325.
1654 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

[143] L. Ängquist, A. Antonopoulos, D. Siemaszko, K. Ilves, M. Vasiladiotis, [163] M. Spichartz, V. Staudt, and A. Steimel, “Analysis of the module-
and H.-P. Nee, “Inner control of modular multilevel converters— voltage fluctuations of the modular multilevel converter at variable
An approach using open-loop estimation of stored energy,” in Proc. speed drive applications,” in Proc. 13th Int. Conf. Optim. Elect.
Int. Power Electron. Conf. (IPEC), Jun. 2010, pp. 1579–1585. Electron. Equip. (OPTIM), May 2012, pp. 751–758.
[144] K. Ilves, A. Antonopoulos, L. Harnefors, S. Norrga, and H.-P. Nee, [164] Y. Okazaki, H. Matsui, M. Hagiwara, and H. Akagi, “Design considera-
“Circulating current control in modular multilevel converters tions on the DC capacitor of each chopper cell in a modular multilevel
with fundamental switching frequency,” in Proc. 7th Int. Power cascade inverters (MMCI-DSCC) for medium-voltage motor drives,”
Electron. Motion Control Conf. (IPEMC), vol. 1. Jun. 2012, in Proc. IEEE Energy Convers. Congr. Expo. (ECCE), Sep. 2014,
pp. 249–256. pp. 3393–3400.
[145] Q. Tu, Z. Xu, and L. Xu, “Reduced switching-frequency modula- [165] J. Kolb, F. Kammerer, and M. Braun, “Dimensioning and design of
tion and circulating current suppression for modular multilevel con- a modular multilevel converter for drive applications,” in Proc. 15th
verters,” IEEE Trans. Power Del., vol. 26, no. 3, pp. 2009–2017, Int. Power Electron. Motion Control Conf. (EPE/PEMC), Sep. 2012,
Jul. 2011. pp. LS1a-1.1-1–LS1a-1.1-8.
[146] B. Bahrani, S. Debnath, and M. Saeedifard, “Circulating current [166] Q. Tu, Z. Xu, Y. Chang, and L. Guan, “Suppressing DC voltage
suppression of the modular multilevel converter in a double-frequency ripples of MMC-HVDC under unbalanced grid conditions,” IEEE
rotating reference frame,” IEEE Trans. Power Electron., vol. 31, no. 1, Trans. Power Del., vol. 27, no. 3, pp. 1332–1338, Jul. 2012.
pp. 783–792, Jan. 2016. [167] K. Ilves, A. Antonopoulos, L. Harnefors, S. Norrga, L. Angquist, and
[147] X. She and A. Huang, “Circulating current control of double-star H.-P. Nee, “Capacitor voltage ripple shaping in modular multilevel
chopper-cell modular multilevel converter for HVDC system,” in Proc. converters allowing for operating region extension,” in Proc. 37th
38th Annu. Conf. IEEE Ind. Electron. Soc. (IECON), Oct. 2012, IEEE-IECON, Melbourne, VIC, Australia, Nov. 2011, pp. 4403–4408.
pp. 1234–1239. [168] J. Pou, S. Ceballos, G. Konstantinou, V. G. Agelidis, R. Picas, and
[148] X. She, A. Huang, X. Ni, and R. Burgos, “AC circulating currents J. Zaragoza, “Circulating current injection methods based on instanta-
suppression in modular multilevel converter,” in Proc. 38th Annu. Conf. neous information for the modular multilevel converter,” IEEE Trans.
IEEE Ind. Electron. Soc. (IECON), Oct. 2012, pp. 191–196. Ind. Electron., vol. 62, no. 2, pp. 777–788, Feb. 2015.
