5808 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 66, NO.

7, JULY 2017

Design of a Single-Stage Inductive-Power-Transfer

Converter for Efficient EV Battery Charging
Zhicong Huang, Student Member, IEEE, Siu-Chung Wong, Senior Member, IEEE, and Chi K. Tse, Fellow, IEEE

Abstract—This paper studies wireless charging of lithium-ion using external reactive elements to form resonant tank circuits
batteries for electric vehicles. The charging profile mandates a is often required for both primary and secondary windings of
constant-current (CC) charging for a discharged battery until the the transformer to enhance the power transfer capability, to min-
battery voltage reaches the cutoff voltage at rated power. The
charging continues at the cutoff voltage with a constant-voltage imize the VA rating of the power supply, to regulate separately
(CV) charging at a power level down to 3% of the rated power the currents in the supply loop and the receiving loop, and to
in order to fully charge the battery. An inductive-power-transfer achieve a higher efficiency. Studies have been carried out for
(IPT) converter should be designed with minimal number of stages selecting the most appropriate compensation topology for IPT
to achieve high efficiency. However, high efficiency for such a wide systems for specific applications [2], [7]–[12].
load range is difficult to achieve. Moreover, the efficiency-to-load
relationship is distinctly different for CC and CV charging opera- Lithium-ion (Li-ion) batteries are widely used in EVs. The
tions, posing difficulties for the single-stage design. This paper de- charging process for Li-ion batteries usually consists of four
scribes the design of a single-stage IPT converter that complies with main stages [13], [14]: trickle charging, constant-current (CC)
the battery charging profile and, at the same time, achieves optimal charging, constant-voltage (CV) charging, and charge termina-
efficiency. Design optimization includes soft switching for the entire tion. A CC charging stage followed by a CV charging stage
battery load range, efficiency optimization for CC and CV modes
of operation, and system efficiency optimization for the whole bat- is the preferred charging algorithm for Li-ion batteries. The
tery charging profile. Measured results of two experimental IPT charging current at the CC stage does not need to be precise
battery chargers are presented for illustration and verification. and a quasi-CC is allowed [15]. With this charging algorithm,
Index Terms—Battery charger, electric vehicles (EVs), system a converter should charge a load with power varying from the
efficiency, wireless power transfer. maximum rated power down to a minimum of about 3%. This
efficiency optimization requirement is challenging for most con-
I. INTRODUCTION verter topologies.
HE direct burning of fossil fuels in combustion engines An IPT power converter is, in general, a resonant converter,
T of vehicles incurs increasing financial and environmental
costs. In recent years, demand for green electric vehicles (EVs)
which consists of an inverter driving a resonant circuit. An ideal
inverter circuit can have the highest efficiency when it is driving
has grown significantly. First generation of EVs uses predom- a pure resistive load. Therefore, the resonant circuit is mostly de-
inantly simple plug-in charging methods, which have safety signed with zero-phase angle (ZPA) at the operating frequency
issues caused by exposed plugs and damaged cables. To elim- driven by the inverter switches, where the inverter switches can
inate plugs and cables, wireless charging methods have been operate at both zero-voltage and zero-current switching [11].
widely studied [1]–[4]. A wireless charging system can have However, application of different power switches may need to
both the primary and secondary sides fully insulated and with- have a slight different adjustment. The popular MOSFET power
out physical contact. Thus, it is versatile in humid and adverse switches have significant parasitic drain-to-source capacitances.
weather conditions. The charging process can be designed to be The resonant circuit is, thus, designed slightly inductive to ab-
automatic, safe, and user convenient. solve all the charges of the capacitances of the inverter switches
Wireless EV battery charging usually uses a loosely cou- before the instance of active turning on (soft switching). Since
pled transformer in an inductive-power-transfer (IPT) process the resonant circuit is slightly inductive, the switches have to
to convert power from the primary side to the secondary side be turned ON at zero-voltage [zero-voltage switching (ZVS)].
separated by a large predefined air gap [5], [6]. Compensation The corresponding currents at switching instance are not zero
and need to be kept small to reduce conduction loss during
the transition of turning ON. It is noted that the depth of di-
Manuscript received July 21, 2016; revised October 24, 2016; accepted rect modulation of the inverter switches should be kept shallow
November 14, 2016. Date of publication November 22, 2016; date of current
version July 14, 2017. This work was supported by Hong Kong Polytechnic to maintain good efficiency of soft switching. The reasons are
University under Central Research Grant G-YBKC. The review of this paper that for pulse width modulation (PWM), the soft switching is
was coordinated by Dr. B. Akin. difficult to maintain, and for frequency modulation (FM), the
The authors are with the Department of Electronic and Information Engi-
neering, The Hong Kong Polytechnic University, Hong Kong (e-mail: zhicong. conduction losses during switching instance at higher operation
[email protected]; [email protected]; [email protected]). frequency can be significant.
Color versions of one or more of the figures in this paper are available online The soft switching IPT converters can further achieve optimal
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TVT.2016.2631596 efficiency at some resonant frequencies with matched input and

0018-9545 © 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications standards/publications/rights/index.html for more information.

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
HUANG et al.: DESIGN OF A SINGLE-STAGE INDUCTIVE-POWER-TRANSFER CONVERTER FOR EFFICIENT EV BATTERY CHARGING 5809

