llesign and GonstruGtion

of the Goreless lniluotion [urnaGe

A thesis

Submitted to the Department of Electrical and Electronics Engineering

University of Technology in Partial Fulfillment of the Requirement for the

Degree of Master of Science in Electrical Engineering

By

ffiJ!*hsffilmru*d ftdf;omr:*s S. &tfr

Supervised By

D J. [.](.m. Al-$h aikhli

1997 s.-r*"all Ab)i

l13;(] : t I i^!r, ..3(.tt;te Lalt
o -rJl
 I certifu that the preparation of this thesis was made under my supervision

at the Electrical and Electronics Engineering in the University of Technology in

partial fulfillment of the requirements for the degree of Master of Science

ur Electrical Engineering

Signature

Name: A-K t4" At-{l,";kLl;

Date: ly-S - 1211
 We certify that we have read this thesis and as an exilmining committee,

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

of \laster of Science in Electrical Engineering

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*'Br ,7, r\qr whM

Date
e).&- {ff
z
 I wish to express my sincere gratitude to my supervisor, Dr. A. K' M' Al-

Shaikhli, for his conscientious, continued encouragement, useful suggestions

&ning this work and his ready generosity wrttr time.
My gratefirl appreciation to all who assisted me and directly or indirectly

contributed to the preparation of this thesis.

Also, I thank gratefully Mr. Jamal. S. Abdul-WQhid for his helpfrrl
srggestions and faithful cooperation.
Finally, thank to the department of electrical and electronic engineering
f,w providing the facilities necessary to accomplish this work.

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.

Many numerical and analytical methods could be applied to induction

heating and melting problems. The most suitable method for the present work
is the equivalent circuit method, a second approach was used which is the

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

operating frequency and the required power must be detennined as a pre-

requisite.

A design procedure was developed as computer programs, these

fograms accomplished a general design which can be utilized to design

coreless furnaces of variable specifications. Then, a design meeting certain

specification was prepared, taking into consideration the available facilities.

Ultimately, the furnace was constructed and the final aim, which is
mhing the metal, was achieved.

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

6, Coil skin depth (m)

6, Charge skin depth (m)
o Emissivity coefficient
0" Coil efficiency

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

List of principal sSrrnbols iii

List of contents \ri
List of figures ............. x
tr-ist of tables............... xii

Chapter one
Introduction................................r............................................r........... l-

1.1 Background to induction furnace ........ 1

1.2 The advantages of the induction furnaces . '....,..,,..2

1.3 The coreless induction furnace.... ...........2
1.4 scopeof the thesis a
J

Chapfer Two
General View of Coreless Induction Furnaces.............,..................... 4
l 1 Introduction..... 4
I 2 Refractory Linings 5

I 3 The Induction coil. 6

I I Flux guides 7

l 5 Pouring systems 7

Vl
2 6 Water cooling system 9

2.7 Power supply systems t0

2.7 .l Mains-frequency supply lt
2.7 .2 Motor-generator sets......... 13

2.7 .3 Solid-state converters l5

2.7 .3.1 Voltage-fed inverter 15

2.7 .3 .2 Cun ent-fed inverter t6

2.7.3.3 Variable mark space ratio inverter...... 18

2.7 .3 .4 Cycloconverter 18

Chapter three
Methods of Analysis of Coreless Induction Furnaces........................ 2A
3.I The numerical methods.. 20

3.2 The analytical methods 22

3.3 The equivalent circuit method... ............... 23

3.3.1 Derivation of the equivalent circuit 23

3.4 The superposition method........... 31

3 .4 .l Derivation of the superposition method... . ....... 3l

3 5 Conclusion 35

Chapter Four

The design procedure of coreless induction furnaces f............... 38

4. I Introduction....... 38
: I Desigr of crucible 39
4.2.1Determination of height to diameter ratio 39

4.2.2 Determination of crucible wall thiclcress. 40

4.3 Frequency selection 40

vl1
4.4 Determination of the required power..... 42

4.4.1Determination of the useful power 42

4.4.2 Determination of heat losses 43

4.5 Induction coil design..... 45

-1 6 Determination of the required cooling 46

Chapter Five
Design implementation and experimental work 48

5.1 Design of the furnace system 5l

5. I . I Preliminary preparation.....,....... 5l
5 .T.2 Optimizing the crucible dimensions 5l
5.1.3 Design of the power supply 55

5.2 Steps of the furnace construction........... 56

5.2.1Construction of the crucible 56

5.2.2 Construction of the induction coil s9

5.2.3 Construction of the power supply 62

5.3 Operation of the induction furnace 62

5.3.1 The practical measurement 62

5.3.2 Melting of the charge 64

5.3.3 The limitation of the power supply..... 66

-< -1 Results and comparisons 67

5.4.I The impedance of the induction furnace 67

5.4.2 The furnace efficiency and the required amper-turns ............ 68
5.4.3 The voltage waveforms 69

vl11
Chapter Six
Conclusions and Suggestions for Fufure'Work...........,..'................... 70
6.1 Conclusions .70

6.2 Suggestions for future work 72

References 73

Appendix 1 ............. ..........77

{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

Fig,2.2 Sections through "lift-swing" and "push-out" furnaces..............9

Fig.2.3 Schematic diagram of water-cooling system. .10
Fig. 2.4 Basic mains-frequency system ...... t2
Frg.2.5 Schematic diagram of phase-balanced power circuit l3
Fig.2.6 Schematic diagram of motor-generator set t4
Frg.2.7 Schematic diagram of voltage - fed inverter t6
- rg 2.8 Load characteristic of voltage-fed inverter...... l6
:;g.2.9 Schematic diagram of current-fed inverter l7
?rg.2.10 Load characteristic of current-fed inverter ...18