[149] Z. Li, P. Wang, Z. Chu, H. Zhu, Y. Luo, and Y. Li, “An inner current [169] R. Picas, S. Ceballos, J. Pou, J. Zaragoza, G. Konstantinou, and
suppressing method for modular multilevel converters,” IEEE Trans. V. G. Agelidis, “Closed-loop discontinuous modulation technique for
Power Electron., vol. 28, no. 11, pp. 4873–4879, Jan. 2013. capacitor voltage ripples and switching losses reduction in modular
[150] M. Zhang, L. Huang, W. Yao, and Z. Lu, “Circulating harmonic current multilevel converters,” IEEE Trans. Power Electron., vol. 30, no. 9,
elimination of a CPS-PWM-based modular multilevel converter with pp. 4714–4725, Sep. 2015.
a plug-in repetitive controller,” IEEE Trans. Power Electron., vol. 29, [170] M. Huang, J. Zou, and X. Ma, “An improved phase-shifted carrier
no. 4, pp. 2083–2097, Apr. 2014. modulation for modular multilevel converter to suppress the influence
[151] L. He, K. Zhang, J. Xiong, and S. Fan, “A repetitive control scheme of fluctuation of capacitor voltage,” IEEE Trans. Power Electron.,
for harmonic suppression of circulating current in modular multilevel vol. 31, no. 10, pp. 7404–7416, Oct. 2016.
converters,” IEEE Trans. Power Electron., vol. 30, no. 1, pp. 471–481, [171] K. Wang, Y. Li, Z. Zheng, and L. Xu, “Voltage balancing and
Jan. 2015. fluctuation-suppression methods of floating capacitors in a new modular
[152] J.-W. Moon, C.-S. Kim, J.-W. Park, D.-W. Kang, and J.-M. Kim, multilevel converter,” IEEE Trans. Ind. Electron., vol. 60, no. 5,
“Circulating current control in MMC under the unbalanced voltage,” pp. 1943–1954, May 2013.
IEEE Trans. Power Del., vol. 28, no. 3, pp. 1952–1959, Jul. 2013. [172] M. Hagiwara, I. Hasegawa, and H. Akagi, “Start-up and low-speed
[153] Y. Zhou, D. Jiang, J. Guo, P. Hu, and Y. Liang, “Analysis and control operation of an electric motor driven by a modular multilevel cascade
of modular multilevel converters under unbalanced conditions,” IEEE inverter,” IEEE Trans. Ind. Appl., vol. 49, no. 4, pp. 1556–1565,
Trans. Power Del., vol. 28, no. 4, pp. 1986–1995, Oct. 2013. Jul. 2013.
[154] Z. Yuebin, J. Daozhuo, G. Jie, H. Pengfei, and L. Zhiyong, “Control [173] A. J. Korn, M. Winkelnkemper, and P. Steimer, “Low output frequency
of modular multilevel converter based on stationary frame under operation of the modular multi-level converter,” in Proc. IEEE-ECCE,
unbalanced AC system,” in Proc. 3rd Int. Conf. Digit. Manuf. Atlanta, GA, USA, Sep. 2010, pp. 3993–3997.
Autom. (ICDMA), Jul. 2012, pp. 293–296. [174] M. Perez and S. Bernet, “Capacitor voltage ripple minimization
[155] S. Li, X. Wang, Z. Yao, T. Li, and Z. Peng, “Circulating current in modular multilevel converters,” in Proc. IEEE Int. Conf. Ind.
suppressing strategy for MMC-HVDC based on nonideal proportional Technol. (ICIT), Mar. 2015, pp. 3022–3027.
resonant controllers under unbalanced grid conditions,” IEEE Trans.
[175] B. Li, Y. Zhang, G. Wang, W. Sun, D. Xu, and W. Wang,
Power Electron., vol. 30, no. 1, pp. 387–397, Jan. 2015.
