Induction Heating

ÓSCAR LUCÍA, JESÚS ACERO,

Appliances
CLAUDIO CARRETERO,
and JOSÉ M. BURDÍO Toward More Flexible Cooking Surfaces

E
fficient energy manage-
ment in residential areas
is a key issue in modern
energy systems. In this
scenario, induction heat-
ing (IH) becomes an alter-
native to classical heating
technologies because of its advantages
such as efficiency, quickness, safety,
and accurate power control.
In this article, the design of modern
flexible cooking surfaces featuring IH
technology is presented. The main ad-
vantages and technical challenges are
given, and the design of the inductor
system and the power electronic con-
verter is detailed. The feasibility of the
proposed system is verified through a
laboratory prototype.
Smart energy distribution and man-
agement has become a priority in most
long-term governmental plans [1], [2]
for both environmental and econo­mic
reasons. Following this trend, special
efforts have been made in residential
areas. Among these developments, ad-
vanced heating techniques [3], includ-
ing IH technology [4]–[6], has emerged
as a new and efficient technology that
may replace classical electrical and
gas heating technology.
IH has already become the
technology of choice for domes-
tic induction cookers in Asia and
Europe because of its advantages
over other classical heating meth-
ods, such as conventional elec-
tric and gas technologies. The main
advantages of the domestic IH tech-
© istockphoto.com/pawel_p
nology are as follows:
■■ Fast heating: IH appliances directly
Digital Object Identifier 10.1109/MIE.2013.2247795 heat the pot, thereby significantly
Date of publication: 19 September 2013 reducing heating time. Figure 1(a)

1932-4529/13/$31.00©2013IEEE September 2013  ■  IEEE industrial electronics magazine  35

 so easily lost to the environment
through the surrounding air and
New design trends use flexible cooking surfaces structural elements. Moreover, the
to improve the user’s experience while providing high-efficiency power converters
make IH the most efficient heat-
an efficient and cost-effective solution. ing technology. Figure 1(b) shows
a comparison of efficiency of the
main heating technologies, includ-
shows the time required to heat that IH technology achieves the ing the energy lost to the ambi-
1.5 L of water from 20 to 90 °C using lowest heating time. ence during the heating process.
the most common domestic heat- ■■ Efficiency: Since IH technology di- IH technology achieves the highest
ing technologies. It can be seen rectly heats the pot, the heat is not efficiency, whereas gas obtains the
lowest values because of the heat
loss to the ambience.
■■ Safety: Unlike other heating technol-
Time to Heat 1.5 L Water (20 − 95 ºC)
ogies, IH does not need to heat the
surface to heat the pot. For this rea-
Resistive Resistive Halogen Gas Induction
Board son, only the residual heat is trans-
9´ 49´´
ferred from the pot to the cooking
8´ 57´´ surface, leading to significantly low-
7´ 53´´ er temperatures. This is an impor-
6´ 02´´ tant safety feature, since it avoids
3´ 54´´ the risk of burning the user or start-
ing fires. Figure 1(c) shows a com-
parison of the temperature profile
achieved during different cooking
processes. The maximum tempera-
(a) ture is higher than 500 °C for classi-
cal resistive cookers, whereas the IH
Efficiency cookers keep the maximum temper-
Induction Microwaves Halogen Resistive Resistive Gas ature under 90 and 160 °C for water
Board and oil, respectively.
77 ■■ Cleanliness: Low surface tempera-
68 tures, as shown in Figures 1(c) and
53 51 2, prevent the food from getting
44 43 burned and stuck to the surface.
This, combined with its compact
built-in construction of vitroceram-
ic glass and its low-profile power
converter, makes IH the cleanest
technology.
(b)
■■ Pot detection: A pot detection sys-
Surface Temperature Comparison tem detects when an appropriate
Resistive (540 ºC Maximum) Induction (90 ºC Water, 160 ºC Oil) pot is placed in the cooking zone
by measuring the current through
the inductor. This feature provides
º
25
0º
18

additional security and ensures

0º

proper power converter operation.

90
54

■■ Advanced control features: Typically,

0º

º
25
15

domestic heating systems control

the power delivered to the load. How-
0º

º
90
50

ever, additional sensors and smart

0º

º
15

25

control of the power converter allow

IH technology to control not only
(c)
the power delivered but also the pot
Figure 1 – A comparison of domestic heating technologies: (a) heating times, (b) efficiency, temperature. This allows for imple-
and (c) surface temperature. mentation of specific temperature

36  IEEE industrial electronics magazine  ■  September 2013

 profiles during the cooking process
or performance of advanced low-
temperature cooking techniques
presently used in high cuisine.
This article analyzes recent devel-
opments in IH appliances and focuses
on the challenges of designing new
highly flexible cooking surfaces.

