energies

Article
Research on Single-Phase PWM Converter with
Reverse Conducting IGBT Based on Loss Threshold
Desaturation Control
Xianjin Huang * ID
, Dengwei Chang, Chao Ling and Trillion Q. Zheng
School of Electrical Engineering, Beijing Jiaotong University, No. 3 Shangyuancun, Beijing 100044, China;
sanyeh@163.com (D.C.); thesuese@163.com (C.L.); tqzheng@bjtu.edu.cn (T.Q.Z.)
* Correspondence: xjhuang@bjtu.edu.cn; Tel.: +86-136-8358-7910

Received: 15 September 2017; Accepted: 5 November 2017; Published: 12 November 2017

Abstract: In the application of vehicle power supply and distributed power generation, there are
strict requirements for the pulse width modulation (PWM) converter regarding power density and
reliability. When compared with the conventional insulated gate bipolar transistor (IGBT) module,
the Reverse Conducting-Insulated Gate Bipolar Transistor (RC-IGBT) with the same package has a
lower thermal resistance and higher current tolerance. By applying the gate desaturation control,
the reverse recovery loss of the RC-IGBT diode may be reduced. In this paper, a loss threshold
desaturation control method is studied to improve the output characteristics of the single-phase
PWM converter with a low switching frequency. The gate desaturation control characteristics of
the RC-IGBT’s diode are studied. A proper current limit is set to avoid the ineffective infliction of
the desaturation pulse, while the bridge arm current crosses zero. The expectation of optimized
loss decrease is obtained, and the better performance for the RC-IGBTs of the single-phase PWM
converter is achieved through the optimized desaturation pulse distribution. Finally, the improved
predictive current control algorithm that is applied to the PWM converter with RC-IGBTs is simulated,
and is operated and tested on the scaled reduced power platform. The results prove that the gate
desaturation control with the improved predictive current algorithm may effectively improve the
RC-IGBT’s characteristics, and realize the stable output of the PWM converter.

Keywords: RC-IGBT; low switching frequency; PWM converter; diode desaturation;

loss threshold control

1. Introduction
The main feature of the pulse width modulation (PWM) converter is ensuring the high power factor
and AC side sine wave current control on the target of the output DC voltage, as well as on the bidirectional
power transmission. With the development of power semiconductor devices, the improvement of
modeling and control strategy and sensing technology, the working performance of PWM converters is
constantly improving. Electric rail transportation, uninterruptible power system (UPS), and a distributed
power generation that is based on wind and solar energy storage have higher requirements on the
power density, working temperature, cycle life, and reliability. The current research on PWM converters
mainly concentrates on three aspects: new circuit topology, universality of the converter mathematical
model, and control strategy [1–5]. Ref. [6] adopted a hybrid clamped five-level topology to successfully
solve the non-uniformity of the DC capacitor voltage of the three-phase PWM converter. Ref. [7] gave
a mathematical model of a single-phase PWM converter based on the unipolar modulation in the
stationary coordinate and d-q coordinate. Refs. [8,9] proposed the predictive current control to achieve
zero steady-state error of the grid current, and greatly reduce its harmonic content. In addition, due to
the fact that power devices are the cornerstone of the converter system, there have also been researches

Energies 2017, 10, 1845; doi:10.3390/en10111845 www.mdpi.com/journal/energies

Energies 2017, 10, 1845 2 of 17

performed on electrical stress, temperature loss, and reliability of PWM converters. Ref. [10] explained
that the Reverse Conducting-Insulated Gate Bipolar Transistor (RC-IGBT) has a uniform current density
and temperature distribution uniformity in insulated gate bipolar transistor (IGBT) or diode mode, and
greatly improves the reliability of the equipment. Ref. [11] effectively realized the two-way flow of the
point current, and improved the efficiency of the inverter by introducing reverse blocking IGBT (RB-IGBT)
to the T-type three-level inverter.
Due to their strong electrical characteristics and gate control capability, IGBT devices have now
fully occupied 500 Hz–100 kHz high voltage and high power applications. In the past 30 years,
IGBT devices have been constantly updated in terms of structure results, leading to the continuous
improvement in performance. The structures of the devices have undergone a technical revolution
of non-punch through (NPT), field stop (FS), and so on, in which, from the earliest punch through
(PT) type, the surface structure developed from a plane grid to a groove gate structure, and various
new structures and derivative devices emerged in an endless stream [12]. As a result, the resistance
junction temperature has become increasingly higher, the device loss is reducing, the switching speed
is becoming faster, and the voltage level and current capacity are improving significantly, gradually
approaching the material performance limit. Commercial IGBT single modules have reached the
6500 V/1000 A and 4500 V/3000 A levels [13]. Reverse conducting IGBT allows for IGBT chips
and freewheeling diode chips to integrate in the same cellular chip within the internal structure and
function design, which improves the thermal efficiency of the chip and reduces the parasitic parameters
of chip connection. It has a higher current tolerance than that of the traditional IGBT with the same
package. The RC-IGBT integrated diode is controlled by desaturation pulse, which can significantly
optimize the diode reverse recovery characteristic. The reverse conducting IGBT device is applied to
the locomotive traction converter, which has a positive effect on reducing the number of devices and
improving the performance of the temperature rise and heat dissipation [14,15].
In this paper, the driving and loss characteristics of RC-IGBT are introduced, and the improved
predictive control method of RC-IGBT for single-phase PWM rectifier is studied. By setting the
desaturation pulse control sequence of four switch devices in full bridge circuit, the diode desaturation
control is realized to reduce reverse recovery loss. Because the inductor current may be zero crossing
repeatedly and the small current detection may be error in the practical operation of single-phase PWM
rectifier, it can cause the desaturation pulse repeated application, the time overlap, logic confusion,
and other problems at the case of the criterion of the direction of the inductor current. A loss reduction
threshold desaturation control method for RC-IGBT single-phase PWM rectifier is proposed. The threshold
value of the current is determined by the threshold of loss variation in order to judge the application of
the desaturation pulse. The loss of threshold desaturation control of RC-IGBT PWM rectifier is simulated
by simulation, and the system loss is verified by analyzing and comparing the system loss with different
threshold values. Finally, the experiment is carried out on a low power experimental platform, and the
results also verify the correctness of the scheme.

2. Control Characteristics of RC-IGBT PWM Converter

2.1. Driving and Loss Characteristics of RC-IGBT

As shown in Figure 1, RC-IGBT allows the free-wheeling diode (FWD) and IGBT to be successfully
integrated into silicon, by partly replacing the P-Collector (p-doped collector) with N-Collector
(n-doped collector) [16,17].
Energies 2017, 10,
Energies 2017, 10, 1845
1845 33 of
of 17
18
Energies 2017, 10, 1845 3 of 18

(a) (b)
(a) (b)
Figure 1.The diagram of Reverse Conducting-Insulated Gate Bipolar Transistor (RC-IGBT): (a) The
Figure 1. The diagram of Reverse Conducting-Insulated Gate Bipolar Transistor (RC-IGBT): (a) The
Figureinternal
1.The structure;
diagram (b)
ofthe electrical
Reverse symbol.
Conducting-Insulated Gate Bipolar Transistor (RC-IGBT): (a) The
internal structure; (b) the electrical symbol.
internal structure; (b) the electrical symbol.
The operating characteristics of the RC-IGBT integrated diode are affected by the gate voltage,
and the
The application
operating of the gate voltage
characteristics of the can changeintegrated
RC-IGBT the flowingdiodepath and arethe recombination
affected by the gateprocess
voltage,
The operating characteristics of the RC-IGBT integrated diode are affected by the gate voltage,
of the carriers. The gate pulse voltage can affect the forward voltage drop
and the application of the gate voltage can change the flowing path and the recombination process of the integrated diode in
and the application
RC-IGBT. As
of the
shown in
gate voltage
Figure 2, the
cancharacteristic
V-I
change the curves
flowing of
path
the
and the recombination
integrated diode at 25 °C
process
and
of the carriers. The gate pulse voltage can affect the forward voltage drop of the integrated diode
of the125
carriers.
°C workingThe temperatures
gate pulse voltage canforaffect thegate
forward voltage dropin ofthe
thefigure
integrated diode in
in RC-IGBT. As shown in Figureare 2, given
the V-I different
characteristic voltages.
curves of The arrow
the integrated diode shows that
at 25 ◦ C and
RC-IGBT.
the gateAs shown
voltages in
are Figure
−15 V, 0 2,
V, the
and V-I
15 V,characteristic
respectively. curves
The of
greater the
the integrated
gate voltage, diode
the at
higher 25 °C and
the
125 ◦ C working temperatures are given for different gate voltages. The arrow in the figure shows that
125 °C working
forward temperatures
voltage drop across arethegiven
samefor different
current level gate voltages.
will be. The level
In the two arrow in the
PWM figuredriving
control, shows that
the gate voltages
signal
are −15 V, IGBT
of the conventional
0 V, andwas 15
15 V,considering
notV,
respectively. The greater
the working
thethe
state of
gate voltage,
diode;
the higher
otherwise,
the
in the the
the gate voltages are −15 V, 0 V, and respectively. The greater the gate voltage, the higher
forward voltageofdrop
application across
driving the same
RC-IGBT, current
the diode level will
conduction be. should
state In the two level PWM control, driving
forward voltage drop across the same current level will be. In the be twodetected
level PWMto impose the low
control, driving
signallevel
of the conventional
gateconventional IGBT
voltage on theIGBT RC-IGBTwas not considering the working state of the diode; otherwise, in the
signal of the wasfornotthe aim of reducing
considering the diodestate
the working conduction loss [18,19].
of the diode; otherwise, in the
application of driving RC-IGBT, the diode conduction state should be detected to impose the low level
application of driving RC-IGBT,
60
V
the diode conduction state should be detected to impose the low
GE =－15V
gate voltage on the RC-IGBT for the aim of reducing the diode conduction loss [18,19].
level gate voltage on the50RC-IGBT for the aim of reducing the diode conduction loss [18,19].
60 VGE =＋15V
diode current-IC/A

