You are on page 1of 9

BIDIRECTIONAL ISOLATED ZVS DC-DC CONVERTER WITH NONPULSATING INPUT & OUTPUT CURRENT

Felix Jauch, Juergen Biela Laboratory for High Power Electronic Systems ETH Zurich, Physikstrasse 3, CH-8092 Zurich, Switzerland Email: jauchf@ethz.ch

This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of ETH Zrichs products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to pubs-permission@ieee.org. By choosing to view this document you agree to all provisions of the copyright laws protecting it.

BIDIRECTIONAL ISOLATED ZVS DC-DC CONVERTER WITH NONPULSATING INPUT & OUTPUT CURRENT
Felix Jauch, Juergen Biela
Laboratory for High Power Electronic Systems ETH Zurich, Physikstrasse 3, CH-8092 Zurich, Switzerland Email: jauchf@ethz.ch

Abstract - A new bidirectional isolated Zero-VoltageSwitching (ZVS) DC-DC converter with nonpulsating input and output current is proposed. This converter is ideally suited for battery charging application and is applied as a submodule in a battery storage system based on a modular multilevel converter (M2 C). The current output of the proposed DC-DC converter enables small current ripple for directly interfacing a battery without the need of additional lter components. The derivation of topology, the operating principle, the analysis of the ZVS area as well as simulation results of a 12.4 kW DC-DC converter and a M2 C submodule are provided. Keywords DC-DC Converter, Nonpulsating Input/Output Current, Zero-Voltage-Switching, Modular Multilevel Converter, Battery Storage I. I NTRODUCTION With increasing use of renewable energy sources like wind power and solar energy, energy storage systems become an important part of future energy distribution systems due to the inherently uctuating and stochastic nature of renewable energy sources. The storage systems are able to compensate for energy shortages during low wind conditions or at nighttime. In order to store considerable amounts of energy and ensure balancing supply and demand, suitable high power electronic equipment interfacing the storage media play an important role. Besides the well known energy storage systems using the potential energy of water, in recent years additional concepts based on compressed air, ywheels, thermal energy storage in molten salt [1] as well as chemical storage in batteries have been investigated. The battery based systems offer the advantage of high energy and power density, high cycle efciency as well as being location-independent and easily scalable. Such systems also become more important in grid applications for load leveling providing fast frequency regulation [2]. At high power levels the battery energy storage system is usually connected to the medium voltage distribution grid. Because of the relatively high voltages, multilevel converter systems are advantageous to use due to lower harmonics, robust operation and reduced switching losses. The multilevel converters [3], and especially the modular multilevel converters (M2 C) [4], additionally offer the benet to easily split the energy storage into smaller modules as proposed in [5, 6], and have the benet to be highly modular and fault tolerant sys-

Phase c Phase b Phase a

Storage Battery
+ -

DC DC

EV Battery
+ -

EV Battery
+ -

Ma,1 Submodule

Storage Battery

DC DC

Ma,n Submodule

+ -

Cdc

La1

MV AC Side Storage Battery


+ -

La2

DC DC

Ma,n+1 Submodule

Storage Battery

DC DC

Ma,2n Submodule

+ -

(a)
S1 D1
+ C1

L1 V1

DC

S2

D2

V2
DC Submodule

(b)
D1 S1 L1 iL1 Cr1 + C1 vC1 L1 iL vp Lm V1 S2 D2 N1 : N2 vC3 L2 vs D4 S4 L2 C4 Cr4
+

C3

D3 S3

Cr3

iL2 V2

Cr2 + C2 vC2

vC4

Submodule

(c)

Fig. 1. M2 C based battery storage system consisting of 2n submodules per phase leg and subsequent DC-DC converters to charge electric vehicles (a), the conventional half-bridge submodule topology with subsequent isolated DC-DC converter (b) and the proposed bidirectional isolated ZVS DC-DC converter with nonpulsating input and output current for use as a submodule (c).