“A modified modular multilevel converter with reduced capacitor
[156] J.-W. Moon, J.-W. Park, D.-W. Kang, and J.-M. Kim, “A control voltage fluctuation,” IEEE Trans. Ind. Electron., vol. 62, no. 10,
method of HVDC-modular multilevel converter based on arm current pp. 6108–6119, Oct. 2015.
under the unbalanced voltage condition,” IEEE Trans. Power Del.,
[176] S. Du, B. Wu, N. R. Zargari, and Z. Cheng, “A flying-capacitor modular
vol. 30, no. 2, pp. 529–536, Apr. 2015.
multilevel converter for medium-voltage motor drive,” IEEE Trans.
[157] M. Guan and Z. Xu, “Modeling and control of a modular multi-
Power Electron., vol. 32, no. 3, pp. 2081–2089, Mar. 2017.
level converter-based HVDC system under unbalanced grid condi-
tions,” IEEE Trans. Power. Electron., vol. 27, no. 12, pp. 4858–4867, [177] S. Du, B. Wu, K. Tian, N. R. Zargari, and Z. Cheng, “An active cross-
Dec. 2012. connected modular multilevel converter (AC-MMC) for a medium-
voltage motor drive,” IEEE Trans. Ind. Electron., vol. 63, no. 8,
[158] G. Bergna et al., “An energy-based controller for HVDC modular
pp. 4707–4717, Aug. 2016.
multilevel converter in decoupled double synchronous reference frame
for voltage oscillation reduction,” IEEE Trans. Ind. Electron., vol. 60, [178] J. Rodríguez et al., “State of the art of finite control set model predictive
no. 6, pp. 2360–2371, Jun. 2013. control in power electronics,” IEEE Trans. Ind. Informat., vol. 9, no. 2,
pp. 1003–1016, May 2013.
[159] Y. Liang, J. Liu, T. Zhang, and Q. Yang, “Arm current control strategy
for MMC-HVDC under unbalanced conditions,” IEEE Trans. Power [179] S. Kouro, M. A. Perez, J. Rodríguez, A. M. Llor, and H. A. Young,
Del., vol. 32, no. 1, pp. 125–134, Feb. 2017. “Model predictive control: MPC’s role in the evolution of power
[160] S. Xiaojie, W. Zhiqiang, L. Bo, L. Yiqi, L. M. Tolbert, and F. Wang, electronics,” IEEE Ind. Electron. Mag., vol. 9, no. 4, pp. 8–21,
“Characteristic investigation and control of a modular multilevel Dec. 2015.
converter-based HVDC system under single-line-to-ground fault con- [180] S. Vazquez et al., “Model predictive control: A review of its applica-
ditions,” IEEE Trans. Power Electron., vol. 30, no. 1, pp. 408–421, tions in power electronics,” IEEE Ind. Electron. Mag., vol. 8, no. 1,
Jan. 2015. pp. 16–31, Mar. 2014.
[161] A. Antonopoulos, L. Angquist, S. Norrga, K. Ilves, L. Harnefors, [181] S. Kouro, P. Cortes, R. Vargas, U. Ammann, and J. Rodríguez, “Model
and H.-P. Nee, “Modular multilevel converter AC motor drives with predictive control—A simple and powerful method to control power
constant torque from zero to nominal speed,” IEEE Trans. Ind. Appl., converters,” IEEE Trans. Ind. Electron., vol. 56, no. 6, pp. 1826–1838,
vol. 50, no. 3, pp. 1982–1993, May 2014. Jun. 2009.
[162] S. Debnath, J. Qin, and M. Saeedifard, “Control and stability analysis [182] J. Rodríguez et al., “Predictive current control of a voltage source
of modular multilevel converter under low-frequency operation,” IEEE inverter,” IEEE Trans. Ind. Electron., vol. 54, no. 1, pp. 495–503,
Trans. Ind. Electron., vol. 62, no. 9, pp. 5329–5339, Sep. 2015. Feb. 2007.