?ig.2.11 Schematic diagram of variable mark space ratio inverter ...18

? ig. 2.12 Schematic diagram of cycloconverter. ...19

: ig 3.1 The flux paths in the coreless furnace 24

?tg.3.2 The equivalent magnetic circuit 24

: rg,3.3 The equivalent electrical circuit ..... 28

- lg 3.4 Conductor near a semi-infinite slab 3t

- Lg. 3.5 The magnetic field intensity on point P 32

rg, 3.6 Number of identical conductors at a specific

air gap from a metallic load ......33

- -g. 4.1 Choice of frequency with furnace capacity ...........41

" ;g,5.1 Flow chart of the progam "Equivalent Circuit" 49

- .g,5.2 Flow chart of the program "Superposition" 50

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

Fig. 5.6 Variation of frrmace efficiency (at pouring temperature)

with L*/D* ratio. ...... .........,,..54

Fig.5.7 The circuit diagram of the power supply

:rg 5.8 The mold

:rg 5.9 Cross-section view of the mold....... .......58

: rg 5 . 10 Cross-section view of the crucible

- rg 5.1 1 The amper-turns, versus frequency,

obtained by the equivalent circuit method 59

I. -< 12 The amper-turns versus frequency obtained

by the superposition method .......... 60

'; 5 13 Cross-section view of the tube........ 6t

.; 5 14 The furnace ...... 6t

'. 5 15 The circuit used to estimate kct........ 63

' _ 5 16 kct as a frurction of the frequency 63

' . -t 17 Temperature rise in the charge for

(0.114 kw, 11.5kHz operation) 64
:.j 18 Temperature rise in the charge for

(0.125 kw,72 kllz operation) ...6s

' - : 19 Impedance as a function of the charge temperature

for (0.092 kw, 10 kl{z) ...65

: -{ l0 P.F as a function of the charge temperature

for (0.092 kw, 10 kHz operation)...... 66

'' - i I1 The efficiency of the power supply.. 66

' - - lrurnace impedanoe ........ 67

' : i Voltage waveforms....... 69

x1
Table 5.1 Optimum values of D./D* ratio 53

liable 5.2Thedimensions ofthe designed crucible (in mm) ..........55

Table 5.3 The furnace efficiency .......68

Table 5.4 The required amper-trns ...... 68

x11
 Introduction

1.1 Background to induction furnace

Elecfomagnetic induction principle was discovered by Faraday in 1831,
dris principle was first applied in motors, generator, ffansformers, etc., and any
heat effect was considered as undesirable side effect and the efforts were
made to minimize this eddy-current heatingfl].

It has shown practically that eddy-current could use to arise the

temperature of conducting materials. This had led to the induction heating

subject. In the early days of induction heating, the developments were held

back by the lack of motor-generator sets and power.factor correction

cqacitors of sufficient size and reliabilrty. By 1927, the first induction furnace

ryplied by motor-generator set,was installed[2]. Inthe fifties, establishing

-fbasfria
meins frequency induction furnaces was shown to be .
Afrer that, the advent of solid-state frequency converter systems have made
- .;tion furnaces to be more efficient and economic.[l]

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.

Induction furnaces have a number of inherent benefits with fuel fired

: --mpetitive furnaces, the induction furnaces advantages have ensured that they

.'.ive achieved a substantial market share of the melting installations.

The advantages of induction furnaces are: [6Jl

l- Very fast melting rates can be achieved.

2- Low running costs as compared with those of the more efficient designs of
fuel furnaces.

3- Automatic stirring which provide mixed melts.

4- Special techniques (such as vacuum melting) are possible.

5- Low maintenance costs on refractory linings.

G Lnproved working conditions and provided environmental benefits.

7- Freedom from deffimental gasses and combustion products.

t- Low noise emission.

13 The coreless induction furnace

The coreless furnace is considered as the most popular type of the

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]

When the coil is excited by an alternating current it produces an axial

&ernating magnetic field. The variation of magnetic field produces an electric
fELd and eddy currents along paths concentric with the axis. The

)
 : : -.:olnagnetic field attenuates rapidly inside the charge. The depth at which

- rrduced density falls to (1/e) of its surface value is usually referred to as

' ,, depth. The skin depth 6 is given by :

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 charge, this phenomenon is known as the proximity effect. [5]

The induction furnace is, therefore, based on four principles:

Electromagnetic induction, skin effect, heat transfer and proximity effect.

1.4 Scope of the thesis i

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

of such furnaces. Furthermore, the coreless furnaces have variable capacities

wting from less than lkg to more than 100 ton.

The basic principle in coreless induction furnace instruction is one and the

asential parts are nearly equivalent. But the differences in capacities and
qlralifications between the manufacfurers have resulted wide varieties of
emeless furnaces [6].

The essential parts of coreless furnace will be described in the following

. - .Jl]S

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.

Thermal : Heat loss should be as low as possible, this may be achieved by

foserting a layer of thermal insulator between the coil and the crucible.
Increasing the insulation thickness reduces heat loss further, but at the same
time the insulation may cause high loads on the crucible lining that may end in
damaging the lining. For economical reasons and longer durability for the lining

hirly high heat losses are accepted.

\letallurgical : According to its cornposition, the crucible may be attacked

: :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

are used. [1]

Maximum wear on the furnace lining occurs at the bottom as a result of

the turbulence within the melt, so it is important to taper the crucible, with a
bottom having a diameter 5-l0o/o smaller than the top. [9]

2.3 The induction coil l

The conventional coreless induction furnace design consists of a water

cooled single layer solenoid coil which sets up the excitation of field and
provides controlled cooling for the refractory crucible.

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

tre inside of the coil. [9]

The electrical insulation of the coil may bp carried out by air spaced,

-,,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

maintaining a steep temperature gradient through the refractory wall. Some

put additional cooling coil (water cooled but not energized) at
manufacturers
the top and bottom of the active coil to provide uniform cooling, in an attempt

to reduce the thermal stress on the lining.

It is important to take into account that the lining will expand and contract

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].

2.4 Flux guides

Flux guides are used only in large furnaces of more than 500 kg capacity,

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

Fig 2.1 Schematic illustration of a coreless furnace

showing the pouring spout and tilting axis

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

body of the furnace as shown n Fig. 2.2 , a separate crucible can be

lowered into each coil altemately. The crucible stands on a hydraulically-
for lifting
operated platform, while one molten charge is raised and be ready
a way then pouring, the other crucible is lowered into the second coil and
power is put on. By alternately charging two crucibles, molten metal is
constantly available.

l- The lift - swing furnace employs one coil can be readily removed from the

crucible and transferred to another prepared base with another crucible,

rr-hile the crucible of the first furnace is ready for pouring , see Fig'

2-2 .tU

:t- clilinder

Lift coil Push-out

Fig.2.2 Sections through "lift-swing" and "push-out" furnaces

- , \\ ater cooling system

,.-.rable and adequate cooling for the coil is indispensable for induction
operation. The transformer and power factor correction capacitor units

:.s: be water cooled to maintain operating temperature withinlimits

of the frulace system may be air cooled (if required)

: - lllponents

- -. 1s a schematic diagram of the type of closed-circuitwatercooling

.: st colrlnonly used. [6]

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

Fig. 2.3 Schematic diagram of water-cooling system

I - Power supply systems

The frurction of the furnace is essentially dependent on the frequency
:: t. So furnaces can be classified according to the frequency into: [3]