Domestic IH Technology
The principle of IH technology [7] was
first described in 1831 by M. Faraday in
his experiments with induced currents.
Later, in the beginning of the 20th cen-
tury, the main applications of IH technol-
ogy were focused on melting and metal
processing. At that time, low-frequen-
cy and spark generators were being
used, which were later substituted with
vacuum tubes. Like many other industri-
al processes, IH technology experienced
major advances during World War II.
Later developments in semiconductor
technology led to the currently reliable
and high-performance IH technology. Figure 2 – Low surface temperatures of IH appliances improves safety and cleanliness.
The first reports about domestic IH ap-
plications appeared in the 1970s when
several U.S. designs and patents were to improve the coupling between the Finally, the inverter block [18]–[22],
reported by GE and Westinghouse [8]– inductor and the pot and shield the elec- which is the main part of the system,
[10]. However, power electronics tech- tronics. Power and control electronics, supplies the inductor with the output
nology was not mature then, leading to which are placed below the inductors, powers, typically up to 4 kW.
the development of low-efficiency, high- generate and control the medium- The structure and advantages dis-
volume, and costly designs with no sig- frequency currents, i.e., 20–100 kHz, re- cussed in this section are related to the
nificant market impact. Development of quired by the inductor. IH technology currently present in the
metal–oxide–semiconductor field-effect Figure 4 shows the main elements of market. However, new design trends use
transistor (MOSFET) and, especially, the power conversion block diagram of flexible cooking surfaces to improve the
insulated-gate bipolar transistor (IGBT) an IH cooktop. First, an electromagnetic user’s experience while providing an ef-
technology motivated the first designs compatibility (EMC) test ensures that ficient and cost-effective solution. This
of low-power rice cookers in Japan, an appliance complies with the regula- article will discuss in detail the main
which met with significant commercial tions. Second, the voltage of mains is benefits and key points to be consid-
success. At the same time, the first com- rectified and filtered to obtain a dc bus. ered while designing such appliances,
mercial units of these cookers appeared
in Europe in the late 1980s, and they be-
came a standard commercial option by
2000. Since then, this technology has Pan Vitroceramic
been evolving rapidly. It is experienc- Glass
ing a major revolution with advances in
power electronics [11]–[16] and flexible
cooking surfaces [17].
Inductor
The structure of a conventional IH
cooktop is presented in Figure 3. A spiral
Power
inductor is used to create a varying mag- Electronics
netic field that generates eddy currents Control
in the base of the pot to heat it. Vitroc-
eramic glass serves as support for the
pan, and some ferrites and aluminum
foils are placed beneath the inductor Figure 3 – IH cooktop structure.

September 2013  ■  IEEE industrial electronics magazine  37

 Rectifier and dc-dc Control

Mains EMC Filter Rectifier and Filter PFC and Inverter Inductor–Pot

v
vF vo

iF io

Figure 4 – Power conversion block diagram of an IH cooktop.

including the inductor and power con- aforementioned benefits of IH, are listed cooking surface. The appliance will
verter and control designs. below. automatically detect the position
■■ Any pot: Classical IH appliances of the pot and selectively activate
Flexible Cooking Surfaces require pots that perfectly match that particular area. This flexibility
Flexible cooking surfaces provide the inductor size to ensure a good results in an improved user expe-
users with a much more adaptable inductor–pot coupling. However, rience not possible with previous
cooking area with additional degrees flexible cooking surfaces allow technologies.
of freedom [23]. This multi-inductor the use of pots of any size, limited ■■ Any shape: Last but not least, flex-
approach has been used in not only only by the size of the cooking sur- ible cooking surfaces allow for
domestic but also many industrial face area. Moreover, there is no lim- efficient heating of pots that are
applications [6], [24], [25]. Figure 5 it to the number of pots that can be noncircular. This feature is espe-
shows the structure of the total active used, which significantly improves cially important in some countries
surface concept, wherein a matrix of the appliance’s usability. where special pots are used to pre-
small spiral inductors allows such ver- ■■ Anywhere: Cooking zones are not pare special dishes.
satility. The main advantages of such delimited and, therefore, the user Although these developments imply
cooking surfaces, in addition to the can place the pot anywhere on the a major breakthrough in the domestic IH
technology, allowing pots of any shape
or size to be used anywhere, they lead to
a technical challenge from the point of
view of inductor and power electronics
design. The main differential character-
istics of this technology are summarized
in the following paragraphs:
First, a high-power-density solution
from the inductor and power elec-
tronics point of view is required. The
implementation of a total active surface
appliance, as shown in Figure 5, requires
a high level of integration, which has to
be one of the design key features.
Along with integration, the cost of
such a complex design must be consid-
Figure 5 – A total active surface IH cooktop. ered when developing a cost-effective

38  IEEE industrial electronics magazine  ■  September 2013

solution that can compete in a con-
strained market. TABLE 1—OUTPUT POWER RATINGS FOR TYPICAL INDUCTION LOADS.
The evolution from fixed defined 2
DIAMETER (cm) AREA (cm2) NOMINAL Po (W) NOMINAL t P (W/cm )
cooking areas to variable undefined
8 50 450 8.95
cooking areas inherently implies sev-
eral design changes. First, the concept 15 176 1,500 8.49
of output power per burner no longer 18 254 1,800 7.07
applies, and the concept of power den- 21 346 2,200 6.35
sity t P emerges 28 616 2,800 4.54

Po
tP = , (1)
r (D 2p /4)
the smaller the pot, the higher the A second major change imposed
where Po is the output power and D p required power density. This implies by undefined cooking surface is the
is the diameter of the pot. Table 1 that the smallest pots will actually fact that the pot can be placed in
summarizes the values obtained for determine the required output power any relative position with respect to
typical IH loads. It can be seen that per inductor. one inductor [Figure 6(a)] or a set of

(a)

CCE CCE
+ vo il1 il2 il 3
Vbus

(b) (c)

FIGURE 6 – Total active surface concept design challenges include different relative positions of (a) an inductor or (b) set of inductors and
(c) undesired emissions.