DEKKA et al.: EVOLUTION OF TOPOLOGIES, MODELING, CONTROL SCHEMES, AND APPLICATIONS OF MMCS 1655

[183] C. Xia, T. Liu, T. Shi, and Z. Song, “A simplified finite-control- [205] D. Soto and T. C. Green, “A comparison of high-power converter
set model-predictive control for power converters,” IEEE Trans. Ind. topologies for the implementation of FACTS controllers,” IEEE Trans.
Electron., vol. 10, no. 2, pp. 991–1002, May 2014. Ind. Electron., vol. 49, no. 5, pp. 1072–1080, Oct. 2002.
[184] V. Yaramasu and B. Wu, “Predictive control of a three-level boost [206] N. Flourentzou, V. G. Agelidis, and G. D. Demetriades, “VSC-based
converter and an NPC inverter for high-power PMSG-based medium HVDC power transmission systems: An overview,” IEEE Trans. Power
voltage wind energy conversion systems,” IEEE Trans. Power Electron., Electron., vol. 24, no. 3, pp. 592–602, Mar. 2009.
vol. 29, no. 10, pp. 5308–5322, Oct. 2014. [207] L. Xu and V. G. Agelidis, “VSC transmission system using flying
[185] T. Geyer and S. Mastellone, “Model predictive direct torque control capacitor multilevel converters and hybrid PWM control,” IEEE Trans.
of a five-level ANPC converter drive system,” IEEE Trans. Ind. Appl., Power Del., vol. 22, no. 1, pp. 693–702, Jan. 2007.
vol. 48, no. 5, pp. 1565–1575, Sep. 2012. [208] C. M. Franck, “HVDC circuit breakers: A review identifying future
[186] P. Lezana, R. Aguilera, and D. E. Quevedo, “Model predictive con- research needs,” IEEE Trans. Power Del., vol. 26, no. 2, pp. 998–1007,
trol of an asymmetric flying capacitor converter,” IEEE Trans. Ind. Apr. 2011.
Electron., vol. 56, no. 6, pp. 1839–1846, Jun. 2009. [209] G. P. Adam, K. H. Ahmed, S. J. Finney, and B. W. Williams, “Modular
[187] M. A. Perez, P. Cortes, and J. Rodríguez, “Predictive control algorithm multilevel converter for medium-voltage applications,” in Proc. IEEE
technique for multilevel asymmetric cascaded H-bridge inverters,” Int. Electric Mach. Drives Conf. (IEMDC), May 2011, pp. 1013–1018.
IEEE Trans. Ind. Electron., vol. 55, no. 12, pp. 4354–4361, Dec. 2008. [210] M. M. C. Merlin et al., “The alternate arm converter: A new hybrid
[188] P. Cortes, A. Wilson, S. Kouro, J. Rodríguez, and H. Abu-Rub, “Model multilevel converter with DC-fault blocking capability,” IEEE Trans.
predictive control of multilevel cascaded H-bridge inverters,” IEEE Power Del., vol. 29, no. 1, pp. 310–317, Feb. 2014.
Trans. Ind. Electron., vol. 57, no. 8, pp. 2691–2699, Aug. 2010. [211] K. Friedrich, “Modern HVDC plus application of VSC in modular
[189] R. Vargas, J. Rodríguez, C. Rojas, and M. Rivera, “Predictive control multilevel converter topology,” in Proc. IEEE Int. Symp. Ind. Electron.,
of an induction machine fed by a matrix converter with increased Jul. 2010, pp. 3807–3810.
efficiency and reduced common-mode voltage,” IEEE Trans. Energy [212] W. Wang and M. Barnes, “Power flow algorithms for multi-terminal
Convers., vol. 29, no. 2, pp. 473–485, Jun. 2014. VSC-HVDC with droop control,” IEEE Trans. Power Syst., vol. 29,
[190] M. Narimani, B. Wu, V. Yaramasu, and N. R. Zargari, “Finite control- no. 4, pp. 1721–1730, Jul. 2014.
set model predictive control (FCS-MPC) of nested neutral point- [213] Z. Zhu, X. Li, H. Rao, W. Wang, and W. Li, “Testing a complete control
clamped (NNPC) converter,” IEEE Trans. Power Electron., vol. 30, and protection system for multi-terminal MMC HVDC links using
no. 12, pp. 7262–7269, Dec. 2015. hardware-in-the-loop simulation,” in Proc. 40th Annu. Conf. IEEE Ind.