- l.lains frequency - 5016AHz

-- l.ledium - frequency - up to lkHz

- ,,rglr - frequency - ll<Jlz and above
-\t first sight, it is desirable to use mains frequency equipment wherever
,!.lle as the first cost is apparently less than the medium and high -
- .3ncy equipments. The fundamental disadvantages of mains frequency
" -ii$ are: [1]
" . 3ry high stirring which means that the power / weight ratio must be kept
_ -i'

: --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

-- 3 a commercial possibility. Possibly the prime influence in establishing

' - ., Hz coreless furnace was the growing cost of raw materials and the need

::.elt low-cost, low-grade scrap convert it to high-quality, high- value, liquid

.:,,-. Also, difficult starting was overcame by using "starting plug" form
'; ".Irs operation of 15 to 20 percent of total capacity. [1]

),ledium frequency furnaces were powered by either motor generator sets

' ':., rnagnetic triplers, developments in solid-state frequency converter have

: l-,re d other forms of converter obsolete. The use of medium frequency

compared to mains frequency equipment.

-'-r;e provides many advantages
; :nairl advantages are:
" -..raller capacity is possible for a given power rating.

-" ,.e ability to melt charges without having to use starling plug.

,ji_eh frequency fuilaces are powered by either motor generator sets or by

:-siate convefters. High frequency allows an even greater power density

- , 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

Trans lormer Water-protective

(if required) devices

\['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----_ _

Fig. 2.4 Basic mains-frequency system

Most coreless furnaces are designed for single phase working, then it is

sffessary to provide each furnace power supply with adequate balancing

fu;itrities, and since the load impedance varies during the heating cycle, so it is
B;essary to provide variable balancing components. Fig. 2.5 illustrates phase-
hursncing system, the two transformers serve different purposes, one being

umei for phase-balancing and the other for conventional voltage transformation

mJ power control. Both transformers are provided with off-load tap-ch*g*g

h;;ltities for phase angle adjustment and power control, the varying coil power

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

Fig. 2.5 Schematic diagram of phase-balanced power circuit

- -,: \Iotor-generator sets

The generation of frequencies, which are multiples of mains frequency by

.
- -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

- .rg cannot be successfully accommodated.

-;re demand for higher frequencies has led to use the inductor-type
,r3. The advantage of the inductor generator is that both armature and

- .,,urdings can stationary and the rotor consists merely of laminations

. - J with open slots, no rotor windings being Llsed or slip rings
. ,. rr1'. [11]

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

. rnotor alternator is 12 kHz. [9]

Eecause the inherent reactance of these machines is high, it is normal to
,.Je a fixed series capacitor. Over the heating cycle, the load (induction

.je automatic control of power factor correction capacitors, which

: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

.', ed by changing the excitation DC voltage of the alternator which vary

- ,.tput voltage. [1]

- rg. 2.6 shows a schematic diagram of motor-generator set typical of most

- ,: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

Fig.2.6 Schematic diagram of motor-generator set

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

- : heary currents and voltages, and fast switching capability. Such a

-: rs the silicon controlled rectifier (SCR)' [1]

-.rere are a large number of rnanufacturers supplying (SCR) frequency

:ricrS, for using in rnetal melting industries. Although each manufacturer's

_- . differs in some way from competitors, all designs stem from four
-'.,;tdamental types i.e. voltage-fed, current-fed, cycloconverter and

, -: rnark space ratio inverter. [12]

: - i,1 \/oltage-fed inverter

---i shown in Fig. 2.7, the three phase mains input converted to a fixed
'.
,:ltage by means of three phase rectifier circuit, then the DC. Voltage
d using a thyristor bridge network with an external oscillator in the firing

to give the desired frequency.

, .: load circuit is a resonant circuit consisting of an induction coil and
: rs. Power control is obtained by varying the output frequency of the

: u.ith respect to the resonant frequency of load circuit as shown in

! changes in resonant frecluency of load circuit during the heating cycle

.,ratically compensated for using an electronic feedback circuit.[13]

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

Fig. 2.7 Schematic diagram Fig 2.8 Load characteristic

of voltage - fed inverter of voltage-fed inverter

This type of inverter can run at an output frequency substantially different

m the resonant frequency of the load circuit. This fact has been used by
ring the inverter to run at a sub-multiple of the load circuit resonant
the frequency limitation that placed on inverter to be
{uency, letting
ueased. [14]

Nthough, voltage-fed method of power confrol has many advantages the

pfoement of switch the thyristor at a frequency different from the resonant

qru€ncy of the load circuit necessitates a large nurnber of auxiliary electronic

trFonents which increase the cost and complexity of the invgrter, and reduce

cfficiency relative to the other fonns of inverters. [12]

-' -1,1 Current-fed inverter

e rnverter, as shown in Fig. 2.9, employs a controlled three phase
: . wherein three phase line frequency is rectified to direct curent,
-J by a filter circuit , and passed on to a parallel bridge inverter. The

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

L-----l L---r L L-----l L I

--J
rectifier filter :|ffi',P 'trrerter [ood

Fig. 2.9 Schematic diagram of current-fed inverter

: ;1 instantaneous positive value in the direction of previotts current flow.

ilunent-fed inverter can not be started by simply commencing to fire
- j:ors at some chosen frequency, because with the parallel'compensated

, - cLrcuit initially uncharged, the voltage across the capacitor is insufficient

---Lreve commutation at starting. A number of startingarrangementsare

, :le , one of starting arrangements appeared in Fi5.2.9.

lLurent-fed inverter always operates at frequency slightly above the

.,;rt frecluency of the load circuit and changes automatically with any

-: in the load circuit resonant. The power control is obtained by varying

, I I voltage fed to the inverter as shown in Fig. 2.10, and the load circuit
commutate the thyristors of the inverter without the need for additional

er components. Thus the current-fed inverter has a high conversion

iency. [2]

t7
 ing
freqmcy

I
l' porer cmtrot
,o I
by dc lottqe
;o vdriotio qt
CL fixed freqrcncY

f requency

Fig. 2.10 Load characteristic of current-fed inverter

!.7-3.3 Variable mark space ratio inverter

This inverter converts the three phase mains input to a fixed DC. Voltage
aJ inverts this to the desired operating frequency. Fig. 2.11 shows
rchematically the operating circuif,y. The essential commutation of the inverter
fu controlled by the load circuit, so that automatic compensation for changes in

&e load circuit resonant frequency is inherent in its operation.U2)

-q4gii,5
--
rL.qc rof IDc

cLDC

Fig 2.1L Schematic diagram of variable mark space ratio inverter.

