Bulging
Bulging
Bulging
Steelmaking
Report
EUR 8963/111 EN
Steelmaking
IRSID
185, rue du Président Roosevelt
F-78105 SAINT-GERMAIN-EN-LAYE
Contract No 7210-CA/306
(1.5.1979-30.4.1982)
FINAL REPORT
Directorate-General
Science, Research and Development
L-2920 LUXEMBOURG
LEGAL NOTICE
Neither the Commission of the European Communities nor any person acting
on behalf of the Commission is responsible for the use which might be made of
the following information
ABSTRACT
CONTENTS
Page
INTRODUCTION 7
1. BIBLIOGRAPHIC STUDY 9
1 Internal cracks 9
1.1 Different types of internal cracks 10
1.2 Influence of operating parameters on the proportion
of internal cracks in the broad faces 11
1.3 Mechanism of formation of internal cracks 13
2 Centre segregation 15
Page
of Plasticine 59
Page
4.6.1 Relationship between centre segregation and
solidification structure 85
4.6.2 Relation between solidification structure and the
bulging deflection 85
4.6.3 Relation between the bulging deflection and the centre
segregation 86
4.7 Synthesis of the results obtained 87
REFERENCES
FIGURES
APPENDICES
7 -
INTRODUCTION
With the object of evaluating the relative effects of various casting para-
meters on the intensity of bulging during continuous casting of slabs, IRSID,
in collaboration with the University of Metz, has made a study of the math-
ematical and physical modelling of the mechanical behaviour of the shell
during bulging. Parallel to this study, IRSID has carried out several cam-
paigns of measurement of bulging on continuous casting machine no. 12 in
USINOR Steelworks I at Dunkerque (Dunkirk).
This paper presents the different results obtained by the two approaches,
theoretical and experimental.
In a second part, after giving an account of the work done at IRSID to deter-
mine the laws of steel behaviour at high temperature, we describe the math-
ematical models of bulging that we have developed as well as the principal
results derived from their application to the case of the casting of slabs
on USINOR machine no. 12 at Dunkerque.
The third part of this report describes the various experiments carried out
on the model that physically simulates bulging, as well as the principal
results that have been obtained.
After a description of the principle and the method of measuring bulging that
we used on USINOR machine no. 12 at Dunkerque, in the fourth part of this
report the experimental results obtained are presented and in particular, the
relation found between bulging deflection, axial segregation, and solidifi-
cation structure.
On the basis of all the results obtained, in the last part of this paper we
present the elements of our thinking that allow definition of the optimal
profile of the roll diameters for a continuous slab casting machine.
- 9
1. BIBLIOGRAPHIC STUDY
Between two adjacent rolls, the shell is the seat of tensile stresses in the
surface and of compressive stresses at the solidification front. The direc-
tion of the stresses is reversed to the right of the axis of the supporting
rolls. The solidification front is then under tension, the surface under
compression. In addition, the bulging of the slab induces tensile stresses
at the liquid/solid interface at the level of the narrow face of the slab
(Fig.6). The longitudinal or transverse tensile stresses are manifested by the
presence of internal cracks, when the cracking threshold at the solidification
(1)
front is exceeded.
The bulging, that in the course of the slab's passage through the continuous
casting machine creates alternating stresses at the solidification front, like-
wise induces movements in the still liquid metal. These movements, when they
develop near to the closing zone of the liquid core, are the origin of certain
defects of continuously cast slabs: linear axial segregation, for example.
1.1.1 ïnternal_çracks
By reason of the deformations of the shell, fissures may appear at the solidi-
(2 3)
fication front, which is very brittle. ' These fissures induce greater
segregation in their axes and form a defect commonly called internal cracking
or segregated thread. According to their importance, these defects are still
present even after subsequent transformation of the semi-finished product.
1.1.1.1 Different_types_of_internal_cracks
• Triple-point cracks.
These defects, the most numerous and most injurious to slab quality, will
be specially discussed below. To simplify the presentation, we call them
"internal cracks in the broad faces" or more simply, "internal cracks".
We present the published results without making any distinction according
to origin of the cracks, whether due to bulging or to other phenomena.
(4)
These results, reported by Fujii et al, show that internal cracks are
the more numerous, the higher the carbon and sulphur contents of the steel
+. (4,6)
cast.
For Al- or Si-killed steels with C ^ 0.10 - 0.2 % and Mn ^ 1.5 % cast on
a bow-type continuous casting machine, the equiaxial zone is more developed
in the outer radius zone than in the inner radius.
This same figure shows clearly that casting under compression on machine
no. 2 at Oita allows a decrease of the internal cracks index, defined by
the ratio of their length to their displacement, especially so far as the
cracks due to bulging are concerned.
This figure shows autoradiographs in which the internal cracks are clearly
revealed. The white part has not been marked by the radioactive element
198
(Au ) , which indicates that when it was added, this part was already solidi-
198
fiad. In contrast, the dark part which contains Au corresponds to a zone
in which there was still liquid steel. The cracks correspond to black threads
and their form is not rectilinear. An extremity of these cracks penetrates
into the solidified phase. This observation indicates that the internal cracks
form near to the solidification interface. These authors (4) have established
that the cracks form, not in a single stage, but progressively (Fig.13).
14 -
Measurements of the thickness of the shell during trials with the radio-isotope
and the distance between the extremity of the cracks and the slab surface show
clearly the progress of cracking at the solidification front (Fig.3).
In addition, during the stoppage the slab has been slowly moved about 10 cm.
The appearance of the cracks (in Fig.14) and their position relative to the
supporting rolls and to the solidus isotherm at the moment of arrest, confirm
that the internal cracks come into being at the solidification front and in
a zone located to the right of the rolls, where the tensile stresses at the
solidification front are at a maximum.
Analysis of the (freshly cast) ingot bending tests allows plotting, on a graph
of deformation as a function of rate of deformation at the solidification front,
of a limiting curve above which cracks form and below which there is no
(12-19) (17)
cracking. An example of a curve of this kind is shown in Fig.16.
-3 -1
For a rate of deformation of the order of 10 s the work reported by Sumitomo
(17)
Metals demonstrates the existence of a critical deformation of the order
of 0.9 % in the case of S 45 C steels.
15 -
f i o \
Trials made by Kobe Steel on the other hand show that when the sulphur
content of the steel rises from 0.014 % to 0.028 %, the critical factor is
-3 -1
halved (from 1.2 to 0.6 %) for a rate of deformation of 210 s . Fujii
(19)
et al point out that the critical deformations for inducing internal
cracking depend on the grade of steel and its value varies from 0.60 % for
medium carbon steel to 0.74 % for extra-soft steel.
Fig.17 shows the results of an analysis of works trials, run on the Mizushima
machine no. 5 and Chiba machine no. 2 of Kawasaki Steel, in the form of
a curve giving the critical deformation for the occurrence of internal cracks
as a function of the square root of the solidification time.
In Fig.18, Kawasaki Steel show that before the rebuilding of the Chiba no. 2
machine to provide it with intense secondary cooling, an increase of the
carbon content of the steel (from 0.14 % to 0.2 %) made it necessary to reduce
the casting speed from 1.6 m/min to 1.2 m/min in order to avoid the formation
of internal cracks.
Finally, it may be noted that, on the basis of findings from the continuous
(12)
casting of billets, Brimacombe estimates 0.2 % to be the critical defor-
mation at which internal cracks form.
rolling, these defects can have an effect prejudicial to the quality of the
final plate. Certain mechanical properties measured in the short transverse
(2i:
direction are clearly weakened in the presence of internal macrosegregatron.
To each side of this central peak there are two zones with slightly decreased
sulphur contents. The central concentration peak is usually accompanied by
a peak of carbon and manganese concentration which is however less pronounced.
Two authors have shown the influence of the bulging deflection on the intensity
(14 ?5)
of centre segregation '~ (Fig.20 and 21).
(14)
Trials run on Mizushima machine no. 5 (Kawasaki Steel ) have shown that
the index of centre segregation increases with the bulging deflection. This
result, shown in Fig.20, is especially clear in the case of a roll spacing
corresponding to withdrawal rolls 280 mm in diameter. The bulging deflection
is a calculated deflection for the corresponding casting conditions.