40VGE =－15V
50 VGE =0V
30
VGE =＋15V
diode current-IC/A

40
20
VGE =0V
25℃
30 10 125℃

20 0
0 2 4 6 8 10 12 14 16 18
diode voltage-VCE/V 25℃
10 125℃
Figure 2. Static characteristic of RC-IGBT in the diode mode [20].
0
0 2 4 6 8 10 12 14 16 18
The low level gate voltage applied todiode the diode mode
voltage-V CE/Vof the RC-IGBT could maintain the low

diode conduction voltage drop, while the high level gate voltage could provide additional compound
carriers for the diode reverse
Figure
Figure 2. recovery
2. Static
Static to reduce
characteristic
characteristic of the reverse
of RC-IGBT
RC-IGBT in recovery
in the
the diode current.
diode mode
mode [20].Before the reverse
[20].
recovery process of the diode, a high level pulse is applied to desaturating the saturated diode
temporarily.
The lowlevelThis
level pulse
gate settingapplied
voltage methodto is the
called the RC-IGBT
diode mode diode
ofRC-IGBT desaturation
the RC-IGBT control (RCDC).
The low gate voltage applied to the diode mode of the couldcould maintain
maintain the low the low
diode
The desaturation pulse control is based on several key factors, such as the imposed pulse time, the
diode conduction
conduction voltagevoltage drop,the
drop, while while
highthe high
level level
gate gate voltage
voltage could provide
could provide additionaladditional
compound compound
carriers
amplitude of the pulse voltage, the pulse duration, the pulse turn-off time, and lock time. Ref. [20]
carriers
for the for the
diode dioderecovery
reverse reverse recovery
to reduce tothereduce
reverse therecovery
reverse current.
recovery Before
current. theBefore
reversetherecovery
reverse
studied the relationship between the single RC-IGBT desaturation pulse control and the device loss
recovery
process process of
of the diode,
(including the diode,
a high
IGBT loss, diodelevel a high
loss,pulse level
steadyisstate pulse
applied is applied
to desaturating
loss, dynamic to desaturating
the saturated
loss, and total the saturated
in thisdiode
diode temporarily.
loss). Our research
temporarily.
This This method
pulse setting pulse setting method
is called is called
the RC-IGBT the RC-IGBT
diode diode
desaturation desaturation
control control
(RCDC). The (RCDC).
desaturation
The desaturation
pulse pulse
control is based on control is based
several key on such
factors, several keyimposed
as the factors, pulse
such time,
as thethe
imposed pulse
amplitude time,
of the the
pulse
amplitude of the pulse voltage, the pulse duration, the pulse turn-off time, and lock time. Ref.
voltage, the pulse duration, the pulse turn-off time, and lock time. Ref. [20] studied the relationship [20]
studied the relationship between the single RC-IGBT desaturation pulse control and the device loss
(including IGBT loss, diode loss, steady state loss, dynamic loss, and total loss). Our research in this
Energies 2017, 10, 1845 4 of 17

between the single RC-IGBT desaturation pulse control and the device loss (including IGBT loss,
diodeEnergies
loss,2017, 10, 1845
steady state loss, dynamic loss, and total loss). Our research in this paper is based 4 of 18on the

fact that the proper desaturation pulse control has been selected on the single RC-IGBT. In a switching
paper is based on the fact that the proper desaturation pulse control has been selected on the single
process of RC-IGBT,
RC-IGBT. the desaturation
In a switching pulse control
process of RC-IGBT, can achieve
the desaturation better
pulse loss suppression.
control can achieve better loss
Figure 3b shows the sequential logic of the desaturation pulse U 0
suppression. ge-T1 . The T1 turns off at the
time t0 , the diode
Figure startsthe
3b shows to sequential
be free-wheeling,
logic of thethe order of the
desaturation pulse T2 U′
turning
ge-T1. The on
T1 is execute
turns off at at
thethe timetime t1 ,
the desaturation
t0, the diode startspulse to starts to be applied
be free-wheeling, to theofT1the
the order atTthe time ton
2 turning 2 , it
is stops
execute being
at theapplied
time t 1 , to
thethe T1
desaturation pulse starts to be applied to the T at the time t , it stops being
at the time t3 , and the T2 turns on actually at the time t4 . Because the time of turning on of the T2
1 2 applied to the T1 at the
timebet3,predicted
cannot and the T2in turns on actually
advance, at the
the gate timepulse
level t4. Because
signalthe of time
T2 isof turning and
delayed, on ofthe
thedelay
T2 cannot
timebeis tdelay .
predicted in advance, the gate level pulse signal of T 2 is delayed, and the delay time is tdelay. The tfree is
The tfree is the system processing time, the tdesat is the desaturation pulse width, and the tlock is the dead
the system processing time, the tdesat is the desaturation pulse width, and the tlock is the dead time.
time. Similarly, when the load current IL is reverse (outflow node), it is necessary to detect the forward
Similarly, when the load current IL is reverse (outflow node), it is necessary to detect the forward
conduction time of the T1 , and to also apply the desaturation pulse to the T2 [21,22].
conduction time of the T1, and to also apply the desaturation pulse to the T2 [21,22].
TheThevalues
valuesof of
thethe
above
abovetime
timehave
have the followingrelationships:
the following relationships:

t == ttfree++t tdesat
tdelay
delay free + t+ tlock
desat lock (1)
(1)

tdelaytdelay
is determined
is determinedby
bythe
thethree
three latter values.
latter values.

Figure 3. Desaturation pulse control of RC-IGBT in half bridge: (a) The structure of RC-IGBT in half
Figure 3. Desaturation pulse control of RC-IGBT in half bridge: (a) The structure of RC-IGBT in half
bridge; (b) The logic of desaturation pulse control.
bridge; (b) The logic of desaturation pulse control.

TheThe
losslossof IGBT
of IGBTconverter
converterisismostly obtainedby
mostly obtained byintegrating
integratingthethe voltage
voltage and and current
current of theof the actual
actual
circuit, and and
circuit, thenthen
obtained according
obtained accordingto the V-IV-I
to the curve of of
curve thethe
device
deviceparameter
parametermanual
manual[23].
[23].Figure
Figure44shows
shows the
the current and current
switch and switch
status onstatus on theoperation
the rectify rectify operation of single-phase
of single-phase PWMPWM rectifier.
rectifier. The total
The total loss of all
loss of
the devices all the
in the
Energies 2017, devices
10,half PEER REVIEWperiod could be calculated based on the working status. status.
in the half
x FORmodulation modulation period could be calculated based on the working 5 of 18
The green region corresponding to one IGBT conduction loss and one diode conduction loss, the
red region corresponding to the conduction loss of two diodes, and the yellow region is
s i
corresponding to the conduction loss of two IGBTs. The dashed area corresponding to the loss of the
one IGBT conducting
one diode conducting
IGBT turning off, and the solid line region is corresponding to the IGBT turn-on loss and diode
two diodes conducting
reverse recovery loss. two IGBTs
conducting
IGBT turning off
IGBT turning on
diode reverse recovery

0 t
2Eon_IGBT Eon, Erec
Eoff
2Eon_diode
Eon_IGBT
Eon_diode

Figure 4. Loss analysis of the single phase PWM rectifier during a modulation period.
Figure 4. Loss analysis of the single phase PWM rectifier during a modulation period.

The green region corresponding to one IGBT conduction loss and one diode conduction loss, the
red region corresponding to the conduction loss of two diodes, and the yellow region is
corresponding to the conduction loss of two IGBTs. The dashed area corresponding to the loss of the
IGBT turning off, and the solid line region is corresponding to the IGBT turn-on loss and diode
Energies 2017, 10, 1845 5 of 17

The green region corresponding to one IGBT conduction loss and one diode conduction loss, the red
region corresponding to the conduction loss of two diodes, and the yellow region is corresponding to the
conduction loss of two IGBTs. The dashed area corresponding to the loss of the IGBT turning off, and the
solid line region is corresponding to the IGBT turn-on loss and diode reverse recovery loss.
In the half modulation cycle, IGBT conduction loss, IGBT turn-on loss, IGBT turn-off loss,
diode conduction loss, and diode reverse recovery loss could be written as follows:
m +1
Econ_ IGBT_ half = ∑ {VCE [ IC (t2i−1 )]· IC (t2i−1 )·TT_ 2i−1 } + 2V CE [ IC (t2 )]· I C (t2 )·T T_ 2 (2)
i =1
m
Eon_ half = ∑ {Eon [ IC_ on (t2i )]} + Eon [ IC_ on (t1 )] (3)
i =2
m n h io h i
Eoff _ half = ∑ Eoff IC_ off (t2i−1 ) + Eoff IC_ off (t2 ) (4)
i =2
m +1 m
Eon_ diode_ half = ∑ {VF [ IF1 (t2i−1 )]· IF (t2i−1 )·TD_ 2i−1 } + 2 ∑ {VF [ IF1 (t2i )]· IF (t2i )· TD_ 2i } (5)
i =1 i =2
m
Erec_ half = ∑ {Erec [ IC_ on (t2i )]} + Erec [ IC_ on (t1 )] (6)
i =2

The total loss during a modulation cycle:

 
Etot = 2 Econ_ IGBT_ half + Eon_ half + Eoff _ half + Eon_ Diode_ half + Erec_ half (7)

where IC (ti ) and TT_i are the average current and the conduction time of the IGBT conduction of the
i-th working state; IC_on (ti ) and IC_off (ti ) are the collector currents of IGBT turn-on and turn-off time
of the i-th working state, IF (ti ) and TD_i is the average current and the conduction time of the diode
conduction of the i-th working state, and m is the ratio of the carrier frequency to the modulation
wave frequency.
The RC-IGBT loss calculation is similar to that of the conventional IGBT, including IGBT switching
loss and conduction loss, diode conduction loss, and reverse recovery loss [24]. The calculation of the
IGBT switching loss and conduction loss for RC-IGBT is as the same as that for conventional IGBT,
without considering if the RC-IGBT worked under RCDC or not. The calculation of the diode loss of
RC-IGBT under saturation control is consistent with that of the conventional IGBT diode. Under the
diode desaturation control, the calculation method is similar, but the selected parameters are different
according to the desaturation pulse. The loss difference between the RC-IGBT under the desaturation
control and the RC-IGBT under the saturation control (the conventional IGBT) is analyzed as following.
During a switching period, the difference of the IGBT turn-on loss is defined as ∆Eon , the conducting
loss and the turn-off loss are the same (constant, if the conventional IGBT), the difference of the diode
conduction loss is defined as ∆Eon _diode , and the difference of the diode reverse recovery loss is defined as
∆Erec . Under the two control conditions, the total loss variation of RC-IGBT, ∆Etot can be expressed as:
∆Etot = ∆Eon + ∆Eon_diode + ∆Erec + const (8)

