Designand Construtionofthe Coreless Induction Furnace
Designand Construtionofthe Coreless Induction Furnace
A thesis
By
Supervised By
ur Electrical Engineering
Signature
eramrned the student in its content and what is related with it, and that in our
oprruon it is adequate with VERY GOOD standing as a thesis for the degree
141tor'
signarure
blame Dr.K"6.(rlKrrr {Chairman}
Date t+ -g g1
: ,:iafure l.4M
' *--1!
l,n - g -9n
Signature
Nrme /tA-L! N kly^f {Member}
Date Lt /f/ff '
: :rafure
Neme A .K f,4' A l^ { ho;kl l; {Supervisor}
Date lt-g- t9)7
4r
Signature I
I
The main task of the present work is to desigu and construct a coreless
induction furnace. At start, the operation principle and the advantages of the
coreless induction furnaces were clarified, then the main parts of the furnaces
were listed and discussed in details.
zuperposition method, to check the results of the first method in one step of the
design.
Prior to the use of thdse two methods the dimensions of the furnace, the
ll
ffi
A Cross-section area (m2)
A.E Crucible cross section area (m2)
A* Charge cross section area (m2)'
I
B, Saturation flux density (Tesla)
C Mass specific heat (whrlkg)
D" Coil inner diameter (m)
D* charge diameter (m)
d Internal diameter of circular tube (m)
E Heat energy (whr)
f Frequency (IIz)
f, Critical frequency (IIz)
H Magnetic field skength (A T/m)
h Distance from the center of th.
surface of the load (m)
conductor to the
h. Effective height (m)
t Coil current (A)
IN Amper-@ (A.T)
J Current density (Nmr)
j Complex operator
k Thermal conductivity c')
kct Current transfonner correction factor
k, E Correction factor-equation 3.17
h Correction factor-equation 3.72
L Iength (m)
Lu Base thickness (m)
L Coil length (m)
u Length of the return flux path (m)
/
L- Charge length (m)
M Mass of charge (kg)
li; Number of coil turns
P. Coil copper loss (w)
Pcond Conduction loss (w)
ID
conv Convection loss (w)
iii
P- Input power (w)
P.ud Radiation loss (w)
Pt Power to be removed by cogling (w)
Pu Usefirl power (w)
P* Power required by the charge
PD Power density (*L')
P.F. Power factor
p Dimensionless oonstant-equation Al.rE
7)"on Convection loss density (*/#)
77raa Radiation loss density (*/*')
a Flow rate of cooling water (rn'/t")
q Dimensionless constant-equation Al . 3 0
R. Coil resistance (o)
R-. Coil reluctance (A/wb)
Rmc Crucible reluctance (A/wb)
R-* Charge reluctance (Aiwb)
R* Charge resistance (Q)
S Coil pitch (m)
T1 Ambient temperatur*-(C')'
T2 Pouring temperature (C') i
t Melting time (hr)
tc Conductor width (m)
U Current per unit length (A/m)
v. Coil voltage (V)
VA Volt-amper (VA)
t) Velocity of the cooling water (n/s)
w The length of water circuit (m)
x. Coil reactance (Ct)
xs crucible reactance (O)
x. Return flux reactance (o)
x* Charge reactance (o)
z" Coil impedance (ct)
Z Distance along the load measured from a point directly
beneath the conductor (m)
Ap Pressure drop (pascal)
At Temperature difference between inlet and outlet water
(c") i
IV
rlr Furnace efificiency
0n Thermal efficiency
lt Absolute permeability (lVm)
Iro Free-space permeability (Wm)
lt, Relative permeability
p Resistivity (o m)
o Charge density (kg/*')
T Conductor thickness (m)
o Flux per unit width (wb/m)
0, Coil flux (*b)
dr Crucible flux (wb)
6 Return flux (wb)
h Total flux (wb)
0" Charge flux (wb)
CI Angular Frequency (radls)
v
Acknowledgments I
Abstract... 1l
Chapter one
Introduction................................r............................................r........... l-
Chapfer Two
General View of Coreless Induction Furnaces.............,..................... 4
l 1 Introduction..... 4
I 2 Refractory Linings 5
l 5 Pouring systems 7
Vl
2 6 Water cooling system 9
2.7 .3 .4 Cycloconverter 18
Chapter three
Methods of Analysis of Coreless Induction Furnaces........................ 2A
3.I The numerical methods.. 20
Chapter Four
vl1
4.4 Determination of the required power..... 42
Chapter Five
Design implementation and experimental work 48
vl11
Chapter Six
Conclusions and Suggestions for Fufure'Work...........,..'................... 70
6.1 Conclusions .70
References 73
{ppendix 2. ..........81
\ppendix 3. ..........85
1X
Wffi
Fig"2.1 Schematic illustration of a coreless furnace
showing the pouring spout and tilting axis.. .......8
x
Fig. 5.3 Variation of P.F. withD/D;.ratio ..52
: rg. 5.4 Variation of the furnace efficiency with D./D* ratio. ..53
x1
Table 5.1 Optimum values of D./D* ratio 53
x11
Introduction
subject. In the early days of induction heating, the developments were held
1
1.2 The advantages of the induction furnaces
The induction furnace is one of the most important applications of the
induction heating, in which both magnetic and non-magnetic metals can be
melted, also the induction furnaces are used for the holding and superheating of
metals.