-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

Fig.2.l2 Schematic diagram of cycloconverter

19
Methods of Analysis of Coreless
Induction Furnaces

To solve the electromagnetic problem of induction heating and melting,

there are two types of solution: numerical solution and analytical solution

3.L The numerical methods

many numerical approaches have been applied to induction and melting
problem; these approaches are:

r Finite Element method

In this method the field region is subdivided into a finite number of discrete

sized subregions or finite elements, the unknown quantities at each element

are represented by suitable interpolation functions that contain node values
of each element. [16]

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

Taylor series approximations. The equations are solved at every point of a

grid constructed over the required space domain' [17]

o Mutually CouPled Circuit method

The region to be investigated is subdivided into subconductors and using

known expressions, the resistance and self inductance of each subconductor,

together with the mutual inductances between each subconductor are
determined. The expressions arc usually presented in the forrn of an
impedance matrix and the application of Kirchhoffs laws lead to a system of
linear equations of the form
(3 1)
lzl[r]:[v]
where lZ) is the square matrix comprising all of coil and charge self and
mufual impedances, [I] is the column matrix of unknown coil and charge
segment curent, and [V] is the column matrix of driving voltages. [18]

t Boundary Element method

This method makes use of the fact that electromagnetic field attenuates
rapidly as it penefrates into the side of conducting material. Volume integral
equations for homogeneous media may be simplified to boundary integral
equations by application of the divergence theorem, and the potential at any
point is then defined in terms of the potential and its normal derivative on

the bounding surface. [19]

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

which the finite element method is applied becomes smaller. [20]

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.

Z-they require specialist knowledge and powerful computational facilities.

3.2 The analytical methods

There are many analyticalmethods, that can be used to design induction
furnaces, these methods are characterized by being simple, and suitable for

practical applications.

I Equivalent Circuit method

The basis of this method is the reduction of the induction coil and the charge
to their equivalent resistances and reactances.l22l

0 Approximate Coil-Design method

This method is a fast and reasonably accurate means of determining the

major coil variables, the basis of the method is the use of design curves,
these curves are primarily theoretical but are colrected by correction factors

based on practical tests. [1,2]

))
 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.

3.3 The equivalent circuit method

The basic assembly of a coreless induction furnace consists of a water
cooled copper coil surround the crucible which contain the charge to be
melted. This assembly is represented by an equivalent circuit.

3.3.1 Derivation of the equivalent circuit

The magnetic flux produced by the induction coil, Qr,hasalternative
parallel paths through either the charge d *, the coil 0
"
or the crucible Q * as
shown in Fig. 3.1.

23
 L

coil

n
0u
tl
d
I

u
charge
l
n
I
n

Fig 3.1 The flux paths in the coreless furnace

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

Fig.3.2 The equivalent magnetic circuit

Generally, the reluctance R* of a fltx path of leng[h L and cross sectional

area A is given by:

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:

Q*: tto4,M*(p- jq) (wb) ..'...(3.6)

And, the charge cross sectional area is:

A*:; D,* (3.7)

substitute equations (3.6) and (3.7) into equation (3.5) to yield.

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]

*=*#(1-i) (wb/m) (3.1 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

The coil reluctance is:

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:

L,=Lc-L**(0.45+k*)D, (*)...... (3'17)

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:

L, D. (0.45 + k, ) (3. 18)

substitute equation (3.18) into equation (3.16) to obtain:

4(0.4s + k*) (3.re)

^'mr -
HofrD
"
now, all the parameter of the magnetic equivalent circuit has been known.
It is easier to deal with the electrical circuit than the magnetic one, the

electrical equivalent circuit can be derived if the relation between the electric
and magnetic parameters are known.

The coil voltage, Vc, is related to total flux by Faraday's law:

! = jarN" /" (v) ..... .., (3.20)

also, the coil voltage is:

Y":ZI" (3.21)

coil current is related to the total flux by Ampere's law:

N" I":/ tRrot
(3.22)

From equation (3.20), (3.21) and (3.22), the relation between the
magnetic circuit reluctance and electric circuit impedance is obtained

Z_N? JO (o)...... (3.23)

" R*,

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

Fig.3.3 The equivalent electrical circuit

The components of the electrical circuit are derived by substituting

equations (3.4), (3.10), (3.15) and (3.19)respectivelyintoequations (3.24)-
(3.27) so as to result :

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 )

where KR = @P"7N? ....... ....... (3.32)

Now, the elecfical equivalent circuit has been constructed, then the

furnace performance quantities, such as the efficiency, power factor, current,

etc., can be grven by simple circuit analysis. As the coil number of turnilis not
known, the circuit components are calculated for a single turn coil. In these
calculations the components of the single turn coil ile represented by adding

the suffix (s) to the symbol.

The single turn circuit impedance Z,,is'.
jx*[jXu. * z*,
Zr= Z"r+ ..... (3.33)
jXrr+jXrr+Zws
Z, is of complex value :

Zu: R, + jX, (3.34)

The coil efficiency is :

R*. (3.35)
4"
R,

and, the power factor is

rr=ft ... (3"36)

The input power can be obtained from the power needed for the charge, P*

29
Pn _P* (w) .. (3.37)
rl,

and, the volt-ampere, VA, of the coil is :

VA: 'q' (vA).... ... (3.38)

P.F.
The voltage of the single turn coil, V",, .ao be calculated as follows :

(3.3e)

-rttiptyi"g by the "ottrge an? ai"iai"g bt N3-

Z ... (3.40)
#=n'. N 2
c

the square root;

.'. Vcs p (vA)Z, ..... (3.41)

The required ampere-turns is

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).

Then, the coil current is :

IN .(3.44)
Ic (A)
Nc
the coil copper loss is:

P" = I3R" (w) .. .... (3.45)

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.

3.4.1.Derivation of the superposition method

The magnetic field strength at any point P along the surface of a semi-
infinite metal rl-rbi produced by a filament carrying current I ata distance

h from the surface of slab can be calculated by assuming an imaginary

h from the surface and carrying culrent
filament within the slab at a distance

-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

Fig. 3.5 The magnetic field intensify on point P.

The axial components of the conductor's magnetic freld srength{and

its image will aid each other, while the radial component H. will cancel each

other. The resultant field on point P, as shown in Fig.3.5, will be:

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

filamentary in nature, so the filament is replaced by a circular conductor of

radius r and the current skin depth into the slab and the conductor is small

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

he Jh' -A ....... (3.s3)

Consider the case illustrated in Fig. 3.6., which includes N identical

conductors parallel to each other and parallel to a metal slab at a distance h

from its surface. Each conductor carries a current I.

