Zhang 2024 J. Phys. Conf. Ser. 2728 012059
Zhang 2024 J. Phys. Conf. Ser. 2728 012059
Zhang 2024 J. Phys. Conf. Ser. 2728 012059
Series
Abstract. Compared to single-phase dual active bridge (1p-DAB) DC-DC converters, three-
phase dual active bridge (3p-DAB) DC-DC converters have advantages such as high single
machine power, high power density, small filtering capacitance, and are more suitable for high-
voltage situations. Therefore, they have good application prospects in DC transmission and
industrial high-power DC-DC converters. However, under traditional single-phase shift (SPS)
control, especially in the low power range, there is a large reactive current and severe soft switch
loss. This article adopts an asymmetric modulation strategy for its larger reactive power, which
reduces its reactive power in the low power range with more degrees of freedom and improves
the efficiency of power transmission. Taking current stress as the optimization objective, the
KKT algorithm is used to solve, and the corresponding analytical solution is ultimately obtained.
Finally, a simulation is built by using MATLAB software to verify the feasibility of the proposed
solution.
1. Introduction
In recent years, new energy vehicles, DC transmission, and energy storage systems have developed
rapidly. The DC converter, as one of the important devices, has also been well developed. DAB is a
well-known high-efficiency DC-DC converter topology with current isolation, bidirectional flow, and
voltage regulation capabilities. With the deepening of research, DAB realizes the soft switching in a
wide power range, which not only improves the efficiency but also reduces the electromagnetic
interference in the circuit. Compared with 1p-DAB, 3p-DAB has better transmission capacity, more
efficient use of magnetic materials, and smaller filter capacitance, which is more suitable for high-
voltage and high-power scenarios [1-2]. Since 3p-DAB requires a smaller filter capacitor, which means
that the volume of the filter capacitor is reduced, 3p-DAB has a higher power density. So, 3p-DAB is
mainly studied in this paper.
SPS is the simplest control strategy, but in light load, the soft switching is lost seriously and the
power reflux is increased. To improve 3p-DAB efficiency in light load, some scholars have proposed to
parallelize the two phases under light load, transform the three-phase structure into a single-phase full
bridge, and adopt the 1p-DAB control method [3]. This scheme improves the efficiency in the low power
range, but, the input and output voltage ripple are increased. Other scholars proposed a control strategy
of primary and secondary duty ratio variation. This method is similar to the TPS of 1p-DAB, and this
modulation strategy can increase the soft switching range and reduce the rms current, thus improving
efficiency [4]. However, the rms current expression is complicated, and taking rms current as the
optimization goal is difficult to calculate. On the contrary, the current stress expression is simple and
easy to optimize, and it is directly related to the transformer design and the power device choice. Peak
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
ICEEPS-2023 IOP Publishing
Journal of Physics: Conference Series 2728 (2024) 012059 doi:10.1088/1742-6596/2728/1/012059
current is often accompanied by switching off. This means that the smaller peak current can reduce the
switching loss to a certain extent. Therefore, this article takes current stress as the optimization objective.
The rest of this article is organized as follows. The second part briefly introduces the topological
structure and working principle of 3p-DAB, and analyzes the typical working waveform and mode
classification under the control of three degrees of freedom (3-DOF). In Section 3, the expressions of
steady-state voltage, peak inductance current, and transmission power for each mode are calculated. In
Section 4, a current stress optimization scheme is proposed. In Section 5, MATLAB simulation software
is used to verify the effectiveness of the proposed scheme. Finally, the conclusion is drawn in the sixth
part.
P1 P3 P5 + u P1 - u
- +
S1
S1 S3 S5
iLa La
+ A Lb a +
V1 B b V2
- C1 Lc C2 -
C c
N n
P2 P4 P6 S2 S4 S6
According to the transmission power direction P and voltage conversion ratio M =Nus1/up1, the
operation of the converter can be divided into the following four types: 1) Forward/boost (P>0, M>1),
2) Forward/buck (P>0, M<1), 3) Backward/boost (P<0, M>1), 4) Backward/buck (P<0, M<1).
The analysis methods of these four work situations are the same, and the forward/boost situation is
sufficient to demonstrate the analyzing procedures and optimization process.
2
ICEEPS-2023 IOP Publishing
Journal of Physics: Conference Series 2728 (2024) 012059 doi:10.1088/1742-6596/2728/1/012059
P1 D1
P3 D0
P5
S1 D2
S3
S5
uP1
u`S1
iLa
t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10t11 t12 t
P1 P1
P3 P3
P5 P5
uP1 uP1
t t
1 1 1
D1 D1
3 3 2
3
ICEEPS-2023 IOP Publishing
Journal of Physics: Conference Series 2728 (2024) 012059 doi:10.1088/1742-6596/2728/1/012059
In conclusion, working methods can be divided into three categories: Type I (0≤D2≤D1≤1/3): There
are zero level states on both the primary and secondary sides, resulting in no transmission power.
Therefore, this mode is suitable for low power range. Type II (0≤D2≤1/3≤D1≤1/2): The high voltage side
still has a zero-level state, and the low voltage side voltage waveform becomes a four-level step wave.