: --mpetitive furnaces, the induction furnaces advantages have ensured that they
induction furnaces because of its suitability for all metal industries [3]. The
coreless induction furnace is a simple unit consisting of a helically wound coil
surrounding a refractory crucible containing the molten charge and connected
to an altemating excitation system. [4]
)
: : -.:olnagnetic field attenuates rapidly inside the charge. The depth at which
rr
Itrr
(-) (1.1)
Eddy currents are concentrated at the outer sur ce of the charge due to
slon effect and heat is developed by the conventional ohmic losses. Part of the
&veloped heat is lost to the surounding medium, and the remaining part is
distributed through the mass of the charge by the effect of thermal
conductivity. tl]
The skin effect is also appeared in the coil conductor. Moreover, the
distribution of the surface current density over the coil is nonuniform, the
greatest current density is to be found in the sections of the coil lying closest to
The aims of this work are preparative generalized design for a coreless
induction furnace, and construction of a coreless induction furnace depending
on the prepared design, with an account on the practical problems that face
zuch a construction.
3
General View of Coreless
Induction Furnaces
2.1 Introduction
The coreless induction furnaces have a large market share in the metal
;lturg industry because of many advantages offered by these
researches and sfudies are continued to improve the performance and operation
asential parts are nearly equivalent. But the differences in capacities and
qlralifications between the manufacfurers have resulted wide varieties of
emeless furnaces [6].
4
The refractory fining(tfrJerucible) is the part that separates the melt from
the coil, so it is an important part of the furnace. Linings of the furnaces
should have satisfactory durability and should fulfill the following
requirements :[7]
\Iechanical : The lining must form a crucible tight and strong enough to take
the melt and to withstand mechanical stresses and thermal shock, it must be
able to withstand not only the loads imposed by charging but also erosion
caused by the powerfrrf melt motion.
: :ectly by both the consistuents of the melt and also by its reaction products.
,.: best requirement of the cmcible composition must neither be react with the
:.t nor get damaged by melt infiltration to the crucible. Selecting refractory
,:erials used in regard to metal melting is possible as stated in reference. [8]
-l
flectrical : For efficient energy transfer, optimum crucible wall thickness must be
chosen, because thicker lining increase the leakage flux, reducing the heating
sf the charge and result in poorer power factor, thinner lining increase both the
risk of mechanical failure and furnace heat loss. [,7]
5
The refractory linings vary according to construction manner. Cast and
pre-fired crucibles are usually employed where frequent changes in melt
compositions are encountered and where contamination between melts must be
avoided (different crucibles are used for different melts), furnabes of cast
crucibles are usually less than 100 kg. For larger capacities, crucibles formed
by ramming refractory materials into an annular gap between the furnace coil
and a metal fonner which can be dismantled after use or more commonly made
out of charge metal and melted out in first melt, this method used extensively
for furnace capacities up to 20 ton. For still larger capacities refractory bricks
The cross section of the coil used can be rectangular, square, rectangular
uith round edge (D-shaped) or circular. Cross section view of any water
eooled conductor franspires that the water channel take the cross section shape
ii.e. all sides of conductor wall have equal thickness. Special coils have
conductors of one side heavier than the other, these coils used only in furnace
operating at mains frequency and the heavier side of the conductor is wound on
-,,pregnated tap or plates inserted between one turn and the another. The main
6
criterion for insulation is to maintain its quality through the life of coil under
hot conditions.
The coil takes an important part in intensi$ring the refractory life through
rn'ith temperature axially and radially, so.the lining must be free to move with
respect to the coil to prevent coil damage if badly wound coil interferes with
this motion. At the szrme time the coil must be rigidlyalampadfboth axially and
radially, to prevent motion that might result in chafing and damage to the coil
insulation [1].
and not in small furnaces. Flux guides zre packets of low-loss electrical sheet
sreel placed behind the coil with the laminations in the radial direction to form
- .,, e11 laminated bath to channel the main flux in the region outside the coil'
The flux guides are also assisted in improving the effrciency of the coil by 1 to
f percent by reducing the sffay losses and reducing the magnetizing curent'
Small furnaces are built without lamination packets, the frame must then be
made of non-magnetic material. [1]
25 Pouring systems
The large furnaces of ramming and bricks refractory crucibles, have to tilt
b pour out the melt, so the frame has to be designed to take the forces at
various angles during pouring. The furnace usually tilts about an axis near the
7
: -'.ring spout to ensure the minimum change of spout position during pouring
.:: Fig.2.1. IU
PorLring spout
Ref roctory
tin.ring
Tiltirrg c:ris
Steel coil
struc!ura
rl[lfi t
Flux guide
Bricli bose
Hy<irouti c roi*n
Some of small coreless furnaces use separate crucible i.e. the refractory is
- rammed and is not part of the furnace. [7] Separate crucible furnace saves
- : to melt metal in a crucible and then to transfer it directly to the point of
-- -ng, rather than transfer it to an intermediate vessel. There are two forms
- ,:parate crucible induction ftirnaces:
8
1- The double - push - out furnace, which has two separate coils fixed in the
l- The lift - swing furnace employs one coil can be readily removed from the
2-2 .tU
:t- clilinder
9
r3 l
I I
1
L
?
(7
I
121
(_
I Furnace: I u'ater-cooled
cable; 3 cooling-rvater monitor: 1 stop-cock; -i temperature-regulating
valve; 6 heat-exchangeri 7 nlanometer (pressure-indicator): 8 pump: 9
header-tank: 10 float-sri'itch lprimarl' make-up riater-inlet): 11 tank
overflow; 12 non-return vaire: 13 secondan' water-inlet: 1,1 *,ater-filter
: --iciency of melt.