- 17 -
(25)
A result of the same type has been obtained by Asano et al by measuring
the thickness of the cold slab at the exit from the machine (Fig.21). These
authors considered that the difference between the desired and actual thick-
nesses of the slab represents the bulging of the slab at the end of solidi-
fication. In this case also, it is found that the intensity of centre segre-
gation increases with the bulging deflection.
The extra-soft steels display centre segregation lines that are more
pronounced than those in medium carbon steels (C > 0.12 % ) . '
In the case of not very intense secondary cooling, the centre segregation
increases with the casting speed.
(34)
1.1.2.2 Mathematical_modelling of_the_formation_of centre_segregation
(34) (32)
Miyazawa and Schwerdtfeger have developed a mathematical model
describing the movements of the liquid steel in the pasty zone at the centre
of the product, during the bulging of the broad faces of a slab and the forma-
tion of centre segregation. The bulging occurs between two perfectly aligned
consecutive rolls, as is shown in the diagram in Fig.28. The maximal deflection
of the external surface of the slab is designated by 6.
Bulging varies the thickness of the slab in the course of time. In Fig.29,
two domains can be distinguished:
Domain A in which the slab expands. A suction of liquid toward the centre
is then produced, to refill the space thus created.
(V'grad Τ)/ε
where
V = the liquid velocity
έ = M?/ít
These results are the only ones available in the literature and this explains
why they have served as a reference to support the majority of the published
models of bulging. It is to be regretted that this unique study has not been
confirmed by other researchers.
These measurements have been made on a continuous slab casting machine, with
a bow radius of 3.9 m, on which various grades of steel, ranging from St 37
to X 60 , have been cast in slabs 150 to 200 mm thick and 550 to 1800 mm wide,
at withdrawal speeds from 0.8 to 2.0 m/min.
The measuring device (Fig.22) has been designed to detect the displacements
at three measuring points relative to the supporting rolls. The chief para-
meters that control bulging have been studied.
The spacing (pitch) of the supporting rolls has a great influence on the
deformation of the shell. For a low ferrostatic pressure (2.8 bars) with
the rolls 860 mm apart, the bulging of the shell is of the order of 6 mm,
increasing to 37 mm when the roll spacing is increased to 1290 mm (Fig.23).
The measured bulgings are less than 3 mm when the roll pitch is less than
550 mm (a frequent situation on industrial machines).
. . . , . (5,35)
- elastic models
, *. , «.. Λ , (32,36,37,38)
- elasto-plastic models
, . (19,32,39,40,41)
- viscous models
Elastic calculations show that in the case where the ratio of slab width and
the spacing between two successive supporting rolls is more than 2, the treat
ment of the slab as a beam gives results close to its treatment as a plate.
f max = pl4/32 E h 3
(1¡
where
ρ = ferrostatic pressure
E = Young's modulus
1 = distance between the rolls
h = thickness of the beam
f = bulging deflection
f = (4 q L 4 A 5 D)<J>(1, L) (2)
D = E h 3 /12(l > 2 )
where
^ = Poisson's ratio
(¡>(1,L) = function of the width and the length of the plate
The majority of elastic calculations have been made assuming uniform tempera
ture of the slab, and it is only with difficulty that the value obtained can
be made to give an acceptable representation of bulging.
Numerous values of Young's modulus for steel at high temperature are proposed
(19 31 32 38)
in the literature. ' ' ' However, the existence of a domain of elastic
deformation of steel at these temperatures has not been definitely established.
The models of bulging based on a law of purely elastic behaviour do not allow
description of all the observed phenomena.
1.3.2 Elastoplastic_models
ε = ε + ε = (σ/Ε) + κ (3)
t e p
where
ε = total strain
ε = elastic deformation
e
ε = plastic deformation
Ρ
σ = stress
E = Young's modulus
η = hardening coefficient
Κ = constant
(31)
Fig.32 shows the results obtained by Kobe Steel. The agreement with
measurements made by the authors seems rather good. It may on the other hand
be noted that the elastic calculation for the same conditions gives bulging
values that are clearly inferior.
1.3.3 Viscous_models
(32)
In the table in Fig.35 (taken from Schwerdtfeger et al are listed the laws
obtained by various authors for the two chief types of creep behaviour:
- 24 -
ii τ> ■ 4- · 4- (19,46)
1) Primary or transient creep
(46 47 48)
2) Secondary or steady creep " ' ' charac
t erised by rate of
strain έ = constant
But the experiments shown in Fig.34 show that there is no instantaneous de
formation, nor steady creep in the period that is of concern to the bulging
phenomenon (around 1 minute). The deformation ε thus arises chiefly from
primary or transient creep.
ε = a0 σ tn (4)
(n *v< 0.5 for low-carbon content steels)
where
a Q is determined from creep tests, with σ = constant
(19)
Fujii et al have derived an expression for the maximum bulging deflection
δ , resulting from a calculation of the bending of plates at uniform tem
perature
δ = (3 t n )/h 3 (6)
max
f = K(T){(q2-0L5-8t°-26)/h4·8} (8)
(32)
and that established by Miyazawa and Schwerdtfeger :
The boundary conditions used in these models ' assume that on the two
supporting rolls, the deflection, the derivative of the deflection, and the
curvature are zero.
In addition, the profiles of bulging between two rolls calculated using primary
(25)
creep data 1 ai
and 2 from Fig.36, reproduce Wunnenberg's measurements rather
well (Fig.37).
- 26 -
We therefore see that the results of these calculations are very sensitive
to the laws of steel behaviour that are used. The problem then arises of
determining the law of stress-strain-rate of strain for the steel under
conditions close to those under which the shell is stressed during continu-
ous casting, that is, by slow deformations (strains) of small amplitude, at
high temperature (between 800 and 1500 C) .
For modelling the bulging of continuously cast slabs, from all the laws of
visco-plastic behaviour described in the literature, we chose a
Norton-Hoff type of law, that can be expressed as:
2 % < ε < 50 %
4 % s -11 << εε << <400 % s
500 °C < θ < 1200 °C
28
In the framework of the present study, the laws of steel behaviour at high
temperature have been established by slow tensile tests carried out on a
modified creep-testing machine (Appendix 1) under conditions of stress near
to those of the bulging of continuously cast slabs.
Level of strain
at which the
From 0.5 to 4
results are
analysed: ε (in %)
Before the tensile test begins, the test-pieces are austenised by heating
to 1300 °C, holding for 5 minutes, and cooling naturally to the tensile-test
temperature.
Two steels for heavy plate have been studied; their chemical compositions by
weight % are given in the following table:
Steel Mn Si Al Nb
For steel 18 M 5 and the similar steel with an added 0.4 % of niobium, the
respective laws obtained are as follows:
144 5 599.3
¿
n o^ , 3521, (0.355 **!,) (0.686 aJ.nJ M0 ,
σ = 0.246 exp( ) ε θ+273 ε θ+273 (12.1
σ+ζ / J
with
σ in daN/mm2
θ in °C
ε in %
ε• ■
in o
% s-1
We have therefore re-analysed the slow tensile tests (54) into the form of a
law of Norton-Hoff type with coefficients η and m as constants. The results
obtained are:
For steel 18 M 5:
with
σ in daN/mm2
θ in °C
ε in %
1
• · c
ε in % s
We will present, successively, two cases studied. The first case involves
a static model of the bulging of slabs, that allows simulation of the defor
mation of the shell during a stoppage of the slab in the machine when casting
is in progress. The second case described concerns a kinetic model in which
we take into account the speed of sagging of the slab on the supporting
rolls, in order to simulate the deformation of the slab during steady state
casting.
31
We therefore assimilate the shell of the steel slab to a beam of small thick
ness compared to its length and its width (Fig.40, drawings 1 and 2 ) .
σ2 = σ.(Τ = 1300 °C) · (1500 - Τ)/200 for 1300 °C < Τ < 1500 °C (17)
(58)
- The.temperature field in the thickness of the slab is assumed to be known.
The symbols we have used are the following. They are used throughout this
report (Fig.40):
uniform temperature)
(19)
σ = g (c, c, y) (21)
(23)
(24)
These parameters are the restraining moment M alone in the static case,
or both the restraining moment M and the restraining force F in the
kinetic case. They are determined by the choice of one or both boundary
conditions at the perpendiculars to the rolls, according to whether we
are concerned with the static or kinetic case. The calculation of M and
o
of F in the kinetic case is done by dichotomy in such a way as to satis
fy the imposed boundary conditions.