For the convenience of calculation, the duty cycle is set 0.5, the theoretical turn-on time of the diode
is half of the switching period. IGBT turn-on loss actually includes the loss caused by normal turn-on
current and the loss that is caused by the superposed reverse recovery current of the corresponding diode.
Under the two control conditions, the turn-on loss variation of RC-IGBT, ∆Etot can be expressed as:
1 1 trr 2 1  trr t A
∆Eon = βV CE · I RM ·trr − βV CE · IRM · − 2β − β2 VCE · IRM · (9)
2 6 ton 6 ton

where ton is the turn-on time of RC-IGBT, the reverse recovery time is trr = tA + tB , tA , and tB are
corresponding to the rise time and the fall recovery time, IRM is the peak current of the diode reverse
 diode is half of the switching period. IGBT turn-on loss actually includes the loss caused by normal
turn-on current and the loss that is caused by the superposed reverse recovery current of the
corresponding diode. Under the two control conditions, the turn-on loss variation of RC-IGBT, ΔEtot
can be expressed as:
1 1 trr 2 1 trr tA (9)
Energies 2017, 10, 1845 ∆Eon = βVCE ·IRM ·trr － βVCE ·IRM · － 2β － β2 VCE ·IRM · 6 of 17
2 6 ton 6 ton
where ton is the turn-on time of RC-IGBT, the reverse recovery time is trr = tA + tB, tA, and tB are
corresponding to the risecurrent,
recovery, IC is the collector time andVthe fall recovery time, IRM is the peak current of the diode reverse
CE is the DC voltage, and β is the influence coefficient with the
recovery, I C is the collector current, VCE is the DC voltage, and β is the influence coefficient with the
desaturation pulse, which is only related to the peak current of the diode reverse recovery, the rise
desaturation pulse, which is only related to the peak current of the diode reverse recovery, the rise
time and the fall recovery time.
time and the fall recovery time.
Under the two control conditions, the diode reverse recovery loss variation of RC-IGBT, ∆Erec can
Under the two control conditions, the diode reverse recovery loss variation of RC-IGBT, ΔErec
be expressed as:
can be expressed as:
1  1 
∆Erec = 1 βV CE · IRM ·trr 1+ 2β − β2 VCE · IRM ·t A (10) (10)
∆Erec = 6 βVCE ·IRM ·trr ＋ 26 β－ β2 VCE ·IRM ·tA
6 6
The diode conduction loss under the desaturation control is less than that of the saturation
The diode conduction loss under the desaturation control is less than that of the saturation
control. Under the two control conditions, the diode conduction loss variation of RC-IGBT, ∆Erec can
control. Under the two control conditions, the diode conduction loss variation of RC-IGBT, ΔErec can
be expressed as: as:
be expressed
∆Erec = ∆VF · IC ·(tdiode_ on − tdes ) (11)
∆Erec = ∆VF ·IC · tdiode_ on － tdes (11)
∆VF is
wherewhere the difference of the diode conduction voltages under the two control conditions with
ΔVF is the difference of the diode conduction voltages under the two control conditions with
different gate voltages,
different gate tdes tis
voltages, the desaturation pulse width, and tdiode_on
des is the desaturation pulse width, and tdiode_on
is diode
is the the diode conduction
conduction time. time.

2.2. Pulse Timing

2.2. Pulse of PWM
Timing Converter
of PWM with
Converter RC-IGBT
with RC-IGBT
Figure 5 is the
Figure 5 ismain circuit
the main structure
circuit of the
structure single-phase
of the voltage-source
single-phase voltage-sourcePWM PWMconverter.
converter. T11~T
~T4 4 are
full controlled devices,
are full controlled Lm and
devices, Lm R m are,
and respectively,
Rm are, respectively,ananinductance
inductance and andequivalent
equivalent resistance
resistance of the
of the
AC side,
AC side, and C and
d isC d is
a a capacitor
capacitor of of
thethe
DCDC side.
side. InIn addition,
addition, u
us and
s andis iare
s the
are grid
the voltage
grid and
voltage current,
and current,
and uand uab and udc are the AC side voltage and the DC output voltage, respectively [25].
ab and udc are the AC side voltage and the DC output voltage, respectively [25].

+
T1 T3

is Lm Rm
a Cd RL
us uab udc
b
T2 T4

Figure 5. Main circuit structure of single-phase voltage-source pulse width modulation (PWM)
Figure 5. Main circuit structure of single-phase voltage-source pulse width modulation (PWM)
converter with RC-IGBT.
converter with RC-IGBT.

In the implementation of energy exchange, the direction of the current of the DC voltage source
flowing into the DC circuit and the direction of the DC voltage connecting to the AC side, namely
the PWM converter in the converter or inverter working conditions, depending on the states of four
switches in the main circuit [26]. With the traction condition as an example, Table 1 gives eight working
states of the desaturation pulse logic of the single-phase PWM converter. “Conventional pulse”
represents four PWM pulses of a conventional PWM converter, and “desaturation pulse” represents
four pulses of an RC-IGBT PWM converter. For example, “0110” means that the gate voltages of
switches 2 and 3 are high, and the gate voltages of switches 1 and 4 are low. The desaturation pulse
logic of RC-IGBT saturation at each state switching point is also given in the table. For example, “3 to
1” means switching from state 3 to state 1, and the T2 turns on and the T1 is reverse recovery. At the
same time, the T1 is applied with a desaturation pulse.
According to the working mode of the PWM converter with RC-IGBT gate desaturation control,
Figure 6 shows the desaturation control process of the single-phase PWM converter in the traction
condition. The solid line boxes represent the alternating work processes of the rectification working
mode and the power supply short circuit mode, while the dashed boxes represent the alternating work
processes of the inverter working mode and the power short circuit mode. The letters on the arrow
represent the devices, which need to be desaturated in the corresponding working mode. At this point,
Energies 2017, 10, 1845 7 of 17

a desaturation pulse is applied depending on the direction of the inductor current and the state of the
switch on the opposite side. The letters around the digits correspond to devices in the diode mode that
need to be applied to the low drive voltage. The operating states 4 to 5 and the working states 8 to
1 will lead to an inductance current zero crossing. In other words, the full devices will switch from
diode mode to IGBT mode. In addition, the desaturation control process of the PWM converter in the
regenerative condition is similar to that in the traction condition.

Table 1. Timing of desaturation pulse for RC-IGBT in different working states.

us is State Conventional Pulse Desaturation Pulse Conducting Switches

1 0110 0110 T2-IGBT , T3-IGBT
2 0101 0100 T2-IGBT , T4-DIODE
3 1010 0010 T1-DIODE , T3-IGBT
4 1001 0000 T1-DIODE , T4-DIODE
>0
3→1 T2 turned on
T1 desaturated
4→2 T1 reverse recovery
2→1 T3 turned on
T4 desaturated
4→3 T4 reverse recovery
5 1001 1001 T1-IGBT , T4-IGBT
6 1010 1000 T1-IGBT , T3-DIODE
7 0101 0001 T2-DIODE , T4-IGBT
8 0110 0000 T2-DIODE , T3-DIODE
<0
7→5 T1 turned on
T2 desaturated
8→6 T2 reverse recovery
6→5 T4 turned on
Energies 2017, 10, 1845 8→7
T3 desaturated
T3 reverse recovery8 of 18

Figure 6. Process of the desaturation control of single-phase PWM converter in traction.

Figure 6. Process of the desaturation control of single-phase PWM converter in traction.
3. Loss Threshold Desaturation Control of RC-IGBT PWM Converter
3. Loss Threshold Desaturation Control of RC-IGBT PWM Converter
3.1. Improved Predictive Current Control
3.1. Improved Predictive
Based Current Control
on the predictive current control of conventional PWM converter and the criterion of
inductance current, an improved predictive current control for the PWM converter, which is suitable
Based on the predictive current control of conventional PWM converter and the criterion of
for the RC-IGBT is obtained [27–31]. The conventional single-phase PWM converter is controlled by
inductance current, an improved predictive current control for the PWM converter, which is suitable
the predictive current method. In a PWM period, the inductance current satisfies the following
for the equation:
RC-IGBT is obtained [27–31]. The conventional single-phase PWM converter is controlled by the
predictive current method. In a PWM period, the inductance *
current satisfies the following equation:
is (tk + Ts ) = i s (tk ) (12)
is (tk + Ts ) = i∗ s (tk ) (12)
According to the analysis of the circuit topology, the following equation is established:

Lm (i* s − is )
u* ab = us − Rmis − (13)
Ts

I * s = Kp (U* dc − udc )+1/Ti  (U* dc − udc )dt (14)