10
Eecause of the disadvantages mentioned before, the use of 50l60Hz
: .re installation was regarded as unsuitable, but in the fifties it was shown
.:,,-. Also, difficult starting was overcame by using "starting plug" form
'; ".Irs operation of 15 to 20 percent of total capacity. [1]
-" ,.e ability to melt charges without having to use starling plug.
- , jrlses even less stirring action than rnedium frequency. For reasons of
- :rent costs, it is likely that high frequency will be used for small furnaces
- ; l.recially to melt laboratory quantities of preciotts rnetals.[3]
: - 1 \Iains-frequency supply
Ihe major colnponents of a typical mains-freqllency supply are shown in
. I J, the transforners are usually included for rnatching low impedance,
.,rltage coil to high voltage supplies. It is usual for several primary taps to
:orporated for a variety of coils and for power control, these are switched
. :J This is a normal method of matching and control. As work coils are
11
r..tilnes operated from standard voltages, such as 2301400 volt, transformer
\{arn suPPIv
- , Je eliminated. [1] l
!_
I
Circuit t-rrelker
I
(il requi red )
__<"______
I
t_
-!
Metering
Contactor
Control circuit
Workhandling
Po* er- lector
correcti on
ca0aci to rs
Power control tecti v e
I erlocks
\['ater.,;coieil
Water .> t rxn sini sst o n
*'atcr- prJt ect i ve
de vices
line s
Temperature
Work mers ure men(
and cont rol
(rf uscd)
l----_ _
Most coreless furnaces are designed for single phase working, then it is
umei for phase-balancing and the other for conventional voltage transformation
L2
,-:Jr can be corrected by remotely controlled contactor switched
..-:citors. [10]
three - phase
three phase.
three, phese
Air-break
co n titc tor
Phase-balance Powc r
lranslorme r i r:1ns l ornlcr
Capaci tors
Induction coil
.
- -i of generators of conventional design involves both high speeds and very
"-; number of poles. Since peripheral speeds are limited by mechanical
. :::iS, the pole pitch becomes uneconomically srnall and indeed the field
13
One of the fundamental factors limiting the increase of the working
: - .3r1c/ of motor-generator sets is the hysteresis and eddy current losses in
; ion of the generator which are increased with the frequency. By making
- iotor out of particularly thin sheet-steel, these losses can be lowered
:1r'. [5] However, the substantial increase in thecostofsuchgenerator
- :: .lS has limited their use in practice, and the maximum frequency obtained
:led in parallel with the working coil. Also,a voltage regulator is usually
- ro keep the voltage constant during the heating cycle. Power control is
- ,:iJl1 fumaces.
Variable power-
lactor correction
Series capaci tors
capac.iror
Motcr nd uction
I Induct<tr
starter molor al ternator Lo,rd
con tacl o r
Field
sr": pp ly'
Control Automi:t!.
circuit voltage
al0 f
L4
: -.3 Solid-state converters
The inherentdisadvantages of rotating machines include fixed frequencies,
':,,r'hat low efficiency and, consequently, high operation costs, gfeat weight
-- ru_rh noise level. The disadvantages of these power sources led to some
. -.rprnent of inverters using high-power multiple series and parallel
:cted transistor, limitations on the current levels and the switching times
. :ted this type of system to a few kilowatts of power.
fSerefore, the need arose for a solid-state switching device that could
_- . differs in some way from competitors, all designs stem from four
-'.,;tdamental types i.e. voltage-fed, current-fed, cycloconverter and
15
Ioqd
chqrqc I eristk
operoting I
tl- l--,--*;- --- 1 r I
I pou/er
levet= i I
---l- I
I Thr
I
I
rl \r
I I
rl 'l
I rl I
I ll r
I tt ll Eo I
I
I ll r] I
.t lr I
I
I
|l I
I
,
_l L------l )
inverter toqd
fi tter
f requency
trFonents which increase the cost and complexity of the invgrter, and reduce
t6
' .rer alternatively switches the direct current through the load via thyristor
: - :s 1-2 and 3-4 respectively. When one thyristor pair is conducting,
'---
" :nutation take place by firing the second pair, since the load voltage
r----- -l -*t T ---'tlr T
I
I
T
I I
I. I I
\ 7.
llt' I I
I
I
30 I
input
I I 1.,
I
I 1 I
I I
,I I
.,;rt frecluency of the load circuit and changes automatically with any
iency. [2]
t7
ing
freqmcy
I
l' porer cmtrot
,o I
by dc lottqe
;o vdriotio qt
CL fixed freqrcncY
f requency
-q4gii,5
--
rL.qc rof IDc
cLDC
-l.J Cycloconverter
, .: cycloconverter converts the three phase mains supply directly to the
: operating frequency. Fig 2.12 shows the basic unitsofthistypeof
supply. An input filter is incorporated to reduce the harmonics generated
: - Jnverter and to irnprove the power factor seen by the mains supply.
18
The inverter section consists of a pair of thyristors connected in anti-
: - ,.1e1 in each line of the three phase supply which feeds into a series
:: - iriltt circuit consisting the parallel tuned load circuit, star connected
:citors and inductance. [15]
f- ts Thr
R
C6 L
3b
input
L5
loqd
C1
L
19
Methods of Analysis of Coreless
Induction Furnaces
there are two types of solution: numerical solution and analytical solution
20
r Finite Difference method
The basis of this method is the conversion of the governing differential
equations of the problem into a set of algebraic equations by the use of the
2l
r Hybrid Finite Element-Boundary Element method
This method makes use of the advantages ofttre finite element and boundary
element methods ta ana\yze the unbounded field problems. It gives more
freedom in the choice of mathematical boundaries, therefore, the interior to
In all above numerical methods, high accuracy is gotten, however, they all
suffer from the following drawbacks. [21]
l-These methods give a solution to a particular problem rather than a general
solution which could be used for different cases.
practical applications.