The maximum output power of this type is greater than that of the first type. Type III (1/3≤D2≤D1≤1/2):
The primary and secondary voltage waveforms are all four-level stepped waves. There is no zero level,
so this type can transmit greater power and is suitable for high-power stages.
The specific voltage waveform for each mode is shown in Figure 4. To simplify the analysis, we can
exclude some modes. For example, in mode d, the voltage waveforms on the primary and secondary
sides are not in a high state at the same time. In this case, energy is first transferred from the primary
side to the inductor, and then from the inductor to the secondary side, which means that a large amount
of energy is stored on the inductor, resulting in increased circuit reactive power and conduction losses.
The efficiency of this mode is greatly reduced, so the optimal mode is not produced in these modes.
Similarly, modes d, i, j, and k are excluded.
3. Steady-state analysis
Due to the large number of modes, only mode a is used as an example to demonstrate the computational
analysis process and make the following assumptions:
1) The converter operates under steady-state conditions and there is no DC bias in the transformer.
2) All switches are ideal, ignoring the effects of parasitic capacitance and dead time.
3) The transformer is an ideal transformer with three-phase symmetry and three equal leakage
inductances, i.e. La=Lb=Lc=L.
4
ICEEPS-2023 IOP Publishing
Journal of Physics: Conference Series 2728 (2024) 012059 doi:10.1088/1742-6596/2728/1/012059
u`S1 u`S1
1 2
uP1
uP1
T/3 2T/3 T t1 t2 t3 t
0 t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 t11 t12 t
Due to three-phase symmetry, only the steady-state characteristics of phase A need to be calculated.
The phase A voltage waveform of mode a is shown in Figure 5, and the current on the inductor can be
calculated based on the voltage at both ends of the inductor:
1 t
L 0
iLa (t ) = iLa (t0 ) + (uP1 ( ) − uS' 1 ( ))d (1)
Combined with the piecewise linear method, the inductor current in one period is integrated as 0, and
the current expression at each time point is as follows:
TV1 ( D1 − D2 M )
TV ( D − D2 M ) iLa (t8 ) = 0
iLa (t0 ) = − iLa (t4 ) = 1 1
3L
iLa (t9 ) = − TV1 (2 D0 + D1 − D2 )
3L
i (t ) = TV1 (2 D0 − D2 + D2 M ) − − + (2)
TV (2 D D D 2 D M ) 6L
La 1 iLa (t5 ) = −
1 0 1 2 2
3L
, 6L ,
TV (2 D0 + D1 + D2 − 2 D2 M )
i (t ) = TV1 (2 D0 + D2 − D2 M ) TV ( 2 D − D + D ) iLa (t10 ) = − 1
6L
La 2 3L iLa (t6 ) = − 1 0 1 2
6L TV1 (2 D0 + D1 + D2 − 2 D2 M )
TV ( D − D2 M )
iLa (t3 ) = 1 1 iLa (t7 ) = 0 iLa (t11 ) = − 6L
3L
The transmission power and current stress are expressed as follows:
1 T 2D0 D2 MTV12 TV (2 D0 − D2 + D2 M )
Pa =
T 0
u P1 (t )iLa (t ) dt =
3L
,I max = iLa (t1 ) = 1
3L
(3)
To simplify analysis and calculation, power and current are standardized as follows, therein
PB=V12T/2πL, IB=V1T/2πL.
3Pa I 2
P' = = 4 D0 D2 M ,I 'max = max = (2 D0 − D2 + D2 M ) (4)
PB IB 3
The above is the expression of power per unit and current per unit of the steady state of mode a.
Similarly, we can obtain the steady-state properties of the remaining modes. It is worth mentioning that
the location of the current stress in mode m is not unique. As shown in Figure 6, when the area of Region
1 is greater than the area of Region 2, the maximum current value is at t3, conversely, when the area of
Region 2 is greater than the area of Region 1, the maximum current value is at t1.
5
ICEEPS-2023 IOP Publishing
Journal of Physics: Conference Series 2728 (2024) 012059 doi:10.1088/1742-6596/2728/1/012059
where P is a unit value of specified power level and gi(D) is composed of the mutual constraints between
the degrees of freedom of each mode. The corresponding Lagrange function is:
n
L = I max
'
( D) + ( P ' ( D) − P) + i gi ( D) (6)
i =1
where λ and μ are the Lagrange multipliers. Then, the KKT conditions are expressed as follows:
I max
'
( P ' ( D) − P) n g
+ + i i = 0( j = 1, 2,3) (7)
D j D j i =1 Dj
( P ( D) − P) = 0, i gi ( D) = 0, 0, i 0
'
According to the different expressions, the full power interval is divided into four sections. The low
power interval runs in mode b, the middle power interval runs in mode m, and the high power interval
runs in mode o. It is worth noting that the mode m is divided into two power intervals. By switching
between these three modes, minimum current stress is achieved in the full power range.