(24)
(26)
(27)
(28)
in • ι m-1 ·
σ = λ (Τ) ε ε (29)
Computer programes have been written for the kinetic and static models
relating to a beam with a heat gradient in its thickness. Below we will
present and discuss the various results obtained using these models.
The calculations made with the kinetic model of bulging have been
carried out in 3 stages :
- Secondly, we applied the model to the casting conditions for which Wünnen-
berg made the only measurements made in works trials that are reported in
the literature, and compared our results with those obtained by other authors
using various models.
Solid thickness 79 mm
Roll pitch 860 mm
Metallurgical height 4 m
Surface temperature 1030 °C
Casting speed 0.85 m/min
Steel grade X 60
In the first portion of the interval between rolls, the deflection of the
slab is positive, and then becomes negative.
39 -
Moreover, only boundary conditions group no. 3 allows the taking into account
of a non-zero deformation at right angles to the rolls, and of a maximum
tensile stress located in the vicinity of these rolls, phenomena revealed
by the experiments of Yamauchi et al (Fig.4).
2.3.1.2 Comparison_of_the_calculations_with_the_avai^
Fig.43 shows the results obtained when using boundary conditions no. 1 for
steels 18 M 5 and 18 M 5 Nb. The calculations were made for the casting
conditions reported by Wünnenberg, assuming constant coefficients m and n
of the behaviour laws.
- At all points on the distance between the supporting rolls, the order of
magnitude of the calculated deflections is the same as that of the measure-
ments .
- 40
- The calculated values are greater than the measured values, the maximum
difference between measured and calculated being 2.5 mm for steel 18 M 5 Nb
and about 5 mm for steel 18 M 5.
The results obtained, using the same assumptions and boundary conditions
no. 3, are shown in Fig.44.
It can be seen that the general form of the curves is independent of the
law of behaviour of the steel and differs considerably from that derived
from Wünnenberg's measurements.
The results presented show that it is for boundary conditions no. 1 that
the calculations best describe the measurements available in the literature,
from a qualitative point of view. The differences between calculated and
measured deflections are, in this case, relatively small.
On the other hand, the various calculations made show that the bulging
deflection is not very dependent on the type of boundary conditions inserted
in the model.
2.3.1.3 Comparison of the calculations with some other results taken from
the literature_
(39)
In Fig.45 are shown the results obtained by Palmaers and Etienne,
calculated for conditions identical to those retained in the previous para-
graph, for a solid thickness of 83 mm.
- 41 -
(39)
The deformation obtained by these authors, using a creep law, has a
profile near to that we obtain with our kinetic model. The derivatives
of deflection are likewise different over the two rolls.
Fig.46 compares the results obtained using our model with those derived,
(32)
for different types of steel, by Schwerdtfeger, using a model that
incorporates the creep laws. As this figure shows the evolution of the
calculated deflection with the spacing of the rolls, we have superimposed
the results of measurements made by Wünnenberg.
It can be seen that the calculations made with our kinetic model using
boundary conditions no. 3 are located in the upper range of Wunnenberg's
measurements, whereas the results using a creep law correspond more closely
to the lower range of these measurements and are even inferior (case of
curve 5).
The characteristic values of the machine that were used for the calculations
are as follows:
*
At each level in the machine, the shell thickness and the shell temperature
have been determined in accordance with heat calculations made using existing
/ COI
models at IRSID.
42 -
VJe compared the evolution of bulging along machine 12 for the two previously
mentioned laws, equations (14) and (15), of visco-plastic behaviour of hot
steel, in which we have assumed the coefficients m and n to be independent
of temperature.
In the upper part of Fig.47, the diameter and the spacing of the supporting
rolls at each point are indicated.
For the two steels studied, it can be seen that the slab deflection increases
suddenly on passing from one zone of rolls to another zone with rolls of
different diameter. Then, in this zone of rolls of the same diameter, the
deflection decreases progressively, in proportion to the increasing thick-
ness of the shell, until the slab enters the next zone.
- 43 -
The maximum values of the deflection are rather small, less than 0.5 mm
at the top of the machine and over a distance of about 7 m from the meniscus,
for roll spacings less than 370 mm.
The maximum bulging deflection is much larger further down the machine and
with each change of roll spacing. It is at a maximum between rolls 76 and
78, that is, where the ferrostatic pressure and the variation of roll spacing
is greatest.
At constant roll spacing (pitch), the decrease of the maximum bulging de-
flection in proportion to descent in the machine shows that an increase of
shell thickness is a very important factor in the limitation of bulging,
and takes precedence over the increase of ferrostatic pressure toward the
base of the machine.
These results show the fundamental part played by roll pitch and the thick-
ness of the shell in the phenomenon of bulging. The smallest deflections
are obtained for steel 18 M 5 Nb (15), which is the more heat-resistant steel.
One of the methods that can be used to limit the bulging of the shell of a
continuously cast slab is control of its cooling. Measurements made by
Wünnenberg ' show that for a slab surface temperature of 1050 C, the
bulging of the shell between the supporting rolls is 7 mm, and when the
temperature is decreased to 900 C, the bulging is only 3 mm (Fig.48).
Using our kinetic model, we have compared the evolution of the bulging, along
machine 12 at USINOR Dunkerque, for two cooling profiles:
In Fig.48 are shown the results obtained for these two settings of the
cooling system in the case of steel 18 M 5. The hot profile curve is
that shown for steel 18 M 5 in Fig.47.
In the case of the cold profile, it is apparent that the bulging deflections
between rolls are less than 0.1 mm over the whole length óf the machine.
The bulging is thus practically suppressed.
t = l/2v (30)
where
1 = the roll pitch (= space between rolls)
v = the casting speed
Fig.49 shows the evolution of the deflection between two supporting rolls,
calculated by using the static model with each of these two laws of behaviour.
- the maximum deflection of the slab is symmetrical between the two rolls
and its maximum is located at the middle of the interval between the two
supports.
a) Static case
In the static case, we have chosen to express the calculated maximum deflection
as a function of the technical parameters of a continuous casting machine
by means of the relationship:
f = K(T){(p~ (32)
max
with
Τ in °K
σ in daN/mm2
1 in mm
h in mm
f in mm
t creep time in s
- 47 -
We present below the results obtained using two laws of behaviour (Table 4)
The high value of the correlation coefficients obtained shows that the
relationship accounts perfectly for all the results obtained.
^Regression Correi,
Vcoefficient coef. a b c d K (T)
Behaviour^v
law ^v
-9
Equation (14) 4.32·10 · exp
steel 18 M 5 0.998 2.00 6.64 0.36 5.23
slow tension (6.655Ί0 - 3 · T)
b) Kinetic case
The calculations in the kinetic case have been made by using the same method
as in the static case, but we chose the following formula for the maximum
bulging deflection:
-12
Steel 18 M 5 7.867 10
boundary conditions 2.25 6.10 0.57 5.26 exp 0.997
group no. 1 (0.00466 T)
For the two laws of behaviour, equation (14) and (15), we have determined
the regression coefficients of equation (33) by a dimensional analysis method.
Regression coefficient a b c d
These results show the essential parts played in bulging by the spacing
(pitch) of the rolls and the thickness of the slab shell.
- 49 -
The dimensional analysis carried out shows that the different regression
coefficients depend solely upon the values of the hardening coefficient n
and viscosity coefficient m in the law of behaviour of the steel.
Thus the pitch of the supporting rolls appears to be the essential parameter
controlling the intensity of bulging; but, with conventional engineering
of the rolls of a continuous casting machine, the diameter of the rolls (and
therefore their spacing) at a given level in the machine cannot be less than
a limiting value, below which the roll deflection, due to the action on it
of the ferrostatic pressure, will be prohibitive. In order to use rolls of
smaller diameter than this minimum value, it is necessary to have recourse
to special techniques, such for example as rolls with intermediate bearingso
The ferrostatic pressure exerted by the liquid steel present in the core of
the product has a substantial influence on the intensity of bulging of the
- 50 -
Equation (33) however shows that, other things being equal, the bulging
deflection decreases when casting speed increases. But an increase of casting
speed is accompanied on a continuous casting machine by a decrease of shell
thickness, and an increase of surface temperature. If all these parameters
are taken into account, it becomes apparent that bulging increases slightly
with the casting speed. Appendix 6 illustrates this phenomenon by means of
simple calculations.