))
t Scale-Model Analogue method
In this rnethod the charge and the induction coil dimensions are scaled
down and the frequency is raised in proportion; the result would enable large
and costly furnaces, drawing very high powers, to be designed from tests
using smaller and more economical furnaces. [1,2]
I Superposition method
This method included the investigation of the surface power density
distribution produced in the charge by a single conductor. Then, the
superposition principle was employed to find the power density produced by
a number of conductors. [23]
The more important methods are the equivalent circuit method and the
superposition method. The details of these two methods are given here.
23
L
coil
n
0u
tl
d
I
u
charge
l
n
I
n
These paths have the following respective magnetic reluctances R**, R*.,
and R*r. The return flux 0a has a reluctance of R r.The assembly total
reluctance equals R*in series with parallel connection R**, R*", and R*, .The
position of R*" with respect to Ro,. does not alter the perforrnance of the
magnetic circuit seriously because, the coil reluctance is generally far higher
than the external reluctance l22l.The magnetic circuit of Fig. 3.2 is formulated
to
R
t +
r mr
A
+
c 'w
mmf R
mc
R
Etw
24
L (3 2)
R* (A/wb)...
ltaltrA
For crucible tr4:1, and the crucible cross-sectional area is:
A, (*') (3.3)
= )
ir:-D'?*
Substitute equation (3 3) into equation (3.2) to obtain:
4L" (3.4)
R*, --
tto r@t -D,*)
The reluctance of the charge is:
N.I. (3 5)
R-*
o-
where /*represents the total flux within the charge, from appendix 1:
4N.I
R** = (3.8)
ttott,Hili,tp - jql
Since, the magnetic field strength is:
H=
N'l' (A.T/m) (3.e)
Lc
the reluctance R^* will be:
4L"
R mw -- (3.1o)
ltolt,il'*(p - jq)
25
The coil reluctance can only be calculated if the spacing between the turns
is ignored for the time being. The coil can then be approximated to a semi-
infinite slab, as it's curvafure is small when compared to the conductor
thickness. The later is greater than the skin depth, the total flux per unit width
being. [1]
As /t, = I for copper coil, and the width is the coil perimeter, the coil flux
can, therefore, be defined as:
/toq,6" H k, (rDc)(l- j)
0"= (3.12)
2
where k' is a correction factor, allowing for the spacing between the turns,
Baker[22] suggested that 1.5> k >1 with l.l5beinga typicalvalue. Later
Riechert [24] could speciff a certain value employing the following equation:
Lc
k, = 0'92 (3.13)
N c tc
where t" is the conductor width
R mc =
N'I"
(3.14)
0"
Substitute equation (3.9) and (3.12) into equation (3.14) to yield
D 2L"
r\mc - lh"k, drD"7a) (3.15)
According to equation (3.2), the reluctance of the return path will be:
4L,
R MI (3 16)
FonD?
26
where I* is the length of the return flux path and it is a function of; coil radius,
the difference between the coil and the charge lengths , and the charge skin
depth L, was found experimentally to be:
where k, is a correction factor depends on the air gap, the difference between
the coil and the charge lenglhs, and the charge skin depth. The value of k* lies
between 0 and 0.1 for the retum flow throu *, ak, and lies between -A.22 and
0.1 if the return flow through iron [24]. For the coil and the charge of the same
length, then equation (3.17) reduces to:
electrical equivalent circuit can be derived if the relation between the electric
and magnetic parameters are known.
Y":ZI" (3.21)
From equation (3.20), (3.21) and (3.22), the relation between the
magnetic circuit reluctance and electric circuit impedance is obtained
27
Then, the electrical equivalent circuit is drived and shown in Fig. 3.3,
where
x Nc
a q ......(3 .24)
o
b
R*u
z*:jN3* .....(3.2s)
z":jN3* (3.26)
x t N?()
"
(3.27)
Rro,
z
c
I
\
Ic t( x
c
v
w
x
w
pi,
Xr:K* n3 - ... (3.28)
4L"
28
z*:R* * jX* = jp) ..(3.2e)
"-*(q*
K,4D.
Z"=R"*jX"=Kn (t *:) (3.30)
2L"
D" (3.31)
X, =KR
0.45 + ko )
Now, the elecfical equivalent circuit has been constructed, then the
R*. (3.35)
4"
R,
The input power can be obtained from the power needed for the charge, P*
29
Pn _P* (w) .. (3.37)
rl,
(3.3e)
Z ... (3.40)
#=n'. N 2
c
IN= .....(3.42)
If there exists a certain preferred voltage V., then the coil turns is :
Nc _v. .. (3.43)
%'
if there doesn't exist a preferred voltage, the coil number of turns is spe cified
according to practical considerations, then the coil voltage is obtained from
equation (3.43).
IN .(3.44)
Ic (A)
Nc
the coil copper loss is:
30
3.4 The superposition method
This method depends on the investigation of the surface power density
distribution produced by a single cuqrent carrying conductor. Then, the
superposition principle has been applied to determine the power density
produced by a number of conductors.
-I . Both the conductor and the image will produce a magretic field H on
the surface, according to Ampere's law, equals to :[25]
I ...... (3.46)
H_
2r h2 +22
where z is the distance along the surface between points P and O, see Fig.3.4
7
TI onduc tor
h
-l
Hp
\
stob I
I
Jp
I
-I
$l*ou.
