conf97(1)
-
Upload
krishnaveni-subramani-s -
Category
Documents
-
view
215 -
download
1
description
Transcript of conf97(1)
APPROXIMATE ANALYSIS OF THE RESONANT LCL DC-DC CONVERTER
Gregory Ivensky, Svetlana Bronstein and Sam Ben-Yaakov Power Electronics Laboratory
Department of Electrical and Computer Engineering Ben-Gurion University of the Negev
P. O. Box 653, Beer-Sheva 84105, ISRAEL, Phone: +972-8-646-1561, Fax: +972-8-647-2949 Email: [email protected], Website: www.ee.bgu.ac.il/~pel
ABSTRACT
An approximate analysis of LCL converters operating above and below resonance is presented. The study based on replacing the rectifier and load by an equivalent parallel resistance, applies the fundamental harmonics method and the assumption that the current through the diodes of the output rectifier has a sinusoidal waveform. Equations obtained for the conduction angle of these diodes (less than π in Discontinuous and equal to π in Continuous Conduction Modes) are used for the developing the output to input voltage ratio as a function of three independent parameters. The theoretical results coincide with good agreement with the experiments.
1. INTRODUCTION LCL converters have a number of advantages: they can operate in both step up and step down modes, they will accommodate a wide range of output to input voltage ratios with a relatively small frequency deviation, and soft switching could be achieved over the entire operating range [1]-[4]. Approximate analysis of LCL converters operating in Continuous Conduction Mode (CCM) above resonance is described in [4].
The objective of the present study was to develop a simple approximate analysis of LCL converters operating above and below resonance. The focus of attention was devoted to the Discontinuous Conduction Mode (DCM), but the results are applicable to the Continuous Conduction Mode (CCM) as well. The study is based on the equivalent circuit concept in which the rectifier and load are replaced by an equivalent parallel resistance [5], [6], and applies the fundamental harmonics method.
2. ANALYSIS The LCL resonant inverter topology can be implemented as a full-bridge, half-bridge or half-way type [7], and the output rectifier could be built as a full-wave bridge or as a full-wave center tapped configuration with capacitive output filter. Fig. 1 represents one of the possible realizations of the LCL converter: L1-C-L2 is the LCL resonant tank, T is a transformer, n=n1:n2 is the turns ratio, D1-D4 are the diodes of the rectifier, Cf is the output filter capacitance, and Ro is the load resistance. The current and
voltage waveforms for the DCM variation are depicted on Fig. 2. 2.1 Main assumptions The analysis is carried out under the following main assumptions: a) Switches, diodes, the transformer and all reactive elements are ideal. b) The current through the diodes D1 ... D4 of the output rectifier has a sinusoidal waveform:
ϑ
λπ
= sinIi pk.DD (1)
where tω=ϑ is the normalized time, counting from the diode turn on instant ϑo, t is the real time, ω is the operating frequency, π≤λ is the conduction angle, ID.pk is the peak value of the diode current. The latter can be expressed as a function of the load current Io or load voltage Vo:
λπ
=λ
π=
2RV
2II
2
o
o2
opk.D (2)
c) The current through the inductor L2 during the conduction angle λ is changing linearly:
λϑ
∆+= 2Lo2L2L IIi (3)
where IL2o is the initial value of the current iL2 and 2LI∆ is the increment of the current iL2 during the conduction angle λ . This current increment can be expressed as a function of the load voltage referred to the primary nVo:
2
o2L L
nVI
ωλ
=∆ (4)
Taking into account (2), we transform (4) to: 2
2
o2
pk.D2L L
Rnn
I2I
πλ
ω=∆ (5)
C
Cf Ro
L 1 L2
n1n2
D1 D2
D4 D3
T
Q1
Q4
Q2
a b
Vo
Vin
Q3
Fig. 1. LCL resonant inverter topology.