':t ? il
0 $ o a
I I
Lzt I
I
I
I
_el
l*z'
I
I
I
I
1s I
I
I I
I I
h Zx
I
I
I

Fig. 3.6 Number of identical conductors at a

specific air gap from a metallic Ioad

According to the superposition principle, the current per unit length U,

induced by these conductors at a point P on the slab, is equal to the sum of
the current per unit length induced by each individual conductor. i.e.

Up = Up, * Upz *" " "*Uplr

whereUnr,Upz ,......,Ur* are the current per unit length induced at point P

from conductors 1,2,.. . .,N respectively.

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

In induction heating , it is more convenient to deal with surface power

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

Substitute equation (3.58) and (3.60) into equation (3.61) to get

34
PD=*[F:u*=r]' .....(3.62)

The above derivation may be applied to a coil srurounding a cylindrical

charge, provided that the distance befween the conductors and the charge
is:[23]
l-Small in comparison with the charge radius. And,
Z-Larye in comparison with the current skin depth in the charge.

3.5 Conclusion

In this research, the equivalent circuit method will be used to predict the

furnace performance. Another method, which is the superposition method, will

be used as ao, aiding method to the first one .

Different forms were available previously to represent the induction

heating systems by an equivalent circuit, but these circuits were applied
according to many conditions and limits l27,zlluntilg-akel ,rl22lputhis method
to represent the induction heating system by comprehensive equivalent circuit,

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

equivalent circuit method to suit more applications.

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

the length of the charge.

as the
2-The assumption of uniform material properties which is not correct
resistivity of the charge changes with the temperature, also the permeability
varies with both the magnetic field strength and temperature. According to
Baker l2ylit is impractical to take into account the varying resistivity and
arithmetic mean value colresponding to the temperature range covered,
could be used successfully. Davies and Simpson [l] used the concept of
..integrated resistivity " of resistivities at starting and finishing of the heating

cycle. And, according to Vaughan and Williamsonl}Tl separate calculations

should be made for maximum and minimum temperature conditions since the

largest value of amper-turns needed by the charge during the heating cycle

exists at one of these conditions.

In accordance with the magnetic properties, the charge could be:

l-Nonmagnetic charge: in this material, neither the magnetic field strength nor
the temperahye affect the relative permeability p,, and for all nonmagnetic
charges the relative permeability is equal to unity'

2-Magnetic charge: below the Curie temperature, the permeability of the

magnetic material varies according to the change of the magnetic field, many
expressions attempt to represent the variation of the relative permeability of

36
 the magnetic material [29], among these expressions is that given by
Baker:1221

1.43 x 106 Bs
Fr- (3.63)
HM

where Bris the saturation flux density and

H, is the peak of magnetic field
sfrength. on the other side, the effect of temperature change
on the
permeability is extremely low and hence it could be neglected.
[30] This is
right until the charge reach the Curie temperature at which
the material loses its
magnetic properties and the relative permeability becomes
equal to I.
Consequently, it
appears that applying the equivalent circuit method on
the nonmagnetic charges directly achieved by using (p, _
l) . while, for
magnetic charges, separated calculations, for above
and below curie
temperature conditions, should be made.

37
 The Design Procedure

of Coreless Induction Furnaces

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.

The equivalent circuit method is regarded as a general method because this

method doesn't determine the dimensions, frequency or the power

requirements of the specific application analyzedand designed by this method.
On the contrary, the use of the equivalent circuit method requires this
informations as a pre-requisite, so to prepare a complete design of the coreless
induction furnace, these informations are deterrnined in the following sections.

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]

Crucible capacity : output (kr/ht) * t (kg) .(4.1)

where t is the melting time in hours

Cruciblevolume : crucible capacity I o (rn') .(4.2)

where o is the charge material density in kg/m3.

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

molten metal of the charge required. [9]

4.2.1Determination of height to diameter ratio

Optimizing the crucible dimensions in relation to a particular size can give
enerry savings up to l0% l4l. Optimum dimensions of crucible depends, first,
on determining a suitable value of height to diameter ratio. Many technical and
economical factors influence this ratio. Increasing height to diameter ratio
prevents uniform electromagnetic stirring and mixing by convection, also the
tilting is more difficult when the height of the furnace is excessive. On the

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

4.2.2 Determination of crucibre wail thickness

The next step to design fte crucible is the determination of the
thickness of the thermal insulator between the coil and the molten metal i.e.
coil diameter to charge diameter (D.lD*) ratio. As indicated
in section 2.2,the
choice of the crucible thickness must satisfi the
following requirements:[l]
l-Thickei lioiog increase leakage flux, reducing the heating
of the charge and
resulting in a poor power factor.
2-Thinner linings increase the furnace conduction
heat loss through the tining
and hence reduce the thermal efficiency.

Ar- mentioned above it would be possible to find an

optimum value for crucible thickness resulting
in highest overall efficiency.[9]

4.3 Frequency selection

Like
linduction heating applications, there is a lower limit for
frequency
which may be used for coreless induction fumace,
this limit is associated with
loss of coupling i.e. below a criticar frequency,
the coil efficiency begin to
decrease at an increasing rate and the transferred
power from coil to the charge
can not increase the temperature to the
desired revel.
Brown, Hoyler and Bierwirth ae.iveaian
expression for the critical
frequency, in the form:
[1] l
, _6.45 p
'"- pt (ru) (4.3)

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.

Initially, the operating frequency must be of a value that gives a

fiansformation efficiency high enough for effective initial heating, that is, to
start the heating cycle. After the charge has reached the liquid stage, the
stirring of the metal is inversely proportional to the frequency. The stirring of
the melted metal must be in a moderate degree and the frequency must be
suited to that requirement.l32l

Consequently,the decision about frequency is, therefore, the usual

engineering balance of these two requirements and also the economic
considerations. [1,32]
Rowan developed a frequency selection chart for coreless induction
furnaces based on practical experience. According to this ehart, the frequency
choice has been related to furnace capacity and the chart is divided into four
zones, as shown in Fig. 4.1.[U

10 000 l0 kHz
.1.1 kHz
-1 kliz
N
I kHz

o
<' kHz
o kHz
kHz
kHz

10
1

Furnace capacity (kg)

Fig. 4.1 Choice of frequency with furnace capacity

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]

4.4 Determination of the required power

The total power required by the charge P*, includes not only the useful
power which is required to melt the charge, but also the accompaniment losses.