Fig. 3.4 Conductor near a semi-infinite slab.
31
\
H Hr
2 Hz\
P
HP 2H z (3.47)
I h
H, =2
2r h2 +22 W (3.48)
HP
Ih (3.4e)
1T h2 +zz
Since, the magnetic field strenglh atthe surface of the slab and the linear
current density U induced per unit length of slab are equal .[1]
i.e. Up:Hp (3.50)
h
IJo=I (A/*) (3.51)
' 1T h2+zz
In induction heating practice, the current carrying conductors a.re far from
32
compared with h and the thickness of the slab, equation (3.50) can be
written as:[26]
Up=I l" ,
' 1r h!+zL
(3.s2)
where
JJ
By substituting equation (3.52) into equation (3.54) to yield:
,n:I[nl
ie.ia. .#e] ...(3.55)
If the distance between the conductors are constant and equals to s, then
ol-ol
22 = Zt* S
. (3.s6)
l
LN _ OI +(N-l)s
and
uP = -!- I t" *
*4 h" *- ""-h"l . (3.57)
tEq *1r, *Y
7t +lz, +G'{ - l)tl' h? -]
or
Ih" N-l 1
(3.58)
U p =- T
x=0 h! +
"' ..
7T @r+ xs;z
density PD rather than the current per unit length U. The power density is a
function of the surface current density J, the relation between PD and U for
semi-infinite slab at point P is given as:[l]
JP
J' H (A/*'). (3.5e)
P
6
Substitute equation (3.50) into equation (3.59), to have:
JP
J' U P
(3.60)
6
and the power density at point P is given by: [1]
I
PD p5 J 2p (w*') (3.61)
2
34
PD=*[F:u*=r]' .....(3.62)
3.5 Conclusion
In this research, the equivalent circuit method will be used to predict the
Baker's method is still followed at the current time because it is easy and its
results are accurate. Afterwards Reichert\ 1241 was able to develop the
Generally, the equivalent circuit method has known defects, to adopt this
method successfully the defects are to be reviewed and discussed regarding
their effects on the specific application in this research.
J5
The equivalent circuit method suffers, mainly, from two
imperfections:[21]
l-The assumption of uniform magnetic field along the length of the charge
hinders this method from being applied to applications which require
nonunifonn densities in the load.
on the
This imperfection is outside the zone of equivalent circuit application
coreless furnaces since in these furnace the magnetic field is uniform
along
largest value of amper-turns needed by the charge during the heating cycle
36
the magnetic material [29], among these expressions is that given by
Baker:1221
1.43 x 106 Bs
Fr- (3.63)
HM
37
The Design Procedure
4.1 Introduction
The equivalent circuit method (also the other methods mentioned in
chapter three) are general methods that can be applied to analysis and design of
all induction heating applications, the coreless furnace is certainly one of them.
38
4.2 Design of crucible
The selection of the proper size of the furnace, and hence the crucible,
depends largely on the desired output, also the proper size follows from the
desired melting time. [9]
Generally, the actual size of the crucible is larger than that required by
molten metal, the excess size depending on the nature of the scrap to be
charged in a typical example the cruciblehas avolume 2\o/olarger,thanthe
other hand, if height to diameter ratio is too small, end effects of the coil result
in excessive reactance which leads, in turn, to increase the cost of power factor
correction capacitors, also small ratio results in greater heat loss from the open
top. [9]
Height to diameter (L*/D*) ratio stated in different references varies,
ranges 1.1 - 1.3 [31], and I - 1.5 [9] are regard as good design.
39
As mentioned in section 2.2, crucible zlre usually not absolutely
cylindrical but are tapered, with bottom having a diamete
r S-l1yosmaller than
the top' The outer surface of the crucible which is in
touch with the coil is
nonnally cylindrical
40
Moreover, there are two special requirements which dictate the selection
of frequency for induction melting, these are.l32l
l-The initial heating of a charge of scrap metal.
2-The stirring of the molten metal by electromagnetic action.
10 000 l0 kHz
.1.1 kHz
-1 kliz
N
I kHz
o
<' kHz
o kHz
kHz
kHz
10
1
4I
I Zone "A" always gives successfi,rl and economical operation.
I Zone "8" represents furnaces that were working, some uneconomically and
some very well, but were caution had to be taken in selecting the
application.
I Zone "C" represents fumaces could melt metal, but sometimes at a great
expense in metal quahty, lining life, and excess operation costs.
I Zone "D" represents furnaces that never should have been built. [33] after
[34]
For any furnace (defined by the mass of charge to be melted), ttre
operating frequency must be chosen from z;one "A" only. It is found that even
zone "A" represents a relatively wide range of operation, the choice within this
zarte must be carefully discussed with the furnace manufachrrer taking into
account the type of metal to be melted, the range of scrap and the production
rate.[1]
where M is the mass of charge i, kg, and C is the mass specific heat in whr&g;
mass specific heat is a physical constant to the charge material. The useful
42
E (4.5)
Pu (w)
t
with the weight of the charge, there is a practical lower limit for the melting
time. Small units may melt their charge in six to eight minutes, but for medium
43
where k is the lining material thennal conductivity. In equation (4.6), the first
term represents the conduction loss from the crucible wall, while the second
This equation does not include the boundary thermal resistance between
charge and insulation or between insulation and coil, equation (4.6) is based
on the assumption that the entire space between charge and coil is filled by
insulation.