tems. With the different submodules a balancing of the batteries is also possible [7, 8]. Fig. 1(a) shows a M2 C where the submodules additionally include an isolated DC-DC converter as shown in Fig. 1(b). Due to the galvanic isolation, the submodule outputs of the three phases a, b and c can be connected in parallel in order to obtain a continuous power ow from the grid to each battery. Additionally, with this parallel connection an interleaving of the three converters is possible, which reduces the current ripple and enables a signicant size reduction of the passive components. As shown by the dashed lines in Fig. 1(a), non-isolated DCDC converters can be connected to the batteries of the submodules and the outputs of these converters can be paralleled in order to perform a high-power charging of electric vehicles. Due to the high charging power an ultra-fast charging of electric vehicles in less than 5 minutes is possible. This charging concept is investigated in the project Ultra-Fast Charging of Electric Vehicles [6]. Usually, half-bridge input stages as shown in Fig. 1(b) are used in the submodule of the M2 C for interfacing the grid. A subsequent isolated DC-DC converter is utilized to perform galvanic isolation, voltage adaptation and charging control. Using for example a dual-active-bridge converter [9] as an isolated DC-DC converter results in a total number of 10 switching devices per submodule. Furthermore, for charging the batteries a low current ripple is required [10] resulting in additional effort for ltering the output current or increasing the output capacitor of the DC-DC converter. Therefore, in this paper a new bidirectional isolated DCDC converter with nonpulsating input and output current as shown in Fig. 1(c) is proposed. The converter is based on an isolated Cuk converter but features soft-switching for all four switching devices. It can be used as a submodule in a M2 C based battery storage system but also as single DC-DC converter for charging batteries. A typical application can be in interconnecting a high voltage DC bus with the 12 V battery in a (hybrid-)electric vehicle. In the following, rst the novel topology is derived in section II and then, the operating principle in grid-to-battery and battery-to-grid operation as well as the ZVS operating area are analyzed in section III. Finally, section IV shows a concept of a 250 kW / 2 MWh battery storage system and simulation results of a 12.4 kW DC-DC converter submodule, which is part of the energy storage system. II. D ERIVATION OF T OPOLOGY The commonly used half-bridge submodule in a M2 C as presented in [4] can be connected to a DC-DC converter to achieve galvanic isolation and voltage adaptation as shown in Fig. 1(b). Nevertheless, for interfacing a battery additional lter components are required to ensure a small current ripple. A possible implementation of an isolated DC-DC converter is a dual-half-bridge [11] or a dual-active-bridge [9] with a high frequency isolation transformer and a subsequent output lter to limit the current ripple. However, the main disadvantage is the high number of switching devices (6 or 10), an increased control complexity, higher losses and reduced reliability.

D1 V1 S1

a1 Cr1 C1
+

b1 Lm b2 S2

L1

L1

N1 : N2

L2

C2 D2 Cr2 L2

V2

a2

Fig. 2. Topology of the basic bidirectional isolated Cuk converter. The bidirectional isolated Cuk converter shown in Fig. 2 basically offers nonpulsating input and output currents and achieves galvanic isolation with only two switching devices [12]. In both power ow directions only one switching device is actively turned on/off during operation. However, the isolated Cuk converter is a hard-switching converter, which results in switching losses, and has the disadvantage that the energy stored in the transformer leakage inductance must be absorbed by the output capacitance of the switching device, resulting in voltage ringing and higher voltage stress [13]. To avoid the voltage ringing across the main switching device and achieve soft-switching at the same time, an auxiliary switch and a clamp capacitor can be connected in parallel to the switching device [14] or the transformer primary side [15]. There, the clamp capacitor absorbs the energy stored in the transformer leakage inductance and the main as well as the auxiliary switching devices are then turned on/off under ZVS conditions. However, these topologies only allow unidirectional power ow. By inserting the active clamp circuit consisting of an auxiliary switch and a clamping capacitor on both the transformer primary (connection a1 a2 in Fig. 2) and secondary side (connection b1 b2 in Fig. 2) the new topology shown in Fig. 1(c) results, whose primary circuit has been presented in [16, 17]. The proposed DC-DC converter features nonpulsating input and output current, bidirectional power ow, galvanic isolation and ZVS of all four switching devices over a wide load range. In the following section, the operating principle of the converter is analyzed in detail. III. O PERATING P RINCIPLE For simplication of circuit analysis, the magnetizing inductance Lm of the transformer is neglected in the following. Thus, the primary and secondary transformer leakage inductance can be summed up to L = L1 + L2 , where L2 is referred to the primary side. The switching devices S1 /S2 and S3 /S4 in the two half-bridge legs are inversely controlled. Hence, there are 4 switching states. One of the capacitors C1 /C2 as well as one of the capacitors C3 /C4 form a resonant circuit together with the transformer leakage inductance L which depends on the switching state of the converter. The switching frequency is chosen to be well above the resonant frequency of this resonant circuit. This paper focuses on controlling the two half-bridge legs with a duty cycle of 50%. However, also duty cycles different than 50% are possible. The phase-shift between the squarewave voltages applied on primary and secondary side of the transformer determines the output power