[191] J. Qin and M. Saeedifard, “Predictive control of a modular multilevel Electron. Soc. (IECON), Oct. 2014, pp. 4402–4408.
converter for a back-to-back HVDC system,” IEEE Trans. Power Del., [214] S. Liu, Z. Xu, W. Hua, G. Tang, and Y. Xue, “Electromechanical tran-
vol. 27, no. 3, pp. 1538–1547, Jul. 2012. sient modeling of modular multilevel converter based multi-terminal
[192] J. Bocker, B. Freudenberg, A. The, and S. Dieckerhoff, “Experimental HVDC systems,” IEEE Trans. Power Syst., vol. 29, no. 1, pp. 72–83,
comparison of model predictive control and cascaded control of the Jan. 2014.
modular multilevel converter,” IEEE Trans. Power Electron., vol. 30, [215] X. Li, Z. Yuan, J. Fu, Y. Wang, T. Liu, and Z. Zhu, “Nanao multi-
no. 1, pp. 422–430, Jan. 2015. terminal VSC-HVDC project for integrating large-scale wind genera-
[193] B. S. Riar, T. Geyer, and U. K. Madawala, “Model predictive direct tion,” in Proc. IEEE PES Gen. Meet. Conf. Expo., Jul. 2014, pp. 1–5.
current control of modular multilevel converters: Modeling, analysis, [216] Y. Wang, Z. Yuan, and J. Fu, “A novel strategy on smooth connection
and experimental evaluation,” IEEE Trans. Power Electron., vol. 30, of an offline MMC station into MTDC systems,” IEEE Trans. Power
no. 1, pp. 431–439, Jan. 2015. Del., vol. 31, no. 2, pp. 568–574, Apr. 2016.
[194] M. Vatani, B. Bahrani, M. Saeedifard, and M. Hovd, “Indirect finite [217] Y. Gao, M. Bazargan, L. Xu, and W. Liang, “DC fault analysis of MMC
control set model predictive control of modular multilevel converters,” based HVDC system for large offshore wind farm integration,” in Proc.
IEEE Trans. Smart Grid, vol. 6, no. 3, pp. 1520–1529, May 2015. 2nd IET Renew. Power Generat. Conf. (RPG), Sep. 2013, pp. 1–4.
[195] P. Liu, Y. Wang, W. Cong, and W. Lei, “Grouping-sorting-optimized [218] L. Jun, J. Tianjun, O. Gomis-Bellmunt, J. Ekanayake, and
model predictive control for modular multilevel converter with reduced N. Jenkins, “Operation and control of multiterminal HVDC transmis-
computational load,” IEEE Trans. Power Electron., vol. 31, no. 3, sion for offshore wind farms,” IEEE Trans. Power Del., vol. 26, no. 4,
pp. 1896–1907, Mar. 2016. pp. 2596–2604, Oct. 2011.
[196] Z. Gong, P. Dai, X. Yuan, X. Wu, and G. Guo, “Design and experimen- [219] H. Liu, K. Ma, Z. Qin, P. C. Loh, and F. Blaabjerg, “Lifetime
tal evaluation of fast model predictive control for modular multilevel estimation of MMC for offshore wind power HVDC application,” IEEE
converters,” IEEE Trans. Ind. Electron., vol. 63, no. 6, pp. 3845–3856, Trans. Emerg. Sel. Topics Power Electron., vol. 4, no. 2, pp. 504–511,
Jun. 2016. Jun. 2016.