output impedances. Deviating from this optimal matching point, TABLE I

CASES OF IMPLEMENTATION
the converter efficiency suffers [16]. In view of the narrow input
impedance and output impedance ranges of resonant converters,
multistage converter topologies with input and/or output power Case Operating point CC charging CV charging
converters connected in-front-of and/or after the IPT converter (C1) LIV MC MC
have been proposed [17]–[19], hoping to improve system effi- (C2) LIC MC MC
(C3) LIV MC native
ciency by operating the resonant converter without modulation (C4) LIC native MC
at its optimal load while allowing losses of extra efficient con- (C5) LIC native
verters. The obvious drawbacks of these multistage converters LIV native
are the following.
note: MC—modulation and control
1) More converters are used. native — minimal modulation for efficiency
2) More complicated control may be needed for the coor-
dination of controls between the primary and secondary with the multistage and hybrid-topology IPT converter at rated
sides of the IPT converter. power, some design options for a single-topology IPT con-
3) These topologies are for general applications as they are verter are available. First, the single-topology IPT converter can
designed for an arbitrary variation of load range without operate near its maximal efficiency point and be controlled for
optimizing the power loss and the charging time of the the required two stages of battery charging. Alternatively, it may
battery charging profile. operate at its optimal efficiency point for the first stage of CC or
Without the additional cascaded power converter(s) for the second stage of CV charging and use a control technique with
tracking of the input and/or load impedance, a single-stage IPT reduced efficiency for the other stages of charging. Furthermore,
converter has less degree of freedom for optimization under the it may operate at its optimal efficiency points at LIC and LIV
specific charging profile. The single-stage IPT converter can for the first and second stages of charging, respectively. The
be designed for maximum efficiency at maximum load power. possible cases of implementations for a single-topology IPT
Thus, the IPT converter can efficiently operate with shallow converter are summarized in Table I.
modulation at either its load-independent voltage (LIV) out- In Table I, cases (C1) and (C2) are mostly achieved by utiliz-
put operation point using series–parallel (S-P) compensation ing different depth and type of modulations. Without modula-
or load-independent current (LIC) output operation point using tion, the converter can be efficiently operated at operation points
series–series (S-S) compensation [11], [12]. At these operation LIV and LIC. As a rule of thumb, efficiency suffers from more
points, the equivalent reactance on the inverter switches is near depth of modulation.
ZPA. To achieve energy-efficient CC and CV charging stages Case (C1) has been used in a SSIPT converter with a sim-
with narrow-range PWM control and/or FM control, switching ple narrow-frequency-range FM control [21]. Normally, a fre-
of IPT compensation topologies from S-S compensation to S-P quency limiter is used to implement the FM control [21] to
compensation is needed [20]. These hybrid- or dual-topology maintain stability in the loosely coupled systems [21], [24],
IPT converters necessitate extra power switches along the main [25]. It is obvious that the single-topology SSIPT converter op-
power path of the converter, incurring additional cost for power erating near the LIV operating point [21] can only be optimized
switches and power loss. Meanwhile, the CV converter is still for efficiency with an operating frequency at a single loading
required to charge the battery load with power varying from the point for the CC and CV modes of charging for an EV battery.
full rated power to a minimum of about 3% of the rated power. Likewise, this also applies to case (C2). However, due to their
In Section II, the conditions for LIC output and LIV output of operations near LIC operating point, cases (C2) and (C4) re-
the single-stage S-S compensated IPT (SSIPT) converter at two quire deep modulation to satisfy the wide current range within
different operating frequencies for CC charging and CV charg- the CV charging stage of the battery charging profile. Thus, a
ing stages of the EV battery are first reviewed. The efficiency large variation of the phase angle between the driving voltage
of the SSIPT converter topology is then analyzed in terms of and the current of a compensated IPT transformer is needed,
the quality factor of Litz-wire windings. Since the system is making soft switching impossible if FM or PWM control is
operating at two fixed frequencies within the full range of an employed for the required load range [26]. In contrast, an IPT
EV battery charging profile, the proposed EV battery charging converter operating at its resonance, i.e., SSIPT converter oper-
method will simplify the control design and improve reliability. ating at LIC operating point, is found to be most power efficient
A nominal load quality factor Qn during the transition from CC [7], [10], [17]–[19], [22], making implementation of case (C3)
to CV operation is determined by optimizing the overall sys- less attractive than case (C5).
tem efficiency of the converter for the entire charging profile in In this paper, a single-stage SSIPT converter operating at
Section III. The analytical results are experimentally evaluated LIC for CC charging and at LIV for CV charging complying
in Section IV. Finally, Section V concludes the paper. with an EV battery profile [i.e., case (C5)] is designed and
optimized for efficiency and ease of control. The LIC operating
point of the SSIPT converter can provide maximum efficiency
II. CHARACTERISTICS OF THE SSIPT SYSTEM
at rated power for charging a significant part of the charging
The SSIPT converter has been extensively studied [7], [8], profile, while the LIV operation point of the SSIPT converter
[10]–[12], [17]–[23]. To eliminate the power loss associated can eliminate the loss due to control during CV charging of the

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
5810 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 66, NO. 7, JULY 2017

TABLE II
PARAMETERS FOR CALCULATION

ZS jωL S + j ω 1C + R S + R L
S
ZP jωL P + j ω 1C + R P
P
ω 2M 2
Zr ZS
v in
iP ZP +Zr
j ω M iP
io ZS
vo io R L
Fig. 1. S-S compensated IPT topology.

vo are calculated with the parameters given in Table II. The

ratio of output current io and input voltage vin is defined as
transconductance G, i.e.,

io jωM
G(ω) = = . (1)
vin ZP ZS + ω 2 M 2
Fig. 2. Equivalent circuit model of Fig. 1.
The ratio of output voltage vo and input voltage vin is defined
EV battery. A minimal amount of control is needed as the LIC as the voltage transfer ratio, E, i.e.,
and LIV operations naturally provide the required CC and CV
outputs, respectively. vo jωM RL
E(ω) = = . (2)
In this section, we highlight the LIC and LIV transfer char- vin ZP ZS + ω 2 M 2
acteristics of the SSIPT converter and analyze the efficiency of
the converter during LIC and LIV operations. The model used The primary resonant angular frequency ωP and the sec-
ignores the switching loss due to the transistor parasitic capaci- ondary resonant angular frequency ωS are defined as
tor, finite rise time, and finite fall time. The practical efficiency
1 1
degradation will be discussed in Section IV. ωP = √ and ωS = √ . (3)
LP CP LS CS
A. Circuit Topology and Equivalent Circuit Model
In previous applications, their ratio
The commonly used loosely coupled transformer model, as
shown in Fig. 1, for the SSIPT converter is adopted in this pa- ωP
μ= (4)
per [11], [12], [23]. Transformer inductances LP in the primary ωS
side, LS in the secondary side, and mutual inductance M are
components of the transformer model shown in Fig. 1. In the is normally set at unity by choosing external compensation ca-
circuit model, RP and RS are the winding resistances of the pacitors CP and CS , i.e., ωP = ωS . We will show in Section II-E
transformer primary and secondary, respectively. Also, CP and that a nonunity μ is necessary to achieve soft switching of the
CS are the primary and secondary external compensation capac- main switches during CC mode of operation.
itors for enhancing energy transfer from an ac source vin to an The characteristics of ideal transconductance and ideal volt-
output loading resistance RL . The ac source is generally taken as age transfer ratio, denoted as Gi and Ei , respectively, are ob-
an equivalent voltage generated from a half-bridge or full-bridge tained by assuming RP = RS = 0. Transfer functions Gi and
switching circuit operating at an angular frequency ω. Ei can be load independent at some operating frequencies. The
As usual, a frequency-domain equivalent circuit is adopted frequencies can be found by setting the coefficients of RL in (1)
and only the fundamental component is considered here for and (2) to zero [27]. The load-independent transconductance Gi
simplicity [3], [11], [12], [23], [24]. Discrepancies in practical can operate at ωP and its magnitude is determined as
applications will be discussed in Section IV.
1
Without introducing ambiguity, voltage and current valuables |Gi (ωP )| = √ . (5)
are considered as phasors for subsequent analyses. Fig. 2 shows ωP k LP LS
an equivalent circuit of Fig. 1 for steady-state analysis. The
Note that μ should be designed close to 1, i.e., ωP ≈ ωS .
dependent source jωM iS in Fig. 2 can be replaced by Zr ,
Otherwise, the converter efficiency described in Section II-C
which is an equivalent impedance reflected from the secondary
might suffer. Likewise, LIV ratio Ei is given by
side to the primary side. Thus, the primary loop is decoupled
from the secondary loop [11].   
LS  k(μ2 + 1 − Δ) 
|Ei (ωL )| = (6)
LP  (2k 2 − 1)μ2 + 1 − Δ 
B. Ideal Transconductance and Voltage Transfer Ratio   
We summarize the basic analysis of an SSIPT converter in this LS  k(μ2 + 1 + Δ) 
|Ei (ωH )| = (7)
subsection [11], [12]. The output current io and output voltage LP  (2k 2 − 1)μ2 + 1 + Δ 

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
HUANG et al.: DESIGN OF A SINGLE-STAGE INDUCTIVE-POWER-TRANSFER CONVERTER FOR EFFICIENT EV BATTERY CHARGING 5811


where Δ = (1 − μ2 )2 + 4k 2 μ2 , and the operating angular
frequencies are given by

μ2 + 1 − Δ
ωL = ωS (8)
2(1 − k 2 )

μ2 + 1 + Δ
ωH = ωS . (9)
2(1 − k 2 )
Hence, the SSIPT converter has an LIC output of |Gi (ωP )|
suitable for CC charging and a LIV output of |Ei (ωH , or ωL )|
suitable for CV charging of an EV battery. Operating just above
the frequency ωH can provide ZVS of the MOSFET main
switches. Thus, operating at and above ωH is usually preferred
over operating near ωL at LIV output of the SSIPT converter
[26], [27].

