AbstractâThis paper proposes a novel ZVS Z-source rectifier, describes the operating principle, control strategy, and superiority. The three-phase ZVS Z-source ...
A Novel ZVS Z-Source Rectifier Xinping Ding1, Zhaoming Qian1, Yeyuan Xie1, Fang Z. Peng1
、2
1
Zhejiang University, Hangzhou, China, 310027 2 Michigan State University Department of Electrical and Computer Engineering Abstract—This paper proposes a novel ZVS Z-source rectifier, describes the operating principle, control strategy, and superiority. The three-phase ZVS Z-source rectifier implements the zero voltage turn-on (ZVS) for power transisitors and zero-current turn-off (ZCOFF) for diodes without any additional circuits. It can buck and boost input voltage, and increase the reliability of the circuit. Thus it provides a low-cost, reliable, and highly efficient single-stage structure for buck or boost power conversion. Index terms—shoot-through, ZVS, buck-boost rectifier, Zsource rectifier.
S
Ⅰ.INTRODUCTION
This paper presents a novel ZVS Z-source rectifier. Actually, it evolved from the recently presented Z-source inverter [1], it is a buck-boost ZVS rectifier that can give a wide range of output voltage. Fig. 1 shows the main circuit configuration of the proposed ZVS Z-source rectifier, where a unique Z-network is coupled between the load and a rectifier circuit. The Z-network is implemented by a split-inductor (L1 and L2) and capacitors (C1 and C2) connected in X shape. This unique Z-network allows the z-source rectifier operating in ZVS condition without any additional circuits to buck or boost its output voltage. 3 − Phase Boost Rectifier
iL
iC
Z − network
VC 2
VC1
ib
iL VC 0
Fig. 3. Sketch of the nonshoot-through states Load
ic SW4 SW6 SW2
The average current flowing through C should equal to zero with a complete switching cycle, we have:
VL 2
= 0,
(1)
Therefore I = D0 I . c 1− D0 L where D0 is the shoot-through duty ratio. During the nonshoot-through states,
(2)
D0 ⋅ ( − I L ) + (1 − D0 ) ⋅ I c
Fig.1. Three-phase z-source rectifier Ⅱ. ZERO VOLTAGE
STATES Unlike the traditional three-phase voltage source rectifier, which has eight states, the three-phase ZVS Z-source rectifier has nine permissible switching states (six active vectors, two zero vectors and one extra shoot-through zero vector). The shoot-through state provides the unique ZVS and buck-boost features of the devices in the rectifier bridge. Fig. 2 is the sketch of the extra shoot-through zero state, where S is an equivalent switch for the rectifier bridge switches, and D is an equivalent diode for the diodes in the rectifier bridge. Because of the circuit symmetry (L1=L2, C1=C2), the following assumptions can be made in the analysis of the circuit: IL1=IL2 VL1=VL2, and IC1=IC2 VC1=VC2. We can further assume that IL1 and IL2 remain continuous and ripple free if L is large enough.
0-7803-9547-6/06/$20.00 ©2006 IEEE.
iC
SW7
VL1
ia
D
Fig. 2. Sketch of the extra zero voltage states
SW1 SW3 SW5
3 LN
During the extra zero voltage state—shoot-through state, the diode D is on, which clamped the voltage of the switch S to the zero. With the view of this extra zero voltage state, the switch S can achieve ZVS condition when it is turned on just after this state.
