100kW Induction Heater Prototype
100kW Induction Heater Prototype
ABSTRACT
Industrial induction heating processes can contribute to the European Union (EU) energy
saving goal of 20 percent before 2020. Induction manufacturers already propose many efficient
solutions available at industrial scale. To improve induction devices for better energy efficiency,
EDF R&D set up a French collaborative project called Innovative Solutions for Induction
Systems (ISIS) with the financial support of the French National Research Agency (ANR).
The objective of ISIS is to promote induction heating as a best available technology
(BAT) and to develop innovative solutions to increase its efficiency. This paper reports the ISIS
project innovations. The paper also discusses about the efficient heat conversion from the
induction heating devices through the use of new auto-adaptive multi-coil power supply with low
losses coils. An important aspect of this project is the recovery of fatally lost heat energy
(cooling of the inductors).
During the prototype testing, first the control algorithms of the multi-coil technology
were successfully tested on a 100 kW 3-coils power supply. A homogenization technique is
proposed to model a multi-strand coil and to use it in industrial setting. The multi-strand coils are
now used only in low power residential applications such as induction cooking. A heat recovery
test bench is built and equipped with a predictable function control (PFC) loop to fit with the
production fluctuations. This paper also presents an analysis of the U.S. potential market for
these new induction heating approaches and their applications in industries.
Figure
F 1. Co
opper Brazing of High--Pressure Stteel Fitting U
Using Induction Heatin
ng—
an Applica
ation That Iss Often Don
ne in a Batcch Oven
The
T inductor is similar to o a transformmer primaryy, and the w workpiece is equivalent tto the
transform
mer secondarry. Thereforre, several off the charactteristics of trransformers are useful iin the
developm ment of guid delines for co oil design. One
O of the m most importan ant features oof transform
mers is
the fact that the effiiciency of coupling
c bettween the w windings is iinversely proportional tto the
square off the distancce between them.
t In adddition, the cuurrent in thee primary off the transformer,
multiplieed by the num mber of prim mary turns, is equal to thhe current in the secondaary, multiplieed by
the numb ber of second dary turns.
Coil
C designs are based on the heating-patternn requiremeents of the applicationn, the
frequency y, and the power-densit
p ty requiremeents. In addittion, the maaterial-handliing techniquues to
be used for
f productio on determinee, to a large extent, the ccoil to be ussed. If a partt is to be insserted
in a coil, moved on a conveyor, or o pushed en nd to end, orr if the coil/hheat station ccombinationn is to
move on nto the partt, the coil design
d mustt take the aappropriate handling reequirementss into
consideraation. Acco ordingly, a variety of specialty c oil designs have evolvved for speecific
applicatioons.
Because
B of itss low resistiv
vity, fully an
nnealed, highh-conductivity copper iss most comm monly
used in thhe fabricatio
on of inductiion heating coils.
c The coopper is typiically in a tuubular form,, with
a minimu um outer diaameter of 0.125 inch (0.32 cm) to aallow for waater cooling. Material oof this
kind is avvailable in a wide range of cross sections (roundd, square, andd rectangulaar) and sizes..
Cooling
C is normally req quired for th he AC poweer supply annd the coilss to removee heat
generatedd by the eq quipment. Due to the close proxim mity to the hhot workpieece, the coills are
Thermal insulation
Air-gap reduction
reduced
Multi-layer or multi- Cooling difficulties
Coil cooling
Coil losses 19 % strand configuration More expensive
circuit
Adaptation to
Magnetic interaction
workpiece size (multi-
between coils
coil)
Workpiece Thermal insulation Electrical efficiency
Coil cooling
radiative 4% increase reduced
Thermal
circuit
losses
This project begun in December 2009 will last almost four years and aims to reach the
industrial prototype level of the innovations.
Project Description
The scientific program is organized in five main tasks, each other independent, involving
at least two or three partners and lasting the all duration of the project.