After describing the physical model developed and the experimental method
used, we will present and then discuss the results obtained.
6
σ = 0.38-10 exp(4620/T) ε 0 ' 2 έ 0 - 0 8 (34)
In steelworks II, the casting machines have rolls of seven different dia
meters . We have therefore divided the machine into seven zones, each zone
corresponding to rolls of uniform diameter. For each zone we have deter
mined the mean shell thickness as well as the maximum and minimum ferro
static pressures.
In order to be able to simulate these seven zones with only one mockup,
we altered the geometric scale for each of the zones studied, by fixing
the roll diameter at 100 mm. The geometric scale therefore varied from
0.227 to 0.555.
53 -
De
3.1.3 scription_of_rhe_mock-up
The various roll spacings (1) range from 107.5 mm to 125 mm. Each roll can
be adjusted + 2 mm around its reference point.
On both sides of the mock-up, and in the interval between each pair of rolls,
are provided supporting plates so shaped thar. retraction is rapid and easy.
By this means, parasitic deformations of the base-plate,that would introduce
errors into the deflection measurements, are avoided.
- 54
3.1.4 Sicilituäe^study
The results of this study show that the dimensioning of the mock-up does
not allow perfect simulation of the phenomenon of bulging on an industrial
continuous casting machine.
(42)
3.2 Realisation of trials on the mock-up
The second series of trials run was concerned with simulation of the travel
of the slab at a speed ν over the supporting rolls. The vessel is, as re
marked, equipped with a system that draws it along at constant speed.
3.2.1.1 Oniformity_of_the_base-plate_thickness
The plates that constitute the plasticine vessels used for the
thials are shaped with a plasticine roller available at IRSID. This roller
(rolling mill) allows an accuracy of the order of 1 micron to be obtained.
Many vessels have been prepared and some trials showed that with
vessels more than 280 mm wide the deflection is practically independent of
the width. We therefore decided to use a vessel 280 mm wide. The vessel
length chosen is 400 mm, representing the distance between the outermost
rolls, and 140 mm was the chosen height of the walls.
56 -
3.2.1.4 Waiting time between tests made with the same vessel
The first tests carried out using the mock-up have revealed a certain
number of "parasitic" phenomena that we had to take into account when
analysing the tests.
3.3.1 Influençe_of_the_true_weight_of_the_plasticine
In order to take into account the intrinsic weight of the plasticine during
the static tests, we have considered that the pressure acting on the plasti-
cine vessel is equal to the height of water increased by twice the thickness
of the vessel. The density of the plasticine is in fact 2.
3.3.2 Influençe_of_the_flasticitY_of_the_Dlasticine
Once the supporting plates are lowered, at the start of each static
test, the base-plate of the vessel deforms instantaneously. This is due
to the partly elastic behaviour of plasticine, as has been reported by
( 61 )
several authors.
The elastic component of the deflection measured on the mock-up has been
determined by carrying out the following test (Fig.53). We filled a vessel
25 mm thick with 100 mm of water in a period of about 900 s, then we quickly
emptied the water from the vessel and observed the evolution of the deflection.
In Fig.53, it can be seen that during loading the deflection takes a quasi-
-instantaneous value of f = 8 0 μπι, then increases with time, following a
curve characteristic of the phenomenon of creep. When the load is removed,
the deflection decreases quasi-instantaneously, then continues to decrease
until it attains an apparently constant value of 67 am. This residual
deflection may reasonably be assigned to the plastic component of the defor
mation.
In all the tests, static or kinetic, we have assumed that the visco-plastic
deflection (fvp) measured on the mock-up is equal to the measured total
deflection (f ) from which we deduct the initial deformation (fe). In the
case of kinetic tests, we have likewise assumed that the elastic deflection
was equal to the instantaneous deflection measured under the same conditions
during the static tests.
59
During our tests, spread over several months, we used two different lots of
plasticine, for which we determined two laws of mechanical behaviour.
The first lot of plasticine (A) was used mostly for static tests, whereas
the second lot (B ) was used for kinetic tests.
3.3.4 Determination_of_the_la.ws_of_5echanical_behaviour_of_glasticine
, , , n «m ,__.
σ = λ (Τ) ε ε (37)
νρ νρ
where
εΛ^ = ε (σ/Ε) (38)
νρ
¿ = σ/Ε (39)
νρ
Ε = Young's modulus
*
cf. Dukes, W.A., "Rheological measurements of lutings", in Rheology of
Disperse Systems, ed. C.C. Mill, Pergamon, London, 1959. Includes
measurements on Plasticine.
60
7
σ = 1.1210 exp(4066/T) ε 0 2 8 1 ¿ ° · 1 0 7 (40)
r
vp vp
with
σ in kg/mm2
Τ in °K
ε is dimensionless
^ . 1
ε in s
vp
a) Static tests
Fig.55 shows the evolution of the profile of deformation between the rolls
as a function of creep time during static tests.
It can be seen that the profile is symmetrical, in good agreement with the
results from the mathematical model.
The upper curve represents the profile of the plasticine plate before the
supporting plates are lowered. In this particular case, the profile is
near to a horizontal line, and thus very satisfactory.
b) Kinetic tests
The kinetic tests have Deen carried out for different speeds of travel of
the vessel, different pressures (loads) of water, and plates of different
thickness. Fig.56 shows the deformation profiles obtained during a kinetic
test with different levels of water. It can be seen that the maximum de-
flection is displaced toward the downstream roll by about 60 % of the dis-
tance between two consecutive rolls. This confirms the result that was
also obtained with the mathematical model in the case of steel (see 2.3.1.1),
The vertical of the upper generatrix of the rolls is also shown in Fig.56,
and it can be seen that the point of contact between roll and plasticine
plate is located upstream of the upper generatrix of the roll, relative to
the direction of travel.
- 62 -
The plasticine plate thus presents, very locally, a positive side, relative
to the horizontal plane tangential to the upper generatrices of the rolls.
Nevertheless, the kinetic tests carried out using the mock-up have never
produced the asymmetric profile of deformation obtained by calculation with
identical boundary conditions on consecutive rolls.
In addition, the kinetic tests show that the tangent to the deformation
at the point of intersection with the vertical plane containing the upper
generatrix of the rolls is not horizontal. Amongst the boundary conditions
retained until then in the modelling, we will retain, for comparison with
the values measured on the mock-up, the symmetrical boundary conditions on
the two adjacent rolls, condition group no. 3. These represent both the
periodicity of the deformation and the fact that the deformation is not
tangentially horizontal to the rolls.
In order to vary the surface temperature of the baseplate the water in the
Y
vessel was heated by a pump thermostatted to 5 above the aimedfor surface
temperature.
64
3.4.3.2 Ças^_of_static_tests
This confirms that the two parameters with the greatest influence on the
bulging deflection are the roll pitch (space between the rolls) and the
thickness of the plasticine plate.
3.4.3.3 ǧf2_2í_lSÍ22y2_Í2EÍs
Ρ in mm
1 in mm
t in seconds
Fig.61 shows the evolution of the measured and calculated total deflections
as a function of the creep time. In all cases, the calculated total de-
flection has been obtained by adding the measured instantaneous deflection
(assigned to the elastic component of the deformation) to the visco-plastic
deflection calculated using the model.
Fig.61 shows a good agreement between measured and calculated values, even
though the calculated values appear, in the majority of cases, to be greater
than the measured values.
3.4.4.2 Çsse_of_the_kinetic_tests
Calculations Calculations
Coefficients Measure- (Dimensional (Statistical analysis of the
ments analysis) results of the mathematical
model)
In Tables 8 and 9 are assembled the principal values obtained for the
regression coefficients of the equation expressing the bulging deflection
as a function of the casting parameters :
67
where
ρ = pressure exerted
1 = spacing of the supporting rolls
h = thickness of the plate
t = residence time (static case), or portion of the residence
time of the plate between two consecutive rolls (kinetic
case)
The tables group together the results obtained for steel 18 M 5 and those
relating to the 2 lots of plasticine used in the mock-ups.
Table 8 relates to the collection of results obtained for the static case,
whereas Table 9 relates to the kinetic results.