Ls *
Energies 2017, 10, 1845 8 of 17

According to the analysis of the circuit topology, the following equation is established:
L m (i ∗ s − i s )
u∗ ab = us − Rm is − (13)
Ts
Z
I ∗ s = Kp (U ∗ dc − udc ) + 1/Ti (U ∗ dc − udc )dt (14)
Ls ∗
u ab (t) = us (t) − [ I s sinωt − is (t)] (15)
Ts
The four channels’ voltage generated by the PWM generator is processed by the logic
of desaturation. The relations of the gate voltages are described as Formulas (16) and (17).
uge _d es1 (t)~uge_ des4 (t) are the four gate voltages after desaturation. The ε(t) is the step function. While the
current is positive, ε[is (t)] = 1, and the current is reverse, ε[is (t)] = 0. f delay and f des are the delay function
and the desaturation processing function, respectively.
" # " # " #
u ge_des1 (t) u ge1 (t) u ge1 (t)
= {1 − ε[is (t)]} ∗ f delay + ε[is (t)] ∗ f des (16)
u ge_des4 (t) u ge4 (t) u ge4 (t)
" # " # " #
u ge_des2 (t) u ge2 (t) u ge2 (t)
= ε[is (t)] ∗ f delay + {1 − ε[is (t)]} ∗ f des (17)
u ge_des3
Energies 2017, 10, 1845
(t) u ge3 (t) u ge3 (t) 9 of 18
Energies 2017,to
According 10,the
1845above equations, the following control charts in Figure 7 are obtained:
9 of 18
U*dc uab(t)
I*s
+* - PWM
U dc PI Sin Lm/TS uab(t)
I+
*s
Generator
-+ PI Sin
-
+ PWM
udc + - Lm/TS
Generator
- ε +
uus(t)
dc -
is(t) ε[is(t)]
us(t) PLL ε
is(t) ε[is(t)]
PLL

+ Uge_des1(t)
1-ε[is(t)] fdelay
uge1(t) + UU
ge_des1(t)(t)
ge_des4
1-ε[is(t)] fdelay +
uge1
uge4 (t)(t) + Uge_des4(t)
uge4(t) ε[is(t)] fdes
ε[is(t)] fdes
+ Uge_des2(t)
ε[is(t)] fdelay + UU
ge_des2(t)
uge2(t) ε[is(t)] fdelay ge_des3(t)
+ Uge_des3(t)
uge2
uge3 (t)(t) +
uge3(t) 1-ε[is(t)] fdes
1-ε[is(t)] fdes

Figure 7. Control block diagram of the improved predictive current control.

Figure 7. Control block diagram of
of the
Figure 7. Control block diagram theimproved
improvedpredictive current
predictive control.
current control.
FromFrom thethe
analysis
analysisofofevery
everystate
stateof
ofthe
the PWM rectifier in
PWM rectifier inSection
Section2.2,
2.2,provided
provided that
that thethe direction
direction
From
of the the
inductor
of the analysis
inductor of is
current
currentevery state ofthe
isdetermined,
determined, theworking
the PWM rectifier
working mode of
mode in
ofeach Section
each 2.2,
bridgearm
bridge provided
armcancan
bebe that the
accurately
accurately direction of
judged.
judged.
Figure
Figure8 shows
8 shows the logic
the logic diagram
diagram of
ofthe
the desaturation
desaturation pulse
pulse control
control of
of
the inductor current is determined, the working mode of each bridge arm can be accurately the
the PWM
PWM converter
converter with
with RC-RC-
judged.
IGBT.
IGBT.
Figure 8 shows the logic diagram of the desaturation pulse control of the PWM converter with RC-IGBT.

Figure 8. Logic diagram of the desaturation pulse control of the PWM converter with RC-IGBT.
Figure 8. Logic
Figure diagram
8. Logic ofofthe
diagram thedesaturation pulsecontrol
desaturation pulse controlofof the
the PWM
PWM converter
converter withwith RC-IGBT.
RC-IGBT.
The inputs T1, T2, T3, and T4 in Figure 8 are four conventional input signals with a dead time of
5 μs,
Theand the inductor
inputs , and T4iLin
T1, T2, T3current, , isFigure
also an8input. Theconventional
are four four output signals are outputs
input signals withTa1, dead
T2, T3,time
and of
T . The direction of the input current i is judged firstly. When its direction is forward,
5 μs, and the inductor current, iL, is also an input. The four output signals are outputs T1, T2, T3, and
4 L the pulse
T4. generations
The directionof the T1 and
of the inputT2 are analyzed
current iL is together. Because
judged firstly. the Tits
When 1 has been in is
direction diode mode the
forward, in this
pulse
process, the driving voltage of the T1 should be low. In addition to meeting the needs of the gate of
generations of the T1 and T2 are analyzed together. Because the T1 has been in diode mode in this
the RC-IGBT
process, in diode
the driving modeofthat
voltage the isT1applied
shouldto bethe low
low. Involtage,
additionthetodesaturation
meeting thepulse
needsmust also
of the be of
gate
Energies 2017, 10, 1845 9 of 17

The inputs T1 , T2 , T3 , and T4 in Figure 8 are four conventional input signals with a dead time
of 5 µs, and the inductor current, iL , is also an input. The four output signals are outputs T1 , T2 , T3 ,
and T4 . The direction of the input current iL is judged firstly. When its direction is forward, the pulse
generations of the T1 and T2 are analyzed together. Because the T1 has been in diode mode in this
process, the driving voltage of the T1 should be low. In addition to meeting the needs of the gate of
the RC-IGBT in diode mode that is applied to the low voltage, the desaturation pulse must also be
imposed to reduce the diode reverse recovery loss. Importantly, the rising edge of input T2 should
be detected, the output T1 is obtained by the desaturation pulse procedure of input T2 . At the same
time, the output T2 can also be obtained by delaying the input T2 for the time tdelay . Similarly, the pulse
generations of T3 and T4 are based on the desaturation pulse procedure and the delay of input T4 .
The process of the reverse input current is treated similarly.

3.2. Desaturation Control of Loss Decrement Threshold

In Energies
the improved
2017, 10, 1845
predictive current control strategy with the desaturation pulse 10 of 18
for the
single-phase PWM converter, the value and direction of the inductor current are the key to RCDC
control,andanddirectly
directly affectaffect
the working and loss characteristics
the working of RC-IGBT. When
and loss characteristics compared toWhen
of RC-IGBT. the operation
compared to
at saturation mode, the loss of RC-IGBT in the RCDC control decreases.
the operation at saturation mode, the loss of RC-IGBT in the RCDC control decreases. In a given In a given modulation period,
the performance of the RCDC control is closely related to the loss variation. When considering the
modulation period, the performance of the RCDC control is closely related to the loss variation.
shock of the inductor current error in zero crossing, as well as the error of the sensor detection, the
When considering
RCDC controlthe in shock
the small of the inductor
current current
condition error
will lead to in zero crossing,
a mutation as well
in the time as theoferror
sequence the of the
sensor detection,
desaturationthe RCDC
pulse, which control
in turn,inmaythecause
smallthecurrent
bridge arm condition
to short will lead
circuit, andto a mutation
pulse blockade in andthe time
sequence of theharmonic.
greater desaturation pulse,the
In addition, which in turn,of
loss variation may the cause
RCDC the bridge
control is notarm to short
obvious under circuit,
the small and pulse
blockade current
and condition. In order to reduce
greater harmonic. the faultthe
In addition, risk loss
that isvariation
caused by of thethe
triggering
RCDCofcontrol
the desaturation
is not obvious
pulse, and to utilize the loss suppression of large current conditions, a desaturation control scheme
under the small current condition. In order to reduce the fault risk that is caused by the triggering of
based on the loss decrement threshold per-unit value is proposed. According to the scheme, as
the desaturation pulse, and to utilize the loss suppression of large current conditions, a desaturation
compared to the RC-IGBT saturation control, the loss decrement of RC-IGBT in the RCDC control is
control changed
scheme littlebased on the
near thezero
loss crossing
decrement threshold
of the current. The per-unit value is proposed.
loss decrement changes largely Accordingin the to the
scheme,moment,
as compared thus the toworking
the RC-IGBTcurrent saturation
is great. By control,
setting the the loss decrement
threshold of the lossof RC-IGBT
decrement in the RCDC
per-unit
control value corresponding
is changed little neara small
the current value, the
zero crossing ofRCDC control isThe
the current. executed when the actual
loss decrement changesper-unit largely in
value of the loss decrement is below the threshold, and is not executed
the moment, thus the working current is great. By setting the threshold of the loss decrement per-unit when it exceeds the threshold.
We set the threshold of the loss decrement per-unit value equal to ΔEth*, when the actual per-
value corresponding a small current value, the RCDC control is executed when the actual per-unit
unit value of the loss decrement is less than or equal to ΔEth*, the inductor current is corresponding
value oftothe loss decrement is below the threshold, and is not executed when it exceeds the threshold.
[−ILmax, −ILth] and [ILth, ILmax], and the RCDC control is performed. When the actual per-unit value of
Wethesetloss
thedecrement
thresholdisof the loss
greater thandecrement
ΔEth*, then theper-unit
inductor value
currentequal to ∆Eth *, when
is corresponding to [−Ithe
Lth, actual
ILth] andper-unit
value ofthethe loss control
RCDC decrement is notisperformed.
less thanILmax or equal to ∆Evalue
is the peak th *, the
of inductor
the inductor current
current, is
andcorresponding
I Lth is the to
current threshold corresponding to the loss threshold. Figure 9 shows the
[−ILmax , −ILth ] and [ILth , ILmax ], and the RCDC control is performed. When the actual per-unit value logic block diagram of the
desaturation
of the loss decrement control for thethan
is greater threshold
∆Eth *, ofthen
the loss
thedecrement per-unit is
inductor current value correspondingtoto[−
corresponding the
ILth , ILth ]
inductance current zero-crossing threshold value.
and the RCDC control is not performed. ILmax is the peak value of the inductor current, and ILth is
For the selection of the loss threshold decrement per-unit value, one should refer to the specific
the current threshold
environment corresponding
in which the device is toapplied
the lossand threshold.
the working Figure
current9 shows
flowingthe logicit.block
through Whendiagramthe of
the desaturation control within
device is operating for thethe threshold of the
rated current lossthe
range, decrement
10% peak per-unit
current is value
selectedcorresponding
as the current to the
inductance current
threshold ILth. zero-crossing threshold value.

Figure 9. Logic
Figure block
9. Logic diagram
block diagramofofthe desaturation
the desaturation control
control forthreshold
for the the threshold ofdecrement
of the loss the loss per-
decrement
per-unitunit
value.
value.