D0 1 I 0 = IC + I L = I L + I = I , 1− D0 L 1− D0 L
(3)
D0 1− 2 D0 I in = I L − I c = I L − I = I . 1− D0 L 1− D0 L
(4)
Due to the lossless nature of the Z-network, the input and output power should be balanced. That is
Vin ⋅ I in = V0 I 0 and V0 = (1 − 2 D0 )Vin . If the three-phase ac input voltages are
951
(5)
⎧van (t ) = VM sin(ωt ) ⎪ 0 (6) ⎨vbn (t ) = VM sin(ωt − 120 ) , ⎪ 0 ⎩vcn (t ) = VM sin(ωt − 240 ) the dc-link voltage of ZVS Z-source rectifier can be expressed as (1− 2 D0 ) 2VM (7) V0 M cosψ
iin = 2 i L
VM is the
peak ac input phase voltage. As shown in Fig. 2, during the shoot-through state, the diode D is on, which clamps the voltage of the switch S to the zero. With the view of this extra zero voltage state, the switch S can achieve ZVS condition when it is turned on/off just after/before the shoot-through state. Ⅲ. TRIANGLE-SHIFT CONTROL STRATEGY Fig.4 illustrates the triangle-shift control sketch, va, vb, vc are the three-phase modulation signals respectively. The operation modes corresponding to the time zones from T0-T9 shown in Fig. 4 are shown in Fig. 5. The Arabic numerals indicated in Fig.5 mean the device serial numbers same as shown in Fig. 1, s1-s6 are the switching sequence of switches sw1-sw6 in the rectifier bridge respectively. Because the buck factor is determined by the shoot-through duty ratio, the shoot-through duty ratio must be kept the same in order to maintain a constant buck. Fig.4 illustrates the triangleshift control sketch, the carrier triangle signal parallel shift to the right, with which the drive signal of sw7 (inverse of shootthrough drive signal) is produced. It is quite similar to the simple boost control method [1]. The shift triangle control method maintains the six active states unchanged as in traditional PWM control, and made the switches, which are in on-state during the free-wheel states, turned on just the moment of sw7 turned off (shoot-through state). That is on the forepart of these free-wheel time zones, the extra shoot-through zero state is operated, then it comes the really free-wheel state. In one switching cycle, the operating states can be described as follows: (a) 0-T0: Prior to T0, the diodes 1, 5, 6 are on, and the switches in the rectifier bridge are off. The ac source transferred power to the load through three diodes. The input current of the Z-network can be expressed as (8) i = i −i , in
L
(9)
the input current of the Z-network mutate to about two times of that in time zone a, which is the reason why all diodes turned on at that moment. (c) Because the diodes clamped switches voltage to zero, the devices sw3, sw5 are turned on in ZVS condition, diode D1 is still in on-state, the other 5 diodes may turn off in zero-current (ZCOFF) condition if we design the appropriate inductor L in znetwork. During the time zone b, each phase is shortened through the input inductor LN, which is called as free wheel state. (d) In the time zone c, the diodes 1, 6 and switch sw5 are in on-state, which is the same to the traditional PWM control. (e) Time zone d is quite similar to time zone a, which also have three diodes in rectifier bridge are on, the only difference is the series diodes which are in on-state. (f) After the time zone d, it is also operated in shootthrough state, which clamped the switch voltage to zero, then the free-wheel state, etc. (g) … As can be seen in Fig.5, if we put the extra zero state just on the forepart of these free-wheel time zones, with the diodeclamp characteristic, the switches will operate in the ZVS condition, and that the diodes will operate in the ZCOFF condition. The operating principle is similar to the traditional PWM control method except the shoot-through states. With the shift-triangle control method, the shoot-through duty cycle D0 can be expressed as (10) 0 < D ≤ 1− M
=
whereψ = arctan ω LN , M is the modulation index, and RN
,
0
V0 2VM
=
(11)
b 1 M cosψ
Carriervc triangle
va
shifttriangle vb
T
1 TD0 2
s1 s3 s5 s4 s6 s2
c
the capacitor current may be ignored compared with the inductor current at the end of the time zone a. (b) T0-T1: At T0, the device sw7 turns-off, all the diodes in the rectifier bridge are in on-states, and the devices in the rectifier bridge are off, the diodes clamp the switch voltage to zero. It operates in shoot-through state. We have
sshoot T0 T1T2
T3
T4 T5T6
T7
T8 T9
Fig. 4. Sketch of shift-triangle control
952
0
5
1
A
1 3
A
B
0
C
(c) T1 -T2
6
(a) 0- T0
A
1
B
3
4
6
2
5
(b)T0-T1 ,(f) T4-T5 (j) T8-T9
Shootthrough 1
3
A 0 6
2
2
(e) T3-T4
2
6
1
1 A
(h) T6-T7
C 6
4
B
B
0
C
(g) T5-T6
C
1
A
B
0
6
(d) T2-T3
A
C
0
B
0
C
5
1
A
B
C
0
5
3
5
A
B
B
C
C
(i) T7-T8
2
k
Fig. 5. Operating series states with the shift-triangle control strategy
we has V0 2 M −1 1 < m = ≤ M cosψ M cosψ 2VM
(12)
The curve of voltage ratio versus modulation index is shown in Fig.7. We can obtain any output voltage greater or smaller than the input voltage with shift-triangle control strategy.