These results show that the static and kinetic mathematical models allow
the influence of casting parameters on the bulging deflection of steel
slabs to be taken into account very well. In fact, the measurements made
on the mock-up, using plasticine with coefficients m and n very close to
those of hot steel 18 M 5 (the case of plasticine B) lead to values of the
regression coefficients of equation (43), iri the case of static and kinetic
tests, that are close to those obtained by calculation with mathematical
models using the law of behaviour of steel 18 M 5.
The fact that the orders of magnitude of the coefficients found from measure
ments on the plasticine mock-up and by the calculations for steel are the
same, seems to confirm that these exponents depend essentially on the coef
ficients of work-hardening n and of viscosity m of the law of behaviour of
the material.
- 68 -
Method of -
Calculations Measurements
determination
Coefficient
a 2 1.3
b 6.64 8.6
c 0.36 0.35
d 5.23 5.8
After describing the principle adopted for the measurements, their execution,
and the method of analysis used, we successively present the results obtained,
namely :
- observations relating to the thickness of the cast slab at exit from the
' machine and the influence of a substantial misalignment of the supporting
rolls on the bulging of the slab.
The slabs cast on this machine have a maximum width of 2090 mm and a thick-
ness between 210 and 300 mm. Various grades of steel are cast, but a large
proportion (80 %) of them are destined to be converted to heavy plates,
especially for welded pipes.
70
All the trials were run on strand 1 of machine no. 12, which is equipped
(28)
with IRSID-CEM "mixing" rolls a technique that allows accurate control
of the solidification structure at the centre of the slabs by the action
of electromagnetic forces.
It should likewise be noted that on the two strands of this machine, as far
as roll 88 located in a zone downstream of the chosen bulging measuring-point,
the spacing of the outer radius and inner radius rolls is fixed by means of
wedges placed between their bearings.
4.1.2 Choice_of_measuring_technology
The roll followers are positioned in contact with the exterior generatrix of
the roll diametrically opposite the generatrix in contact with the slab, and
providing a continuous measurement of the displacement of the exterior genera-
trix. The third follower is located at the middle of the space between the
rolls (Fig.63), continuously recording the displacement of the slab. The
follower is provided with a system that allows it to be withdrawn so that the
71 -
roller will not be heated too much. In addition, the three sensors are
water-cooled. During the last campaign, we have been able to put a fourth
follower in an analogous position to the two roll followers. The follower
has been installed above roll 75.
Ma
4.1.3 JSÌn2 *^ e measurement
a) Experimental method
We made the majority of our measurements between rolls no. 76 and no. 78
on the outer radius face of the slab.
2) Rolls 76 and 78 are located in the zone where the liquid core closes,
under the usual casting conditions, which allows the base of the liquid
core to be brought down to the level of the measuring point by slightly
altering the rate of withdrawal.
The three sensors on the outer radius are fixed on a rigid beam that is
built into the framework of the machine (Fig.63). We assumed that this
beam could serve as the fixed reference-point for our measurements, which
amounts to neglecting the deformation of the framework of the machine when
casting is in progress.
roll followers were then referred to this (f'l and f'3). These three values,
f'l, f'2, f'3 together constitute the values called the calibration of the
followers. Calibration is carried out, rather frequently, between the trial
castings, in order to follow the evolution with time of the calibration con-
stant E = f'2 - ^(f'l + f'3) (eqn. 46). For this purpose we used a gauge
1500 mm long and of 12 m radius fixed to rolls 76 and 78. A geometric model
shows that the determination of this constant is not dependent on the relative
angular position of the two adjacent rolls.
The measurement signals from the position sensors are recorded on a chart and,
after digitalisation, in an H.P. 1000 computer. At the same time as bulging
was measured, we continuously recorded the casting speed and the surface tem-
perature of the product at the level of the measuring point by means of a
monochromatic optical pyrometer aimed at the inner radius of the slab. In
the course of our trials, we have been able to improve the accuracy of the
measurement to 0.05 mm.
4.2.1 Examole_of_the_direct_recording_of_data
The speed of withdrawal has likewise been recorded in this figure, which
shows steady state casting followed by a decrease of speed due to casting
coming to an end. We noted that the three position signals have a periodicity
that corresponds to the time taken for a complete rotation of the rolls.
Their mean value, whilst the speed is steady, tends to decrease regularly.
The slab signal amplitude is about 1 mm whereas those of rolls 76 and 78
are respectively 0.25 mm and 0.5 mm. The decrease of speed and the momentary
stoppage of casting bring about a decrease of the deflections of the two
rolls, followed by heating of the generatrix in contact with the slab. The
slab signal likewise decreases by an amount equal to that of the deflection
of the rolls.
4.2.2 Çai2HÍ2Íi22_2Í_íí}2_inEÍ25í2D22HE_21íi2iri9_2^2Íi2c£i0íí
The measurements made during calibration serving as the zero reference, the
instantaneous bulging deflection "f" has been derived using the following
formula:
f = f2 - M f l + f3) - E + d (47)
where
f = bulging deflection in mm
fl, f2, f3 = instantaneous values of measurements from the
three sensors
d = expansion of the rolls over their diameter
E = calibration constant
74 -
The value of d has been determined by simple calculation of the heat expansion
of a roll, and estimated to be 0.4 mm.
Two examples of the results obtained are shown in Fig.67. The curves in the
upper part of the figure (Fig.67a and 67c), after filtering at 0.028 Hz,
show the evolution with time of the bulging deflection, measured under the
following casting conditions:
Casting no. 2: steel for heavy plate, micro alloyed with Nb; casting
speed 0.86 m/min.
The two cases shown relate to the casting of slabs 210 mm thick. In the lower
part of the figure are shown plots of the energy spectra relating to measure-
ments made in the course of these two casts of steel. These spectra were
obtained by numerical treatment, by a computer, after digitalisation of the
bulging deflection signals filtered at 0.028 Hz.
These results show that the measurement of the bulging deflection between
rolls 76 and 78 of the casting machine is substantially influenced by the
deformations of the slab during its passage over the rolls upstream and on
both sides of the measuring point.
This phenomenon, linked to the permanent set of the rolls (called "eccentricity")
is not taken into account in the various mathematical models for the calcu
lation of bulging that have been developed during this study.
During the analysis of our measurements, we have adjusted the angular position
of rolls 76 and 78 according to their eccentricity. We have designated as
maximum (or minimum) position of each of these rolls, the angular position
such that the exterior generatrix is the most distant (respectively, the
closest) to the inner radius face of the machine.
Case 1 :
Rolls 76 and 78 are in the maximum position at the same instant (i.e. they
are in phase). At each rotation of the roll, the bulging deflection then
displays a maximum value before the rolls are in the maximum position.
The angular displacement between these two phenomena corresponds to the
time taken by the rolls to make a quarter-rotation.
Case 2:
Rolls 76 and 78 are in opposite phase: in this case, the bulging deflection
displays two maxima during one rotation of the rolls. One of them is a
quarter-rotation of the roll in advance, relative to the maximum position
of the roll presenting the greatest eccentricity.
Case 3:
Rolls 76 and 78 are in any position: the maximum of the bulging deflection
then appears when the roll displaying the greatest eccentricity is in the
maximum position.
All our measurements were made with constant machine geomerty, that is,
the roll spacings and the roll diameters have not been modified. However,
for one test, a 4 mm misalignment of roll 78 towards the outer radius was
arranged. Likewise, the adjustment of the secondary cooling has always
been kept at its various settings according to the casting speed and the
grade of steel being cast.
The parameters whose influence on the bulging deflection have been studied
by us are:
the grade of steel cast (the whole range of steels cast at USINOR
Dunkerque)
the thickness of the slab cast: 210, 250, and 300 mm.
The tests proper have consisted of recording, during the castings, the
signals from the various installed sensors, and of taking samples (cross-
sections) for sulphur prints, during each run of casting at a steady speed
of a length greater than 20 m.
The signals emitted by the sensors of roll displacement show that during
steady state casting, rolls 76 and 78 oscillate in the direction normal to
the broad face of the slab with an amplitude when under load varying from
0.5 mm to 1 mm (see Fig.66). The eccentricity when not under load is usu-
ally slightly different to the eccentricity when under load, but no precise
correlation has been found between these two values. The following comments
can be made on the subject of the eccentricity of rolls:
Fig.68 shows the evolution over time of the signals emitted by the sensors
relating to rolls 76 and 78. The curve shows that when they are put under
load, that is, when the foot of the slab passes, the action of the ferro-
static pressure causes these rolls to suddenly develop a deflection of 1.2 mm
for roll 76 and 1.0 mm for roll 78. After the head of the slab passes, at
the end of casting, the deflection of the rolls disappears.