[197] A. Dekka, B. Wu, V. Yaramsu, and N. R. Zargari, “Dual-stage model [220] A. K. Gupta and A. M. Khambadkone, “A space vector modulation
predictive control of modular multilevel converter,” in Proc. IEEE Int. scheme to reduce common mode voltage for cascaded multilevel
Conf. Ind. Technol. (ICIT), Mar. 2016, pp. 1073–1078. inverters,” IEEE Trans. Power Electron., vol. 22, no. 5, pp. 1672–1681,
[198] J.-W. Moon, J.-S. Gwon, J.-W. Park, D.-W. Kang, and J.-M. Kim, Sep. 2007.
“Model predictive control with a reduced number of considered states [221] M. Winkelnkemper, A. Korn, and P. Steimer, “A modular direct
in a modular multilevel converter for HVDC system,” IEEE Trans. converter for transformerless rail interties,” in Proc. IEEE Int. Symp.
Power Del., vol. 30, no. 2, pp. 608–617, Apr. 2015. Ind. Electron. (ISIE), Jul. 2010, pp. 562–567.
[199] L. Ben-Brahim, A. Gastli, M. Trabelsi, K. A. Ghazi, M. Houchati, [222] F. Ma et al., “A railway traction power conditioner using modular
and H. Abu-Rub, “Modular multilevel converter circulating current multilevel converter and its control strategy for high-speed railway
reduction using model predictive control,” IEEE Trans. Ind. Electron., system,” IEEE Trans. Transport. Electrific., vol. 2, no. 1, pp. 96–109,
vol. 63, no. 6, pp. 3857–3866, Jun. 2016. Mar. 2016.
[200] A. Dekka, B. Wu, and N. R. Zargari, “Minimization of DC-bus [223] H. M. Pirouz and M. T. Bina, “Extended modular multilevel converters
current ripple in modular multilevel converter under unbalanced con- suitable for medium-voltage and large-current STATCOM applica-
ditions,” IEEE Trans. Power Electron., vol. 32, no. 6, pp. 4125–4131, tions,” in Proc. Conf. IPEC, Oct. 2010, pp. 487–492.
Jun. 2017. [224] Z. Liu et al., “The start control strategy design of unified power quality
[201] A. Dekka, B. Wu, V. Yaramasu, and N. R. Zargari, “Dual-stage model conditioner based on modular multilevel converter,” in Proc. IEEE Int.
predictive control with improved harmonic performance for modular Elect. Mach. Drives Conf. (IEMDC), May 2013, pp. 933–937.
multilevel converter,” IEEE Trans. Ind. Electron., vol. 63, no. 10, [225] V. Khadkikar, “Enhancing electric power quality using UPQC:
pp. 6010–6019, Oct. 2016. A comprehensive overview,” IEEE Trans. Power Electron., vol. 27,
[202] R. Adapa, “High-wire act: HVdc technology: The state of the art,” no. 5, pp. 2284–2297, May 2012.
IEEE Power Energy Mag., vol. 10, no. 6, pp. 18–29, Nov. 2012. [226] J. A. Ferreira, “The multilevel modular DC converter,” IEEE Trans.
[203] P. Bresesti, W. L. Kling, R. L. Hendriks, and R. Vailati, “HVDC Power Electron., vol. 28, no. 10, pp. 4460–4465, Oct. 2013.
connection of offshore wind farms to the transmission system,” IEEE [227] S. Du, B. Wu, D. Xu, and N. R. Zargari, “A transformerless bipo-
Trans. Energy Convers., vol. 22, no. 1, pp. 37–43, Mar. 2007. lar multistring DC–DC converter based on series-connected mod-
[204] W. Long and S. Nilsson, “HVDC transmission: Yesterday and today,” ules,” IEEE Trans. Power Electron., vol. 32, no. 2, pp. 1006–1017,
IEEE Power Energy Mag., vol. 5, no. 2, pp. 22–31, Mar. 2007. Feb. 2017.