4. Simulation and Experiment

For the selection of the loss threshold decrement per-unit value, one should refer to the specific
environment in which
4.1. Simulation the device is applied and the working current flowing through it. When the
Analysis
In order to verify the feasibility of the desaturation pulse control in the PWM converter with
conventional IGBT, the simulation model was built in Matlab Simulink 8.6, R2015B (MathWorks,
Natick, MA, USA). The controller adopted a double closed loop approach that was based on the
improved predictive current control. The Table 2 shows the simulation parameters of the main circuit
originated from the design parameters of the locomotive four quadrant converter of China Railway
Energies 2017, 10, 1845 10 of 17

device is operating within the rated current range, the 10% peak current is selected as the current
threshold ILth .

4. Simulation and Experiment

4.1. Simulation Analysis

In order to verify the feasibility of the desaturation pulse control in the PWM converter with
conventional IGBT, the simulation model was built in Matlab Simulink 8.6, R2015B (MathWorks,
Natick, MA, USA). The controller adopted a double closed loop approach that was based on the
improved predictive current control. The Table 2 shows the simulation parameters of the main circuit
originated from the design parameters of the locomotive four quadrant converter of China Railway
High-speed 2 (CRH2):

Table 2. Design parameters of locomotive four quadrant converter of China Railway High-speed 2 (CRH2).
Energies 2017, 10, 1845 11 of 18

Table 2.Parameter Name of locomotive four quadrant

Design parameters Signal Value
converter of China Railway High-speed 2
(CRH2).
AC side rated voltage UN 1950 V
AC sideParameter
rated frequency
Name fN
Signal Value50 Hz
DC side output
AC side ratedvoltage
voltage UNUd 1950 3500
V V
Operating frequency
AC side rated frequency fN fs 50 Hz450 Hz
ACDCside
sideinductance
output voltage UdLm 35002.195
V mH
AC side resistor
Operating frequency fsRm 450 Hz0.01 Ω
DC side support capacitance
AC side inductance LmCd 2.195 mH5 mF
FilterACinductance
side resistor RmL2 0.011.27
Ω mH
Filter
DC sidecapacitance
support capacitance CdC2 5 mF 2 mF
Filter
Load inductance
resistor L2 R 1.27 mH8.75 Ω
Filter capacitance C2 2 mF
Load resistor R 8.75 Ω
Figure 10 shows the models of the predictive current control of the conventional IGBT PWM
rectifier andFigure 10 shows the
the improved models of
predictive the predictive
current controlcurrent control of the conventional
with desaturation pulses of the IGBT PWM PWM
RC-IGBT
rectifier and the improved predictive current control with desaturation pulses of the RC-IGBT PWM
rectifier under the same simulation parameters.
rectifier under the same simulation parameters.

Figure 10. The

Figure 10.simulation circuit
The simulation of of
circuit thethesingle-phase PWMrectifier
single-phase PWM rectifier under
under saturation
saturation controlcontrol
and and
desaturation control.
desaturation control.

The simulation results are shown in Figure 11. The output voltage udc, the grid voltage us and the
inductor current is are shown under different operations: (a) the conventional IGBT PWM rectifier;
(b) the RC-IGBT PWM rectifier under the desaturation control of the current threshold 10%; and, (c)
the RC-IGBT PWM rectifier under the all period desaturation control. It can be seen from the diagram
that the three control methods can all ensure the DC output stable at 3500 V. The voltage and current
of the grid keep synchronous, using fast Fourier transformation (FFT) analysis, the grid current total
harmonic distortion (THDs) are 2.02% under saturation control, 2.01% under desaturation control of
Energies 2017, 10, 1845 11 of 17

The simulation results are shown in Figure 11. The output voltage udc , the grid voltage us and the
inductor current is are shown under different operations: (a) the conventional IGBT PWM rectifier;
(b) the RC-IGBT PWM rectifier under the desaturation control of the current threshold 10%; and, (c)
the RC-IGBT PWM rectifier under the all period desaturation control. It can be seen from the diagram
that the three control methods can all ensure the DC output stable at 3500 V. The voltage and current
of the grid keep synchronous, using fast Fourier transformation (FFT) analysis, the grid current total
harmonic distortion (THDs) are 2.02% under saturation control, 2.01% under desaturation control of
the current threshold 15%, and 2.01% under all period desaturation control. The system works stably
Energies 2017, 10, 1845 12 of 18
in the rectifier state.

(a)

us/V is/A udc/V

4000
3000
2000
1000
0
－1000
－2000
－3000
0.4 0.42 0.44 0.46 0.48 0.5
t/s

(b) (c)
11. Steady-state
Figure Figure waveforms:
11. Steady-state waveforms:output voltage uu
output voltage , grid
, grid
dcdc voltage
voltage us and
us and the inductor
the inductor current current
is is
under (a) conventional IGBT; (b) all period desaturation control; (c) desaturation control of the current
under (a) conventional IGBT; (b) all period desaturation control; (c) desaturation control of the current
threshold
threshold 10%. 10%.

According
According to thetosimulation,
the simulation, the theaverage
average current
current value
value of of
thethe
device
devicein a switching cycle could
in a switching cycle could
be calculated. Referring to the device parameter manual, the total loss of a modulation cycle under
be calculated. Referring to the device parameter manual, the total loss of a modulation cycle under
rated conditions can be calculated. The following Table 3 shows the loss data of the four conditions:
rated conditions
conventional can be calculated.
IGBT, The following
RC-IGBT saturation Table 3 shows
control, RC-IGBT current the loss data
threshold of the four
desaturation conditions:
control,
conventional IGBT, full
and RC-IGBT RC-IGBT saturationcontrol.
cycle desaturation control,
DueRC-IGBT current
to the different threshold desaturation
characteristics control, and
of the conventional
RC-IGBT fullall
IGBT, cycle
of thedesaturation
partial loss andcontrol.
the total Due
lossto
of the different characteristics
the desaturation controlled RC-IGBT of theareconventional
decreased. IGBT,
The
all of the IGBT loss
partial turn-on
andloss,
thethe diode
total lossconduction loss and reverse
of the desaturation recovery
controlled loss of the
RC-IGBT areRC-IGBT
decreased.under
The IGBT
saturation control are increased. Additionally, there is no difference in the total loss between the
turn-on loss, the diode conduction loss and reverse recovery loss of the RC-IGBT under saturation
desaturation control of the current threshold 10% and the whole cycle desaturation control.
control are increased. Additionally, there is no difference in the total loss between the desaturation
control of the current
Table threshold
3. Datasheet of loss10% and the whole
for conventional cycle
insulated gatedesaturation
bipolar transistor control.
(IGBT) and RC-IGBT
As shownPWM in Figure 12, the partial loss and the total loss of the system are compared under
converters.
different loads.PWM When compared to
Converter the conventional
Econ_IGBT /mJ Eon/mJ IGBT, the total
Eoff/mJ loss/mJ
Eon_diode of RC-IGBT
Erec/mJ single-phase
Etot/mJ PWM
rectifier under the saturation control increases by about 5.2% under three different loads, respectively.
Con-IGBT *1 3520 24,160 15,680 11,950 13,040 68,350
The total lossRC-IGBT(SAT)
of RC-IGBT *2
single-phase3420
PWM 26,880
rectifier15,620
under all12,510 of the period 13,480 71,910
desaturation control
RC-IGBT(RCDC) *3 3420 17,340 15,620 11,510 12,430 60,320
decreases byRC-IGBT(THDC)
11.7% under*4the rated3420 load, 10.8% under15,620
17,500 the half load, 11,590and 12.5% 12,590 under the 1.5 times
60,520
load. In otherNote: words, under
*1 expresses the condition
conventional of a RC-IGBT
IGBT; *2 expresses large load andcontrol
saturation high (SAT);
load *current, the loss of the
3 expresses RC-

RC-IGBT under IGBT alltheperiod

desaturation
desaturationcontrol
control; decreased more significantly.
*4 expresses RC-IGBT Whenof the
desaturation control the current
current threshold
threshold (THDC) 10%.
is different, the loss suppression is different. So it is necessary to adjust the setting according to the
actual workingAs shown in Figure 12, the partial loss and the total loss of the system are compared under
conditions.
different loads. When compared to the conventional IGBT, the total loss of RC-IGBT single-phase
PWM rectifier under the saturation control increases by about 5.2% under three different loads,
respectively. The total loss of RC-IGBT single-phase PWM rectifier under all of the period
desaturation control decreases by 11.7% under the rated load, 10.8% under the half load, and 12.5%
Energies 2017, 10, 1845 12 of 17

Table 3. Datasheet of loss for conventional insulated gate bipolar transistor (IGBT) and RC-IGBT
PWM converters.

PWM Converter Econ_IGBT /mJ Eon /mJ Eoff /mJ Eon_diode /mJ Erec /mJ Etot /mJ
Con-IGBT *1 3520 24,160 15,680 11,950 13,040 68,350
Energies 2017, 10, 18452 3420 26,880 15,620 12,510 13,480 13 of 18
71,910
RC-IGBT(SAT) *
RC-IGBT(RCDC) * 3 3420 17,340 15,620 11,510 12,430 60,320
under the 1.5 times
RC-IGBT(THDC) *4 load. In other words,
3420 under the condition
17,500 15,620 of a11,590
large load and high load current,
12,590 60,520
the loss of the RC-IGBT under the desaturation control decreased more significantly. When the
Note: *1 expresses conventional IGBT; *2 expresses RC-IGBT saturation control (SAT); *3 expresses RC-IGBT all
current
period threshold
desaturation is different,
control; the loss
*4 expresses suppression
RC-IGBT is different.
desaturation controlSo
ofitthe
is current
necessary to adjust
threshold the setting
(THDC) 10%.
according to the actual working conditions.
120,000
Erec 105.3%
Eon_diode 100%
100,000 Eoff 87.7%
87.5%
Eon
80,000 Econ_IGBT 105.2%
100% 88.5%
88.3%
60,000
105.1%
40,000 100% 89.5%
89.2%