The first task “analysis of induction potential diffusion” aims to identify processes for
which an induction solution may be relevant in term of energy efficiency. Diffusion of
innovative solutions developed in the other project tasks is also analyzed.
The next two tasks study technical developments of induction heating solution regarding
the improvement of energy consumption. In that way, the task “realization of an auto-adaptive
multi-coils power supply” aims to design a flexible power supply able to energize many coils in
mutual interaction and driving the coils’ current to adapt the heater to a large range of heated
pieces. Preliminary works were done before, both on the inverter technology (Manot 2013;
Souley et al. 2009; Souley et al. 2010 a) and on a multi-coil configuration (Forzan et al. 2010).
The aim of this task is to reach a semi-industrial scale. The task “development of low losses
inductors” develops multi-strands solutions having better performance than conventional coils.
This work aims to extend the relevance of the CELINE™ concept to higher frequency range (up
to 400 kHz).
The last two tasks deal with heat recovery on and around induction process. The task
“heat recovery on induction heating losses” studies the pieces of equipment to be added on
existing cooling systems to recover the fatal coil losses. Preliminary studies (Paya 2008; Paya,
Gheorghe & Tudorache 2009) show that it is possible to use high fluid temperature (above 90
Laboratory Testing
g: Descriptio
on and Resu
ults
Auto-adaptive multti-coil powerr supply pro ototype des cription. Too control thee current injeected,
each coill is energizeed by a speccific inverterr. All the inv
nverters workk at the sam me frequencyy, but
current amplitude
a annd phase may y vary. The control of th
the multi-coiil system is organized inn two
interlinkeed loops (seee Figure 2).
T outer loop controls the temperrature profille and deterrmines the rreference cuurrent
The
value forr each coil. The
T inner loo op controls th
he current innjected in eaach coil by comparison wwith a
referencee value giveen by the ou uter loop. To
T achieve thhat, a full eelectromagneetic, thermaal and
power eleectronic mod deling was developed
d by
y LAPLACE E (Souley et al.2010 b).
The
T current control
c requ
uires the kno owledge of the impedaance matrix of the multti-coil
system thhat can be evvaluated eith
her by measu urement or bby a 2D or 3D electromaagnetic numeerical
model. This
T impedaance matrix is introduceed into the ppower electtronic model which givves in
return th
he driving seequences off switches or o gate contrrol signals. These sequuences have been
successfuully implemeented into a 100 kW ind duction heateer. Figure 3 shows the ppicture of thee 100
kW 3 coil power sup pply prototyype built and d operationaal at the EDFF laboratoryy that implemments
the auto-adaptive mu ulti-coil conttrol strategy.
Figure 4 shows the experimental results of the auto-adaptive multi-coil control power
supply testing. The x-axis represents the time in milli-seconds while the y-axis represents the
magnitude of current going through each coil. The source current, phase – 1 and phase – 3
current magnitudes are shown in Figure 4. Key highlight of the waveforms shown in Figure 4 is
that the current magnitude in each of the pairs of phases are varied to meet the specific heat
requirements of the object to be heated which is not obtained in traditional control strategies.
The temperature control requires the knowledge of the normalized induced currents
generated by each coil separately and evaluated by a 2D or 3D electromagnetic model; the global
induced current distribution is then determined by superimposition of the elementary induced
current’s distribution. The thermal problem is modeled directly inside the power electronic
model and gives in return the values of the reference currents to be put into each coil. The
temperature control loop has been already tested numerically.
Experimental validation in a large setting is in progress since January 2013 on a 600 kW
facility implemented at Fives Celes premises. For that purpose, Fives Celes has developed a new
driving card able to control the inverter bridges. Numerical capacities (memory, fast
calculations) have been increased to host the control loops software. Each slave card dedicated to
an inverter hosts the corresponding current loop and the master card hosts the temperature loop.
Fast communication busses at basket bottom allow data transfer between the cards.