C. Efficiency and Control

It is commonly believed that the SSIPT converter should
operate at resonance frequency ωS , which is the LIC op-
erating point, for best converter efficiency at rated load
[7], [10], [17]–[19], [22]. However, a more detailed study has
revealed that the SSIPT converter operating at ωH can be more
efficient than the converter operating at ωS at light loads, while
the converter operating at ωH is less efficient than the converter
operating at ωS at rated and heavy loads [11]. This feature
will facilitate our implementation (C5) of a single-topology IPT
charger complying with the battery charging profile described Fig. 3. Comparison of efficiency η versus Q L operating at ω S (CC mode) and
in Section III. The efficiencies ηP of the primary side and ηS of ω H (CV mode) for various k by using the model of (a) constant IPT transformer
winding resistance and (b) constant IPT transformer winding quality factor.
the secondary side are calculated as
(Zr )
ηP = (10)
RP + (Zr ) ω = ωH (CV mode) are calculated using (12), as shown in
Fig. 3. In the calculation, we use QP m ax = QS m ax = QP =
RL
ηS = (11) QS = 100 and k = 0.3, 0.4, and 0.5. Fig. 3 shows that the peak
RS + RL efficiency in the CC operation appears at a higher QL than that of
where (Zr ) is the real component of Zr shown in Table II. the CV operation. This trend supports the implementation case
The overall efficiency of the SSIPT is given by (C5) of battery charging described in Section III. Hence, the
ω 2 k 2 L P L S (R S +R L ) location of QL has practical importance for the peak efficiency
(R S +R L ) 2 +X S2 RL and deserves further analysis.
η = ηP ηS = . (12)
RP + ω 2 k 2 L P L S (R S +R L ) RS + RL For constant RP and RS , the maximum efficiencies
(R S +R L ) 2 +X S2
ηm ax R (ωP ) and ηm ax R (ωH ) with μ = 1 can be calculated using
In general, resistances RP and RS of the primary windings (12), and the corresponding load quality factors can be approx-
and the secondary windings are frequency dependent. The re- imated as
sistances are normally represented as quality factors in resonant
circuits as follows: 1
QL CCR ≈ , for QP , QS  1 (16)
ωLP k
QP (ω) = (13) 
RP (ω) 1 1−k
QL CVR ≈ , for QS  1. (17)
ωLS k 1+ Q Q
P

QS (ω) = (14) S

RS (ω)
Likewise, for constant QP and QS , we have
ωS LS
QL = (15)
RL 1
QL CCQ ≈ , for QP m ax , QS m ax  1 (18)
where QL is the quality factor of a series compensated circuit k

with an equivalent loading resistance RL . 1 1−k
As an illustration, the efficiencies versus load quality fac- QL CVQ ≈ , for QS m ax  1. (19)
k 1+ Q P m ax
tor for μ = 1, operating frequencies ω = ωP (CC mode) and Q S m ax

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
5812 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 66, NO. 7, JULY 2017

Fig. 4. Comparison of Q L versus k at peak efficiency operating at ω S

(CC mode) and ω H (CV mode). Both models of constant IPT transformer
winding resistance and constant IPT transformer winding quality factor give
near identical result.

Fig. 6. Efficiency comparison of SSIPT converter operating at ω H (CV mode).

(a) Peak efficiency versus μ. (b) Efficiency versus Q L .

of an SSIPT converter in CV mode where the maximum effi-

ciency increases with decreasing μ, as shown in Fig. 6(a). More-
over, Fig. 6(b) shows the efficiency versus QL . When QL > 2,
the efficiency improves with decreasing μ, and when QL < 2,
the efficiency degrades slightly with decreasing μ. The over-
all degradation of efficiency for μ < 1 in the CV mode for the
normal load range is thus insignificant. The efficiency trend of
the converter designed with μ < 1 for the CC mode is plotted
in Fig. 7(a). The efficiency degrades significantly with decreas-
ing μ. The degradation of peak efficiency versus μ is shown in
Fig. 7(b). Therefore, in our design, μ will be restricted to a few
percent below 1 to facilitate soft switching during CC mode of
operation.
D. Practical Transconductance and Voltage Transfer Ratio
The operating frequencies for ideal load-independent
Fig. 5. Comparison of η m a x versus k operating at ω S (CC mode) and transconductance G and voltage transfer ratio E are studied
ω H (CV mode) using model of (a) constant IPT transformer winding resis-
tance and (b) constant IPT transformer winding quality factor. in Section II-B by assuming zero power loss. As the converter is
of high power efficiency, operating at the frequencies found in
Section II-B will still be subject to small variations of G and E
Using (16)–(19), plots of QL and ηm ax versus k at peak due to load variation. The practical load-independent transcon-
efficiency can be obtained, as shown in Figs. 4 and 5. The QL of ductance G(ωP ) can be found by substituting ωP into (1). The
the battery is high during CC charging, while it is low during CV normalized error Δg is defined as
charging. Therefore, the system efficiency for implementation
|Gi (ωP )| − |G(ωP )|
case (C5) will follow the solid curve during the CC mode of Δg =
charging and the dashed curve during the CV mode of charging, |Gi (ωP )|
 
as shown in Fig. 3. In Section III-A, the location Qn where the  μ2 k 2 
 
system switches from CC mode to CV mode will be chosen = 1 −  μ 2 −1  (20)
j + μ 2
+ μ
+ μ2 k2 
for optimizing the overall efficiency of the EV battery charging QP QP QS QP QL
profile. which can be simplified by putting μ ≈ 1, i.e.,
In Section II-E, μ will be designed smaller than 1 to facilitate
ZVS of MOSFET switches during CC mode of operation. In pre- 1
Δg(QL ) ≈ , for QS  QL . (21)
vious study [23], μ < 1 has been reported to improve efficiency 1+ k 2 QP QL

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
HUANG et al.: DESIGN OF A SINGLE-STAGE INDUCTIVE-POWER-TRANSFER CONVERTER FOR EFFICIENT EV BATTERY CHARGING 5813

Fig. 8. Plots of (a) Δg and (b) Δe versus Q L .