Be
4.4.1.3 haviour_of_the_rolls_during_a_casting_stoDgage
The behaviour of the rolls thus described shows that the slab suffers defor-
mations that are not negligible, arising from the roll movements during the
casting.
4.4.1.4 Relation between the deflection of the rolls and the width of
the_slab_
The experimental points are plotted in Fig.69 and compared with the results
of a calculation of deflection in which the roll was treated as an elastic
beam resting on two simple supports (see paragraph 5.2.2). It is apparent
that for a distance between bearings of 2800 mm, and slab widths close to
1800 mm, the mean measured deflection of the rolls is between 1.05 and 1.7 mm,
whereas the calculated deflection is only 0.85 mm. However, it appears that
the values calculated in this way are in all cases in relatively good agree-
ment with the minimum measured values.
4.4.2 Influ2nÇe_of_various„Parameters_2£_the_bulging_deflection
4.4.2.1 Influence_of_casting_soeed
In the case of extra-soft steels, the bulging deflection varies from 0,75 mm
to 1.25 mm with casting speeds varying from 0.6 m/min to 0.9 m/min.
So far as steels for heavy plate are concerned, the measurements show the
great influence of the thickness of the slab being cast. However, for each
thickness, the bulging deflection increases with the casting speed.
The mean values of the deflection, measured during the casting of steels
for heavy plate, are classified in the following table:
0.85 210 2
0.65 250 0.9
0.65 300 1.6
81 -
When comparing the measurements made of the two types of steel studied,
it is apparent that:
During the casting of an Nb alloyed steel for heavy plate, the amplitude
of the measured deflection is greater than that recorded during the
casting of extra-soft steel for sheet. At a thickness of 210 mm, and a
set casting speed, either 0.8 m/min for slabs for heavy plate, or 1.0 m/min
for slabs for sheet, the respective measured amplitudes are 0.8 and 0.4 mm.
4.4.3 Çomparison_of_measured_and_calculated_bulging_deflecti
The measured bulging deflections have been compared with the values
calculated, for the same conditions, by means of the kinetic visco-plastic
model of bulging.
The comparison has been made for the two steady states of casting (casting
no. 1 and casting no. 2) defined in 4.2.3.
Fig.74 shows the measurements made during periods of steady casting speeds
of these two castings, together with the bulging deflection calculated using
the symmetrical boundary conditions over the rolls (condition group no. 3)
with the law of behaviour of steel 18 M 5. For the casting conditions employed,
the model allows calculation of a deflection of 1.04 mm for casting no. 2
and of 1.07 mm for casting no. 1. In both cases, the calculated values are
in good agreement with the mean values of the measured bulging deflection.
These results establish the validity of the model with boundary conditions
group no. 3.
4.5 Study of the behaviour of the slab during the casting process
Fig.75 shows, for the case of a steel micro alloyed with niobium for rolling
to heavy plate, the evolution of the slab thickness measured at the exit
from the machine. Also shown in Fig.75 is the bulging deflection of the
slab during its passage between rolls 76 and 78. In order to make the
bulging deflection correspond to the measured thickness of the same portion
of slab, we have displaced the recordings by the time that the slab takes
to travel the distance separating the two measuring points.
It can be seen that slab thickness varies periodically, with a period corre-
sponding to the time per rotation of the supporting rolls in the zone of
the machine in which the slab solidifies. The amplitude of slab thickness
is 1 mm on average, the order of magnitude of the corresponding bulging
deflection.
During the transitional stage, the passage of the liquid core past roll 78
was clearly detected. At that precise moment, the deflection of roll 78
suddenly decreased, the solid slab thereafter being in contact with the roll
only intermittently (a phenomenon due to the expansion of the rolls as a
result of the radiation from the slab) (Fig.76, shaded zones).
- 84 -
If the results obtained by means of the thermal model are accurate, we can
put forward the hypothesis that there exists a zone in the base of the liquid
core that is sufficiently compact, even though partly liquid, as not to
transmit the ferrostatic pressure.
From these sulphur prints we have rated the centre segregation according to
a chart drawn up by IRSID that uses an index of segregation ranging from 1,
for very slight segregation, to 10, for the most pronounced segregation
(Fig. 77).
(28)
In fact, the use of electromagnetic agitation ensures that even with high
superheats of the steel in the tundish, a large equiaxial zone is present
at the centre of the product.
- 85
It can be seen that the use of agitation, and the consequent presence of a
large equiaxial zone at the centre of the product, allows the index of segre-
gation observed in the slabs to be limited to 6, whereas in the absence of
agitation, and therefore usually with a columnar structure in the centre of
the product, the segregation index attains 10, the maximum value. Likewise,
the mean index of segregation, 7 in the absence of agitation, is only 4 for
the agitated castings.
The results obtained with and without agitation are shown in Fig.79, in the
form of an experimental relationship between the bulging deflection and the
casting speed.
4.6.3 52i5íi22_22í^22D_!E}2_2HÍ9Í22_§2íi22Íi2í?_aí}2^_í^e_22í}t£2_Ee2£e2at'on
Fig.80 contains, for 210 and 250 mm thick slabs, as a function of the calcu
lated bulging deflection, histograms indicating the severity of the centre
segregations seen on sulphur prints, in the absence of electromagnetic
agitation during casting. To simplify the presentation of the results,
these have been grouped in three classes: bad (i > 6), medium (3 < i =ζ 6),
and good (i ^ 3).
It is apparent that for the 210 mm thickness, the majority of sulphur prints
made for casting conditions such that the calculated bulging deflection is
within the range 1 to 1.5 mm have segregation indices below 6. In addition,
for bulging deflections greater than 1.5 mm, 80 % of the sulphur prints have
been rated higher than 6. In this case, we consider that the segregation
index is the greater, the more the bulging deflection. The maximum permiss
ible bulging deflection seems to be less than or equal to 1.5 mm.
- 87 -
At 250 mm thickness, the calculated bulging deflections are all less than
1 mm, for the casting conditions used on machine no. 12. In this case,
less than 25 % of the sulphur prints have a segregation index higher than 6.
Two types of results have been obtained during the measurements made
at USINOR Dunkerque: those of the first type describe the mechanical behav-
iour of the slab during its passage through the machine whereas those of the
second allow an analysis of the interdependencies between centre segregation,
solidification structure, and bulging of the slabs.
Besides bulging, over the length of the machine the slab in the course of
solidifying is subjected to a certain number of deformations related to
faults in the geometry of the supporting system: in particular, eccentricity
and misalignment of the rolls. The order of magnitude of the resultant
deformations of the slabs is sometimes equal and even greater than the defor-
mation by bulging. Between two rolls, the shape of the slab at each instant
depends simultaneously on the position of each of these rolls, but likewise
on the deformations imposed on the slab upstream.
On the basis of the results presented in the first four parts of this
report, a critical study has been made of the chief solutions to the prob-
lem of bulging proposed in the literature. For this, we have used the
developed kinetic visco-plastic model of bulging, as well as a model of
roll deformation due to ferrostatic pressure developed on this occasion.
4) One of the best means of supporting the shell during its descent in the
continuous casting machine, and of minimising its bulging, is to ensure
support over the largest possible surface of the slab faces. A solution
(66)
adopted by Kobe Steel, and used over a limited length below the mould,
consists of using "marching bars" consisting schematically of two groups
of shoes that are alternately applied to the slab. This is then perma-
nently supported over its length.
90
5) Under this heading we group together all the techniques which, sometimes
without limiting bulging itself, limit its metallurgical and mechanical
consequences. An example of this is the technique of electromagnetic
/ oft)
agitation, which perceptibly limits the streaks of centre segregation.
For this, we have defined the criteria of allowable bulging of the slab and
of maximum deformation of the rolls.
5.2.1 Assumptions_made_f2£_the_bulging_calculations
All the calculations of bulging have been made using the following
conditions:
4) The boundary conditions used for the calculations are the boundary con-
ditions group no. 1 defined in 2.2.3.