4.4.1Determination of the useful power

The heat enerry E, required to arise the temperature of the charge from
the ambient to the pouring temperature, equals to

E:CxM (*hr)...... ....."(4.4)

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

power P, observed by the charge, then, is:

42
 E (4.5)
Pu (w)
t

To determine the melting time, some limits have to be considered. A

lower limit is given by practical considerations, short melting times imply
frequent and very rapid charging; also if melting is too rapid, it would be
difficult to make the several necessary changes of voltage, frequency and

capacitance (when no automation is available); finalty, the power source

frequently stands idle while the furnace is charged. Since short melting times
require a relatively high input, and the charging time does not increase linearly

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

and large units no melting time shorter than20 minutes is practical.

The upper limit is determined by the danger of excessive heat losses.

Melting time longer than two hours are ordinarily not recommended. [9]

4.4.2 Determination of heat Iosses

The heat losses from the furnace can be divided into ttre following items;
l-conduction loss through the lining;}-radration loss from the open top and; 3-
convection loss.
Conduction lps$ is, normally, the largest heat loss and can be obtained
by:[8]

2r L c k(r, - T,) , aD*k(T, - T,) (w). (4.6)

P"ona -t,
ln(D Dw )
"l

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

term represents the conduction loss from the crucible base.

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]

?*=#4tr, +273)a - G +zn)al (w / m2) ............. .(4.7)

where e:emissivity coefficient of surface. The power loss due to radiation is

then given as [1]

D
r
nDl ?,ud (w) (4.8)
rad
4

Convectio! loss, generally is small in induction furnace, as the charge is

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]

?"on, = 1.54(T2 -T,)oo (w/m2)... ...... (4.9)

then, the power loss due to radiation is given by:

nD'* ?",nu (*) .......

P^^-..
'conv -- 4
...(4.10)

The total power required by the charge is given by:

P* = & + P*na * P*o * P*ou (*) ....... ...... (4.11)

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

efficiency I r,,where. [9]

...(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

minimum value of 2.25 6c in order to minimize the copper losses, but

conversely, a great wall thickness increases the thermal resistance between
the hottest part of the coil and the cooling water. While this may not be
dangerous in normal operation, it is important in case of current surges. Thus,

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]

4.6 Determination of the required cooling

The exciting coil consists of a number of turns of water cooled conductor,

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

may occur due to insufficient cooling, so it is important to circulate a sufficient

amount of water through the coil to insure safeguard to the insulators and to
prolong the life of the lining. [6]

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)

where P, is the power to be removed by cooling, which is equal to:

Pr = P" * P*,a (w) ........ ...(4.14)

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

use of the concept of equivalent diameter, where

a' g.Iott- t99@1 qtu (-) ... ..

Equivalent diameter,6 = - weffed perimeter
. ..(4.16)

High water velocity is desirable with minimum outlet water temperature,

but this may produce excessive pressure across the water circuit. A simplified
expression for the pressure drop is given by: [1]

Ap: s.s22*# +746.25 o2 (pascal) (4.17)

where W is the length of the water circuit, and it is given by:

W:rN.@,+d) (m)....... ...(4.18)

The pressure drop, across the water circuit, determine the kind of pump
needed to circulate the water. When the pressure drop is high, the water circuit
is normally split into a number of parallel paths, this leads to a lower pressure,
in this case the flow rate per path is given by: [l]
'<P no.ofpaths
"""""
and, the value of w obtained from equation (a.18) is divided by the
number of paths before substituting in equation (4.17).

47
 Design ImPlementation

and ExPerimentalWork
t

in the last two chapters'

In accordance with the theoretical basis discussed
progfam 'Required Power" calculates
three progfams were developed, the first
melt and also calculates the thermal
the required power by the charge to be
stated in sec' 4'4'The second
efficiency of the furnace, based on the equations
equivalent circuit method' as
program ,,Equivalent circuit" is based on the
designing technique which was
previously mentioned, this method is the main
is supplemented with a subprogram
used in this thesis. The second progra,
constants p
developed compute the Bessel functions of the dimensionless
to
functions were obtained from
and q. The polynomial expressions of the Bessel
based on the superposition
reference [36 ].The third program "superposition",
of amper-turns required to melt
method was developed to calculate the number
the charge

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

results of the equivalent circuit method.

The flow charts of "Equivalent Circuit" and "Superposition" programs are

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

calculate 0c, Ew, p, q

tum
es

rl c,
YA,IN, VCS

does
no yEs
there edst a
certain preferred
voltage

Fig 5.L Flow chart of the program'Equivalent Circuit'

49
 start

read input data

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

Fig 5.2 FIow chart of the program 'superposition.

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

5.L.1 Preliminary preparation

A variety of the metal were studied to chose the charge,
the limited
facilities available strongry affected the options,
as improper workshop and
lack of step-down fransformer to supply the required
high cunent. So the best
furnace capacity obtained was 400 grurm, and the
most proper charge metar
was tin.

5.1.2 Optimizing the crucible dimensions

* As previously mentioned in sec. 4.2,

determining the crucible dimensions
depends upon the L*lD* and D./D*ratios. To get the optimum
design ofthe
crucible' the effects of these two ratios
on the furnace performance were
discussed.

For different values of ratio L*D', the curves

of the furnace power factor
as a fi.rnction offfi,ratio is shown in Fig. 5.3

--

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

Fig. 5.3 Variation of P.F. with D.lD* ratio

The furnace total efficiency qp (where rlr.:ec x Ir,) as a fi.rnction of the

D./D* rafio is shown in Fig. 5.4

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

Fig. 5.4 Variation of the furnace efficiency with D./D* ratio.

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

Fig. 5.5 Variation of the furnace P.F with L,"D,o ratio

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

Fig. 5.6 Variation of furnace efficiency

(at pouring temperature) with L,,D,, ratio.

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

Charge diameter (at the top) 37.3r

Charge diameter (at the bottom) 3s.14
Charge height 53.25
Coil inner diameter st.22
Coil height s3.2s
Base thickness ./.J
Table 5.2 The dimensions of the d esigned crucible (in mm)
t 5.1.3 Design of the p ower supply
The critical frequency, for the furnace to be construct in this work, is

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

is able to generate frequencies above the critical frequency, the solution is to

construct a static-inverter as a power supply. In sec. 2.7 .3, details were glven

of the industrially used static inverters for industrial process meeting,

constructing any type of these static inverters is very expensive, so the

alternative is to design a simplified inverter circuit of an acceptable efficiency

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.