Fig. 7. Efficiency comparison of SSIPT converter operating at ω P (CC mode).
(a) Efficiency versus Q L . (b) Peak efficiency versus μ. within a few percent, permitting soft switching by using phase
shift PWM control to regulate the desired output current and
voltage of the IPT converter.
The practical LIV transfer ratio E(ωH ) can be found by
substituting ωH into (2). The normalized error Δe(ωH ) can be E. Input Phase Angle and Soft Switching
defined as
  Inductive input phase angle is important for the MOSFET
|Ei (ωH )| − |E(ωH )|  E1 
Δe = =1−   (22) main switches to achieve zero voltage turn-on. The input
|Ei (ωH )| E2 − E 3  impedance of the equivalent circuit model shown in Fig. 2 with
where parameters in Table II is given by
ωH
ν= (23) Zin = ZP + Zr . (28)
ωS
 2  The corresponding input phase angle is given by
ν jν
E1 = − 2 + 1 −ν 2 + 1 + − E3 (24)  
μ QL (Zin )
 2  θin = tan−1 . (29)
ν ν2 (Zin )
E2 = − 2 + 1 + j 2 E4 (25)
μ μ QP To achieve zero voltage turn-on of MOSFET switches, induc-
tive input impedance is expected for both CC and CV modes of
ν4 2
E3 = k (26) operation. For the CC mode, the input phase angle is given by
μ2

  −k 2 1 − μ12
−1 
E4 = −ν 2 + 1 + jν
ν
+
1
. (27) θC C = tan (30)
1
A2 + (1 − 1 )2 + k 2 A
QS QL QP 1 μ 2 1

From (20) and (22), Δg increases with decreasing QL while   

1
Δe increases with increasing QL , i.e., G(ωP ) and E(ωH ) de- = tan−1 QL −μ , for QP ≈ QS  QL (31)
μ
crease with increasing load. For illustration, Fig. 8(a) shows the
percentage error on Δg and Δe versus QL . The load quality where A1 = ( Q1S + μQ1 L ). Fig. 9(a) shows the input phase angle
factor QL CCR or QL CCQ to achieve maximum efficiency at a θCC versus QL in CC mode. When μ = 1, θCC is exactly zero,
particular k in CC mode is shown in Fig. 4. The value of QL which leaves no room for PWM control with soft switching.
read from Fig. 4 can be used to estimate the errors of Δg and Δe When μ is designed to be slightly smaller than 1, positive input
using Fig. 8. The input voltage variation of the prototype IPT phase angle can be achieved to have PWM control with soft
converter described in Section IV can thus be designed to be switching.

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
5814 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 66, NO. 7, JULY 2017

TABLE III
CHARGING PARAMETERS

Time (h) [point in Fig. 11] Current (A) Voltage (V)

0 [a] yC 3x
0.125 [b] yC 3.725x
0.375 [c] yC 4x
0.625 [d] yC 4x
0.875 [e] yC 4.2x
1.25 [f] 0.6y C 4.2x
1.75 [g] 0.27y C 4.2x
2.25 [h] 0.133y C 4.2x
2.75 [i] 0.05y C 4.2x

control. The fundamental component of vin is modeled as

4 θ 4
vin = UIN cos or vin = UIN D (34)
π 2 π
where UIN is the dc input volatge, θ ∈ (0, π) is the phase shift
angle of the H-bridge, and D = cos θ/2 ∈ [0, 1] is the equiv-
alent duty cycle. The driving signals are generated by a DSP
controller. Information of the charging voltage and current in
the secondary side is collected and transmitted wirelessly to the
controller at the primary side.
Fig. 9. Input phase angle versus Q L for (a) CC mode and (b) CV mode of
operation. III. DESIGN CONSIDERATIONS
A. Maximizing Efficiency
For the CV mode, the input phase angle is given by

v2   2
Suppose a battery pack consists of y parallel-connected bat-
−1 μ 2
− 1 A22 + A23 − μv 2 k 2 A3 teries, and each battery consists of x series-connected cells [28].
θCV = tan  2  v2 (32) The nominal values of voltage, current, and resistance of a bat-
μ 2 Q P A2 + A3 + μ 2 k A2
v2 2 2
tery pack are given by
where A2 = ( Q1S + v Q1 L ) and A3 = (1 − v12 ). Fig. 9(b) shows
Un = 4.2 × V (35)
the input phase angle θCV versus QL in CV mode, where the
positive input phase angle guarantees PWM control with soft In = yC A (36)
switching.
Un 4.2x
Rn = = Ω (37)
In yC
F. Control Scheme
where In is the current at CC charging, Un is the voltage at CV
The main circuit of the SSIPT system for EV battery charging
charging, Rn is the dc nominal resistance at the point of switch
with a control scheme is illustrated in Fig. 10. The dc voltage UIN
over from CC to CV charging, and C is the maximum current
is modulated as a high-frequency ac voltage vin by the set of H-
the battery can supply for 1 h. Depending on the values of x and
bridge power MOSFET switches that drive the transmitting coil
y, different battery packs have different specifications.
through a series compensation network. Since battery charging
For simplicity of calculating the averaged charging effi-
is a slow process for the converter, the battery can be modeled
ciency, the battery charging profile is approximated by several
as a resistor Rbattery , which indicates the amount of power
piecewise-linear segments, as shown in Fig. 11(a). From the
acquired from the charger and varies slowly according to the
charging parameters given in Table III, the equivalent dc resis-
battery charging profile. It should be noted that Rbattery has no
tance (Ω) of the battery Rbattery ranges from 0.714Rn to 20Rn ,
relation to the battery internal resistance, which decreases with
i.e., from y3xC to 0.05y
4.2x
C , as shown in Fig. 11(b). We define the
increasing state of charge of the battery. The rectifier with the
load quality factor at the point of switch-over from CC to CV
battery can be modeled as an equivalent resistor RL , i.e.,
charging as Qn , which is ω8S RL S according to (15), (33), and
n
π2
8
RL = Rbattery . (33) (37), and this gives a charging profile ranging from 0.05Qn to
π2 1.4Qn , as shown in Fig. 11(b).
The system will operate at ωP , which is slightly lower than The equivalent resistance of the battery varies with time
ωS , to provide a CC and at ωH to provide a CV. Since errors are within the whole charging profile. Therefore, the charging ef-
inevitable in the output current and voltage, the input voltage ficiency varies with time. From Figs. 3 and 11(b), the choice
is regulated by an H-bridge inverter using a phase-shift PWM of Qn affects the charging efficiency versus time as shown in

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
HUANG et al.: DESIGN OF A SINGLE-STAGE INDUCTIVE-POWER-TRANSFER CONVERTER FOR EFFICIENT EV BATTERY CHARGING 5815

Fig. 10. Main circuit and control scheme.

Fig. 12. Charging efficiency η profile for different values of Q n . Overall

efficiency should be maximum at a particular Q n denoted as Q n , o .

Fig. 11. (a) Piecewise-linear battery charging profile. (b) Piecewise-linear

resistance of the battery during the whole charging profile.

Fig. 12. The optimization of the converter efficiency for the

whole charging profile is thus simplified to choosing Qn (or Fig. 13. Design plot of Q n , o versus k (red), and corresponding max(η ove ra ll )
(blue), showing that max(η ove ra ll ) increases with increasing k, Q P ,
ωS LS since Rn is fixed) that maximizes the charging efficiency and/or Q S .
for the charging profile, i.e.,
outputenergy
ηoverall = max(ηoverall ). Interestingly, there is no observable change in
inputenergy
T Qn ,o for a range of QP and QS from 10 to 5000. Therefore, it
UC (t)IC (t)dt is safe to omit QP and QS in obtaining Qn ,o . To facilitate the
=  T 1 U (t)I0 (t)  T U C (t)I C (t) (38)
design, Fig. 13 presents Qn ,o versus k graphically. In Fig. 13,
0 η CC (Q L (t)) dt + T 1 η CV (Q L (t)) dt
C C

the red curve plots Qn ,o versus k, showing that Qn ,o is not a

where T1 is the CC charging time, T is the total charging time, function of QP or QS . Moreover, when comparing Qn ,o with
and QL (T1 ) = Qn . QL CC ’s in (16) and (18), Qn ,o is a bit higher than k1 for all
We have done extensive numerical calculations to ob- values of k. The blue curves show max(ηoverall ) versus k for
tain an optimum Qn ,o that corresponds to ηoverall (Qn ,o ) = different values of QP = QS . From Fig. 13, a higher overall