N1 V1 V2 d(1 2d) P = N2 fs L

(1)

turns on at t4 . Then, switch S4 can be turned on at nearly zero voltage. 5) Mode 5 (t4 < t < t5 ): The input current iL1 and output current iL2 are linearly decreasing. Depending on the output voltage V2 referred to the primary side in comparison to the input voltage V1 , current iL increases (V2 < V1 ), remains constant (V2 = V1 ) or decreases (V2 > V1 ). 6) Mode 6 (t5 < t < t6 ): Switch S1 is turned off at t5 and capacitances Cr1 /Cr2 provide ZVS conditions. Cr1 /Cr2 is charged/discharged by the current driven by L and inductor L1 until Cr2 is completely discharged and diode D2 starts to conduct at t6 . Now, switch S2 can be turned on at nearly zero voltage. The input current iL1 reaches its minimum at (t6 t5 )/2 and starts to increase linearly again. 7) Mode 7 (t6 < t < t7 ): The current iL decreases linearly and at t7 , iL = iL1 , so that diode D2 stops conducting and the current commutates to S2 after t7 . 8) Mode 8 (t7 < t < t8 ): The current iL1 is increasing linearly and iL2 as well as iL are decreasing linearly. Furthermore, iL changes its direction. 9) Mode 9 (t8 < t < t9 ): At t8 , the current through diode D4 reaches zero, so that D4 turns off and switch S4 starts conducting. 10) Mode 10 (t9 < t < t10 ): At t9 , switch S4 is turned off and capacitances Cr3 /Cr4 provide ZVS conditions. Cr3 /Cr4 is discharged/charged by the current driven by L and L2 . Capacitor Cr3 is discharged until diode D3 turns on, so that switch S3 can be turned on at nearly zero voltage. The output current iL2 reaches its minimum at (t10 t9 )/2 and starts increasing linearly.

where V1 is the input voltage, V2 the output voltage, N1 /N2 the transformer turns ratio, d the relative phase-shift between the two square-wave voltages on primary and secondary side of the transformer and fs the switching frequency. At a xed output voltage, the converter behaves like a phase-shift controlled current source with average output current I2 = N1 V1 d(1 2d) . N2 fs L (2)

For a relative phase-shift of d = 0.25 the output power reaches its maximum at Pmax = N1 V1 V2 . N2 8fs L (3)