September 2013  ■  IEEE industrial electronics magazine  39

 these systems. All these aspects
must be considered when designing
The cost of such a complex design must be considered any kind of inductor. This section fo-
when developing a cost-effective solution that can cuses on the differential aspects of
inductors for modern flexible cook-
compete in a constrained market. ing surfaces.
The elements considered in the elec-
tromagnetic analysis of the standard
inductors [Figure 6(b)]. Moreover, Inductor System Description arrangement of an inductor system are
uncovered inductors might create and Design shown in Figure 7. Other elements, such
undesired near-field emissions if not The inductor system of an IH cook- as the vitroceramic glass or electrical
controlled properly [Figure 6(c)]. As top plays two main roles. First, it insulators, are not considered because,
a consequence, a precise power con- supplies power to the pot; second, it in the desired frequency range, they
verter control is required to ensure represents the load of the power elec- exhibit electromagnetic properties
the desired output power control, out- tronics ­inverter. The inductor system similar to those of air. The flux concen-
put power distribution in the pot, and should fulfill several requirements, trators improve the electromagnetic
avoidance of undesired emissions. including those of high efficiency of coupling between the inductor and the
Finally, the placement of inductors in electromagnetic power transference, pot. Bars of ferrite, a high-permeability
the vicinity of other inductors produces low cost, and size constraints [26]– low-dissipative material are generally
cross-coupling effects, which may affect [28]. Owing to these reasons, analy- used as concentrators. The shielding
the induction efficiency and the effective sis tools, based either on analytical is a sheet of a low-dissipative metallic
output power control, further compli- [29] or finite-­element analysis [30], material, usually aluminum. The same
cating the overall system design. have often been used for designing materials are used in total active surface
cooktops, but these are adapted to the
characteristics of the application.
The inductor is made of a copper
wire and, considering the usual current
Pot
levels (up to tens of amperes), nonneg-
Inductor ligible power dissipation occurs. Let TR
and R W be two resistances representing
Flux Concentrators the power dissipation in the pot and the
inductor wires, respectively. The induc-
Shielding tion efficiency of the system in Figure 7
a1
an is defined as the power transferred to
the pot (which is the power spent on
heating the bottom of the pot) with re-
FIGURE 7 – Elements considered in the analysis of a standard inductor system. spect to the electrical power supplied
by the power electronics inverter to the
inductor [5]. Therefore,
102 TR . (2)
h ind =
Plane (nr = 1,000) TR + R W
Bars ( nr = 2)
The resistances TR and R W can
Air ( nr = 1)
be obtained by using the analysis
101 tools mentioned previously. These
variables depend on the parameters
DR (X)

of the system in Figure 7: the number

of turns of the coil (n t), the frequency
100 of the currents (~), the electromagnetic
properties (electrical conductivity v
and magnetic permeability n) of the
materials, and geometrical parameters,
such as the diameter of the inductor and
10–1
10 100 the distances between the materials. The
nt resistance R W mainly depends on the
characteristics of the wires, parameters
FIGURE 8 – Values of TR with respect to the number of turns for several ferrite arrangements. n t and ~, and the geometry of the coil.

40  IEEE industrial electronics magazine  ■  September 2013

Design of the Coils
The design of the individual small
coil is addressed considering the out-
Inductors of conventional IH arrangements are wound
put power ratings of Table 1. As men- with multistranded wires, also called Litz wires.
tioned previously, the variable TR
represents the power transferred to
the pot, which depends on various bars, and air. The permeabilities con- spiral coils is required. Therefore, the
parameters. However, some of them sidered in the calculation of TR are inductors of the total active surface
are restricted by the characteristics of also shown in Figure 8. cooktops are wound without holder,
the application. Considering the total As shown in Figure 8, the minimum i.e., the turns are closely packed and
active surface cooktops, the diameter number of turns required to achieve in several layers. A schematic rep-
and rated power are design specifica- TR = 20 X is n t = 50. Considering resentation of this coil is shown in
tions, as shown in Table 1. Similarly, the size restrictions of the coils and Figure 9. Hence, the conventional flat-
the frequency range is also restricted the cable cross section necessary to type spiral inductors become new
to between 20 and 150 kHz. Other geo- carry the above-mentioned current, ring-type coils in total active surfaces.
metrical parameters are also deter- an arrangement completely different Another significant change with re-
mined by the vitroceramic glass and from the conventional single-layer spect to conventional inductors is that
flux concentrators (ferrites) available
in the market. Therefore, the main pa-
rameters of the coil that should be de-
fined are the number of turns and the Bobbin Windings
type and characteristics of the cable.
The resonant inverter block of
Figure 4 is equivalent to a voltage
source that feeds the inductor–
pot system. Close to the resonant
frequency, a sinusoidal current can
be assumed to be supplying power to
the inductor–pot system. At this point,
the maximum output power is ob-
tained, which is an important design
r int = 10
specification. Therefore,
2
V inv r ext = 20
Po = , (3)
TR

where Vinv is the root mean square

FIGURE 9 – A schematic representation of the coil. Dimensions are in mm.
(rms) value of the voltage supplied
by the resonant inverter at the reso-
nant condition. This voltage de-
pends on the voltage of the mains
1
and the inverter topology. For the
basic half-bridge resonant inverter,
Vinv = ^ 2 r h Vmains [31], where Vmains 0.95
is the rms value of the voltage of the
mains. Considering Vmains = 230 V and
0.9
the data in Table 1, we can obtain an
hind

estimation of the required TR by us-

ing the last formula and (3). The value 0.85
of TR = 20 X ensures a maximum
output power per inductor of at least
500 W. For this value, the current in 0.8
the inductor at the rated output power
is up to 5 A.
0.75 3
Figure 8 shows the values of TR 10 104 105 106
with respect to the number of turns f (Hz)
at 65 kHz for several ferrite arrange-
ments, including ferrite plane, ferrite FIGURE 10 – Induction efficiency with respect to the frequency.