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
5816 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 66, NO. 7, JULY 2017

C. Converter Design
Given an input voltage UIN , the converter as shown in Fig. 10
provides the required charging current In and voltage Un ac-
cording to (35), (36), and the battery profile shown in Fig. 11.
Here, QL varies from 1.4Qn to Qn during CC mode of charging
Fig. 14. Circular unipolar coupled transformer with constant inner and outer and from Qn to 0.05Qn during CV mode of charging. There-
radii and evenly distributed wire distance for (a) N = 10, (b) N = 20, and fore, the design should satisfy (5) and (7), which are practically
(c) N = 40. equivalent to

π 2 In

= |Gi (ωP )| 1 − Δg(QL )
8Di (QL )UIN
1 − Δg(QL )
= √ (39)
ωP k LP LS

Un

= |Ei (ωH )| 1 − Δe(QL )
Dv (QL )UIN

Fig. 15. Coupling coefficient k and inductance L S (or L P ) versus number of
 LS
turns N . = ΔH 1 − Δe(QL ) (40)
LP

 
 +1+Δ ) 
where ΔH =  (2kk2(μ
2
efficiency can be achieved by using a transformer with higher
−1)μ 2 +1+Δ  is a function of k and μ only
k, QP and/or QS .
and duty cycles Di and Dv varying with QL are given in (34)
for operation with LIC and LIV outputs, respectively. At rated
B. Loosely Coupled Transformer power QL = Qn , which corresponds to switching of CC mode
The loosely coupled transformer for stationary EV charg- to CV mode, both (39) and (40) should be satisfied, giving
ing can be designed with a circular pad, a double-D pad, a
double-D quadrature pad, or a bipolar pad [29], [30]. The pop- 
π 2 In LS
ular primary and secondary circular pad structures shown in Qn = (41)
8 Un CS
Fig. 14 will be adopted. In this paper, the circular pads of equal  
size have an outer diameter of do , inner diameter of di , and 1 Di (Qn ) 1 − Δg(Qn )
= (42)
a separation gap of h. For a given structure, a higher k can kμ Dv (Qn ) ΔH (1 − Δe(Qn ))
be achieved with a larger dho and/or a larger ferrite section
area [29]. According to the overall charging efficiency indi- where Qn should be designed close to Qn ,o , as obtained from
cated in Fig. 13, using a larger pad diameter and/or better mag- (38) for the maximum overall charging efficiency (see Fig. 13).
netic and/or conducting material, a higher ηoverall (Qn ,o ) can be As Rn = UI nn in (41) is fixed for a given battery, the converter
achieved.
can be designed with a suitable value of CL SS for achieving
In Fig. 14, the simplified pad has two layers. The top layer
Qn = Qn ,o . According to the simulated results shown in Fig. 13,
contains the coil that generates magnetic field and the second
max(ηoverall ) increases with k at a reducing rate (saturates as k
layer contains the ferrite to reduce the reluctance. The two pads
becomes large), and from Fig. 15, k also increases with LS at a
are arranged with a magnetic linking path h. It is assumed that
reducing rate. Hence, increasing LS will offer diminishing re-
the secondary pad is attached to the underside of an EV, while
turn of max(ηoverall ). We may therefore use X = ΔΔNk = 0.001
the primary pad is buried under the ground. Once an EV has
as an indicator for choosing an initial value of N or LS .
stopped over the charging system, power is transferred across
Other indicators may also be adopted for design [29], [30].
the air gap via magnetic coupling from the primary pad to the
The value of ωS = √L 1 C obtained in (41) should be verified
secondary pad. S S

The structure and the dimension of the loosely coupled trans- as being within the efficient operating range of the magnetics,
former are usually designed according to some expected ranges switches, etc. Otherwise, a better choice of N or LS should be
of k, LP , and LS . We have performed Ansoft Maxwell simu- used.
lations using the transformer structures shown in Fig. 14 with Duty cycles Di (Qn ) and Dv (Qn ) at rated power loaded by
do = 500 mm, di = 100 mm, h = 100 mm, and the number of Qn should be designed close to 1 for best efficiency. For the
turns N = NP = NS of the coils varying from 5 to 40. Design reasons analyzed in Section II, the LIC operating point has a
curves shown in Fig. 15 show how k and LS vary with N , which much tighter tolerance than the LIV operating point for soft
are consistent with the results shown in [31]. switching implementation. Therefore, Di (Qn ) should be given

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
HUANG et al.: DESIGN OF A SINGLE-STAGE INDUCTIVE-POWER-TRANSFER CONVERTER FOR EFFICIENT EV BATTERY CHARGING 5817

TABLE IV
DESIGN SPECIFICATION

Name Parameter Value

Input voltage U IN ≈190 V

Nominal charging voltage Un 175 V
Nominal charging current In 6.4 A
Air gap distance h 100 mm
Transformer outer diameter do 500 mm

TABLE V
COMPARISON OF CALCULATED AND MEASURED TRANSFORMER AND
CIRCUIT PARAMETERS
Fig. 16. Experimental prototype of the IPT system.

priority and set as 1. Thus, from (39) and (40), we have

In 8 1 − Δg(Qn )
= 2 √ (43)
UIN π ωP k LP LS

Un
 LS
= Dv (Qn )ΔH 1 − Δe(Qn ) . (44)
UIN LP
The values of duty cycles Di (1.4Qn ) at the beginning of
the CC charging mode and Dv (0.05Qn ) at the end of the CV
charging mode are thus given by
1 − Δg(Qn )
Di (1.4Qn ) = (45)
1 − Δg(1.4Qn )
1 − Δe(Qn )
Dv (0.05Qn ) = Dv (Qn ) . (46)
1 − Δe(0.05Qn )
According to Fig. 7(b), the implementation of soft switch-
ing by using μ < 1 in CC mode comes with an efficiency
penalty, as studied in Section II-C. In practice, μ can be as-
signed as 0.96. Using (30), θCC (Qn , μ) and θCC (1.4Qn , μ) can
be obtained which should be larger than 2 cos−1 {Di (Qn )}
and 2 cos−1 {Di (1.4Qn )}, respectively. Otherwise, a smaller
μ should be assigned with additional efficiency penalty. Thus,
soft switching during CC mode of charging is guaranteed. Using
(32), θCV (Qn ) and θCV (0.05Qn ) can be obtained, which should
be larger than 2 cos−1 {Dv (Qn )} and 2 cos−1 {Dv (0.05Qn )}
respectively. Thus, soft switching during CV mode can also
be guaranteed. A more detailed illustration will be given in rameters will be used for the subsequent design. The winding
Section IV. resistances are measured near the calculated operating frequen-
cies. i.e., RP ω = 293 mΩ and RS ω = 298 mΩ at fP = ω2πP , and
IV. EXPERIMENTAL EVALUATION RP ω = 379 mΩ and RS ω = 378 mΩ at fH = ω2πH . These give
the winding quality factors of approximately 174 at 50 kHz and
A. Experiment
185 at 68.5 kHz. The equivalent resistance RP for the model can
Two prototypes of an 1.5-kW IPT system with and without be obtained as RP = RP ω + 2Ron [23], which gives QP = 136
soft switching during CC mode of operation for EV battery at fP = 50 kHz and 152 at fH = 68.5 kHz. In contrast, QS is
charging, as shown in Fig. 16, are built with design speci- determined by the windings resistance only. Moreover, the loss
fications given in Table IV. Switching devices used are In- due to the rectifier diodes can be modeled as a voltage source of
fineon CoolMOS IPW60R199CP with Ron = 0.041 Ω. Rec- 2VF connected in series with the battery. Essentially, the quality
tifier devices are STMicroelectronics STTH60AC06C with factors of this transformer stay between the models of constant
VF = 0.8 V at rated power. The loosely coupled transformer resistance and constant quality factor described in Section II-C.
is constructed according to Section III-B using Litz wire With the parameters k, QP , and QS measured, QL CC can be
AWG38 with NP : NS = 20 : 20. The calculated and measured obtained from (16) or (18), and QL CV can be obtained from (17)
transformer parameters are shown in Table V. The measured pa- or (19).