Low losses multi-sttrand coils.. The multi--strand conffiguration off coil conduuctor, comm monly
called Liitz wire, is widely used d in small power
p devicces such as domestic innduction coooking
systems.
Itts extrapolatiion to high power
p industtrial applicattions requirees paying atttention to thee coil
cooling. Thus, it is neecessary to have
h a good evaluation oof the Joule losses in thee windings. F Finite
elements numerical models
m quicckly reach thheir limits iff several thoousands of sstrands are ffinely
meshed.
First simulatiions were realized by SIIMAP with a small num mber of strannds. 2D moddeling
(see Figuure 5 and Fiigure 6) sho ows a curren nt distributioon among thhe conductoors looking llike a
global sk
kin effect. Inn the Figures 5 and 6 th he colors corrrespond to the temperaature of the coils,
blue reprresenting colld and red reepresenting hot.
h
To evaluate the global Joule losses in the composite wire, SIMAP has proposed a
homogenization technique (Scapolan, Gagnoud & Du Terrail 2012) the multi-stands coil is
approximated by an equivalent electrical resistivity applied to the global cross section of the
wire. Simulations of an aluminum bar heating with a homogeneous model and with the fine
multi-strand coil models are in agreement in a wide frequency range (see Figure 7). The
significant conclusion drawn from this simulation as see in Figure 7 is that the multi-stranded
coils can be approximated to a solid metal coil with equal cross section area.
Heat recovery. A heat energy recovering system using heat exchangers is in operation in EDF
laboratory since 2010.
The main goal of the heat recovery test bench is to test and quantify in industrial
conditions the heat energy to be recovered coming from the coil cooling circuit. Figure 8 shows
the schematics of the test bench and provide the main functions of the bench. The test bench is
made up of two heat exchangers which can be by-passed according to the control and regulation
system. The “recovery heat exchanger” collects the heat to be recovered and transfers it to the
reuse, the recovery circuit simulator. For our test bench, this simulator is a 300 liters (79 liquid
gallons) water tank (“domestic hot water storage”) which can store the collected energy.
RECOVERY
DOMESTIC HOT CIRCUIT
WATER CIRCUIT
25°C 30°C
20°C
DOMESTIC HOT
WATER STORAGE
50°C
INDUCTOR
CIRCUIT
80°C
16°C
75°C
19°C 30°C
The “waste heat exchanger” evacuates the remained energy to ensure a safe use of the
induction heating device. These two exchangers are designed for 100 kW. The water flow is
obtained by two pumps and their associated valve, the smallest for the range 0.3 – 2 cubic meters
per hour or m3/h (equal to 1.32 – 8.81 gallons per minute or gpm) and the biggest for 1.5 – 9
A new application of induction heating is a non-contact system for the heating barrels
used with plastics molding and extrusion machines developed by Xaloy, Inc. (EPRI 2007). They
have demonstrated that induction barrel heating with an interposed insulating layer increases
barrel heating efficiency to around 95% from typically 40 to 60% with band-heaters. By virtually
eliminating the thermal mass of the heating system, induction accelerates temperature response
to seconds (versus minutes with band-heaters), which greatly improves control predictability and
reduces the sensitivity of control performance to thermocouple depth. In addition, induction
heaters are more reliable than band heaters, and can typically provide three times the heat flow
into the barrel.
Depending upon the size and application, the capital cost of the induction heating systems
can range from under $8,000 to over $1 million. The smallest systems generally cost around
$8,000/kW while the largest systems are perhaps $1,000/kW (EPRI 2007). Maintenance costs
are usually lower than the alternatives due to the reliability of solid-state electronics and minimal
Induction 35 Industrial Park Circle Induction Atmospheres designs and builds turnkey
Atmospheres Rochester, NY 14624 induction heating solutions for continuous flow
LLC 585.368.2120 manufacturing. They evaluate customer parts and
www.inductionatmospheres.com process requirements in their laboratory to determine
the optimum induction heating system that matches
the process requirements.