20,000

0
*1 *2 *3 *4 *1 *2 *3 *4 *1 *2 *3 *4
0.7MW 1.4MW 2.1MW

Figure 12. Total

Figure 12.loss comparison
Total between
loss comparison conventional
between conventional IGBT
IGBTand
andRC-IGBT
RC-IGBTat at different loads.(Note:
different loads. (Note: PWM
Converters under four type 1 2 3 4
PWM Converters underconditions * , * , * and
four type conditions * , * ,** in
1 2 3 andFigure 12 are
* in Figure
4 12corresponding
are corresponding to to
those
thoseofofTable 3).
Table 3).
By changing the current threshold, the single-phase PWM converter is simulated under different
load currents,By changing
and thethe current threshold,
average loss in the thepositive
single-phase
halfPWM converter
cycle is simulated
is calculated. under
Table differentthat the
4 shows
load currents, and the average loss in the positive half cycle is calculated. Table 4 shows that the load
load current IL was, respectively, 0.2, 0.5, 0.8, and 1.5 IN , the percentage of the current threshold
current IL was, respectively, 0.2, 0.5, 0.8, and 1.5 IN, the percentage of the current threshold IPU (the
IPU (theratio
ratio
of of
thethe current
current threshold
threshold to the
to the peak peak was
current) current)
equal was equal
to zero to zero
(complete (complete
judgment judgment
on zero-
on zero-crossing),
crossing), 0.1 0.1 to 10.8,
to 0.8, 1 (saturation
(saturation control),control), theloss
the average average lossper-unit
decrement decrement per-unit
value in the half value
cycle in the
half cycle is ∆E*
is ΔE* (theofratio
(the ratio of the difference
the difference of the
of the average lossaverage loss
between the between
saturation the saturation
control control and
and desaturation
control to saturation control). The IN is 500 A.
desaturation control to saturation control). The IN is 500 A.
Table 4. Average loss decrement per-unit value in half cycle of single-phase PWM converter under
Table 4. Average loss decrement per-unit value in half cycle of single-phase PWM converter under
different current thresholds and different load currents.
different current thresholds and different load currents.
IL\IPU 0 10% 20% 30% 40% 50% 60% 70% 80% 1
0.2 IN 0.1400 0.1317 0.1080 0.0894 0.0867 0.0834 0.0795 0.0603 0.0516 0
IL \IPU 0 10% 20% 30% 40% 50% 60% 70% 80% 1
0.5 IN 0.1513 0.1437 0.1319 0.1205 0.1177 0.1119 0.1040 0.0764 0.0608 0
0.2 IN 0.8 IN0.1400
0.15450.1317
0.1506 0.1080
0.1459 0.0894
0.1343 0.0867
0.1211 0.0834
0.1212 0.0795
0.1155 0.06030.0625
0.0820 00.0516 0
0.5 IN IN 0.1513 0.16060.1437
0.1544 0.1319
0.1493 0.1205
0.1384 0.1177
0.1332 0.1119
0.1238 0.1040
0.1207 0.07640.0681
0.0868 00.0608 0
0.8 IN 1.5 IN0.1545
0.16900.1506
0.1619 0.1459
0.1515 0.1343
0.1396 0.1211
0.1338 0.1212
0.1261 0.1155
0.1233 0.08200.0669
0.0859 00.0625 0
IN 0.1606 0.1544 0.1493 0.1384 0.1332 0.1238 0.1207 0.0868 0.0681 0
1.5 IN By0.1690
using the0.1619
MATLAB0.1515 0.1396
three spline 0.1338 to set
interpolation 0.1261 0.1233 relation
up the binary 0.0859among 0.0669
IL, IPU and 0
ΔE*, the three-diensional (3D) graphs of which are shown in Figure 13. When the IPU is less than 0.1,
the ΔE* does not change significantly; when the IPU is higher than 0.1, with the decrease of IPU, the
By using the MATLAB three spline interpolation to set up the binary relation among IL , IPU and
ΔE* increases significantly. Taking the rated load as an example, when the IPU is 0.1, the maximum
∆E*, thevalue
three-diensional
0.16 of ΔE* is (3D) graphsWhen
obtained. of which are shown
considering in Figureand
repeatability 13. When the IPUlow
small current is less
lossthan 0.1,
the ∆E*characteristics
does not changenear significantly;
the zero currentwhen
of thethe
loadIPU is higher
current, than 0.1,
the optimal with
graph ofthe
the decrease of IPUof, the ∆E*
loss decrement
increasesthesignificantly.
PWM converter Taking the namely
is drawn, rated load as an
IPU 10%, theexample,
curve of ΔE*when
andthe IPU
IL. In is 0.1, the
addition, maximum
Figure 13 also value
0.16 of ∆E*
shows the curve ofWhen
is obtained. ΔE* and IPU when ILrepeatability
considering is IN. The intersection
and small point of thelow
current two loss
curves is the best
characteristics near
working point of RC-IGBT single-phase PWM rectifier under rated conditions.
the zero current of the load current, the optimal graph of the loss decrement of the PWM converter is
drawn, namely IPU 10%, the curve of ∆E* and IL . In addition, Figure 13 also shows the curve of ∆E*
and IPU when IL is IN . The intersection point of the two curves is the best working point of RC-IGBT
single-phase PWM rectifier under rated conditions.
Energies 2017, 10, 1845 13 of 17
Energies 2017, 10, 1845 14 of 18

Energies 2017, 10, 1845 14 of 18

Figure 13. Surface graph of the variety of average loss decrement per unit value with load current and
Figure 13. Surface graph of the variety of average loss decrement per unit value with load current and
current threshold percentage.
current threshold percentage.
4.2. Experimental Results
Figure 13. Surface graph of the variety of average loss decrement per unit value with load current and
4.2. Experimental
In
Results
current
orderthreshold percentage.
to further verify the design method above, the experimental study is carried out. The
grid voltage
In order is reduced
to further verifybythe
thedesign
voltagemethod
regulatorabove,
to 195 V thefrom 220 V. The experimental
experimental study is carried parameters
out. The grid
4.2. Experimental Results
are Us = 195 V, fN = 50 Hz, Ud* = 350 V, and switching frequency fS = 1 kHz. The PWM converter model
voltage is reduced by the voltage regulator to 195 V from 220 V. The experimental parameters are
is made up ofto
In order a half bridge
further switching
verify devices
the design for Infineon
method above, the FZ1000R65KR3, and another
experimental study is carriedhalfout.
bridge
The
Us = 195for
V, fvoltage
gridInfineon
Hz, Ud * =by
N = 50FZ750R65KE3
is reduced
350 V,The
[32]. andinductance
the voltage
switching frequency
regulatorLmtois195
12.8VmHfrom
fS220
and =the
1V.kHz. The are
capacitors PWM twoconverter
The experimental parallel andmodel is
parameters
made up areofUseries
two as =half bridge
V, fN = 50 switching
195capacitors, namely
Hz, 350devices
Ud* =470 μF/450 forThe
V. Infineon
V, and switching digital FZ1000R65KR3,
signal
frequency fSprocessing and
(DSP)
= 1 kHz. The PWM inanother half
the controller
converter bridge
model is for
InfineonTexas
isFZ750R65KE3
made Instruments
up of a half [32].
(TI) Theswitching
inductance
TMS320LF28335
bridge Lm isfor
(Texas
devices 12.8 mH and
Instruments
Infineon (TI), the capacitors
Dallas,
FZ1000R65KR3, TX, USA), are twohalf
to mainly
and another parallel
bridgeand two
achieve
the improved
for Infineon
series capacitors, current470
FZ750R65KE3
namely predictive
[32]. The
µF/450 control
Thealgorithm,
inductance
V. andmH
Lm issignal
digital 12.8 the complex
and programmable
the capacitors
processing (DSP) are logic
twocontroller
in the device
parallel andis Texas
(CPLD)
two series is ALTERA’s
capacitors, EPM7128AETC100-7
namely 470 μF/450 V. (Altera,
The digital Silicon
signal Valley,
processing
Instruments (TI) TMS320LF28335 (Texas Instruments (TI), Dallas, TX, USA), to mainly achieve CA, USA)
(DSP) in to
the realize
controller the
is the
desaturation
Texas Instruments control algorithm.
(TI) Figure 14(Texas
TMS320LF28335 showsInstruments
the test platform.
(TI), Dallas, TX, USA), to mainly achieve
improved current predictive control algorithm, and the complex programmable logic device (CPLD) is
the improved current predictive control algorithm, and the complex programmable logic device
ALTERA’s(CPLD)EPM7128AETC100-7 (Altera, Silicon(Altera,
is ALTERA’s EPM7128AETC100-7 Valley, Silicon
CA, USA) to realize
Valley, CA, USA)the desaturation
to realize the control
algorithm. Figure 14 shows the test platform.
desaturation control algorithm. Figure 14 shows the test platform.

Figure 14. Test platform for RC-IGBT PWM convertor.

The steady state experimental waveform PWM converter is under saturation control and all
period desaturation control. As shown
Figure 14. Testin Figure for
platform 15, RC-IGBT
the outputPWMvoltage udc under conventional control
convertor.
Figure 14. Test platform for RC-IGBT PWM convertor.
is about 350 V, which is 1.79 times the grid voltage, and the ripple voltage is small. Although the
switching frequency
The steady stateisexperimental
low, the AC side currentPWM
waveform waveform is stillisclose
converter under to saturation
the sinusoidal waveform,
control and all
The steady
with
period
state
a power experimental
factor
desaturation equal toAs
control. Atwaveform
1. shownthe in
same PWM
time,
Figure 15,under
converter
saturation
the output voltage
iscontrol,
under saturation
the
udc under
control
output voltage
conventional and and all
control
period desaturation
output 350 V,control.
is aboutcurrent which isAs
waveform 1.79shown
of the
timesPWMin Figure
the 15, the
converter
grid output
areand
voltage, almost voltage
the
the ripple same u under
asdcthose
voltage is conventional
under
small. the control
saturation
Although
is aboutcontrol.
350 V,Through
switching which is
the1.79
frequency grid
is times
low,side theside
thecurrent
AC grid voltage,
waveform
current dataand the ripple
are analyzed
waveform is still voltage
by to
close thethe
MATLAB is small. Although the
FFTwaveform,
sinusoidal function,
with
switching a power factor
frequency equal
is low, thetoAC1. Atside
the current
same time, under saturation
waveform is stillcontrol,
close tothetheoutput voltage waveform,
sinusoidal and
output current waveform of the PWM converter are almost the same as
with a power factor equal to 1. At the same time, under saturation control, the output voltage and those under saturation
control. Through the grid side current waveform data are analyzed by the MATLAB FFT function,
output current waveform of the PWM converter are almost the same as those under saturation control.
Through the grid side current waveform data are analyzed by the MATLAB FFT function, the THD of
the gird side current is 4.76% under saturation control. Those of desaturation pulse width 50 µs are
Energies 2017,
Energies 2017, 10,
10, 1845
1845 15 of
15 of 18
18
Energies 2017,
Energies 2017, 10,
10, 1845
1845 15 of
14 of 17
18

the THD
the THD of the
the gird side
side current isis 4.76% under
under saturation control.
control. Those of
of desaturation pulse
pulse width
the THD of of the gird
gird side current
current is 4.76%
4.76% under saturation
saturation control. Those
Those of desaturation
desaturation pulsewidth
width
50
50 μs are
are 4.60% under all period desaturation
desaturation control. The results prove that the
the desaturation control
50 μs
4.60%μs are 4.60%
under
4.60%all under
under all
period period
desaturation
all control.
control. The
period desaturation resultsThe
control. results
prove
The thatprove
results that
that the desaturation
the desaturation
prove control
control is effective
desaturation control
is
is effective and feasible.
feasible.
is effective
and and
feasible.
effective and feasible.