September 2013  ■  IEEE industrial electronics magazine  41

a ferrite plane is placed beneath Coupling Between Inductors
the windings, instead of dis- Z11 The coils of the total active sur-
crete ferrite bars, to achieve the face cooktops are contiguously
maximum TR with the minimum + + placed, and several coils can
number of turns. i1 be covered by the same pot, as
v1 v12 = Z12 · i2
– – shown in Figure 6. In this case, the
Design of the Cable role of the pot is similar to that of
Inductors of conventional IH ar- the core in a transformer, lead-
rangements are wound with mul- ing to some coupling between
FIGURE 11 – An equivalent circuit of a coil coupled with other
tistranded wires, also called Litz coil. the coils [33]. Although coupling
wires [28]. Litz wires are also used between the coils has several
in total active surface cooktops, disadvantages, some of which
as a satisfactory tradeoff between are mentioned in the “Domestic
efficiency, cost, and packing fac- 1.25 IH Technology” section, it does
tor is achieved. 1.2 allow utilization of a new power
Litz wires consist of n s 1.15 control variable.
strands of diameter z s and un- 1.1 Assuming that the coupling
P1/Prated

dergo some process of braiding can be represented by pure in-

1.05
or twisting. The parameters n s ductance, we can define voltage
and z s can be defined to achieve 1 of the circuit in Figure 11 as
the maximum induction efficien- 0.95
cy at a particular frequency. The 0.9 v 1 = Z 11 i 1 + j~L 12 i 2,(4)
diameter should be less than 0.85
the skin depth of the fields at the 0 50 100 150 200 250 300 350 where i 1,and i 2 are the currents in
considered frequency, and the α the coils. These currents can be
number of strands is defined by shifted to an angle a and there-
the rms current of the inductor FIGURE 12 – A power variation of a coil with respect to the fore can be expressed as follows:
phase shift between the currents of two coils.
at the nominal power. The skin
depth of copper at 100 kHz is %
i 1 = i 1 e j~t
d = 189 nm and the criterion for cal- design [32]. Therefore, for these coils, t
i 2 = i2 e (j~t + a) . (5)
culating the cross section of the cable z s = 150 nm and n s = 35. Figure 10
is that the current density should be shows the induction efficiency with Figure 11 represents the equivalent
lower than 15 A/mm2, which is an ap- respect to the frequency for the de- circuit of a coil, where the coupling
propriate value for compact winding signed Litz wire cable. with a contiguous coil has been consid-
ered by means of the voltage source v 2 .
It is obtained from (4) and (5) when a is
... varied from 0 to r.
i bus + + +
Cr,1 Cr,3 Cr,n – 1 By definition, the power de-
livered by the voltage source is
vl,1 L vl,3 L vl,n – 1 L
eq,1 eq,3 eq,n – 1 P1 = Real (v 1 $ i 1); therefore, the usual
Sh values of the impedances for the 48-
Req,1 Req,3 Req,n – 1
ch Cs – – – turn coil and the rated power of the
Th D S 1 i l,1 S 3 il,3 Sn – 1 i l,n – 1 coil Prated can be considered to calcu-
h
late the ratio P1 /Prated with respect to
c1 c3 cn – 1
+ the phase shift a, and the results are
io T1 D1 T3 D3 Tn – 1 Dn – 1
Vbus ... shown in Figure 12. As can be seen in
+ T2 D2 T4 D4 Tn Dn
the figure, a variation in the phase of
vo c2 c4 cn coil 2 causes a change of up to !15% in
Sl S2 S4 Sn the rated power of coil 1.
i l,2 il,4 il,n
cl Cs + + +
Tl Req,2 Req,4 Req,n Power Converter and
Dl
Control Strategy Design
vl,2 Leq,2 vl,4 Leq,4 vl,n Leq,n To supply the inductor system proposed
in the “Inductor System Description and
– – Cr,2 – Cr,4 – Cr,n
... Design” section, a multiple-output reso-
nant inverter is required. To obtain the
FIGURE 13 – A series resonant multi-inverter. desired performance and cost-effective

42  IEEE industrial electronics magazine  ■  September 2013

solution, the series resonant multi-in-
verter was proposed as a well-balanced Continuous Mode Discontinuous Mode
solution [34], [35].
• Variable Frequency Duty Cycle • Regenerative Control (RC)
The series resonant multi-inverter Control (VFDC) • Direct Conduction Control (DCC)
(Figure 13) comprises a common in- • High-Frequency Pulse Density
verter block (CIB), composed of the Modulation (HF-PDM)
• Pulse Delay Control (PDC)
common switches S h and S i and a
resonant load block (RLB), which com- • Low-Frequency Pulse Density Modulation (LF-PDM)
prises the specific switches S i, and
the resonant loads R eq, i - L eq, i - C r, i .
FIGURE 14 – Control strategies classification for the series resonant multi-inverter.
R eq,i and L eq,i model the ith inductor
system, whereas C r,i is the resonant
capacitor added to form the resonant resonant multi-inverter. The top part phase shifts { i . It is important to note
tank. The main advantage of this topol- of the figure shows the signals cor- that all of these control parameters
ogy is the reduced number of switch- responding to the CIB, whereas the provide the designer with many de-
ing devices compared with classical bottom part shows the RLB signals of grees of freedom that allow controlling
topologies, which allows achieving a loads 1–4. The complete set of control independently the power delivered to
high-performance, cost-effective solu- parameters is depicted in the figure, each inductor. In the next subsections,
tion. In addition, each load can be in- including the CIB switching period each modulation strategy of Figure 13
dividually disconnected, avoiding the Thb, duty cycle D hb, dead times t d, hb1, is explained briefly and a compari-
undesired emission issue stated in the t d, hb2, the RLB switching periods Ti, son between them is presented in the
“Domestic IH Technology” section. duty cycles D i, delay times t d,i, and ­“Experimental Results” section.
In IH appliances, the output power,
which is selected by the user according
to his or her cooking needs, is the vari-
able that needs to be controlled. Since Thb
the dynamics of the electrical system are Dhb
much faster than those of the thermal
ch
system, the transient response of the td,hb2 td,hb1
power converter is not usually an issue,
cl
and the attention is focused on the mod-
ulation strategy applied. In addition to vbus
ensuring the desired output power, the
modulation strategy must avoid audible vo,io
noise emissions, ensure fulfillment of
EMC standards [19], and deal with cou-
pling effects. Depending on the CIB op- T1
eration mode, the modulation strategies {1 td1 D1
c1
can be classified as continuous [31] and T3
{3 D3
discontinuous modes [17] (Figure 14). c3
The continuous operation mode
is achieved when the output current i1
i o is different from zero at any time,
whereas the discontinuous mode is i3
achieved when i o becomes zero during
certain time intervals. The latter reduc-
es the current through the converter c2
and, therefore, increases the efficiency T4
D4
of the converter. The continuous and c4
the discontinuous modulation strat-
i2
egies are complemented by low-fre-
quency pulse density modulation [19],
which is based on turning the modula- i4
tion on and off periodically and can be t
applied to both modulation strategies.
Figure 15 shows the main control FIGURE 15 – Main control signals (gray background) and waveforms (white background). Top
signals and waveforms of the series part corresponds to the CIB, whereas bottom part corresponds to loads 1–4 of the RLB.