Radiation loss from charge surface can play an important role lat high
temperature. Temperature below about 850 C'usually results in very little loss
in most metal. Above this temperature, radiation loss becomes significant.
The radiation power density is given by the Stefan-Boltzmann law:[l]
D
r
nDl ?,ud (w) (4.8)
rad
4
contained in a crucible, which does not permit air movement. For the part of
the charge in open air, i.e. the surface of the charge. A good approximation for
convection loss density is given by: [1]
44
Now, it is customary to describe the thermal performance of the furnace
by means of an efficiency figure. Such figwe defined usually as the thermal
...(4.12)
"'l=*
4.5 Induction coil design
The induction coils are the most important elements of the coreless
induction furnaces, the efficiency and operational qualities of the furnace
depend, to a considerable extent, upon the geometric dimensions and the
design of the coil.[5]
The first step to design the coil is the determination of the coil height.
Generally the height of the coil is equal to the height of the charge. [] There
are some exception, for example coils of shorter than the charges are used to
reduce the stirring in low frequency furnaces. [9]
The next step is the selecting of the number of turns of the exciting coil.
The choice of coil number of turns must be bowed to the following limits; large
number of turns of small conductor results in small water passages, giving risk
of clogging. This may impose a limit on the maximum number of tums. On the
other hand, for small number of furns of large conductors, insulation problem
may appear. The elecffical insulation must be adequate for the maximum value
of r.m.s voltage applied continuously and for transient of five to seven times
this value, which occurs when switching the furnace load and the capacitor
units. Thus, the insulation criterion may impose a limit on the minimum number
of turns.
If there exists a certain preferred coil voltage, the coil turns could be
obtained directly from equation (3.43),while the limits mentioned above must
always be considered. [1 ]
45
Under the influence of proximity effect and skin effect, onlya small layer,
parallel to the inside surface of the coil and with a thickness of skin depth, is
carrying the current. The rest of the cross-section of the conductor assists to
support the current-carrying part and serves as a wall for hollow through which
the cooling water flows. Principally, the conductor thickness r should have a
the tlrickness is frequently selected between 26c and 2.256c. [9] In addition,
Edgerley regarded a value of r around 1.576c as a good choice.[4]
in addition to the coil copper losses, heat which reach the coil from the melt via
the crucible wall have to be cooled away by the cooling system.fl]
trxperience has shown that failure of coil insulator and refractory lining
The flow rate of cooling water needed to remove a certain amount of heat
is: [ 1]
PT
a 1.15 At
1m3 Itr) (4.13)
46
and, At is the temperature difference between inlet and outlet water. Inlet
temperature must be less than 32Co, but excessive cold water can cause
condensation which, in turn, can cause damage to coil insulation. [35]
The velocity of the water inside the conductor is given by: [1]
4Q
u- '--'-'-'-'------ (r/s) (4.15) I
3600 r d'
where d is the internal diameter for circular tube .E has been found
practically that equation (4.15) could be used for non-circular sections by the
47
Design ImPlementation
and ExPerimentalWork
t
48
The first two progrurms complete the design of the fumace, taking into
notice that this design can be applied to conclude designs of furnaces with
different capacities and able to melt different charges. While the third program
and it's results were used only to compare the number of amper-turns with the
shown in Figs. 5.1 and 5.2 respectively. Also the listing of the three progrzuxs
are given in Appen dlx 2.
start
nrput
t
tum
es
rl c,
YA,IN, VCS
does
no yEs
there edst a
certain preferred
voltage
49
start
calculate lh/rc,
h3, 8",h, P",,, AL
n=2
EO L=0
calculate ap n=n+ 1
p=p+
t
L=L+AL
yes
L< Lc
P>P
yes
print lN
stop
50
5.1 Design of the furnace system
The desis of the furnace obviously depended on the computer
programs
to achieve the optimum performance possible for
the furnace
--
51
P.F.
0.4
_LWDW=1
-+LMDw =1 .2
.4
-X-LMDw=1
*Lw/Dw=1 .6
#LwlDw=1 .8
#Lw/Dw=2
0.3
0.2
0.1
0 Dc/Dw
1.1 1.'t5 1.2 1.25 1.3 1.35 1.4 1.45't.5 1.55 1.6 1.65',t.7 1.75 1.8
52
nF
0.35 _LW/DW=I
-+LWDw =1 .2
.4
-X-LMDw=1
0.3
-#Lw/Dw='l .6
-& LwlDw=1 .8
--{-LwiDw=2
0.25
o.2
0.15
0.1
0.05
*
0 Dc/Dw
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
The above figures indicate that the increase in D.lD* ratio results in
increasing the furnace efficiency and decreasing the power factor, optimum
values of D./D* are tabulated in table 5.1.
L*/D* D./D*
1 1.1 23
1.2 1.13
1.4 1.15
1.6 1.1s
1.8 1.16
2 r.t7
Table 5.1 Optimum values of Dc/I)w ratio
53
According to the above table, it is better to make a thin-walled crucible.
However, to guarantee that the crucible does not break during it's
manufacturing and to guarantee a good mechanical performance of the crucible
during the furnace operation. The thickness of the crucible must, at least, be
7.5 mm and determining the crucible wall .thickness by this limit make the
optimum design of the crucible depends solely on the L*D* ratio. For crucible
wall thickness equals to 7.5 mm, Fig. 5.5 shows the variation of the furnace
power factor with L.,/D* ratio, Fig. 5.6 shows the variation of the furnace
efficiency with L*lD* ratio.