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
5818 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 66, NO. 7, JULY 2017

Fig. 17. Waveforms of v A B , iP , and IO at (a) start and (b) end of CC charging Fig. 19. Waveforms of v A B , iP , and IO at (a) start and (b) end of CV mode
mode of the IPT converter designed with μ = 1. of charging of the IPT converter designed with μ = 1.

Fig. 20. Waveforms of v A B , iP , and IO at (a) start and (b) end of CV mode
Fig. 18. Waveforms of v A B , iP , and IO at (a) start and (b) end of CC charging of charging of the IPT converter designed with μ = 0.96.
mode of the IPT converter designed with μ = 0.96.

Two prototype converters are designed with μ = 0.96 and

1. Since the physical transformers have k = 0.447, Qn ,o can
be read from Fig. 13 as 2.4. From (41), CS is calculated as
59.04 nF for the two converters. Compensation capacitors CS
and CP are measured. Their values are shown in Table V. Using
Δg and Δe determined from (20) and (22), Di (Qn ,o ) can be
assigned as 1 and Dv (Qn ,o ) can be obtained as shown in Table V
according to (42).
The rated output voltage Un and current In can be scaled up
or down with a suitable value of UIN by using (21), (22) and
(43), (44). The value of Uin,m in can be uniquely determined Fig. 21. Measured charging efficiency η versus load quality factor Q L .
in this design. However, if (43) and (44) give two significantly
distinct values of Uin,m in for CC and CV charging, the de-
sign should be reiterated with some appropriated values of LP corresponding waveforms are shown inFig. 18(a) and (b), which
and LS . clearly show ZVS operation.
Using (30), inequality θC C > 2 cos−1 {Di (Qn )} is checked
for Qn varying from Qn ,o to 1.4Qn ,o . Hence, soft switching C. Soft Switching in LIV Operation for CV Charging
during LIC operation is guaranteed for the converter designed
with μ = 0.96. Using (32), inequality θCV > 2 cos−1 {Dv } is Using the prototype IPT converter designed with μ = 1,
verified for Qn varying from 0.005Qn ,o to Qn ,o . Hence, soft waveforms of the converter at the start and end of the CC charg-
switching during LIV operation is also guaranteed for the con- ing stage are shown in Fig. 19(a) and (b), respectively. Also,
verter designed with μ = 0.96. The values calculated are shown for μ = 0.96, corresponding waveforms are shown in Fig. 20(a)
in Table V, where it is observed that the converter with μ = 1 and (b). All waveforms of LIV operation show ZVS.
does not have soft switching during LIC output operation.
D. Efficiency
A comparison of the calculated and measured efficiencies
B. Soft Switching in LIC Operation for CC Charging of the converter with μ = 0.96 for the load quality factor
Using the prototype IPT converter designed with μ = 1, wave- varying from 0.11 to 3.08 of the battery charging profile is
forms of the converter at the start and end of the CC charging shown in Fig. 21. The efficiencies have been measured using
stage are shown in Fig. 17(a) and (b), respectively. It can be a two-channel Voltech PM100 power analyzer, one channel for
observed that soft switching is still achievable in Fig. 17(b), the dc input power and the other channel for the dc output
but Fig. 17(a) shows hard switching. Moreover, with μ = 0.96, power. Theoretical calculations are shown in solid curve for CC

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
HUANG et al.: DESIGN OF A SINGLE-STAGE INDUCTIVE-POWER-TRANSFER CONVERTER FOR EFFICIENT EV BATTERY CHARGING 5819

Fig. 22. Measured charging efficiency η versus time t.

output and dashed curve for CV output. Practical measurements

are marked with “◦” for CC mode and with “ ” for CV mode.
The reasons for the practical efficiency being lower than the
theoretical efficiency for both LIC and LIV operations are the
following.
1) Turn-on switching loss is not completely eliminated after
employing ZVS.
2) Turn-off switching loss is not considered in the model.
3) The output rectifying diode-bridge has nonlinearity, which
cannot be fully represented by a single loading resistor
RL .
4) The model only considers the fundamental frequency
component and omits all other components.
5) The rectifying diodes have forward voltage VF varying
with the load, which has not been considered in the loss
calculation.
The experimental charging efficiency versus time of the two
prototype converters is shown in Fig. 22. The converter with μ =
0.96 achieving ZVS gives a higher charging efficiency (marked
with “◦”), while the hard switching converter with μ = 1 gives
Fig. 23. Loss component estimation. (a) Illustrative voltage and current wave-
lower charging efficiency (marked with “”) in CC charging forms at the two phase legs of the H-bridge. (b) Measured charging efficiency
mode. When the LIC output converter is controlled to operate in η versus duty cycle D with LIC operation with μ = 1. The LIC output is reg-
CV mode, the measured efficiency (marked with “
”) decreases ulated by varying D to provide the required constant output current for CC
mode and the required constant output voltage for CV mode. (c) Estimated loss
rapidly due to the required fast decrease of duty cycle at lighter components versus D, where P S w itch = P S w itch O n + P S w itch O ff .
load conditions and the converter ends up with hard switching.
In Fig. 22, only four hard switching “
” data points are shown
for the LIC operation providing a constant output voltage. The converter with μ = 1 is given in Fig. 23(b). Since the duty
measured efficiency points (marked with “ ”) are the same data cycles required for CV output by using the LIC operating point
points appeared in Fig. 21. There is no observable change in are well below the ZVS limit of D = 0.995 for the prototype
measured efficiency of the LIV output for the two prototype converter with μ = 0.96, a converter with a much smaller
converters. μ than 0.96 is needed to achieve soft switching. Fig. 23(c)
The losses in hard switching can be analyzed in terms of shows the loss components versus D, which verifies that
the typical square voltage waveform vin after modulation the additional losses, i.e., PSW + PRR , due to hard switching
and the current waveform iP of the H-bridge inverter, as increases significantly with decreasing D. In contrary, the
shown in Fig. 23(a). The switching loss and diode reverse efficiency gained from soft switching is unable to counteract
recovery loss during switching are estimated assuming a the extra loss incurred by using a much smaller μ, as explained
phase angle of θ2 . Power losses included in the calculation in Section II-C. Hence, the converter operating with LIC output
are switching loss PSwitch , diode reverse recovery loss PRR , has a narrow range of control by using PWM modulation and
conduction loss PCond , primary winding loss PP , secondary is unsuitable for use as a CV output converter.
winding loss PS , and forward voltage drop loss PF of the As a final remark, the converter operating efficiently with
rectifier diodes [32]. The power efficiency is thus estimated as LIC output has a tight modulation margin which is only good for
η = P O +P S w i t c h +P CondP+P
O
RR +P P +P S +P F
. A comparison of the achieving soft switching against load variation. This supports the
calculated and previously measured efficiency of the prototype view that implementation cases (C2) and (C4) of battery charg-