6
ICEEPS-2023 IOP Publishing
Journal of Physics: Conference Series 2728 (2024) 012059 doi:10.1088/1742-6596/2728/1/012059
a b c e f g
h l m n o
1.0
0.8
I`max
0.6
0.4
0.2
5. Experimental result
A 3p-DAB simulation model is built by using MATLAB simulation software. The converter usually has
maximum transmission efficiency when the transformer ratio is 1. According to the power expression,
when other variables are constant, the power is inversely proportional to the auxiliary inductor. The
value is taken according to the rated power and a certain margin is left to meet the full power output.
The simulation parameters are shown in Table 1.
The current stress of 3-DOF control and SPS control is compared under the same transmission power.
The unit power values are taken as 0.05, 0.5, and 1, respectively. The simulation waveform is shown in
Figure 8. When P=0.05, the current stress of the SPS is 23.02 A, while the 3-DOF is 16.71 A. When
P=0.5, the current stress of the SPS is 36.06 A, while the 3-DOF is 30.02 A. When P=1, the current
stress of both is equal to 50.05 A.
500 400
0
uP1 200 uP1
0
-200
-500
500
u`S1
600
400 u`S1
0 200
0
-500 -200
-400
20
Imax=23.02A Imax=16.71A
10
10
0
-10
iLa 0 iLa
-10
-20
8.8 8.85 8.9 8.95 5.88 5.9 5.92 5.94 5.96 5.98 6 6.02 6.04
×10-4 ×10-4
SPS 3-DOF
(a) P=0.05
7
ICEEPS-2023 IOP Publishing
Journal of Physics: Conference Series 2728 (2024) 012059 doi:10.1088/1742-6596/2728/1/012059
500 500
0
uP1 0
uP1
-500 -500
500 u`S1 500 u`S1
0 0
-500 -500
30 Imax=36.06A 30 Imax=30.02A
20 20
10 10
0 0
-10
-20
-30
iLa -10
-20 iLa
-30
9.6 9.65 9.7 9.75 -4 1.29 1.295 1.3 1.305
×10 ×10-3
SPS 3-DOF
(b) P=0.5
500 500
0
uP1 0
uP1
-500 -500
500 u`S1 500 u`S1
0 0
-500 -500
60
40
Imax=50.05A 60
40
Imax=50.05A
20 20
0
-20 iLa 0
-20
iLa
-40 -40
-60 -60 6.5
5.4 5.42 5.44 5.46 5.48 5.5 5.52 5.54 5.56 6.52 6.54 6.56 6.58 6.6 6.62 6.64 6.66
×10-4 ×10-4
SPS 3-DOF
(c) P=1
Figure 8. Simulation waveform under different transmission power.
A SPS
Current stress Optimized
50 RMS Current Optimized
40
Imax
30
20
10
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
P
Figure 9. Comparison of Current Stress between different modulation strategies.
Figure 9 shows the comparison of current stress between the proposed scheme, rms current
optimized control, and SPS at different power unit values. The current stress of SPS is the greatest in
the whole power range. In the low power range, the rms current-optimized control has the same
current stress as the current stress-optimized control. In the medium power range, the current stress
of the current stress optimization scheme is the least, while the current stress of the rms current
optimization scheme is between the two. In the high power range, the three current stresses are the
same. The simulation results achieve the expected goal and effectively reduce the current stress of 3p-
DAB.
6. Conclusion
This paper proposes an optimal modulation strategy based on 3p-DAB topology. This modulation
strategy allows the converter to be operated with minimal inductance current stress, which reduces turn-
off losses and conduction losses, thereby increasing efficiency. Compared with other schemes proposed
before, two new control variables are added to the modulation strategy proposed in this article. Although
8
ICEEPS-2023 IOP Publishing
Journal of Physics: Conference Series 2728 (2024) 012059 doi:10.1088/1742-6596/2728/1/012059
more degrees of freedom increase the mode of operation of the converter and make the analytical
calculation complicated, the TDA is proposed to ensure the analysis is accurate. Each mode is divided
according to the shape of the inverter bridge voltage, and then the steady-state expression is solved by
the volt-second balance of the inductor. Then the optimized solution is obtained according to the KKT
algorithm. Finally, by comparing the current stresses under the three control schemes, the effectiveness
of the proposed optimal modulation strategy is verified.
References
[1] N. H. Baars, J. Everts, C. G. E. Wijnands and E. A. Lomonova. (2016) IEEE Trans. Power Electron.
pp. 6814–6823.
[2] H. van Hoek, M. Neubert and R. De Doncker. (2013) IEEE Trans. Power Electron. pp. 5499–5507.
[3] H. van Hoek. (2017) Ph.D. dissertation, Inst. Power Electron. Elect. Drives, RWTH Aachen Univ.
Aachen.
[4] Z. Li, Y . Wang, L. Shi, J. Huang and W. Lei. (2016) In: Proc. IEEE 8th Int. Power Electron. Motion
Control Conf. pp. 2787–2791.
[5] D. Mouetal. (2021) IEEE Trans. Power Electron. pp. 4572–4584.
[6] D. Mou, Q. Luo, J. Li, Y. Wei and P. Sun. (2021) In: IEEE Transactions on Power Electronics. pp.
10584-10601.