The load circuit is a resonant circuit consisting of the induction

coil and the
load matching capacitors. The power control is obtained by varying
the three

phase voltage by using a three phase autotransformer.

[35,37,3g,3g].

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:-;;;, !

Fig. 5.7 The circuit diagram of the power supply.

5.2 Steps of the furnace construction

5.2.1Construction of the cruciblB
The material used to construct the crucible was the fireclay. This material
have very acceptable thernal, mechanical, metallurgical qualities for this task,

also there is available experience in handling this material.

50
 fl

An iron mold was designed and constructed, consisting of three parts; a

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

Fig. 5.8 The mold.

57
 {
,r ll

male

ttl

l{;
5$

,
female
?s

l{

pusher

:I

Fig. 5.9 Cross-section view of the mold.

58
 IU
$s.2
.(
l{f S

i] i '-S,

Fig. 5.10 Cross-section view of the crucible.

5.2.2 Construction of the induction coil

I
Neither the equivalent circuit method nor the superposition method
specifii the number of tums of the induction coil, these methods can only obtain

amper-tums required by the charge to be melt. The numbers of amper-turns for

various frequencies obtained by the equivalent circuit and the superposition
methods me shown in Figs. 5.11 and 5.12

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

Fig. 5.11 The amper-turns, yersus frequency, obtained

by the equivalent circuit method

59
 IN
T
2000

1500

1000
500

0
f
123 4 5 6 7 8 910 '11 ,i,z t, 14 15 krlz

Fig. 5.12 The amper-turns versus frequency obtained

by the superposition method

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.

Determination of the number of turns of the coil is exclusively influenced

by practical considerations. Since the designed power supply of this work

doesn't have the ability to produce high currents, so it was compensated by
using a coil of large number of turns of small diameter tube.
Another parameter consider here, was the water-cooling performance of
the tube conductor, this was done by theoretical calculation andpractical test.

The wounded induction coil consists of 16 turns of a copper tube of 3mm

diameter. The tube insulated by means of sleeve insulator. Cross-section view

of the insulated tube is shown in Fig. 5.13. To fixthe turns of the coil, an ,.
adhesive tape was used, in two layers internal and external.

60
 3.S

3"4-*-_+

Fig. 5.13 Cross-section view of the tube.

The coil water-cooling system consists of a water pump and proper

i
tubing. Fig. 5.14 shows the furnace and it's associate water tubes

t \#
t !t
I
tt ^l r "*.-. il r-l

'\a q
4

"1* -r,i} L.
I
rd

Fig. 5.14 The furnace.

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

to make the losses as low as possible.

The main problem that appears in conskucting and operating the power
supply, was the load matching capacitors, because the capacitor needed in this
work must have the ability to operate within high frequencies (i.e. the

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

obtain the proper capacitor for this work.

For maximum dissipation of heat, the MOSFET was fixed on a heat sink,
also a small fanwas used to cool the elements of the circuit

t
5.3 Operation of the induction furnace
5.3.1 The practical measurement
For observation of the furnace perflormance during it's operation,

medsuring instruments connected, seeFig. 5.1 ,taklnginto considerationthat

the instruments connected to the high frequency stage of the power supply
have operating frequency range higher than the frequencies used in the furnace
operation.

Since the furnace operation requires high currents, so a current

transformer should be used, the available transfonner is specified for 50Hz
I
only, so there was some error in the readings of the instruments connected to

the secondary of the transformer. A correction factor (kct), was then estimated
-

experimentally, to compensate the ratio effor. The circuit used to estimate the

correction factor is shown in Fig. 5.15, the correction factor relation to

different frequencies is illusfated in Fig. 5.16.

62
 signal RL

Fig. 5.15 The circuit used to estimate kct

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

Fig. 5.16 kct as a function of the frequency

The temperature measurement of the charge, during the heating cycles

was done by a digital thermometer (tWe cormak 5000) and Ni-Al
a

thermocouple. During each test of melting the charge, measurements of the

insffuments were taken every 5 minutes throughout the heating cycle. Minor

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

melting the charge was achieved.

i
Temperature
c
250

no
150

100

50
time
0
min
0 5 10 15 20 25 30 35 & 45 50 55 60

Fig 5.17 Temperature rise in the charge for

(0.114 kw, 11.5kHz operation) a

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

Fig. 5.18 Temperature rise in the charge for

(0.125 lnil, 12 kHz oPeration)

During the heating cycle, the used instruments could not sense the

variances in physical properties of the charge, because these instruments don't

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

Figs. 5.19 and 5.20 respectivelY .

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

Fig 5.19 Impedance as a function of the charge temperature for

(0.092 lrw, 10 kHr)

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

Fig. 5.20 P.F as a function of the charge temperature for

(0.092 kw, l0 kIIz operation)

5.3.3 The Iimitation of the power supply

Although, the power supply circuit used in this work to convert the
t
frequency was a simple, still it's operation is good and of an acceptable
efficiency. Fig. 5.21 shows the relation betweenthe efficiency ofthe power
supply with the output frequency of the power supply.

T1

0.5

o.4

0.3

o.2

0.1

0 f
10 10.5 11 11.5 12 12.5 13
kIIz

Fig. 5.21 The efficiency of the power supply.

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

up to 13 kllzand not above.

5.4 Results and comparisons

5.4.1The impedance of the induction furnace

Practically, the impedance was measured by two methods, the first one
was by the instruments connected to the coil terminals, and the second method

was by using LCZ meter (tlpe IP 42764).

I The furnace impedance, is shown in Fig. 5.22.

lzl *LCZ meter

o -f instruments
o.7 --tstheoretical

0.65

0.6

0.55

0.5

0.45

0.4
f
10 10.5 '11 11.5 12 12.5 i3 kJIz

Big. 5.22 Furnace impedance

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.

input power frequency Furnace efficiency(%) output power melting time

(K!v) (kIIr) practical theoretical (Kw) (min)

t
0.12s t2 16.93 10.14 0.021 30

0.155 t2"5 21.0t 18.02 0.032 t5

f, 0.183 13 33.07 24.37 0.06 10

Table 5.3 The furnace efficiency

The required mmber of amper-turns to melt the charge practically

measured and theoretically calculated , is illustrated in table 5.4

mput power frequency amper-turns (A.T)

(Kw) (kHr) practical theoretical

0.125 t2 761.22 966.34

a
0.155 t2.s 899.2 10t4.7
0.1 83 13 t00t.t2 1059.9

Table 5.4 The required-turns

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.