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
5820 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 66, NO. 7, JULY 2017

[8] A. J. Moradewicz and M. P. Kazmierkowski, “Contactless energy trans-

fer system with FPGA-controlled resonant converter,” IEEE Trans. Ind.
Electron., vol. 57, no. 9, pp. 3181–3190, Sep. 2010.
[9] S. Valtchev, B. Borges, K. Brandisky, and J. Klaassens, “Resonant con-
tactless energy transfer with improved efficiency,” IEEE Trans. Power
Electron., vol. 24, no. 3, pp. 685–699, Mar. 2009.
[10] J. Shin et al., “Design and implementation of shaped magnetic-resonance-
based wireless power transfer system for roadway-powered moving elec-
tric vehicles,” IEEE Trans. Ind. Electron., vol. 61, no. 3, pp. 1179–1192,
Mar. 2014.
[11] W. Zhang, S.-C. Wong, C. K. Tse, and Q. Chen, “Analysis and comparison
of secondary series- and parallel-compensated inductive power transfer
systems operating for optimal efficiency and load-independent voltage-
transfer ratio,” IEEE Trans. Power Electron., vol. 29, no. 6, pp. 2979–2990,
Jun. 2014.
Fig. 24. Measured charging profiles of voltage U C and current IC . [12] W. Zhang, S.-C. Wong, C. K. Tse, and Q. Chen, “Load-independent duality
of current and voltage outputs of a series or parallel compensated inductive
power transfer converter with optimized efficiency,” IEEE J. Emerg. Sel.
ing cannot achieve high efficiency, as described in Section II. Topics Power Electron., vol. 3, no. 1, pp. 137–146, Mar. 2015.
[13] A. Khaligh and Z. Li, “Battery, ultracapacitor, fuel cell, and hybrid energy
Fortunately, the charging current needs not be precise [15], and storage systems for electric, hybrid electric, fuel cell, and plug-in hybrid
thus, as long as the current rating of the battery has been taken electric vehicles: State of the art,” IEEE Trans. Veh. Technol., vol. 59,
care of, the converter can be left uncontrolled at its maximum no. 6, pp. 2806–2814, Jul. 2010.
[14] J. Deng, S. Li, S. Hu, C. C. Mi, and R. Ma, “Design methodology of LLC
duty cycle. If a precise output current is needed against input resonant converters for electric vehicle battery chargers,” IEEE Trans. Veh.
voltage variation, a fixed-frequency ON–OFF control can be uti- Technol., vol. 63, no. 4, pp. 1581–1592, May 2014.
lized for this converter [33]. Finally, the experimental charging [15] H. J. Bergveld, W. S. Kruijt, and P. H. Notten, “Battery charging algo-
rithms,” in Battery Management Systems. New York, NY, USA: Springer,
voltage and charging current versus time for the ZVS converter 2002, pp. 169–192.
with μ = 0.96 are shown in Fig. 24. The overall experimental [16] M. K. Kazimierczuk and D. Czarkowski, Resonant Power Converters.
charging efficiency is found to be 94.5%. Hoboken, NJ, USA: Wiley, 2011.
[17] M. Fu, H. Yin, X. Zhu, and C. Ma, “Analysis and tracking of optimal load
in wireless power transfer systems,” IEEE Trans. Power Electron., vol. 30,
no. 7, pp. 3952–3963, Jul. 2015.
V. CONCLUSION [18] R. Bosshard, J. W. Kolar, J. Muehlethaler, I. Stevanovic, B. Wunsch, and
F. Canales, “Modeling and η-α-pareto optimization of inductive power
An IPT EV charging system, which is based on a single transfer coils for electric vehicles,” IEEE J. Emerg. Sel. Topics Power
compensated topology and two fixed operating frequencies, has Electron., vol. 3, no. 1, pp. 50–64, Mar. 2015.
[19] W. X. Zhong and S. Y. R. Hui, “Maximum energy efficiency tracking for
been described in this paper. The S-S capacitor compensation wireless power transfer systems,” IEEE Trans. Power Electron., vol. 30,
topology has the characteristics of LIC output and LIV output at no. 7, pp. 4025–4034, Jul. 2015.
two different operating frequencies, which are suitable for CC [20] X. Qu, H. Han, S.-C. Wong, C. K. Tse, and W. Chen, “Hybrid IPT
topologies with constant-current or constant-voltage output for battery
charging and CV charging of the EV battery, respectively. Anal- charging applications,” IEEE Trans. Power Electron., vol. 30, no. 11,
ysis and design for efficiency optimization have been studied in pp. 6239–6337, Nov. 2015.
depth, and experimental evaluation of a 1.5-kW IPT EV battery [21] C. Zheng et al., “High-efficiency contactless power transfer system for
electric vehicle battery charging application,” IEEE J. Emerg. Sel. Topics
charger has been reported. Power Electron., vol. 3, no. 1, pp. 65–74, Mar. 2015.
[22] H. Li, J. Li, K. Wang, W. Chen, and X. Yang, “A maximum efficiency
point tracking control scheme for wireless power transfer systems using
REFERENCES magnetic resonant coupling,” IEEE Trans. Power Electron., vol. 30, no. 7,
pp. 3998–4008, Jul. 2015.
[1] J. G. Hayes, M. G. Egan, J. M. D. Murphy, S. E. Schulz, and J. T. Hall, [23] W. Zhang, S.-C. Wong, C. Tse, and Q. Chen, “Design for efficiency
“Wide-load-range resonant converter supplying the SAE J-1773 electric optimization and voltage controllability of series–series compensated in-
vehicle inductive charging interface,” IEEE Trans. Ind. Appl., vol. 35, ductive power transfer systems,” IEEE Trans. Power Electron., vol. 29,
no. 4, pp. 884–895, Jul. 1999. no. 1, pp. 191–200, Jan. 2014.
[2] C. S. Wang, O. H. Stielau, and G. A. Covic, “Design considerations for a [24] C. S. Wang, G. A. Covic, and O. H. Stielau, “Power transfer capability and
contactless electric vehicle battery charger,” IEEE Trans. Ind. Electron., bifurcation phenomena of loosely coupled inductive power transfer sys-
vol. 52, no. 5, pp. 1308–1314, Oct. 2005. tems,” IEEE Trans. Ind. Electron., vol. 51, no. 1, pp. 148–157, Feb. 2004.
[3] G. A. Covic, J. T. Boys, M. L. G. Kissin, and H. G. Lu, “A three-phase [25] S. Li and C. C. Mi, “Wireless power transfer for electric vehicle applica-
inductive power transfer system for roadway-powered vehicles,” IEEE tions,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 3, no. 1, pp. 4–17,
Trans. Ind. Electron., vol. 54, no. 6, pp. 3370–3378, Dec. 2007. Mar. 2015.
[4] U. K. Madawala and D. J. Thrimawithana, “A bidirectional inductive [26] G. B. Joung and B. H. Cho, “An energy transmission system for an
power interface for electric vehicles in V2G systems,” IEEE Trans. Ind. artificial heart using leakage inductance compensation of transcutaneous
Electron., vol. 58, no. 10, pp. 4789–4796, Oct. 2011. transformers,” IEEE Trans. Power Electron., vol. 13, no. 6, pp. 1013–1022,
[5] H. Sakamoto, K. Harada, S. Washimiya, K. Takehara, Y. Matsuo, and Nov. 1998.
F. Nakao, “Large air-gap coupler for inductive charger [for electric vehi- [27] Q. Chen, S. C. Wong, C. K. Tse, and X. Ruan, “Analysis, design and
cles],” IEEE Trans. Magn., vol. 35, no. 5, pp. 3526–3528, Sep. 1999. control of a transcutaneous power regulator for artificial hearts,” IEEE
[6] C. Duan, C. Jiang, A. Taylor, and K. Bai, “Design of a zero-voltage- Trans. Biomed. Circuits Syst., vol. 3, no. 1, pp. 23–31, Feb. 2009.
switching large-air-gap wireless charger with low electrical stress for [28] A. Affanni, A. Bellini, G. Franceschini, P. Guglielmi, and C. Tassoni,
plugin hybrid electric vehicles,” in Proc., IEEE Transp. Electrific. Conf. “Battery choice and management for new-generation electric vehicles,”
Expo., Jun. 2013, pp. 1–5. IEEE Trans. Ind. Electron., vol. 52, no. 5, pp. 1343–1349, Oct. 2005.
[7] J. Sallan, J. L. Villa, A. Llombart, and J. F. Sanz, “Optimal design of [29] M. Budhia, G. A. Covic, and J. T. Boys, “Design and optimization of
ICPT systems applied to electric vehicle battery charge,” IEEE Trans. Ind. circular magnetic structures for lumped inductive power transfer systems,”
Electron., vol. 56, no. 6, pp. 2140–2149, Jun. 2009. IEEE Trans. Power Electron., vol. 26, no. 11, pp. 3096–3108, Nov. 2011.