5) The maximum casting speed used in these calculations is 1.6 m/min. This
speed will be used for the dimensioning calculations of the continuous
casting machine, whereas a mean speed of 1.3 m/min will be used to define
the usual casting conditions.
- 92 -
5.2.2 The_model_of_deformation_of_the_rolls
where
fr = roll deflection (mm)
e = 60 mm , L = 2150 mm , 1 1600 mm
o o
with
L = metallurgical length (m)
5.2.3 B u l g i n g _ p r o f i l e s _oh2tainecl_on_aLn_inclus t £ i a l _ m a c h i n e
The roll spacings are likewise indicated in Fig.81. We can see that for
this casting speed, the maximum bulging deflection is obtained at the last
change of roll size and is 1.17 mm.
The profile of roll deflections along the machine shows (contrary to obser-
vations concerning the bulging deflection) a decrease with each change of
roll diameters, then a regular increase with constant roll spacing.
For this profile of roll spacings, we find that the roll deflection is always
less than 1 mm.
We have seen that bulging is the origin of two major types of defect:
internal cracks and centre segregation. The mechanism of formation of these
94 -
defects, described in the same part of this report, allows the use of two
criteria that can characterise the intensity of bulging: the bulging de-
flection, connected with the formation of centre segregation, and deformation
at the solidification front, connected with the formation of internal cracks.
The study of the literature has allowed us to hold that the deformations at
the solidification front (generally considered responsible for the occurrence
of internal cracks) range from 0.2 % to 1 % according to the type of steel
being cast and the rates of strain encountered. In the majority of the
calculations, we will therefore retain an allowable deformation of 0.5 %.
With regard to the bulging deflection, the findings reported in the literature,
as well as our own results, show that if the deflection exceeds 1 mm, the
segregation index reaches values indicative of a very bad defect. The deflec-
tion value of 1 mm will be retained as a median criterion for the majority
of our calculations in which a criterion of bulging deflection is used. The
two values of the chosen criteria are in agreement with the results shown in
Fig.81.
Taking into account the calculated values of roll deflections in the case
of machine no. 1 (Fig.81), the mean value of the mechanical criterion adopted
corresponds to a maximum allowable roll deflection of 1 mm. We will likewise
study the influence of the value of this criterion, as well as that of the
distance between bearings, and that of slab width, on the profile of optimal
roll spacing in a continuous casting machine.
- 95
The influence of slab width and also of the distance between bearings are
shown in Fig.83, for the case of a 1 mm criterion of roll deflection.
These curves show that the diameters of hollow rolls ought to be greater,
the wider the slab or larger the distance between bearings of the rolls,
and the smaller the value of the criterion of deflection.
Taking into account the assumptions made for the purpose of calculation, these
results are independent of the casting conditions practiced, especially the
casting speed.
The extreme criteria lead to roll spacings which differ by about 200 mm
over the whole length of the machine. All these profiles of roll spacing
are almost parallel, capable of being roughly derived from one another by
vertical translation. At a given metallurgical length, the roll spacings
are the greater, the higher the value of the criterion.
The roll diameters, taking into account the minimum distance necessary between
two rolls to allow passage of the water sprays for secondary cooling, are
deduced from these profiles by subtracting this distance from the spacings
of the corresponding rolls.
97
Note that the increasing gradient of these profiles results from competition
between the increase of ferrostatic pressure and the increasing thickness
of the shell, a phenomenon of preponderant influence in the case of a machine
of 12.2 m radius.
Fig.84 also gives rise to the comment that the criteria ε = 0.5 % and δ = 1 mm,
on the one hand, and the criteria ε = 0.3 % and δ = 0.5 mm on the other, are
substantially equivalent, because they lead to profiles of roll spacing that
are very close together.
We observed that in the upper part of the machine (L < 5 m ) the bulging
deflection is usually much greater than in the bottom of the machine, and
that the bulging deflection seems to pass through a minimum in a zone between
12 and 15 m, before again increasing slightly in the bottom of the machine.
The large values of deflection obtained at a small metallurgical length at
a given strain ε, indicate that the bulging in the upper part of the machine
ought to be controlled by a deformation (strain) criterion and, in contrast,
by a bulging deflection criterion in the lower part of the machine.
The sensitivity of the results to the casting speed has been studied
for a criterion of bulging deformation (strain) ε = 0.3 % and a constant slab
surface temperature profile of 1000 C. Fig.86 shows the three profiles of
roll spacing corresponding to the same criterion of deformation ε = 0.3 %,
for the case of three different casting speeds, ν = 1.0 m/min; ν = 1.3 m/min;
ν = 1.6 m/min.
98 -
We verified that the roll spacings are the smaller, the higher the casting
speed. From this we deduce that the dimensioning of the rolls of a con
tinuous casting machine ought to be done for the conditions most favourable
to bulging and thus, in particular, the highest envisaged casting speed.
5.3.4 Influence_of_surface_temperature
The influence of the surface temperature of the slab has been made
evident in Fig.87, where we have plotted, for machine no.l, the profiles
of bulging deformation (ε) for two constant surface temperatures, one 1000 C,
and the other 900 C. It can be seen that on this machine the profile of
roll spacing leads to a bulging deformation that is always less than 0.25 %,
if the surface temperature is 900 C, whereas the deformation (strain) ε
reaches 0.4 % in the lower part of the machine for a constant surface tem
perature of 1000 C.
The results that we will present are however relative to a given machine
geometry, and greatly influenced by the assumptions made as a basis for
calculations. According to the assigned values used in our study of the
metallurgical and mechanical criteria, two situations may occur:
Situation 1 :
The curve giving the profile of roll diameters giving constant deformation
at the solidification front does not intersect and is located below that
of the roll diameters offering the maximum imposed deflection.
This is the case in Fig.88, where we have plotted the profile of roll dia
meters corresponding to fr = 1 mm (L = 2150 mm and 1 = 1600 mm) with
o o
respect to the two profiles corresponding to ε = 0.5 % and ε = 0.3 %. For
these conditions, in the domain situated between these curves and the curve
fr = 1 mm, it is possible to envisage all the feasible combinations of roll
diameters according to their position in the machine. One such example is
given in Fig.88 for an allowable deformation (strain) at the solidification
front of 0.3 % and for a maximum allowable roll deflection of 1 mm.
Situation 2:
1) To retain the use of undivided rolls, choosing roll diameters such that
their deflection remains below or equal to 1 mm, and decrease the slab
surface temperature sufficiently to limit its bulging to a solidification
front deformation (strain) of ε = 0.5 %.
100
2) strains at the junctions of roll trains (ε.) arising from the different
deflections under load of adjacent rolls of different diameters.
Fig.90 shows the contribution of ε, and ε. to the total strain. For our
d 3
calculations, we have assumed a constant misalignment of 0.5 mm between
all the adjacent rolls of the machine whose characteristics are given in
Fig.88 (machine no.l). We can consider that in these assumptions, the mis
alignment strain includes the strain due to roll eccentricity.
The misalignment strain alone has been calculated using the following
equation:
εα = 4 6 h/12 (51)
where
δ = misalignment (mm)
h = solid thickness (mm)
1 = roll spacing (mm)
where
ε = % strain
E = slab thickness (mm)
h = solidified thickness (mm)
The straightening and bending strains have been calculated for two types
of machine, no.l and no.2, for two withdrawal speeds and a constant surface
temperature of 1000 C.
The table shows that if the allowable maximum strain at the solidification
front is set at 0.5 %, the bending of the slab ought not to be done at a
single point. Assuming that, because of the viscosity of the steel, the
successive strains imposed on the solidification front are not additive,
bending at many points may be envisaged. If bending is imposed at 4 or 2
points, the strain imposed on the solidification front can then be estimated
as 0.22 % or 0.42 % respectively at each bending point.
In all cases, taking into account the strains arising from the machine
geometry in the expression for the criterion of allowable strain ε , imposes,
at all points in the machine, an allowable roll diameter that is smaller than
those previously presented, that were calculated by taking only bulging into
account.
Fig.91 shows that taking into account a misalignment strain of 0.1 % over
the length of the machine requires the roll diameters to be 20 mm less than
those obtained in the case where misalignment is neglected.
The following tables show the roll diameters that, for a casting speed of
1.6 m/min, allow a criterion of total strain of 0.5 % to be met in the bending
and straightening zones.