P.F
0.1
0.08
0.06
0.04
( 0.02
0 Lw/Dw
1 1.2 1.4 't.6 1.8 2
qF
0.3
0.26
o.22
0.'t8
o.'t4
0.1
0.06
o.o2
LilDw
+-
1 1.2 1.4 1.6 1.8 2
54
The optimum value of L*/D* was concluded from Figs. 5.5 and 5.6 to be
1.47 . As previously mentioned, it is preferable for the crucible to be tapered,
so the inner diameter D* at the top was 60/o gleater than the diameter of the
bottom. Finally, since there is nothing against increasing the thickness of the
base of the crucible, on the contrary, increasing the thickness decreases the
conduction heat loss, so it was chosen to increase the thickness of the base of
the crucible. The dimensions of the designed crucible are given intable 5.2
equal fc:462 Hz (see equation 4.3), so that the use of main-frequency power
supply is not possible. Also, because of the lack of a motor-generator set which
construct a static-inverter as a power supply. In sec. 2.7 .3, details were glven
The overall circuit is made up of three parts including rectifier, chopper and
load circuit, a complete circuit diagram of the power supply is shown in Fig.
5.7 the three phase input voltage is converted to DC. voltage by means of three
phase rectifier circuit. The DC voltage is then chopped using the MOSFET
5)
with an external oscillatoi in the firing circuit to give the desired frequency.
i:liii';'r,!i
:.rl i]-: f i*
Tfu::e*
1,'
Sltrrc{
Ari:;
k fr
i-H Jitillii-
,. ;(: i
I;'$ tr*"{t $ip?',*i Fr.**ii,i,ll.!*l
Erl'E {:a ij: { $-1.*t'i+1. *}'
t
-
{"a1.'1qiqt$e;r
t,
f'**::r-i!.*:*
i:-;;;, !
50
fl
male, female to demarcate the shape of the cruoible, and pusher to press the
fireclay within the mold between the first two parts.
The stages of consffucting the crucible included putting the fireclay in the
mold then pressing this material by a press of 25 ton via the pusher.
Then the shapen fireclay was exfracted from the mold. After that, it was
fired in a furnace of about 1600 Co for 14 hours.
During the firing process, the fireclay would definitely contract, this
contraction was taken into consideration during the design of the mold, there
was a limited difference between the obtained crucible and the preliminary
design, the maximum difference was less than I .75% the mold and its cross-
section view are shown in Figs 5.8 and 5.9 respectively, cross-section view of
the obtained crucible is shown in Fig. 5.10.
,
r
rt
57
{
,r ll
male
ttl
l{;
5$
,
female
?s
l{
pusher
:I
58
IU
$s.2
.(
l{f S
i] i '-S,
IN
A.T
2500
2000
1500
1000
500
0
f
1 2 3 4 s 6 7 B 9 10 r 12 19 14 1s kl{z
59
IN
T
2000
1500
1000
500
0
f
123 4 5 6 7 8 910 '11 ,i,z t, 14 15 krlz
From those above figures, it was obvious that the results of the first
method are higher than the second method, with a difference up to 12.5%. The
t results of the superposition method is omiued and the resultant numbers of
amper-turns by the equivalent circuit method was chosen as a safer option.
60
3.S
3"4-*-_+
t \#
t !t
I
tt ^l r "*.-. il r-l
'\a q
4
"1* -r,i} L.
I
rd
6l
5.2.3 Construction of the po\Mer supply
The power supply circuit was connected according to Fig. 5.7. A
conductors of heavy diameter were used to connect the elements of the circuit
capacitors must have a very small loss factor). Because of the lack of data
sheets for the available capacitors, these capacitors underwent many tests to
t
5.3 Operation of the induction furnace
5.3.1 The practical measurement
For observation of the furnace perflormance during it's operation,
the secondary of the transformer. A correction factor (kct), was then estimated
-
experimentally, to compensate the ratio effor. The circuit used to estimate the
62
signal RL
kct
1.5
I 1.4
1.3
1.2
1.1
1
i
o 15oo 3ooo 4s0o 6000 Tsoo gooo io50o 12ooo 13s0 150co Hz
fluctuations were detected (due to the fluctuations in the voltage of the public
supply network).
63
5.3.2 Melting of the charge
The experiments were carried out with a tin charge of 300 gram, the
power supplied to the furnace was limited by the rating current of the
autoffansfofiner used. During each test of charge melting, the frequency used
was constant, and many test were undertaken, with different frequencies. In the
experiments of frequencies below lzk}Iz, the temperature of the charge was
rising, but not enough to melt the charge. Melting the charge was achieved at
frequencies of lzkllz and above, the measuring melting point of the charge
was 265C" .Fig. 5.17 indicates the rise in the charge temperahre throughout a
heating cycle when melting the charge was not achieved and Fig. 5. 18 when
i
Temperature
c
250
no
150
100
50
time
0
min
0 5 10 15 20 25 30 35 & 45 50 55 60
64
temperature
C melting point
300
zfi
200
150
100
50
time
0 mint
o 5'lo 15 20 25 30 35 N 45 50 55 60
During the heating cycle, the used instruments could not sense the
I have the sensitivity to detect the mild variation (for tin, the variation is only in
the resistivity and the relative permeability is l). The measuring impedance and
the power factor of the furnace as a function of the temperature are shown in
Z
O
0.49
0.48
o.47
0.46
0.45
o.44
0.43
temperahre
a
30 50 70 90 110 130 150 'l7o 190 c
65
P.F
0.13
o.128
o.126
o.124
o.122
temperature
o.12
30 50 70 90 'l1o 130 150 170 190 C
T1
0.5
o.4
0.3
o.2
0.1
0 f
10 10.5 11 11.5 12 12.5 13
kIIz
66
The above figure indicates that the efficiency of the power supply
decreases with raising frequency, and the reason was that the electrical losses
in the elements and the conductors of the power supply increase with the
r\alsing frequency of the losses, the experiments was continued for frequencies
0.65
0.6
0.55
0.5
0.45
0.4
f
10 10.5 '11 11.5 12 12.5 i3 kJIz
67
5.4.2 The furnace efficiency and the required amper-turns
On practical basis, it is difficult to estimate the induced power in the
chargel2gl and the heat losses associated with the melting process. So it was
impossible to identi$r the electrical and therrnal efficiencies of the furnace, but
the overall furnace efficiency "T1u" (where Tlr: T-1. x tlr,) can be obtained by
measuring the input power to the furnace and calculating the ou@ut power
required to melt the charge Table 5.3 illusfrates the function efficiency
obtained theoretically and measured practically.