(a)
(a) (b)
(b)
(a) (b)
Figure
Figure 15. Steady-state
Steady-stateexperimental
experimental waveforms:
waveforms: the
thethe grid
gridgrid voltage
voltage uuusss,,, output
us , output output
voltagevoltage
udc anduuugrid and grid
current
Figure 15.
Figure dc and
15. Steady-state
Steady-state experimental
experimental waveforms:
waveforms: the grid voltage
voltage output voltage
voltage dc
dc and grid
grid
current i , (a)
is , (a) saturation
current s saturation control, THD
control, THD =THD = 4.76%; (b)
4.76%; (b) all(b) all period
periodperiod desaturation
desaturation control, control, t
tdesat =tdesat
desat = 50
50 µs, tμs, μs, = 3 ==µs,
t lock 33
current iiss,, (a)
(a) saturation
saturation control,
control, THD == 4.76%;
4.76%; (b) all
all period desaturation
desaturation control,control, tdesat == 50 μs, ttlock
50 lock lock = 3
μs,
THD
μs, THD =
= 4.60%. 4.60%.
μs, THD
THD == 4.60%.
4.60%.

Figure 16
Figure 16 compares the the gate voltage waveforms of of T 1 and T T2 under saturation
saturation control
under saturation and all
control and
Figure 16 compares
compares the gate gate voltage
voltage waveforms
waveforms of TT111 and
and T222 under
under saturation control
control and allall
period
period desaturation
desaturation
desaturation control.
control.
control. It
It
It can
can
can be
be
be seen
seen
seen from
from
from Figure
Figure
Figure 16b
16b
16b that
that
that T1T
T isis desaturated
desaturated and
and thethe gate
gate pulse
pulse
1 is desaturated and the gate pulse
1 of
period desaturation control. It can be seen from Figure 16b that T1 is desaturated and the gate pulse
of
T
of T
is is unchanged
unchanged under under desaturation
desaturation control
control when when
the the
grid grid
current current
is is
positive.positive.
When theWhen
grid the grid
current is
of2 TT22 is
is unchanged
unchanged underunder desaturation
desaturation control
control when
when the
the grid
grid current
current isis positive.
positive. When
When the the grid
2
grid
current
reverse,is
current isT reverse, T2 is desaturated
is desaturated desaturated
and the gateandpulse
the gate
gate
of Tpulse of T
T1 is unchanged.
1 is unchanged. unchanged.
current is2reverse,
reverse, TT22 is
is desaturated and
and the
the gate pulse
pulse of
of T11 is
is unchanged.

uuuge-T1
ge-T1
ge-T1
(10V/div)
ge-T1
(10V/div)
uge-T1
(10V/div)(10V/div)
ge-T1
ge-T2 uu
(10V/div)
uge-T2

iiiss
(5A/div)(10V/div)
isiiss uuge-T2

s
(5A/div)
(5A/div)

uuuge-T2
ge-T2
ge-T2
ttt(4ms/div)
(4ms/div)
(4ms/div)
(a)
(a) (b)
(b)
(a) (b)
Figure 16.
Figure
Figure
Steady-state experimental
16. Steady-state experimental waveforms:
waveforms: the
the gate voltagesuuuge
gate voltages ge of the
of the T
the T and T
T111 and
and T andandthe
T222and
the inductor
inductor
Figure16.
16.
current,
Steady-state
(a)Steady-state
saturation
experimental
experimental
control; (b) all
waveforms:
waveforms:
period
the
thegate
gatevoltages
desaturation voltagestugegeof
control, of=the
50 T the inductor
1 and T2 and the inductor
µs, t = 3 µs.
current, (a)
current, (a) saturation control;
control; (b) all
all period desaturation
desaturation control, ttdesat
desat == 50
50 μs,
μs, ttlock == 33 μs.
lock
lock μs.
current, (a) saturation
saturation control; (b)
(b) all period
period desaturation control,
control, tdesat
desat = 50 μs, tlock = 3 μs.

Figure 17
Figure 17
17isisan
an
anenlargement
enlargementofof the positive
the half
positive period
half period of the
the current
of the in Figure
current Figure 17b. 17b.
in Figure The gate
gate
The level
gate
Figure 17is is anenlargement
enlargementof ofthe
thepositive
positivehalf
halfperiod
periodof of thecurrent
currentin
in Figure17b.17b.TheThe gatelevel
level
of the
level
of T has a short
of the Ta1 short pulse before
has a short before gate voltage
pulse beforevoltage of the
gate voltage T is high
of thehigh every
T2 isevery time, i.e.,
high everyi.e., the desaturation
time, i.e., pulse.
the desaturation
of the
the TT11 has
has a short pulse
pulse before gate
gate voltage ofof the
the TT22 is every time,
time, i.e., the
the desaturation
desaturation pulse.
1 2
is high pulse.
Figure 17a shows
pulse. Figure
Figure shows that the
17a shows the desaturation
that pulse width
the desaturation width of 50
pulse width 50 μs,
μs, and
of 50 Figure 17b shows
and Figure
µs,Figure shows the interlock
17b shows interlock time
the interlock
Figure 17a shows that the desaturation pulse width of 50 μs, and Figure 17b shows the interlock time
17a that desaturation pulse of and 17b the time
of
of 3
time
3 μs.
of
μs. Therefore,
3 µs.
Therefore, the
Therefore,
the experimental
the
experimental parameters
experimental
parameters are
are completely
parameters
completelyare consistent
completely
consistent with those
consistent
with those in
with
in the
the theoretical
those in the
theoretical
of 3 μs. Therefore, the experimental parameters are completely consistent with those in the theoretical
design.
theoretical design.
design.
design.

(a)
(a) (b)
(b)
(a) (b)
Figure
Figure 17. Parameters
Parameters under
underdesaturation
desaturationcontrol:
control: (a) desaturation pulse width 50 (b)
μs;lock
(b) lock
lock time
time
Figure 17.
Figure Parameters under
17. Parameters under desaturation
desaturation control:(a)(a)
control: desaturation
desaturation
(a) pulse
pulse
desaturation width
width
pulse 50
50 μs;
50 µs;
width μs; (b)
(b) time
lock 3time
µs.
3 μs.
33 μs.
μs.
Energies 2017, 10, 1845 15 of 17
Energies 2017, 10, 1845 16 of 18

Figure
Figure 18 shows the
18 shows the gate
gate voltage
voltage waveforms
waveforms of of T
T1 and T2 under the condition of (a) the current
1 and T2 under the condition of (a) the current
threshold
threshold == 22 A
A corresponding
corresponding to to aa loss
loss decrement
decrement threshold
threshold per-unit
per-unit value
value of of 0.13,
0.13, (b)
(b) the current
the current
threshold = 10 A corresponding to a loss decrement threshold per-unit value of 0.08,
threshold = 10 A corresponding to a loss decrement threshold per-unit value of 0.08, and (c) the current and (c) the
current
threshold = 13 A corresponding to a loss decrement threshold per-unit value of 0.05. The greaterThe
threshold = 13 A corresponding to a loss decrement threshold per-unit value of 0.05. the
greater
current the currentis,threshold
threshold is, then the
then the shorter the desaturation
shorter the desaturation times
times of T1 and T2ofwill
T1 and T2 will
be. The PWMbe. converters
The PWM
converters
can all workcan all work
steadily understeadily undercurrent
the different the different current
thresholds. thresholds.
As the thresholdAs the threshold
current increases, current
the loss
increases,
decrement threshold per-unit value gradually decreases. Figure 18a is the waveform of the18a
the loss decrement threshold per-unit value gradually decreases. Figure is the
condition
waveform
that has theofbest
the loss
condition that has the best loss characteristics.
characteristics.

(a) (b)

(c)
Figure 18.
Figure Steady-stateexperimental
18. Steady-state experimental waveforms
waveforms under
under different
different lossloss thresholds:
thresholds: (a) Ith(a)
= 2IthA,=ΔE*
2 A,
=
∆E* = 0.13; (b) I = 10 A, ∆E* = 0.08; (c) I
th ΔE* = 0.08; (c) Ith = 13 A,
0.13; (b) Ith = 10 A, = 13 A, ∆E*
th ΔE* = 0.05. = 0.05.