1656 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 5, NO. 4, DECEMBER 2017

[228] S. Du, B. Wu, K. Tian, D. Xu, and N. R. Zargari, “A novel medium- Ricardo Lizana Fuentes (S’11–M’16) received the
voltage modular multilevel DC–DC converter,” IEEE Trans. Ind. M.Sc. and D.Sc. degrees in electronic engineering
Electron., vol. 63, no. 12, pp. 7939–7949, Dec. 2016. from Universidad Técnica Federico Santa Maria,
[229] M. A. Perez, D. Arancibia, S. Kouro, and J. Rodríguez, “Modular Valparaiso, Chile, in 2011 and 2015, respectively.
multilevel converter with integrated storage for solar photovoltaic appli- He joined the Universidad Católica de la Santisima
cations,” in Proc. 39th Annu. Conf. IEEE Ind. Electron. Soc. (IECON), Concepción, Concepción, Chile, in 2015, where he
Nov. 2013, pp. 6993–6998. is currently an Assistant Professor with the Departa-
[230] B. N. Alajmi, K. H. Ahmed, G. P. Adam, S. J. Finney, and mento de Medio Ambiente y Energia. His current
B. W. Williams, “Modular multilevel inverter with maximum power research interests include high-power converters,
point tracking for grid connected photovoltaic application,” in Proc. HVdc transmission systems, and renewable energy
IEEE Int. Symp. Ind. Electron., Jun. 2011, pp. 2057–2062. systems.
[231] S. Debnath and M. Saeedifard, “A new hybrid modular multilevel Dr. Ricardo was a recipient of the Chilean National Research, Science
converter for grid connection of large wind turbines,” IEEE Trans. and Technology Committee scholarship in 2011, for his Ph.D. studies at
Sustain. Energy, vol. 4, no. 4, pp. 1051–1064, Oct. 2013. Universidad Técnica Federico Santa Maria.
[232] M. Vasiladiotis and A. Rufer, “Analysis and control of modular
multilevel converters with integrated battery energy storage,” IEEE
Trans. Power Electron., vol. 30, no. 1, pp. 163–175, Jan. 2015.
[233] E. M. Farr et al., “Alternate arm converter (AAC) operation under
faulted AC-grid conditions,” in Proc. 7th IET Int. Conf. Power
Electron., Mach. Drives (PEMD), Apr. 2014, pp. 1–6.
[234] F. Kammerer, J. Kolb, and M. Braun, “Fully decoupled current control
and energy balancing of the modular multilevel matrix converter,” in
Proc. 15th Int. Power Electron. Motion Control Conf. (EPE/PEMC),
Sep. 2012, pp. LS2a.3-1–LS2a.3-8. Marcelo Perez (M’07–SM’14) received the Engi-
[235] L. Baruschka and A. Mertens, “A new three-phase AC/AC modular neer degree in electronic engineering, the M.Sc.
multilevel converter with six branches in hexagonal configuration,” degree in electrical engineering, and the D.Sc.
IEEE Trans. Ind. Appl., vol. 49, no. 3, pp. 1400–1410, May 2013. degree in electrical engineering from the University
of Concepción, Concepción, Chile, in 2000, 2003,
and 2006, respectively.
Apparao Dekka (S’10) received the B.Tech. degree From 2006 to 2009, he was a Post-Doctoral
in electrical and electronics engineering from Jawa- Research Fellow with Universidad Técnica Federico
harlal Nehru Technological University, Hyderabad, Santa Maria, Valparaiso, Chile, where he was an
India, in 2007, the M.Tech. degree in electrical Associate Researcher in 2009 and an Assistant
engineering from the Indian Institute of Technology Professor in 2015. He is currently an Associate
Bombay, Mumbai, India, in 2010, and the Ph.D. Researcher with the Solar Energy Research Center, Santiago, Chile, and
degree in electrical and computer engineering from a main Researcher with the Advanced Center for Electrical and Elec-
Ryerson University, Toronto, ON, Canada, in 2017. tronics Engineering, Universidad Técnica Federico Santa Maria. He has
From 2010 to 2012, he was a Research Engineer co-authored two book chapters and more than 90 journal and conference
with the Power Electronics Laboratory, The Petro- papers. His current research interests include power converter topologies and
leum Institute, Abu Dhabi, United Arab Emirates. control, multilevel converters, simulation, electro-mobility, smart grids, and
He is currently a Post-Doctoral Fellow with Ryerson University. He has HVdc systems.