, For practical reasons, the load voltage waveform could not be

photographed, nevertheless, the load voltage could simply be concluded from
Fig. 5.23, where the load voltage would have a square wave with it's upper
value equals the DC voltage value, and it's lower value equals zero, and with
frequency equals the frequency of the firing signal.

a-

s :,-- ,-

DC voltage from rectifier

(20 psec, 10 Volt / Div)

b-

r
Voltage acrossde

(20 psec, 10 Volt / Dlv)

Fig. 5.23 Voltage waveforms

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

between the theoretical calculations and the practical requirements of the

crucible, so on theoretical basis, the thickness of the crucible wall must be
as' thin as possible (see table 5.1), but this is not possible practically, since
there is a minimum limit of thickness for the crucible wall, below which the

crucible will fail.

2- In the early stages of constructing the crucible, the mold was composed of a

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

pressure of the press to all over the shapen crucible equally.

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.

4- Observing Fig . 5.22 , it was evident that there some differences

between the theoretical calculations and the measurements of the
measuring insffuments of the furnace
impedance. If the lssdings of the QCZ meter) were regarded as a
l-
I
reference, it can be conoluded that :

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

readingsof the (LCZ meter) was up to 25o/o for the impedance

This differences were due to the percentage effor of
the measuring instruments, also due to the neglect of the effect of the
phase-angle error of the current transformer (because this ffansformer was

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

\ pouring). In the theoretical calculation, itwas assumed that the maximum

heating losses occur throughout the heating cycle, so it was observed from
tables 5.3 and 5.4, that the theoretical efficiency of the furnace was less
than the actual efficiency and the theoretical amper-turns was higher than
the actual value.

6.2 Suggestions for future work

Induction coreless furnaces play an essential role in a wide variety of
industries that require metal metting. This work is a practical step in a long
way to build such an industry in our country and enter this new field to make a
t
progress in our heavy indusffies. Suggestions for developing this work may
include.

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

harmful effects induced by connecting induction furnaces to the electricity

supply.
4-Insimilar applications, and to improve the accuracy of the ptactical

measurements, an alternative must be usedforthe current transfoflner as a high

current, low value resistor

72
 l- Davies, iJ. and Simpson, P., "Induetion Heating Handboo,k", McGraw-Hill

2- Simpsofl, P., "Induction Heatingi', McGraw-Hill, New York, 1960.

3- Hastings, B. and Quickenden, C., "Electric Induction Furnace
Developments snd Trend,x", Metallurgia, Vol. 50, 1983, pp. 146-150.
4- Edgerley, C, "Improvements in the Melting Efficiency of Coreless
t Furnac es", mEE conference publicati on l49, 197 7, pp. 2 8-33
5- Lozinskii, M., *fndustrial Appltcations of Induction Heoting", Pergamon
press, London, 1969.
\
6- Harrison, W., "Engineering and Installation of Coreless Induction
Furnuce Plunf', Electrowiirme Int., Vol. 37, 1979,pp.209-214.
7- Plockinger, E. and Etterich, O. ,"Electric Furntace Steel Production",lohn
Wiley & Sons Ltd.,1985..
8- Campbell, I. and Sherwood, E., *High*temperature Materiak and
Technologlt", John W,illy & Sons Inc.,1967 .

9- Paschkis, V. and Presson, J., *Industrial Electrtc Furnaces ond

Appliance,s", Interscience publishers Inc., 2nd Edition, New York, 1960.
l0- Tremayne, 1., "fmpedance and Phuse Baloncing of Mains-Frequency
Inductton Furnace,i", IEE proc., Vol. 130, 1983, pp. 16l-170.
ll- Draper, A., *Electrtcal Machine,s", Longman Grop limited, 2"d Edition,
London, 1979.

73
 -\
li- Hobson, L., *Static Inverters for Industrial Process Heatingi', IJEEE,
'
Vol. 20, 1983, pp.23-31 .

13- Pelly, 8., "Lastesl Developments in Static High Frequency Power

, ,saarces for Induction Heating!" IEEE ffans., IEC l-17, 1970,
pp. 297 -312.

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.

2O-Salon, S. and D'Angelo, J., "Appltcattons of the Hybrid Finite Elemtent'

I
Boundary Element Method In Electromagnetics", IEEE frans., Mag-24,
1988, pp.80-85.

21-A1-Shaikhli, A. and Hobson, L, "Method of Analysis for Induction

Heating Work Coils",20th UPEC, Hadderfield polytechnic, 1985.
22-Baker, R., "Design snd Calculatton of Inductton Heating Coils", AIEE
trans., Vol. 76, 1957 , pp.31-40.

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

Conducting Crucibles", Electrotechnik, Y ol. 49, 1965, pp. 37 6-397 .

25-Hobson, L. and Al-Shaikhli, A.,"Illustrating Electromugnetics Ustng an

Industrtal Process", IIEEE, Y oI. 23, 1 986.

26-Ha5rt, W ., " E ngin eering E lectr omugnetics", McGraw-Hill book company,

4ft Edition, 1981.

27-Yau$tffi, J. and Williamson, J., "Design of Inductton Heuttng Coilfor
Cylindricol Nonmagnetic Loods", AIEE trans, Vol. 64, 1945,
pp.587-592.
28-Baker, R., "Heating of Nonmagnetic Electric Conductors by Mugnetic
Induction Logitudinul Fhtx",AIEE trans. Vol. 63, Ig44,pp.273-278.
t
29-A1-Shaikhli, A. and Makki, M.,"PFacttcal Evaluation of the Equtvalent
Circuit Metltodfor the Inductton Heattng of Magnetic Lout',2nd kaqi
technological conference, 15-16 Muy, 1996.
3O-Thornton, C., "Resistance Heating of Mild Steel Contatner at Power

Frequencies",IEE proc., Vol. 99, 1952,pp. 85-93.

31-May, E., *fndustriul-High Frequency Electric PoweF", Chapman and
Hall, London, L949.
32-Stansel, N., "fnduction Heottng-Selection of Frequency", AIEE trans.,
Vol. 63, 1944, pp. 7 55-7 59.
33-Razzak, A., "Analysis und Study of Design Criteriu of Inductton Heating
Furnuces", M.Sc. thesis, University of Baghdad,1994.
34-Rowan, H., *A New Frequency Selection Chartfor Coreless Induction
Furnaces", Foundry management and technology , A penton publication,
1987.
35 -Rancocas, N., " Ittductoth erm Corp oration-M aintenance M unu af', I97 7 .

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.

4l-Holman , J ., "Heat Transfer", McGraw-Hill, 5ft Edition, 1981 .

76
-