September 2013  ■  IEEE industrial electronics magazine  43

Continuous-Mode Modulation as shown in Figure 16. As a conse- The continuous-mode modulation
Strategies quence, the output power of each strategies are classified according to the
The main advantage of the continu- load can be accurately controlled, parameters used to control the output
ous-mode modulation strategies is regardless of the operation mode of power. Variable-frequency duty cycle
that all the resonant loads operate the remaining loads, minimizing the modulation (VFDC) [31], [36] adjusts Thb
in parallel with a voltage source, coupling effects. and D hb (see Figure  15) to control the
overall output power, whereas the high-
frequency pulse density modulation (HF-
PDM) additionally controls Ti,, D i,, and { i
Si + and Si – ON Tj + and Tj – OFF
to provide the output power control for
io ... ... each load (Figure 15, loads 3 and 4). Final-
+ il,i + il,i – Dj + Dj –
+ ly, pulse delay control (PDC) uses t d,i to
+
Vbus Vbus perform a fine tuning of the output power
_ –
+ vo + + in each load (Figure 15, load 1).
Req,i + Req,i – Req,j + Req,j –
kiVbus
vl,i + vl,i – vl,j+ vl,j –
Leq,i + Leq,i – Leq,j + Leq,j – Discontinuous-Mode Modulation
Cr,i + Cr,i – Cres,j + Strategies
_ + – + _ Cres,j –
... ... The discontinuous-mode modula-
tion strategies are used under light-
load conditions to improve efficiency
FIGURE 16 – Continuous-mode operation: equivalent circuit. [17], [37]. The strategy based on us-
ing the available capacitor charge in
two loads connected to the same
bus side [Figure 17(a)] is called regen-
Cr,2p + 1
erative control (RC), and the strategy
Leq,2p + 1 based on using two loads connected
il,2p il,2q
Req,2p + 1
to different bus sides [Figure 17(b)] is
+
il,2p+1 called direct conduction control. In both
Cr,2p Cr,2q
Vbus cases, the result is a reduced current
il,2p
Vo Leq,2p Leq,2q + in the CIB, reducing conduction losses
Req,2p and, therefore, improving power con-
Vo
– Req,2p Req,2q Leq,2p verter efficiency. The main drawback of
the strategy is the cross-dependence of
– Cr,2p
the operating conditions of the loads,
(a) (b) making it more complex to manage.

FIGURE 17 – Discontinuous-mode operation. (a) Regenerative mode. (b) Direct conduction mode. Low-Frequency Pulse
Density Modulation
Finally, low-frequency pulse density
tpdm modulation (LF-PDM) [19] is used to
tpdm,1 tpdm,2 tpdm,m complement the control strategy in
Po,1 those situations in which the continu-
ous mode and/or discontinuous-mode
Po,1 Ppdm,1,1 Ppdm,1,2 Ppdm,1,m Ppdm,1,1 control strategies cannot achieve the
desired performance. The modulation
Po,2
is based on periodically changing the
Po,2 Ppdm,2,1 Ppdm,2,2 Ppdm,2,m Ppdm,2,1 control parameter set } i with a time
scale of seconds. The mean power in
Po,n each interval i for each load j, Ppdm, i, j,
may change, but the mean output
Po,n Ppdm,n,1 Ppdm,n,2 Ppdm,n,m Ppdm,n,1 power per load Po,i is ensured to be
the desired one (Figure 18). This al-
DPpdm
lows a great flexibility, since any con-
RPo,i trol strategy can be applied in each
}1 RPpdm,i,1 }2 RPpdm,i,2 }m RPpdm,i,m }1 RPpdm,i,1 t
interval t pdm,i, and the coupling effects
can be avoided by selectively deacti-
FIGURE 18 – LF-PDM. vating loads in the neighborhood. The

44  IEEE industrial electronics magazine  ■  September 2013

output power when using LF-PDM is
100
calculated as follows:
98
m m
/ t pdm,j Ppdm,i,j / t pdm,j Ppdm,i,j 96
j= 1 j= 1
Po, i = = $
m
Tpdm
/ t pdm,j 94
j= 1
(6) 92

h (%)
90
The maximum power step TPpdm
is selected to minimize flicker emis- 88
sions, whereas the maximum time
step t pdm,i is selected to minimize user 86 VFDC
performance impact. HF-PDM
84 PDC
Finally, Figure 19 shows a compari- RCi
son of the efficiency obtained by the 82 RCj
proposed control strategies. It can be DCC
80
seen that the proposed modulation 0 100 200 300 400 500 600 700 800
strategies can be combined to achieve Po,i (W)
any desired output power with good
FIGURE 19 – Control strategies simulation results.
efficiency.