authored more than 25 technical papers, and co-authored a book Modu- Dr. Perez received the Best Paper Award from the IEEE T RANSACTIONS
lar Multilevel Converters: Analysis, Control, and Applications (Wiley-IEEE ON I NDUSTRY A PPLICATIONS M AGAZINE in 2012 and the Second Best Paper
Press, 2018). His current research interests include high-power converters, Award from the IEEE T RANSACTIONS ON P OWER E LECTRONICS in 2015.
pulsewidth modulation schemes, renewable energy systems, and advanced He is an Associate Editor of the IEEE T RANSACTIONS ON P OWER E LEC -
control schemes. TRONICS and the IEEE T RANSACTIONS ON I NDUSTRIAL E LECTRONICS .
Dr. Dekka was a recipient of the Student Research Award from Toronto He is the Coordinator of the IEEE-IES Region-9 Latin-America Chapter and
Hydro and Connect Canada in 2013 and 2015, respectively, and the Research the President of the IEEE-IES Chapter Chile.
Excellence Award from the Department of Electrical and Computer Engineer-
ing, Ryerson University, in 2014, 2015, and 2016, the Outstanding Reviewer
Award in 2015 from the IEEE T RANSACTIONS ON P OWER E LECTRONICS
in 2016, and the Student Paper Travel Award for IEEE-ICIT’16 from the IEEE
Industrial Electronics Society in 2016.

Bin Wu (S’89–M’92–SM’99–F’08) received the

M.A.Sc. and Ph.D. degrees in electrical and
computer engineering from the University of Navid R. Zargari (M’94–SM’08–F’15) received
Toronto, Toronto, ON, Canada, in 1989 and 1993, the B.Eng. degree from Tehran University, Tehran,
respectively. Iran, in 1987, and the M.A.Sc. and Ph.D. degrees
He joined Ryerson University, Toronto, in 1993, from Concordia University, Montreal, QC, Canada,
where he is currently a Professor and the Senior in 1991 and 1995, respectively.
NSERC/Rockwell Automation Industrial Research Since 1994, he has been a Senior Designer and
Chair in power electronics and electric drives. He has then the Manager with the Medium Voltage R&D
authored or co-authored more than 400 technical Department, Rockwell Automation, Cambridge, ON,
papers and four Wiley-IEEE Press books. He holds Canada, where he is currently a Product Architect.
more than 30 granted/pending U.S./European patents in power conversion, For the past 19 years, he has been involved in the
medium voltage drives, and renewable energy systems. simulation, analysis, and design of power converters
Prof. Wu is a fellow of The Engineering Institute of Canada and the for medium voltage ac drives. He has co-authored more than 100 research
Canadian Academy of Engineering. He was a recipient of the Gold Medal of papers and two Wiley-IEEE Press books. He holds 30 U.S. granted/pending
the Governor General of Canada in 1993, the Premiers Research Excellence patents in power converters and medium voltage applications. His current
Award in 2001, the NSERC Synergy Award for Innovation in 2002, the research interests include power converter topologies and their control aspects,
Ryerson Distinguished Scholar Award in 2003, the Ryerson YSGS Outstand- power semiconductors, and renewable energy sources.
ing Contribution to Graduate Education Award, and the Professional Engineers Dr. Zargari was a recipient of the Premiers Award for the Innovator in 2009
Ontario Engineering Excellence Medal in 2014. from the Province of Ontario, Canada.