5. Conclusions
5. Conclusions
Due to
Due to the
the fact
fact that
that it
it has
has its
its own
own low
low thermal
thermal resistance
resistance andand high
high current
current density,
density, the
the RC-IGBT
RC-IGBT
satisfied the high capacity single-phase PWM converter requirements, namely
satisfied the high capacity single-phase PWM converter requirements, namely high power density high power density and
reliability. In this paper, the gate desaturation control of the diode integrated in the
and reliability. In this paper, the gate desaturation control of the diode integrated in the RC-IGBT is RC-IGBT is studied
and the desaturation
studied control ofcontrol
and the desaturation switching devices ofdevices
of switching the fullofbridge circuit
the full bridgeis realized.
circuit isThe improved
realized. The
predictive current control of the PWM converter is studied. The loss decrement
improved predictive current control of the PWM converter is studied. The loss decrement threshold threshold desaturation
control method
desaturation applied
control to theapplied
method RC-IGBT to single-phase
the RC-IGBTPWM converter
single-phase PWMis proposed.
converterInisorder to avoid
proposed. In
the desaturation pulse chaos in the current zero crossing, by setting the loss decrement
order to avoid the desaturation pulse chaos in the current zero crossing, by setting the loss decrement threshold to
limit the current threshold corresponding to the desaturation pulse applied to
threshold to limit the current threshold corresponding to the desaturation pulse applied to RC-IGBT, RC-IGBT, the operating
performance
the operatingofperformance
the PWM converter in lowconverter
of the PWM switchinginfrequency is improved.
low switching frequencyFinally, the modeling
is improved. and
Finally,
simulation of the PWM converter of the RC-IGBT are completed,
the modeling and simulation of the PWM converter of the RC-IGBT are completed, and theand the simulation results confirm
the effectiveness
simulation results ofconfirm
the desaturation control. Experiments
the effectiveness are carried
of the desaturation out on
control. a small power
Experiments are experiment
carried out
platform
on a smalltopower
verify experiment
the effectiveness of the
platform to control strategy.
verify the effectiveness of the control strategy.

Acknowledgments: In
Acknowledgments: In this
this paper,
paper,thethesamples
samplesand
andtest data
test areare
data recorded with
recorded thethe
with strong support
strong of Toshiba
support and
of Toshiba
Infineon Beijing Co. We wish to express our heartfelt thanks to them. The research is also supported by “the
and Infineon Beijing
Fundamental Co.Funds
Research We wish
for to
theexpress
Centralour heartfelt(2017JBM061)”.
University thanks to them. The research is also supported by “the
Fundamental Research Funds for the Central University (2017JBM061)”.

Author Contributions: Xianjin HUANG and Trillion Q. Zheng conceived and designed the experiments;
Dengwei Chang and Chao Ling performed the experiments; Xianjin HUANG analyzed the data; Trillion Q.
Zheng contributed materials and analysis tools; Xianjin HUANG wrote the paper.
Energies 2017, 10, 1845 16 of 17

Author Contributions: Xianjin Huang and Trillion Q. Zheng conceived and designed the experiments; Dengwei
Chang and Chao Ling performed the experiments; Xianjin Huang analyzed the data; Trillion Q. Zheng contributed
materials and analysis tools; Xianjin Huang wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Wen, J.P. Research on the Control Strategies of Single-Phase Three-Level Pwm Rectifier; Beijing Jiaotong University:
Beijing, China, 2012; pp. 1–96.
2. Song, K.J. Research and Implementation of Three Phase Current Source Rectifier; Nanjing University of Aeronautics
and Astronautics: Nanjing, China, 2014; pp. 1–99.
3. Liu, B. Constant-Frequency-Hysteresis Control Strategy for Single Phase PWM Rectifiers; Southwest Jiaotong
University: Chengdu, China, 2016; pp. 1–75.
4. Zhang, Y.C.; Peng, Y.B.; Qu, C.Q. Predictive current control of PWM rectifier under unbalanced grid voltage
condition. Trans. China Electrotech. Soc. 2016, 31, 88–94.
5. Song, W.S.; Deng, Z.X. Model predictive power control scheme for single-phase PWM rectifiers with constant
switching frequency. Electr. Mach. Control 2016, 20, 93–100.
6. Yin, K.; Ma, M.Y.; Cao, M.H.; He, Y.M.; Deng, Y.; Hen, X.N. Research on three-phase hybrid-clamped
five-level PWM rectifier. Proc. CSEE 2012, 32, 59–66.
7. Huang, H. The Research of Single-Phase PWM Rectifier and Control Strategy; Huazhong University of
Science & Technology: Wuhan, China, 2012; pp. 1–66.
8. Yang, L.Y.; Yang, S.; Zhang, W.P.; Chen, Z.G.; Long, X.U. An improved deadbeat predictive current control
method for single-phase PWM rectifiers. Proc. CSEE 2015, 35, 5842–5850.
9. Deng, Z.X.; Song, W.S.; Cao, M.H. A model predictive current control scheme for single-phase PWM rectifiers.
Proc. CSEE 2016, 36, 2992–3004.
10. Wei-Zhong, C.; Ze-Hong, L.; Bo, Z.; Min, R.; Jin-Ping, Z.; Yong, L.; Zhao-Ji, L. A snapback suppressed
reverse-conducting IGBT with uniform temperature distribution. Chin. Phys. B 2014, 23, 513–518.
11. Qiu, H. T-Type Three-Level Inverter with Reverse Blocking IGBT; Beijing Jiaotong University: Beijing, China,
2015; pp. 1–80.
12. Li, X.P.; Liu, J.; Cao, G.; Gao, M.C.; Wang, Y.H.; Jin, R.; Wen, J.L. Development and application of reverse
conducting-IGBT in smart grid. Smart Grid 2017, 5, 1–8.
13. Mitsubishi Datasheet: CM1000HG-130XA IGBT Module Data Sheet; Mitsubishi Electric Corporation:
Tokyo, Japan, 2017.
14. Huang, X.J.; Ling, C.; Sun, H.; You, X.J.; Xiao, C.J. Working characteristics of Reverse-Conducting IGBT
applied in vehicle converters. Semiconduct. Technol. 2017, 5, 363–370.
15. Fock-Sui-Too, J.L.; Chauchat, B.; Nicolau, S.; Madrid, F.; Austin, P.; Tounsi, P.; Mermet-Guyennet, M.
Experimental study of power IGBT technologies at large range temperature. In Proceedings of the
International Conference on Mixed Design of Integrated Circuits and Systems IET, Poznan, Poland,
19–21 June 2008; pp. 519–524.
16. Chen, W.Z. Theoretical Model and Novel Structures for RC-IGBT; University of Electronic Science and Technology
of China: Chengdu, China, 2014; pp. 1–94.
17. Zhu, L.H. Study on Novel Reverse-Conducting IGBT; University of Electronic Science and Technology of China:
Chengdu, China, 2014; pp. 1–101.
18. Werber, D.; Pfirsch, F.; Gutt, T.; Komarnitskyy, V.; Schaeffer, C.; Hunger, T.; Domes, D. 6.5 kV RCDC: For increased
power density in IGBT-modules. In Proceedings of the IEEE International Symposium on Power Semiconductor
Devices & IC’s, Waikoloa, HI, USA, 15–19 June 2014; pp. 35–38.
19. Werber, D.; Hunger, T.; Wissen, M.; Schuetze, T. A 1000 A 6.5 kV power module enabled by reverse-conducting
trench-IGBT-technology. In Proceedings of the International Exhibition and Conference for Power Electronics,
Intelligent Motion, Renewable Energy and Energy Management (PCIM), Shanghai, China, 28–30 June 2016;
pp. 351–358.
20. Huang, X.J.; Ling, C.; Chang, D.W.; You, X.J.; Zheng, T.Q. Loss characteristics of 6.5kV RC-IGBT applied to a
traction converter. Energies 2017, 10, 891. [CrossRef]
Energies 2017, 10, 1845 17 of 17

21. Eckel, H.; Wigger, D. Comparison of the output power of RC-IGBT and IGBT/diode inverters. In Proceedings
of the Preliminary Conference Program PCIM Europe 2011, Nuremberg, Germany, 17–19 May 2011; pp. 38–43.
22. Hermann, R.; Krafft, E.U.; Marz, A. Reverse-conducting-IGBTs—A new IGBT technology setting new
benchmarks in traction converters. In Proceedings of the IEEE European Conference on Power Electronics
and Applications, Lille, France, 2–6 September 2013; pp. 1–8.
23. Huang, X.J.; You, X.J.; Zheng, Q.L. Traction converter with desaturated diode control for reverse conducting
IGBT. Trans. China Electrotech. Soc. 2016, 31 (Suppl. 2), 58–66.
24. Zhang, Z.R.; Xu, Z.; Xue, Y.L. Valve losses evaluation based on piecewise analytical method for MMC–HVDC
links. IEEE Trans. Power Deliv. 2014, 29, 1354–1362. [CrossRef]
25. Liu, Z.G. Power Electronics; Tsinghua University Press; Beijing University Press: Beijing, China, 2004; pp. 182–191.
26. Du, X.H. Modeling and Characteristic Research of High Power IGBTs Applied in Traction; Beijing Jiaotong
University: Beijing, China, 2016; pp. 1–75.
27. Zhang, Y.; Liu, J. An improved model-free predictive current control of PWM rectifiers. In Proceedings of
the 2017 20th International Conference on Electrical Machines and Systems (ICEMS), Sydney, Australia,
11–14 August 2017; pp. 1–5.
28. Kwak, S.; Kim, S.E.; Park, J.C. Predictive current control methods with reduced current errors and ripples for
single-phase voltage source inverters. IEEE Trans. Ind. Inform. 2015, 5, 1006–1016. [CrossRef]
29. Luo, Y.; Liu, C.; Yu, F. Predictive current control of a new three-phase voltage source inverter with phase
shift compensation. IET Electr. Power Appl. 2017, 5, 740–748. [CrossRef]
30. Garcia, C.F.; Rivera, M.E.; Rodríguez, J.R.; Wheeler, P.W.; Peña, R.S. Predictive current control with instantaneous
reactive power minimization for a four-Leg indirect matrix converter. IEEE Trans. Ind. Electron. 2017, 2, 922–929.
[CrossRef]
31. Liu, Y.C.; Ge, X.; Tang, Q.; Deng, Z.; Gou, B. Model predictive current control for four-switch three-phase
rectifiers in balanced grids. Electron. Lett. 2017, 1, 44–46. [CrossRef]
32. Dütemeyer, T.; Schütze, T. Mitsubishi Datasheet: FZ750R65KE3 IGBT Module Data Sheet; Infineon Corporation:
Munich, Germany, 2009.

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).