Experimental Results those obtained from the simulation.

The main experimental results to Moreover, the combination of control TABLE 2—POWER CONVERTER DESIGN
prove the feasibility of the proposed strategies proposed allows obtaining PARAMETERS.
solutions are presented. To test the IH efficiencies above 95% in the com- PARAMETER
system, a versatile power electronic plete range of operation and induc- Input voltage 230 Vrms
test bench was implemented [38] tion loads. Continuous-mode control Load number 4
along with 4–8-cm inductors. Table 2 strategies achieve high-efficiency at
Output power per load 500 W
­summarizes the main power converter high output power levels, whereas
Switching frequency range fo to 150 kHz
specifications. The system was con- discontinuous-mode control strate-
trolled by a Xilinx Spartan 3 FPGA [39], gies allow improving efficiency under Resonant capacitor 44 nF

which performs the required mea- low-load conditions. The feasibil- Snubber capacitor 3.3 nF
surements and generates the control ity of the proposed implementation IGBT/diode HGTG20N60D
signals for the power converter. Ex-
perimental measurements were per-
formed for two different commercial
IH pots, Pot1 and Pot2, which present
different electrical equivalents shown
in Table 3, i.e., different output power
and efficiency, and are good samples
of the commercially available pots.
Figure 20 shows the modular power
converter implemented along with the
inductor system.
The main evaluation criteria are the
output power control and the system ef-
ficiency. Both have been measured with
the power analyzer Yokogawa PZ4000,
and the main results are summarized in
Figure 21.
The experimental results show
good output power controllability, al-
lowing for control of the output power
from the maximum of 500 W down to
50 W, regardless of the IH load used.
These results are consistent with FIGURE 20 – Experimental test bench.

September 2013  ■  IEEE industrial electronics magazine  45

 Pot1 VFDC Pot1 HF-PDM Di = 0.66 Pot1 VFDC Pot1 HF-PDM Di = 0.66
Pot1 HF-PDM Di = 0.33 Pot1 PDC Pot1 HF-PDM Di = 0.33 Pot1 PDC
Pot1 RCi Pot1 RCj Pot1 DCC Pot1 RCi Pot1 RCj Pot1 DCC
800 100
700 98
600 96
94
500
92
Po,i (W)

h (%)
400 90
300 88
86
200
td,i = 25% 84
100 td,i = 50% 82
td,i = 75%
0 80
20 40 60 80 100 120 140 0 100 200 300 400 500 600 700 800
fs,hb (kHz) Po,i (W)
(a) (b)

Pot2 VFDC Pot2 HF-PDM Di = 0.66 Pot2 VFDC Pot2 HF-PDM Di = 0.66
Pot2 HF-PDM Di = 0.33 Pot2 PDC Pot2 HF-PDM Di = 0.33 Pot2 PDC
Pot2 RCi Pot2 RCj Pot2 DCC Pot2 RCi Pot2 RCj Pot2 DCC

800 100
98
700
96
600 94
500 92
Po,i (W)

h (%)

400 90
88
300
86
200
td,i = 25% 84
100 td,i = 50% 82
td,i = 75%
0 80
20 40 60 80 100 120 140 0 100 200 300 400 500 600 700 800
fs,hb (kHz) Po,i (W)
(c) (d)

FIGURE 21 – Experimental results: output power control and efficiency for [(a) and (b)] Pot1 and [(c) and (d)] Pot2 commercial induction pots.

of modern IH appliances featuring flex- Acknowledgments

TABLE 3—INDUCTION LOADS PARAMETERS ible cooking surfaces is presented. This work was partly supported by
AT RESONANT FREQUENCY. These appliances significantly im- the Spanish MICINN under Project
PARAMETER POT1 POT2 prove the user’s performance but, at TEC2010-19207, Project CSD2009-00046,
the same time, require major techno- and Project IPT-2011-1158-920000, by
Req 13.5 X 14 X
logical developments. A novel induc- the DGA-FSE, by the DGA-grant B019/12,
Leq 98 µH 125 µH
tor design has been proposed along and by the Bosch and Siemens Home
fo 76 kHz 68 kHz with a multiple-output power convert- Appliances Group.
er with improved control strategies.
The combination of these two spe- Biographies
of the total active surface concept cifically designed elements provides Óscar Lucía (olucia@unizar.es) re-
is proved. a cost-effective solution with a proper ceived his M.Sc. and Ph.D. degrees in
output power control and efficiency electrical engineering from the Uni-
Conclusion in the complete operating range. In versity of Zaragoza, Spain, in 2006 and
IH technology is expected to play a conclusion, a total active surface 2010, respectively. He is currently an
major role in modern heating systems appliance has been developed that assistant professor with the Depart-
because of its advantages regarding ef- significantly improves the current in- ment of Electronic Engineering and
ficiency, quickness, safety, and output duction technology and will lead re- Communications at the University of
power control. In this article, the design search and development in the future. Zaragoza. He has been involved in the