t
0.12s t2 16.93 10.14 0.021 30
68
5.4.3 The voltage waveforms
The operation of the power supply can be demonstrated from fig. 5.23,where
Fig. 5.23-A represents the output voltage from therectifierandFig.5.23-8
rephesents the voltage across the chopper.
a-
s :,-- ,-
b-
r
Voltage acrossde
69
r
Conclusions and Suggestions
for Future Work
I
6.L Conclusions
From the stages of design, construction and operation of the induction
furnace, it can be concluded that :
1- It appeared clearly, during the stages of designing the furnace, the contradict
two parts only that demarcated the shape of the crucible. The obstacle
observed then, was the break of the shapen crucible during it's extraction
from the mold. This was treated firstbymakingtheinnersurfaceofthe
female part tapered, also the inner surfaces of the mold smoothed well. Still
70
with these maneuvers, the improvement was partial. The final solution was
to add a third piece to the mold, which is the pusher, to distribute the
3- During the firing process, the crucible would conhact, the degree of
contraction depends on the density of the fireclay material (which is
detenrrined by the pressure of the press) and on the water content of the
fireclay. So many trial were required to manipulate these factors to get a
crucible with the most proper dimensions in regard to the preliminary
design.
A- The results of the equivalent circuit method was of good accuracy, and it's
error didn't exceed 3.77% for the furnace impedance
This provide , that in this work , the equivalent circuit
method was a suitable choice to design the induction furnace.
B- The differences between the readings of the measuring instruments and the
used out of it's specific frequency), and finally due to the neglect of the
effect of the harmonics.
5- Regarding the firmace efficiency and the required amper-turns, it appeared
that there were differences between the theoretical calculations and the
practical measurements. The factors mentioned in -4- above contribute in
7t
these differences, with additional factor, which is not taking into
consideration in the theoretical calculations the gradual increase in the
losses during the heating cycle (nil at the start and maximum at the
l- As, the general design of this work proved it's success with the construct of
furnace done. It is worlhy to try to apply this design to consfruct fumaces
of larger capacities, and tomelting metals with higher melting point.
2- By applying the scale-model analogue method , designs can be prepared for
furnaces of larger capacities, by using the consffucted furnace as a model
for practical tests, and by doing the practical tests on this model expenses
are decreased.
3- The constructed furnace can be used to study the effects of the coreless
induction furnaces on electricity supplies and, then, suggest solutions for I
73
-\
li- Hobson, L., *Static Inverters for Industrial Process Heatingi', IJEEE,
'
Vol. 20, 1983, pp.23-31 .
14- Londis, J.,"A Stafic Power Supplytor Induction Heattngi', IEEE ffans.,
mC l-17 ,197A, pp. 313-320.
. l5-Havas, G. and Sommer, R., '7 High Frequency Power Supply for
Induction Heating and Meltingi', IEEE ffans., WC l-17,1970,pp.321'
326.
16-Nemoto, K. and Taubuchi, M., "Thermal Analysts of Inductton Heating
by the Finite Element Using o Computet", 10th congress of the
international union for electroheat, Stocktrolm, 1984.
t
l7-Reichert, K., "A Numerical Methodtor Calculating Induction Heating
InstullatiozC', Elecfrowtirme Int., Vol. 26, 7968,pp. I 13-123.
l8-Dudley, R. and Burke, P., "The Prediction of Current Distribution in
Induction Heating Installattow", IEEE trans. IA-8, 1972, pp.565-571.
l9-Fawa, T . , Ali, K. and Burke ,P., "Bottndary Integtal Equations Analysis
of Induction Devices With Rotational Symmetty", IFEE trans., Mag-19,
1983, pp.36-44.
74
23-Al-Shaikhli, A. and Hobson, L., "Novel Technique for the Design of
Induction Billet Heaters", IEE proc., Vol.l33, 1986, pp.323-330.
24-Reichert, K., "The Calculatton of Coreless Furnaces l|/ith Electrically
75
36-Mclachian, N., '?essel Functionfor Engtneers", Clarendo press ,1955.
37-Hobson, L., "Laboratory Experiments Incorporating a Solid State
Inductton Heating Power Source",IJEEE, Vol.20, 1983, pp. 351-359.
t 38-Edwin, S., "Power FETS und Their Applicutions",Prentice-Hal1, 1982.
39-Landef , C., "PoweF Electronics", McGraw-Hill Book company (UK), 2'd
Edition, 1987.
40-Al-shaikhli, A., "A Novel Method fot the Design of Induction Heattng
Work Cotls", Ph.D. thesis, Loughborough University of Technology,
1985.
76
-