The control strategy leads to capacitor voltages vC 1 /vC 2 equal to the input voltage V1 and capacitor voltages vC 3 /vC 4 equal to the output voltage V2 . Therefore, the voltage stress on the primary switching devices is double the input voltage and double the output voltage on the secondary switching devices. The voltage stress can be reduced by controlling the primary half-bridge with a duty cycle above 50% and the secondary half-bridge with a duty cycle below 50% as will be discussed in a future paper. A. Grid-to-Battery Operation In grid-to-battery operation, the square-wave voltage applied to the transformer primary side leads the square-wave voltage applied to the transformer secondary side similar to a phase-shift controlled dual-active-bridge. The operation over a whole switching period Ts can be described using 12 modes which are shown in Fig. 4 with the corresponding key waveforms in Fig. 3. At time t0 , diodes D1 and D3 are conducting. Switches S1 and S3 are turned on, S2 as well as S4 are turned off. The voltage vC 1 /vC 2 across the capacitors C1 /C2 equals V1 , vC 3 /vC 4 across the capacitors C3 /C4 equals V2 . 1) Mode 1 (t0 < t < t1 ): The absolute value of the leakage inductance current iL is larger than iL2 , so that diode D3 is conducting. Furthermore, iL and the output current iL2 are linearly increasing while the input current iL1 is linearly decreasing. 2) Mode 2 (t1 < t < t2 ): At t1 , |iL | = iL2 , so that the current through diode D3 reverses its polarity and therefore commutates to switch S3 . 3) Mode 3 (t2 < t < t3 ): Since iL > iL1 , the current through diode D1 reaches zero at time t2 , so that switch S1 starts to conduct. 4) Mode 4 (t3 < t < t4 ): At t3 , switch S3 is turned off and capacitances Cr3 /Cr4 provide ZVS conditions. Cr3 /Cr4 is charged/discharged by the current driven by leakage inductance L and the output inductor L2 . At (t4 t3 )/2 the output current iL2 reaches its maximum and starts to decrease linearly. As soon as Cr4 is completely discharged, diode D4

Mode 1

3 4

67

9 10

11

12

S1 S2 S3 S4 vp

1 0 1 0 1 0 1 0

vC1=V1 0 -vC2=-V1 vC4 =V2 V2>V1 0 -vC3 =-V2

vs

iL

V2<V1 V2=V1

iL1 iL2 dTs t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10

I1 I2 t

t11 t12=Ts

Fig. 3. Key waveforms of the proposed DC-DC converter in


grid-to-battery operation.

D1

Current Reversal

L1 i L1

L1 i L Lm

L2 L2 i L2 N1 : N2
+ C4

Current Reversal

S1

Cr1 + C1

+ C3

D3 S3

Cr3

D1 S1 L1 i L1

Cr1 + C1 L1 i L Lm L2

+ C3

D3 S3

Cr3

V1 S2

D2

Cr2 + C2

D4 S4

V1 S2

L2 i L2 N1 : N2
+ C4

Cr4

D2

V2

Cr2 + C2

D4 S4

Cr4

V2

(a) Mode 1 (t0 < t < t1 ).


D1

(b) Mode 2 (t1 < t < t2 ).


D3 S3

L1 i L1

L1 i L Lm

L2 L2 i L2 N1 : N2
+ C4

Resonant Transition

Current Reversal

S1

Cr1 + C1

+ C3

Cr3

D1 S1 L1 i L1

Cr1 + C1 L1 i L Lm L2

+ C3

D3 S3

Cr3

V1 S2

D2

Cr2 + C2

D4 S4

V1 S2

L2 i L2 N1 : N2
+ C4

Cr4

D2

V2

Cr2 + C2

D4 S4

Cr4

V2

(c) Mode 3 (t2 < t < t3 ).


D1 S1 L1 i L1

(d) Mode 4 (t3 < t < t4 ).


D3 S3

L1 i L Lm

L2 L2 i L2 N1 : N2
+ C4

Resonant Transition

Cr1 + C1

+ C3

Cr3

D1 S1 L1 i L1

Cr1 + C1 L1 i L Lm L2

+ C3

D3 S3

Cr3

Active

V1 S2

D2

Cr2 + C2

D4 S4

V1 S2

L2 i L2 N1 : N2
+ C4

Cr4

D2

V2

Cr2 + C2

D4 S4

Cr4

V2

(e) Mode 5 (t4 < t < t5 ).


D1 S1 L1 i L1 Cr1 + C1 L1 i L Lm V1 S2 D2 N1 : N2 D4 S4 L2 L2 i L2
+ C4 + C3

(f) Mode 6 (t5 < t < t6 ).


D3 S3 Cr3 D1 S1 L1 i L1 Cr1 + C1 L1 i L Lm V1 S2 D2 N1 : N2 D4 S4 L2 L2 i L2
+ C4 + C3

D3 S3

Cr3

Current Reversal

Cr2 + C2

Current Reversal

Cr4

V2

Cr2 + C2

Cr4

V2

(g) Mode 7 (t6 < t < t7 ).


D1

(h) Mode 8 (t7 < t < t8 ).