46  IEEE industrial electronics magazine  ■  September 2013

design of IH appliances since 2005. He [2] B. Jinsung, H. Insung, K. Byeongkwan, and P. [23] I. Millán, J. M. Burdío, J. Acero, O. Lucía, and
Sehyun, “A smart energy distribution and man- D. Palacios, “Resonant inverter topologies for
has published more than 70 technical agement system for renewable energy distribu- three concentric planar windings applied to
papers and 15 patent applications. He tion and context-aware services based on user domestic induction heating,” Electron. Lett.,
patterns and load forecasting,” IEEE Trans. Con- vol. 46, no. 17, pp. 1225–1226, 2010.
is a Member of the IEEE. His research sum. Electron., vol. 57, no. 2, pp. 436–444, 2011. [24] H. Pham, H. Fujita, K. Ozaki, and N. Uchida,
interests include multiple-output con- [3] R. Narita, “A new detection method for rice “Phase angle control of high-frequency reso-
cookers,” IEEE Trans. Consum. Electron., vol. nant currents in a multiple inverter system for
verters, digital control, and resonant CE-33, no. 3, pp. 136–141, 1987. zone-control induction heating,” IEEE Trans.
power conversion for IH applications. [4] T. Tanaka, “A new induction cooking range for Power Electron., vol. 26, no. 11, pp. 3357–3366,
heating any kind of metal vessels,” IEEE Trans. 2011.
Jesús Acero received his M.Sc. and Consum. Electron., vol. 35, no. 3, pp. 635–641, 1989. [25] H. N. Pham, H. Fujita, K. Ozaki, and N. Uchida,
Ph.D. degrees in electrical engineering [5] J. Acero, J. M. Burdio, L. A. Barragan, D. “Dynamic analysis and control for resonant
­Navarro, R. Alonso, J. R. Garcia, F. Monterde, P. currents in a zone-control induction heating
from the University of Zaragoza, Spain, Hernandez, S. Llorente, and I. Garde, “Domes- system,” IEEE Trans. Power Electron., vol. 28,
in 1992 and 2005, respectively. From tic induction appliances,” IEEE Ind. Appl. Mag., no. 3, pp. 1297–1307, Mar. 2013.
vol. 16, no. 2, pp. 39–47, Mar.–Apr. 2010. [26] J. Acero, C. Carretero, R. Alonso, and J. M.
1992 to 2000, he worked in several in- [6] J. Egalon, S. Caux, P. Maussion, M. Souley, and Burdio, “Quantitative evaluation of induc-
dustry projects, specifically in custom O. Pateau, “Multiphase system for metal disc tion efficiency in domestic induction heating
induction heating: Modeling and RMS current ­applications,” IEEE Trans. Magn., vol. 49, no. 4,
power supplies for research laborato- control,” IEEE Trans. Ind. Appl., vol. 48, no. 5, pp. 1382–1389, 2013.
ries. Since 2001, he has been an asso- pp. 1692–1699, Sept.–Oct. 2012. [27] V. M. Primiani, S. Kovyryalov, and G. Cerri,
[7] A. Mühlbauer, History of Induction Heating and Melt- “Rigorous electromagnetic model of an
ciate professor with the University of ing. Essen, Germany: Vulkan-Verlag GmbH, 2008. induction cooking system,” IET Proc. Sci.,
Zaragoza. He has been involved in the [8] H. Yamamura, K. Matsuo, and N. Nagai, “Induction Meas., Technol. (1994–2006), vol. 6, no. 4, pp.
heating apparatus,” U.S. Patent 4 115 677, 1978. 238–246, 2012.
design of IH appliances since 2001. He [9] R. W. Mackenzie, P. Wood, T. M. Heinrich, and [28] J. Acero, R. Alonso, J. M. Burdío, L. A. Barragán,
has published more than 80 technical R. M. Oates, “Frequency controlled induction and D. Puyal, “Frequency-dependent resistance
heating apparatus,” U.S. Patent 4 085 300, 1976. in Litz-wire planar windings for domestic induc-
papers and 20 patent applications. He [10] R. L. Steigerwald, “Constant duty cycle con- tion heating appliances,” IEEE Trans. Power
is a Member of the IEEE. His research trol of induction cooking inverter,” U.S. Patent Electron., vol. 21, no. 4, pp. 856–866, 2006.
3 781 505, 1972. [29] C. Carretero, J. Acero, R. Alonso, J. M.
interests include resonant converters [11] M. Gaurnieri, “Trailblazers in solid-state elec- Burdio, and F. Monterde, “Embedded ring-
for IH applications, inductive-type load tronics,” IEEE Ind. Electron. Mag., vol. 5, no. 4, type inductors modeling with application to
pp. 46–47, Dec. 2011. induction heating systems,” IEEE Trans. Magn.,
modeling, and magnetics. [12] J. Holtz, “Power electronics: A continuing chal- vol. 45, no. 12, pp. 5333–5343, 2009.
Claudio Carretero received his lenge,” IEEE Ind. Electron. Mag., vol. 5, no. 2, pp. [30] J.-K. Byun, K. Choi, H.-S. Roh, and S.-Y. Hahn,
6–15, June 2011. “Optimal design procedure for a practical
M.Sc. degree in physics and his Ph.D. [13] J. Rabkowski, D. Peftitsis, and H. Nee, “Silicon induction heating cooker,” IEEE Trans. Magn.,
degree in electrical engineering from carbide power transistors: A new era in power vol. 36, no. 4, pp. 1390–1393, 2000.
electronics is initiated,” IEEE Ind. Electron. [31] O. Lucía, J. M. Burdío, I. Millán, J. Acero, and
the University of Zaragoza, Spain, in Mag., vol. 6, no. 2, pp. 17–26, June 2012. L. A. Barragán, “Efficiency oriented design