off in the zero current condition if we guarantee iin decreased to iLm in the end of the shoot-through state. The inductor current and voltage are shown in Fig. 9. In the time of T1, inductor current is increasing and the inductor is in its charging time; during the shoot-through time, the inductor current is decreasing and the inductor is in its discharging time. So the average current of L (iLav) is equal to the load current I0. From Fig. 9, we have (13) V I Lm − I LM = ∆I L = L (1 − D0 )T L where ILm and ILM are the maximum current and minimum current respectively, D0 is the shoot-through duty ratio, and L is the inductor of Z-network. In the Z-source rectifier, when operated in the shoot-through states, the voltage of the inductor is expressed as: 1 − D0 (14) VL = VC = − Vo , 1 − 2 D0 while during the nonshoot-through states, one has D0 (15) V = V . L
1 − 2 D0
o
with (9) and (11), we have Fig. 6. Voltage ratio versus modulation index M
I Lm − I LM
(1 − D0 ) D0 V = ∆I L = L (1 − D0 )T = TV0 L L(1 − 2 D0 )
(16)
In order to make it working in the CCM, it must have
Ⅳ. THE DIODE ZCOFF CONDITION AND DESIGN OF THE ZNETWORK INDUCTOR
With shift triangle control strategy, the ZVS switching condition is fulfilled. From the Fig.10, the input current of znetwork, iin, is two times of the inductor current iLm at the beginning of the shoot-through state. The diodes may be turned
I LM
953
V T (1 − D0 ) D0 = I0 − 0 > 0, 2L(1 − 2 D0 )
(17)
in the ZVS condition. With the felicitous design of the inductor in Z-network, the diodes of the rectifier bridge can operate in zero-current turn-off (ZCOFF) condition, that is to say the current through diodes decreased to zero at the end of the shootthrough states.
iL I Lm
I LM
0 D0V0 1 − 2 D0
VL
−
(1 − D0 )V0 1 − 2 D0
0 (1 − D0 )T
I0
t
t
D0T
Fig. 7. Voltage and current of the inductor in Z-network
Therefore L>
V0T (1 − D0 ) D0 , 2 I 0 (1 − 2 D0 )
(18)
where v0 is the output of the Z-network, I0 is the dc-link current. As Fig. 9 shown, the input current of the z-network iin can range from ILm to 2ILm in the moment from nonshoot-through to shoot-through states, ignored the capacitors (C1 and C2) current. Therefore if we guarantee iin decreased to ILm in the end of the shoot-through state, the diodes may be turned off in the zero current (ZCOFF) condition. We have I (19) ∆I ≥ Lm , L 2 that is (20)
Fig. 8. Simulation results for sudden change from full rectification to full inversion (M=0.75, D0=0.2)
V0T (1 − D0 ) D0 I 0 V0T (1 − D0 ) D0 ≥ + . L(1 − 2D0 ) 2 4 L(1 − 2 D0 )
Based on (12), (14), and (15), the range of the inductor L must be satisfied with (21) V0T (1 − D0 ) D0 3V T (1 − D0 ) D0