D3 S3

L1 i L1

L1 i L Lm

L2 L2 i L2 N1 : N2
+ C4

Resonant Transition

Current Reversal

S1

Cr1 + C1

+ C3

Cr3

D1 S1 L1 i L1

Cr1 + C1 L1 i L Lm L2

+ C3

D3 S3

Cr3

V1 S2

D2

Cr2 + C2

D4 S4

V1 S2

L2 i L2 N1 : N2
+ C4

Cr4

D2

V2

Cr2 + C2

D4 S4

Cr4

V2

(i) Mode 9 (t8 < t < t9 ).


D1 S1 L1 i L1

(j) Mode 10 (t9 < t < t10 ).


D3 S3

L1 i L Lm

L2 L2 i L2 N1 : N2
+ C4

Resonant Transition

Cr1 + C1

+ C3

Cr3

D1 S1 L1 i L1

Cr1 + C1 L1 i L Lm L2

+ C3

D3 S3

Cr3

Active

V1 S2

D2

Cr2 + C2

D4 S4

V1 S2

L2 i L2 N1 : N2
+ C4

Cr4

D2

V2

Cr2 + C2

D4 S4

Cr4

V2

(k) Mode 11 (t10 < t < t11 ).

(l) Mode 12 (t11 < t < t12 = Ts ).

Fig. 4. Operating modes of the proposed DC-DC converter over one switching period Ts in grid-to-battery operation.

Bounda ry

12) Mode 12 (t11 < t < t12 = Ts ): At t11 , switch S2 is turned off and capacitances Cr1 /Cr2 provide ZVS conditions. Cr1 /Cr2 is discharged/charged by the current driven by L and L1 . Cr1 is discharged until diode D1 turns on, so that switch S1 can be turned on at nearly zero voltage. Current iL1 reaches its maximum at (t12 t11 )/2 and starts decreasing linearly. At the end of mode 12, i.e. after time t12 = Ts , mode 1 starts again. B. Battery-to-Grid Operation In battery-to-grid operation the control signals of S1 and S2 have to be exchanged with the ones of S3 and S4 compared to grid-to-battery operation, i.e. the primary and secondary side circuit in Fig. 1(c) are exchanged. The square-wave voltage applied to the transformer primary side lags the square-wave voltage applied to the transformer secondary side. C. Zero-Voltage-Switching Area To ensure ZVS conditions, the current owing through the switching device at turn-off has to be large enough to charge/discharge the two capacitors parallel to the switches in a bridge leg. The conditions in grid-to-battery operation are given by N1 iL (t3 ) + iL2 (t3 ) N2

1.4 Voltage ratio V2/V1 1.2 1 0.8 0.6 0.4 0.2

Bo

und

ary E

11) Mode 11 (t10 < t < t11 ): The currents iL1 and iL2 are linearly increasing and iL increases (V2 > V1 ), remains constant (V2 = V1 ) or decreases (V2 < V1 ) depending on the output voltage V2 in comparison to V1 .

1.8
5) q. (

1.6

Eq. (7)

ZVS area

un Bo

nda Bou 4) q. ( ry E

E ry da q. (6 )

0.1

0.2 0.3 Relative phase-shift d

0.4

0.5

Fig. 5. ZVS area of the proposed DC-DC converter in grid-to-battery operation for the parameters given in Table I. secondary resonant transitions are L + L1 , L L1 Ceq1 L N2 2 + L2
1

0,1

(11)

N2

0,2 r,1 = Z0,2 V > 2V2 (4) (5) (6) (7)

L N2 2 L2 Ceq 2
1

N2

(12)

r,3 V

r,2 = Z0,1 (iL (t5 ) iL1 (t5 )) > 2V1 V N1 = Z0,2 iL (t9 ) iL2 (t9 ) > 2V2 N2 r,4 = Z0,1 (iL (t11 ) + iL1 (t11 )) > 2V1 V

In Fig. 5 the ZVS area determined with the equations above is given. With a relative phase-shift in the range of d [0, 0.25] a much smaller ZVS area results than with d [0.25, 0.5]. The reason is the much smaller peak current iL which limits the ZVS area signicantly. IV. 250 kW / 2 MWh BATTERY S TORAGE S YSTEM

with

Z0,1

L L1 (L + L1 )Ceq1 L N2 2 L2
1

(8)