ID.pk IoiD1=iD3 iD2=iD4iD
Ioϑ
ϑ
ϑ
ϑ
iL2 ∆IL2
λ λπ π
vab
VovT2Vo
Vin
ϑo
Fig. 2. Currents and voltages waveforms. DCM. 2.2 The diodes’ conduction angle The diodes’ conduction angle λ can be obtained from the following equation that is derived by Kirchoff’s law:
ϑω++ϑ
+
ω+
ϑ
+
ω= ∫ ddi
LVdin
iC
1d
in
id
Lv 2L2Co2L
D2L
D
1ab (6)
where vab is the instant voltage across the diagonal ab of the inverter bridge (Fig. 1) and VCo is the initial value of the capacitor’s C voltage. We consider a time interval where the voltage vab is constant: vab=Vin (Vin is the input voltage) and when identical diodes of the rectifier continue to conduct. In this time interval the voltage across the inductor L2 is equal to the reflected output voltage:
o2L
2 nVd
diL =
ϑω (7)
and the voltage across the inductor L1 has a constant component proportional to the reflected output voltage:
2
1o
2L1 L
LnV
ddi
L =ϑ
ω (8)
Taking into account (7) – (8) we simplify (6) to:
Co2LD
D
1o2
1in Vdi
ni
C1
dn
id
LnVLL
1V +ϑ
+
ω+
ϑω+
+= ∫ (9)
Taking the derivative of (9) we find:
+ω
+ϑ
ω= 2LD
2
D2
1 ini
C1
dni
dL0 (10)
Applying (1) and (3) we transform (10):
λϑ
∆+ω
+
ϑ
λπ
ω+
ϑ
λπ
λπ
ω−=
2Lo2L
pk.D2
1pk.D
IIC
1
sinn
IC
1sinLn
I0 (11)
0.4
0.6
0.8
1
1 20.80.60.4
L1
L2
=0.5
ωωrs
λπ
B=2
520
1 20.80.60.4 ω
ωrs
0.4
0.6
0.8
1
λπ
L1
L2=1
0.23
(a) (b)
Fig. 3. The ratio λ/π versus normalized operating frequency ω/ωρs for different values B (a) and L1/L2 (b) (dots – simulation, solid lines – theoretical prediction).
From which:
0I
nII
nIsin1
pk.D
2L
pk.D
o2L22
rs=ϑ
λπ
π∆
++
ϑ
λπ
+
λπ
ωω
− (12)
In the last equation:
CL1
1rs =ω (13)
is the resonant frequency of the series circuit L1C. Applying the assumption:
nI
|I| pk.Do2L << (14)
and taking into account (5) we obtain from (12) the approximate expression of the conduction angle λ as half sum of two angles:
( )215.0 λ+λ=λ (15) The angle λ1 is solution of (12) for the case λ<ϑ 1.0 , i. e
when ϑλπ
≈
ϑ
λπsin :
2
rso2
22
o2
22
o2
21
RnL
2RnL
16RnL
4
ωωωπ
+
ωπ+
ωπ−=
πλ (16a)
while the angle 2λ is solution of (12) for the case λ=ϑ 5.0 :
2
rso2
22
o2
22
o2
22
RnL
RnL
4RnL
2
ωωπω
+
ωπ+
ωπ−=
πλ (16b)
The value of πλ / as a function of the normalized operating frequency rs/ ωω , calculated by (15), (16a), (16b) for different values of load coefficient
1o
2
LCRnB = and for different inductances ratio L1/L2
is depicted on Fig. 3. a,b. The conduction angle λ gets higher for smaller values of B and L1/L2. When π=λ , DCM is replaced by CCM. The plots (Fig. 3) also include points that were obtained by PSPICE simulation. The agreement between theoretical prediction and simulation points is good. 2.3 The output to input voltage ratio To obtain the output to input voltage ratio
in
oo V
Vk = (17)
we replace the transformer and the rectifier by an equivalent resistance '
eqR (Fig. 4a) calculated by energy balance consideration:
2
'eq2
pk)1(o2o n
RI
21RI = (18)
where pk)1(I is the peak of the fundamental harmonic of the rectifier input current (to simplify the analysis we don’t take into account higher harmonics). Under the
second main assumption, the following equation describes the dependence I(1)pk on the load current Io and the conduction angle λ:
2opk)1(
1
2cos
I2I
πλ
−
λ
= (19)
Applying (18) and (19) we define the expression of 'eqR :
2
2
2
o2
'eq
2cos
1
2Rn
R
λ
πλ
−
= (20)
Note that transition to CCM ( π=λ ) translates (20) to the known expression [5]:
o2
2'eq Rn8R
π= (21)
As a next step of analysis we replace the parallel network 2
'eq LR − by the equivalent series network
"2
"eq LR − (Fig. 4b) using the following equations:
ω+
=
ω+
=
2
'eq
2
2''2
2
2
'eq
'eq''
eq
RL
1
LL
LR
1
RR
(22)
Hence, the complete reactance of the inverter is: "21 L
C1LX ω+
ω−ω=Σ
(23)
b
Vab(1)
C
Req
L1
L2
a
' Vab(1)
CL1a
b
XΣ
L''2
R''eq
I(1)
(a) (b)
Fig. 4. Series-parallel (a) and series (b) equivalent circuits.