1998 and 2010, respectively. Since 2010, [14] R. Singh and M. Pecht, “Commercial impact of of ZVS half-bridge series resonant inverter
silicon carbide,” IEEE Ind. Electron. Mag., vol. 2, with variable frequency duty cycle control,”
he has been a postdoctoral researcher no. 3, pp. 19–31, Feb. 2008. IEEE Trans. Power Electron., vol. 25, no. 7,
in the Department of Electronic En- [15] V. Esteve, E. Sanchis-Kilders, J. Jordan, pp. 1671–1674, July 2010.
E. J. Dede, C. Cases, E. Maset, J. B. Ejea, and [32] A. V. d. Bossche and V. C. Valchev, Inductors
gineering and Communications, Uni- A. Ferreres, “Improving the efficiency of IGBT and Transformers for Power Electronics. Boca
versity of Zaragoza. He has published series-resonant inverters using pulse density Raton, FL: CRC Press, 2005.
modulation,” IEEE Trans. Ind. Electron., vol. 58, [33] J. Acero, C. Carretero, R. Alonso, O. Lucía, and
more than 40 technical papers and 12 no. 3, pp. 979–987, Mar. 2011. J. M. Burdío, “Mutual impedance of small ring-
patent applications. He is currently [16] T. Hirokawa, M. Okamoto, E. Hiraki, T. Tanaka, type coils for multi-winding induction heating
and M. Nakaoka, “A novel type time-sharing appliances,” IEEE Trans. Power Electron., vol.
working in the field of electromagnetic high-frequency resonant soft-switching in- 28, no. 2, pp. 1025–1035, Feb. 2013.
modeling of IH systems. He has been verter for all metal IH cooking appliances,” in [34] O. Lucía, J. M. Burdío, L. A. Barragán, J. Acero,
Proc. 37th Annu. Conf. IEEE Industrial Electron- and I. Millán, “Series-resonant multiinverter for
involved in the design of IH appliances ics Society IECON11, pp. 2526–2532. multiple induction heaters,” IEEE Trans. Power
since 2005. He is a Member of the IEEE. [17] O. Lucía, J. M. Burdío, L. A. Barragán, J. Acero, Electron., vol. 24, no. 11, pp. 2860–2868, Nov.
and C. Carretero, “Series resonant multi-inverter 2010.
José M. Burdío received his M.Sc. and with discontinuous-mode control for improved [35] O. Lucía, J. M. Burdío, I. Millán, J. Acero, and D.
Ph.D. degrees in electrical engineering light-load operation,” IEEE Trans. Ind. Electron., Puyal, “Multiple-output resonant inverter to-
vol. 58, no. 11, pp. 5163–5171, Nov. 2011. pology for multi-inductor loads,” in Proc. IEEE
from the University of Zaragoza, Spain, [18] S. Chudjuarjeen, A. Sangswang, and C. Koompai, Applied Power Electronics Conf. Expo., 2010, pp.
in 1991 and 1995, respectively. He is cur- “An improved LLC resonant inverter for induc- 1328–1333.
tion-heating applications with asymmetrical [36] E. R. C. da Silva, E. Cipriano dos Santos, and
rently a professor with the Department control,” IEEE Trans. Ind. Electron., vol. 58, no. C. B. Jacobina, “Pulsewidth modulation strate-
of Electronic Engineering and Communi- 7, pp. 2915–2925, July 2011. gies,” IEEE Ind. Electron. Mag., vol. 5, no. 2, pp.
[19] O. Lucía, J. M. Burdío, I. Millán, J. Acero, and D. 37–45, June 2011.
cations at the University of Zaragoza. He Puyal, “Load-adaptive control algorithm of half- [37] H. Sarnago, O. Lucía, A. Mediano, and J. M.
has published more than 140 technical bridge series resonant inverter for domestic in- Burdío, “Modulation scheme for improved
duction heating,” IEEE Trans. Ind. Electron., vol. operation of a RB-IGBT based resonant inverter
papers and 25 patent applications. He is a 56, no. 8, pp. 3106–3116, Aug. 2009. applied to domestic induction heating,” IEEE
Senior Member of the IEEE. His research [20] N. A. Ahmed, “High-frequency soft-switching Trans. Ind. Electron., vol. 60, no. 5, pp. 2066–
ac conversion circuit with dual-mode PWM/ 2073, May 2013.
interests include modeling of switching PDM control strategy for high-power IH ap- [38] O. Lucía, L. A. Barragán, J. M. Burdío, O.
converters and resonant power conver- plications,” IEEE Trans. Ind. Electron., vol. 58, ­Jiménez, D. Navarro, and I. Urriza, “A versa-
no. 4, pp. 1440–1448, Apr. 2011. tile power electronics test-bench architecture
sion for IH applications. [21] J. I. Rodriguez and S. B. Leeb, “Nonresonant applied to domestic induction heating,” IEEE
and resonant frequency-selectable induction- Trans. Ind. Electron., vol. 58, no. 3, pp. 998–1007,
References heating targets,” IEEE Trans. Ind. Electron., Mar. 2011.
[1] D. Villa, C. Martin, F. J. Villanueva, F. Moya, vol. 57, no. 9, pp. 3095–3108, Sept. 2010. [39] E. Monmasson, L. Idkhajine, and M. W. Naouar,
and J. C. Lopez, “A dynamically reconfigu- [22] M. Kamli, S. Yamamoto, and M. Abe, “A 50–150 “FPGA-based controllers,” IEEE Ind. Electron.
rable architecture for smart grids,” IEEE kHz half-bridge inverter for induction heating Mag., vol. 5, no. 1, pp. 14–26, 2011.
Trans. Consum. Electron., vol. 57, no. 2, pp. applications,” IEEE Trans. Ind. Electron., vol. 43,
411–419, 2011. no. 1, pp. 163–172, Feb. 1996. 

September 2013  ■  IEEE industrial electronics magazine  47