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.
HUANG et al.: DESIGN OF A SINGLE-STAGE INDUCTIVE-POWER-TRANSFER CONVERTER FOR EFFICIENT EV BATTERY CHARGING 5821

[30] A. Zaheer, H. Hao, G. Covic, and D. Kacprzak, “Investigation of multiple Chi K. Tse (M’90–SM’97–F’06) received the B.Eng.
decoupled coil primary pad topologies in lumped IPT systems for inter- (Hons.) degree (with first class honors) in electrical
operable electric vehicle charging,” IEEE Trans. Power Electron., vol. 30, engineering and Ph.D. degree from the University of
no. 4, pp. 1937–1955, Apr. 2015. Melbourne, Melbourne, Australia, in 1987 and 1991,
[31] C. Fernández, O. Garcia, R. Prieto, J. Cobos, S. Gabriels, and G. Van respectively.
Der Borght, “Design issues of a core-less transformer for a contact-less He is currently a Chair Professor with the Hong
application,” in Proc. 17th Annu. IEEE Appl. Power Electron. Conf. Expo., Kong Polytechnic University, Hong Kong, where he
2002, vol. 1, pp. 339–345. was the Head of the Department of Electronic and
[32] J. Klein, “Synchronous buck MOSFET loss calculations with Excel Information Engineering from 2005 to 2012. He is
model,” Fairchild Semiconductor, San Jose, CA, USA, Appl. Note the author/coauthor of ten books, 20 book chapters,
AN6005, 2006. and more than 500 papers in research journals and
[33] Y. S. Lee and Y. C. Cheng, “Design of switching regulator with combined conference proceedings and holds five U.S. patents.
FM and on-off control,” IEEE Trans. Aerosp. Electron. Syst., vol. AES-22, Dr. Tse has received a number of research and industry awards, including
no. 6, pp. 725–731, Nov. 1986. Prize Paper Awards by the IEEE TRANSACTIONS ON POWER ELECTRONICS in
2001 and 2015, the Best Paper Award by the International Journal of Circuit
Theory and Applications in 2003, RISP Journal of Signal Processing Best Paper
Award in 2014, two Gold Medals at the International Inventions Exhibition in
Zhicong Huang (S’14) received the B.Sc. degree in Geneva, Switzerland, in 2009 and 2013, and a number of recognitions by the
electrical engineering and automation and M.Sc. de- academic and research communities, including honorary professorship by sev-
gree in mechanical and electronic engineering from eral Chinese and Australian universities, Chang Jiang Scholar Chair Professor-
Huazhong University of Science and Technology, ship, IEEE Distinguished Lectureship, Distinguished Research Fellowship by
Wuhan, China, in 2010 and 2013, respectively. He is the University of Calgary, Gledden Fellowship and International Distinguished
currently working toward the Ph.D. degree in power Professorship-at-Large by the University of Western Australia. While with the
electronics at the Hong Kong Polytechnic University, Hong Kong Polytechnic University, he received the President’s Award for Out-
Kowloon, Hong Kong. standing Research Performance twice, the Faculty Research Grant Achievement
His current research interests include wireless Award twice, the Faculty Best Researcher Award, and several teaching awards.
power transfer and power electronics. He serves/has served as the Editor-in-Chief for the IEEE TRANSACTIONS ON CIR-
CUITS AND SYSTEMS II: EXPRESS BRIEFS (2016–2017), IEEE Circuits and Sys-
tems Magazine (2012–2015), and IEEE Circuits and Systems Society Newsletter
(since 2007), an Associate Editor for three IEEE Journal/Transactions, an Editor
for the International Journal of Circuit Theory and Applications, and is on the
Siu-Chung Wong (M’01–SM’09) received the B.Sc. editorial boards of a few other journals.
degree in physics from the University of Hong Kong,
Hong Kong, in 1986, the M.Phil. degree in electron-
ics from the Chinese University of Hong Kong, Hong
Kong, in 1989, and the Ph.D. degree from the Univer-
sity of Southampton, Southampton, U.K., in 1997.
He joined the Hong Kong Polytechnic in 1988 as
an Assistant Lecturer. He is currently an Associate
Professor with the Department of Electronic and In-
formation Engineering, The Hong Kong Polytechnic
University, Hong Kong, where he conducts research
in power electronics. From 2012 to 2015, he was appointed as a Chutian Scholar
Chair Professor by the Hubei Provincial Department of Education, China, and
the appointment was hosted by Wuhan University of Science and Technology,
Wuhan, China. In 2013, he was appointed as Guest Professor by the School
of Electrical Engineering, Southeast University, Nanjing, China. He was a Vis-
iting Scholar with the Center for Power Electronics Systems, Virginia Tech,
Blacksburg, VA, USA, in November 2008, Aero-Power Sci-tech Center, Nan-
jing University of Aeronautics and Astronautics, Nanjing, in January 2009, and
the School of Electrical Engineering, Southeast University, in March 2012.
Dr. Wong is a member of the Electrical College, The Institution of Engi-
neers, Australia. He is an editor of the Energy and Power Engineering journal
and a member of the Editorial Board of the Journal of Electrical and Control
Engineering. He serves as a Guest Associate Editor for the IEEE JOURNAL OF
EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, Special Issue on
Power Electronics for Biomedical Applications, 2014. He serves as an Associate
Editor for the IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II: EXPRESS
BRIEFS.

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 02,2023 at 11:41:27 UTC from IEEE Xplore. Restrictions apply.