N2

Z0,2

N2 (L N2 2 1

(9)

+ L2 )Ceq2

The capacitance Ceq1 /Ceq2 is the constant, energy equivalent capacitance for the two nonlinear output capacitances of the switching devices plus the auxiliary capacitance in parallel to the switch. To enable ZVS turn-on of the switches, the two capacitances must be completely charged/discharged within the interlocking delay. The resonant transition time is given by 2V 1 arcsin . (10) tr, = r, 0, V {1, 2} denotes the bridge leg, {1, 2, 3, 4} the resonant transition. The angular frequencies of the primary and

Due to the nonpulsating input and output current, the softswitching DC-DC converter presented in the previous sections is ideally suited for the battery storage system shown in Fig. 1(a). The system consists of 38+2 submodules per leg/phase and is connected to the 6.6 kV medium voltage AC grid. Each submodule has a maximum output power of 12.4 kW. The total system power is 250 kW. The DC link of the battery storage system is kept at a constant voltage of 13 kV. The outputs of three submodules each in a different phase are connected in parallel and interleaved in order to further reduce the current ripple and the size of the output inductors. Furthermore, a continuous power ow from the grid to the Lithium-ion battery pack with an average voltage of 300 V and an energy capacity of 50 kWh is obtained with this parallel connection. On the primary side of the submodules, the input inductors are replaced by a common inductor La1 /La2 reducing the system volume. In the following, detailed results for a submodule are presented.

A. Simulation Results of the Submodule


Voltage (V)

The proposed circuit topology is simulated as DC-DC converter and as single submodule in the M2 C battery storage system shown in Fig. 1(a). The simulation model corresponds with Fig. 1(c) but neglects the magnetizing inductance Lm of the transformer as well as the nonlinearity of the output capacitances of the switching devices. The DC-DC converter is simulated with a constant input voltage of 600 V, a constant output voltage of 300 V, a switching frequency of 20 kHz and the parameters given in Table I. The simulation was carried out at a relative phase-shift of d = 0.15. TABLE I Simulation parameters
Parameter V1,nom V2,nom fs L1 L2 N1 /N2 L 1 L 2 Lm C1 , C 2 C3 , C 4 Cr 1 , C r 2 , C r 3 , C r 4 Ceq1 , Ceq2 Value 50 V. . . 600 V 240 V. . . 360 V 20 kHz 10 mH 6 mH (in M2 C: 2 mH) 2 90 H 22.5 H neglected 15 F (in M2 C: 2.5 F) 60 F 1 nF 2 nF

500 0 500

vp

vs

Current (A)

20 0 20

iL

Current (A)

40 20 0

iL2 iL1

0.05

0.1 Time (ms)

0.15

Fig. 6. Simulated key waveforms of the proposed DC-DC converter


in grid-to-battery operation (d = 0.15).
vp

Voltage (V)

500 vs 0 500

Current (A)

20 0 20 0

iL

Fig. 6 shows the simulated key waveforms in grid-to-battery operation whereas Fig. 7 shows the key waveforms in batteryto-grid operation. ZVS operation of all switching devices is achieved as shown in Fig. 8. Additionally, the submodule is simulated with a sinusoidal input voltage from 50 V to 600 V as it is the case in the M2 C battery storage system. Simulation results of the input current and the interleaved output power to the battery are shown in Fig. 9. Compared with the simulation parameters for the DC-DC converter, the inductance L2 can be decreased by a factor of 3 for a constant current ripple due to interleaved control of the submodules in the three different phases (see Fig. 1(a)). Furthermore, the capacitances C1 and C2 are reduced to 2.5 F to limit the reactive power ow needed to charge and discharge the primary capacitors with the grid frequency. The gate control signals are only mathematically calculated. In a next step, a controller will be implemented. B. Submodule Prototype In the M2 C based battery storage system, the input voltage of a submodule varies sinusoidal from 50 V to 600 V. Due to the series connection of the submodules in a phase leg, the balancing of the batteries needs to be done by controlling the input voltages of the submodules. Therefore, the submodule prototype is designed for an input voltage range up to 800 V to provide power control reserve. A battery voltage swing of 300 V 20% is considered. The submodule is operated at a