ϕ
ψ
VL2(1)
Vab(1)
I(1)XΣ
I(1)ωL''2
eqI(1)R'' Fig. 5. Vector diagram of the voltages in Fig. 4b.
The ratio between the peaks of the fundamental harmonics of the voltages across the inductor L2 and across the diagonal ab of the inverter bridge is defined from the vector diagram (Fig. 5):
( )
( ) 2
"eq
2
"eq
"2
pk1ab
pk12L
RX
1
RL
1
coscos
VV
+
ω+
=ψϕ
=
Σ
(24)
where ϕ is the angle between the vectors Vab(1)pk and "eq)1( RI , while ψ is the angle between the vectors VL2(1)pk
and "eq)1( RI ; I(1) is the peak of the fundamental current
through the network (Fig.4b). To obtain the input to output voltage ratio, the peak
voltage ( )pk1abV in (24) needs to be expressed as a function of the inverter’s input voltage Vin while VL2(1)pk in the same equation should be expressed as a function of the reflected output voltage nVo.
The instant voltage across ab has a square waveform with steps inV± (Fig. 2). Therefore, under the first harmonics assumption Vab(1)pk=4Vin/π is the fundamental harmonics of Vin.
The expression ( )pk12LV as a function of nVo is more complicated because the equation o2L nVv ±= is valid only during the rectifier conduction angles π<λ (Fig. 2). We obtain the expression ( ) ( )opk12L nVV ϕ= using following additional assumptions:
1. The symmetrical axis of the waveform of the inductor L2 voltage coincides with the middle of the rectifier conduction angle.
2. The waveform of the inductor L2 voltage coincides with its fundamental harmonic during the time intervals when both diodes of the rectifier are not conducting. Under these assumptions:
( ) λ+λ
λ
=sin
2sin4
nVV opk12L (25)
Applying (17), (24) and (25) we derive the expression of the output to input voltage ratio as:
2
"eq
2
"eq
"2
o
RX
1
RL
1
2sinn
sink
+
ω+
λ
π
λ+λ=
Σ
(26)
The reflected output to input voltage ratio nko=nVo/Vin as a function of the normalized operating frequency ω/ωr, load coefficient B and inductances ratio L1/L2, calculated using the obtained equations, is depicted on Fig. 6, a, b.
3. EXPERIMENTAL The experimental set up included a half-bridge inverter and a full bridge output rectifier. The inverter was built around a half bridge configuration with two power MOSFETs (IRF840), LCL resonant tank, and planar transformer (200W T250DC-4-24 by Payton, n=3:1). Inductances of the tank were constructed on ETD29 3F3 cores using Litz wire. The rectifier diodes were MUR460. The input voltage was kept constant around 100V. Load resistance varied from 10Ω to 250Ω. Parallel inductance L2 was 78µH. Series inductance was varied: 16µH, 38.5µH, and 95µH. Operating frequency was varied from 40kHz to 500kHz, covering all operating modes. Maximum power involved in experiment was 70W. It was limited by high hard-switching losses when operating at low frequencies. Experimental voltage ratios nko as a function of the load resistance Ro for different frequencies (Fig. 7), and as a function of the operating frequency for different load resistances (Fig. 8, a,b), show good agreement with the theoretical predictions.