Induction 209 Travis Lane Induction Systems offers the latest technology in heat
Systems, Inc. Waukesha, WI 53189 treating scanners, power supplies, billet heating,
888.856.2096 custom heat treating systems, water systems, and
www.inductionsystemsinc.com quality monitor systems, as well as servicing all types
and makes of induction equipment.
Inductoheat, Inc. 32251 North Avis Dr. Inductoheat builds induction heating equipment for
Madison Heights, MI 48071 case hardening, tempering, annealing, bonding,
800.624.6297 brazing, strip/slab heating, galvanic annealing, and
www.ihs-usa.com other applications.
Conclusion
Technological innovations of efficient induction solutions are discussed in this paper,
namely, flexible multi-coil power supply with current control for each coil, development of a
multi-strand conductor for industrial induction heating, energy recovery on coil cooling water
with “Predictable Function Control” (PFC). New numerical models are developed and adapted to
the various needs: simplified multi-coil model for power electronic control, homogenization
technique for multi-strand wire. Work is on going to reach industrial applications within 2014.
References
EPRI 2007. “Efficient Electric Technologies for Industrial Heating – Emerging Technologies” EPRI, Palo
Alto, Calif.: 2012. 1024338.
Forzan Michele, Lupi Sergio, Spagnolo Aristide, Pateau Olivier, Paya Bernard. 2010. “Space control of
multi-coil transverse flux inductors”. Heat Processing (8) Issue 4, 2010, ISSN 1611-616X
Manot Gilbert. 2003. “Modélisation couplée des dispositifs électromagnétiques associés à des circuits
d'électronique de puissance. Intégration de la commande des convertisseurs - aide a la conception
d'un dispositif de chauffage par induction à flux transverse”, PhD Thesis, LAPLACE Laboratories,
Toulouse (2003), vol.22, n.1, 134-148
Paya Bernard, Gheorghe Felicia, Tudorache Tiberiu. 2009. “Recovering energy in an induction coil:
Impact on the water and coil temperature”. EPM 2009, 6th International Conference on
Electromagnetic Processing of Materials, Dresden (Germany), Oct. 19-23, 2009, ISBN 978-3-
936104-65-3
Paya Bernard, Pateau Olivier, Neau Yves. 2010. “Development of Energy Saving Solutions for Induction
Heating Devices”, International symposium on heating by electromagnetic sources, Padua (Italy),
May 18-21, 2010, 271-276 ISBN 978-88-89884-13-3.
Scapolan Raphaël, Gagnoud Annie, Du Terrail Yves. 2012. “Modélisation de conducteurs multibrins par
conductivité électrique équivalente”. NUMELEC 2012, 7th conference on numerical methods in
Electromagnetism, Marseille (France) July 3-5, 2012.
Souley Majid, Spagnolo Aristide, Pateau Olivier, Paya Bernard, Hapiot J.C., Ladoux Philippe, Maussion
Pascal. 2009. “Methodology to characterize the impedance matrix of multi-coil induction heating
device”. EPM 2009, 6th International Conference on Electromagnetic Processing of Materials,
Dresden (Germany), Oct. 19-23, 2009, ISBN 978-3-936104-65-3
Souley Majid, Maussion Pascal, Ladoux Philippe, Pateau Olivier. 2010 a. “Simplified model of a metal
disc induction heating system”. HES-10, International symposium on heating by electromagnetic
sources, Padua (Italy), May 18-21, 2010, 137-144, ISBN 978-88-89884-13-3.
Souley Majid, Eagalon Julie, Caux Stéphane, Pateasu Olivier, Maussion Pascal. 2010 b. “Modeling and
control of a multi phase induction system for metal disc heating “. IECON’ 2010, 36th Annual
Conference of the IEEE Industrial Electronics Society, Phoenix, Ariz. (USA), Nov. 7-10, 2010