Current (A)

20 40 0

iL1 iL2 0.05 0.1 Time (ms) 0.15

Fig. 7. Simulated key waveforms of the proposed DC-DC converter


in battery-to-grid operation (d = 0.15).

switching frequency of 20 kHz. Using a three-level neutral point clamped (NPC) input stage for the submodule prototype as shown in Fig. 10 divides the required blocking voltage of the primary switching devices. The primary and secondary switching devices are chosen to be 1.2 kV silicon carbide MOSFETs from CREE [18]. In Table II the components of the submodule prototype are given while Table III shows the estimated losses of the switching devices and the passive components at an output power of 12.4 kW. Fig. 11 shows a possible hardware realization of the submodule.

Voltage (V)

1500 0 1500 0 800 0 S1 vS2

vS1 1 0
L1

D1a S1a

Cr1a
+

C1 vC1 vC3 L1 iL vp Lm L2 vs N1 : N2
+

S1b iL1 D1b

Cr1b

Dc1

C3

D3 S3

Cr3

Voltage (V)

S2 vS3

1 0

L2 C4 D4 S4 Cr4

iL2 V2

D2a V1 S2a

Cr2a

Dc2
+

vC4

Voltage (V)

C2 vC2

S2b

Cr2b D2b
Submodule Prototype

S3

1 0

Voltage (V)

800 v S4 0 0 S4 0.05 0.1 Time (ms) 0.15 1 0

Fig. 10. Submodule Topology with a three-level neutral point clamped (NPC) input stage for dividing the blocking voltage of the primary switching devices.
Inductor L2 MOSFETs 254 mm

Fig. 8. ZVS of all switching devices of the proposed DC-DC converter in simulation (d = 0.15). vS 1 to vS 4 depict the voltages across the switching devices.
600 400 200 0 20 Current (A) 0 20 Output Power (kW) 10 p2 5 0 iL1 v1

Transformer

Voltage (V)

96 mm

Capacitors C3,C4 Capacitors C1,C2

260 mm

Fig. 11. Possible hardware realization of a submodule. TABLE III Estimated losses at 12.4 kW output power
Component Switching devices
0 5 10 15 20 25 Time (ms) 30 35 40 45

Inductors Transformer

Fig. 9. Simulated input current of the submodule and the


interleaved output power p2 to the battery in the M2 C battery storage system. The voltage v1 corresponds to the voltage seen by a single submodule. The input current is only pre-controlled.
Capacitors Estimated efciency

Type of losses Switching losses Conduction losses Core losses Copper losses Core losses Copper losses AC losses

Value 25 W 150 W 15 W 35 W 75 W 55 W 2W 97.1%

TABLE II List of components of the submodule prototype


Component Inductor core L1 Inductor core L2 Transformer core Capacitor C1 /C2 Capacitor C3 /C4 Switching devices Estimated volume Type Metglas AMCC-100 (Alloy 2605SA1) Metglas AMCC-80 (Alloy 2605SA1) 2x Metglas AMCC-320 (Alloy 2605SA1) Electronicon E53.H59-252T10 Electronicon E53.M59-603T20 CREE CMF20120D 6.3 dm3

V. C ONCLUSION In this paper, a new bidirectional isolated DC-DC converter with nonpulsating input and output current and ZVS over a wide load range is presented. The new topology is derived from a Cuk converter and the operating details are discussed. Due to the low input and output current ripple, this converter is ideally suited for battery charging applications, which is demonstrated with a 250 kW / 2 MWh medium voltage battery energy storage system that can also be used for ultra-fast