4. DISCUSSION AND CONCLUSIONS An approximate analysis of the resonant LCL converter was carried out, based on the equivalent circuit concept and applying the fundamental harmonics method. The convenient approximate closed formed equations derived for the conduction angle and output to input voltage ratio as a function of three independent normalized parameters (frequency, load resistance and inductances ratio) allow to define the operating regime of the converter for the given circuit parameters and operating frequency and can be used in the design purpose. The analytical results were examined by simulation and experiment and found to be in a good agreement.
Although this study gave focus attention to the DCM, analytical expressions obtained here are applicable to the CCM as well by setting λ=π in the calculations.
Notwithstanding the fact that some of the expressions developed in this study looks complex, their use is obviously simple once implemented into computer mathematical packages such as MATLAB, Mathematica or MathCad. Hence, the analytical expressions developed in this study could help the designer to study the behavior of a given LCL converter and to optimize the values of the elements per the target specifications.
01234
B=1
3
5
10
1 20.80.60.4
nVo
Vin
ωωrs
L1
L2
=0.5
13-10
00.511.522.53
1 20.80.60.4 ωωrs
nVo
Vin
B=3
0.25
0.5
1
=1.5L1
L2
0.25
1.5
(a) (b)
Fig. 6. Input to output voltage ratio versus normalized operating frequency for different values of B (a) and L1/L2 (b).
0
1
2
3
4
5
6
7
Ro10 100
nVo
Vin
70kHz
96kHz
150kHz
fr=117kHz
L1
L2
=1.22
20 50 200 [Ω]
84kHzf=
Fig. 7. Input to output voltage ratio as a function of the load
resistance for different operation frequencies (dots – experiment, solid line – theoretical prediction).
10
1
100 200 300 400 f, kHz
L1
L2
=0.5
nVo
Vin
fr=183.8kHz
100Ω
Ro=10Ω
100 200 300 f, kHz
10
1
nVo
Vin
Ro=10Ω
100ΩL1
L2
=1.22
fr=117kHz
(a) (b)
Fig. 8. Input to output voltage ratio as a function of the operating frequency for different load resistances (dots – experiment, solid line – theoretical prediction).
ACKNOWLEDGMENT
This research was supported by THE ISRAEL SCIENCE FOUNDATION (grant No. 113/02) and by the Paul Ivanier Center for Robotics and Production management.
5. REFERENCES [1] J. F. Lazar and R. Martinelli, “Steady-state analysis of LLC
series resonant converter,” IEEE APEC’2001 Record, pp. 728-735.
[2] H. Jiang, G. Maggetto, and P. Lataire, “Steady-state analysis of the series resonant DC-DC converter in conjunction with loosely coupled transformer – above resonance operation,” IEEE Trans. Power Electronics, vol. 14, no. 3, May 1999, pp. 469-480.
[3] Bo Yang and F. C. Lee, “LLC resonant converter for front end DC/DC conversion,” IEEE APEC’2002 Record, pp. 1108-1112
[4] A. K. S. Bhat, “Analysis and design of LCL-type series resonant converter,” IEEE Trans. Industrial Electronics, vol. 41, no. 1, Feb. 1994, pp. 118-124.
[5] R. L. Steigerwald, “A comparison of half-bridge resonant converter topologies,” IEEE Trans. Power Electronics, vol. 3, no. 2, pp. 174-182, Apr. 1988.
[6] G. Ivensky, A. Kats, and S. Ben-Yaakov, “An RC load model of parallel and series-parallel resonant converters with capacitive output filter,” IEEE Trans. Power Electronics, vol. 14, no. 3, pp. 515-521, May 1999.
[7] M. K. Kazimerczuk and D. Czarkowski, “Resonant power converters,” John Wiley & Sons, Inc., 1995.