charging of electric vehicles. The storage system is based on a M2 C structure that utilizes the proposed converter topology as submodules. For validating the proposed concept, a 12.4 kW prototype of a submodule is investigated. ACKNOWLEDGEMENT The authors would like to thank Swisselectric Research and the Competence Center Energy and Mobility (CCEM) very much for their strong nancial support of the research work. REFERENCES [1] J. Nemecek, D. Simmons, and T. Chubb, Demand sensitive energy storage in molten salts, Solar Energy, vol. 20, no. 3, pp. 213217, 1978. [2] A. Oudalov, D. Chartouni, and C. Ohler, Optimizing a battery energy storage system for primary frequency control, IEEE Transactions on Power Systems, vol. 22, no. 3, pp. 12591266, 2007. [3] L. M. Tolbert, F. Z. Peng, and T. G. Habetler, Multilevel converters for large electric drives, IEEE Transactions on Industry Applications, vol. 35, no. 1, pp. 3644, 1999. [4] A. Lesnicar and R. Marquardt, An innovative modular multilevel converter topology suitable for a wide power range, in Proc. IEEE Power Tech Conference, vol. 3, 2003. [5] L. Maharjan, S. Inoue, and H. Akagi, A transformerless energy storage system based on a cascade multilevel PWM converter with star conguration, IEEE Transactions on Industry Applications, vol. 44, no. 5, pp. 1621 1630, 2008. [6] SUNISA project by EPFL-LEI in Ultra FastCharging of Electric Vehicles (UFCEV), Annual Activity Report of the Competence Center Energy and Mobility (CCEM), pp. 2022, 2010. [Online]. Available: http://www.ccem.ch/MediaBoard/ CCEM Annual Activity Report 2010.pdf [7] H. Akagi and L. Maharjan, A battery energy storage system based on a multilevel cascade PWM converter, in Proc. Brazilian Power Electronics Conference (COBEP), 2009, pp. 918. [8] L. Maharjan, S. Inoue, H. Akagi, and J. Asakura, Stateof-charge (SOC)-balancing control of a battery energy storage system based on a cascade PWM converter, IEEE Transactions on Power Electronics, vol. 24, no. 6, pp. 16281636, 2009. [9] F. Krismer, J. Biela, and J. W. Kolar, A comparative evaluation of isolated bi-directional DC/DC converters with wide input and output voltage range, in Proc. Industry Applications Conference, 40th IAS Annual Meeting, vol. 1, 2005, pp. 599606. [10] D. Aggeler, F. Canales, H. Zelaya-De La Parra, A. Coccia, N. Butcher, and O. Apeldoorn, Ultra-fast DC-charge infrastructures for EV-mobility and future smart grids, in Proc. Innovative Smart Grid Technologies Conference Europe (ISGT), 2010, pp. 18. [11] H. Fan and H. Li, High frequency high efciency bidirectional DC-DC converter module design for 10 kVA solid state transformer, in Proc. 25th Applied Power

[12]

[13]

[14]

[15]

[16]

[17]

[18]

Electronics Conference and Exposition (APEC), 2010, pp. 210215. A. A. Aboulnaga and A. Emadi, High performance bidirectional Cuk converter for telecommunication systems, in Proc. 26th International Telecommunications Energy Conference (INTELEC), 2004, pp. 182189. R. W. Erickson and D. Maksimovic, Fundamentals of Power Electronics (Second Edition). Kluwer Academic Publishers, 2004. B.-R. Lin, C.-L. Huang, and J.-F. Wan, Analysis of a zero voltage switching Cuk converter, in Proc. 33rd Industrial Electronics Society Conference (IECON), 2007, pp. 19721977. B.-R. Lin and Y.-S. Huang, ZVS double-ended Cuk converter, IEEE Transactions on Circuits and Systems II: Express Briefs, vol. 57, no. 11, pp. 908912, 2010. H. Li, F. Z. Peng, and J. S. Lawler, A natural ZVS medium-power bidirectional DC-DC converter with minimum number of devices, IEEE Transactions on Industry Applications, vol. 39, no. 2, pp. 525535, 2003. F. Z. Peng, H. Li, G.-J. Su, and J. S. Lawler, A new ZVS bidirectional DC-DC converter for fuel cell and battery application, IEEE Transactions on Power Electronics, vol. 19, no. 1, pp. 5465, 2004. 1.2 kV SiC MOSFET CMF20120D. CREE. [Online]. Available: http://www.cree.com/products/pdf/ CMF20120D.pdf

You might also like