103 -
a) Straightening zone
Fig.91 and Table 15 (? 13) show that assuming a misalignment of 0.5 mm over
the whole length of the machine and straightening at a single point, the
rolls in the straightening zone can be of the undivided type if the allow
able total strain is 0.5 %. In contrast, if the misalignment exceeds 0.5 mm,
it then becomes necessary to provide divided rolls with intermediate bearings
in the straightening zone, or else to carry out straightening at several points.
b) Bending zone
For a total strain of 0,5 %, Table 14 (? 12) shows that bending at four points,
assuming a misalignment strain of 0.5 mm, allows undivided 240 0 rolls to be
104 -
1) The roll spacing being the preponderant parameter that controls the
degree of bulging, it is expedient to optimise its profile over the
length of the continuous casting machine.
3) In view of the values that have been assigned to the metallurgical and
mechanical criteria, the use of intense cooling of the slab will only
allow the retention of roll diameters and spacings sufficiently large
to avoid the use of divided rolls as a solution to the problem of bulging
if the slab surface temperature is kept below 900 C. A solution of this
last-mentioned type can give rise to metallurgical problems (surface
cracking on straightening of the slab) that demand complete mastery of
the procedure.
- 105
6. CONCLUSION
The bulging deflection and the strain induced at the solidification front
depend chiefly on the geometry of the supporting system, on the thermal
state of the slab, and on the casting speed. In the case of a steel for
heavy plate, the form of the bulging deflection (δ) may be written as a
function of the roll spacing (1) in mm, the metallurgical height (H) in m,
the casting speed (v) in m/min, the solidified thickness (h) in mm, and the
surface temperature (T) in K:
This series of experiments has likewise revealed that the bulging deflection,
at a given point in the installation, depends on the history of strains sus
tained upstream by the slab.
In addition, it appears that besides bulging, during its passage through the
machine the slab suffers substantial strains connected with defects in the
geometry of the system of support: eccentricity of the rolls, misalignment
of the rolls. The amplitude of these strains may, in some cases, be greater
than those due to bulging.
- 107 -
The mathematical model of bulging developed has then been used to define
the geometry of continuous slab casters, allowing minimisation of the strains
suffered by the product. In the majority of the cases studied, it is evident
that the use of monobloc (undivided) supporting rolls of judiciously chosen
diameters allows an acceptable limitation of the bulging of the slab and the
deformation (strain) of the rolls.
However, the results obtained by such a theoretical approach are very depen-
dent on the assumptions made concerning the mechanical behaviour at high
temperature of the steel: the law of behaviour and criterion of rupture in
the vicinity of the solidus. In order to help the makers and users of con-
tinuous casting plants to define the technology and the maintenance policy
of the machines, improvement of knowledge in this field ought to be con-
sidered as an objective having priority.
108
BIBLIOGRAPHIE
REFERENCES
[6] M. HASEGAWA et al
On the internal cracks in continuously cast slabs SUS 430 slabs
SHUNAN Works Nisshin Steel Co, Ltd
Tetsu to Hagané août 1982 76 A 115-118
[10] H. M I Z U K A M I , T. K I T A G A W A , K. M U R A K A M I , T. K A W A W A , Y. MIYASHITA
Influence de la composition et de la structure de solidification sur les
propriétés mécaniques des brames de coulée continue près de la tempé
rature de solidus.
Tetsu to Hagané, 1978, 64, S 149, traduction IRSID T.3607C, juin 1978
[11] H. B IAUSSER
Revue bibliographique sur la "transition ductile fragile" à haute tempé
rature des aciers C Mn
Texte non publié
[14] T. UEDA et al
High speed casting on verticalB ending type slabcaster.
Kawasaki Seitetsu Giho 12 (1980) n° 3
[15] S. KOJIMA et al
Condition of internal cracks in continuouscasting of steel slabs.
Kawasaki Seitetsu Giho 12 (1980) n° 3 p. 101109
[16] J. MIYAZAKI et al
lOOème ISIJ MEETING, 1980
[24] K. K U M A I et al
Study on solidification behavour, solute segregation and fluid flow in
continuously cast slab.
Tetsu to Hagané 1974 _60 n°7 p. 156-175
[26] KAWAMOTO et al
Wire Journal, Juillet 1977, p. 64-70
[29] K. WÜNNENBERG
Strangausbauchung zwischen Stützvollen beim Stransggiessen von Brammen
Stahl und Eisen, n° 6, 1978
[33] T. K A W A W A , H. SATO
Macro and microstructures of continuously cast slabs.
Tetsu to Hagané, 1974, 60 (5), 486-500
[35] M I Z U K A M I et al
Analytical calculation of bulging of the solidification shell formed at
continuous slab casting. NKK
9 7 t h ISIJ MEETING Avril 1979 S 169
@ . R5ID
Θ. R S I D
[51] C. ROS5ARD
Méthodes expérimentales d'essais de matériaux à chaud. Mise en forme
des matériaux et alliage. Ecole d'été de métallurgie physique Villars/Ollon
1975 Editions du CNRS 1976
[53] C. ROSSARD
Comportement mécanique et structural dans les conditions de déformation
à chaud
Rapport CESSID RFP 194, Janvier 1977
[55] G. B ERNARD
Détermination des caractéristiques mécaniques de l'austénite à haute tempé
rature pour des déformations lentes et de faible amplitude
Texte non publié
[56] R. LAMB S
Communication personnelle
ÇJ i R S i a ACI 82 RC
[60] P. RATTE
Laminage des poutrelles
Recherche CECA n° 7210 EA/3/304 (1978) IRSID
FIGURE CAPTIONS
Fig. 9 Influence of casting speed and cooling rate on inner cracks [4]
Fig. 28 Sketch showing the bulging zone used for computing the formation
of axial segregation [34]
Fig. 4-7 Bulging deflection profile along the slab caster No. 12 at
USINOR Dunkerque (boundary conditions No. 1)
Fig. 4-9 Influence of steel grade on the bulging profile between two
rolls (static model)
Fig. 58 Influence of water height (p) and casting speed (v) on the
total and visco-plastic deflection (cinematic experiment)
IRSID
coeur liquide
Liquid core
Shell
croûte solide gonflement
Bulging
WÊÈSÉÈÈÊÈÈÈÊ
^φ^^β00Φ^00Ε^.
mim y&BaÊkii&íià*
I 50mm.
Figure 3 : Autoradiographie et empreinte B aumann d'une même
coupe longitudinale de brame (d'après [4])
IrAl* : *
\*V l ¡y ■£
:.>
Casting direction
sens de coulée
I II1 I
Narrow f a c e
petite face
«/
ι"· I
ι 1 I r ' ι ι ' ï γ ν b/
j·:,'./ Í ». ; . . » I ι.' Μ ;ϊ
1
κ
ΊΗ
Figure 7 : Schéma décrivant les divers types de criques
dues au gonflement
123
' · η *¿JL
ι— —'
Ch" o o . ^ î* * *
60
50
o'f
8 o o °' <m ? ¿ 40 _"_! (si? /.ej
ι ·°· 8>
30
y
Ό
/
/
ë 0.6 ?n
S. i0 > " 1
CO Q4
0 m rfm · / ·
5 10 15 20 25
0.4 ι 0.5 0.6 0.7 0.8 0.9 [C] « 10' or 'S] < 10' (%)
Casting speed (m/min.)
No. o f i n t e r n a l c r a c k s
nombre de criques internes Internai cracks
in inner radius
criques internes
'en intrados
criques internes
en extrados
I n t e r n a i cracks
p o n f i f i t i f t c i f ^ f i ^ f i n q *
17 1? 21 23 in outer radius
R o l l n o . / N° du rouleau
Figure 11 : Initiation des criques internes le long d'une machine de coulée
continue (Répartition dans l'épaisseur de brame)
(d'après [5])
I n d e x of i n t e r n a l . c r a c k s
indice de criques internes
• 5
0,« 60 — . ie
a'' 40
_¿f?^ !
-/ y
20
:
0.5
.*«'" !
0>^
0.5 1.0 .5 2.0 2.5 3.0 3.5 4.0
//"(min*í Unbending point7
/ 0 0.3 I 2 3 4 5 6 7 8 910 12 14 16
Distance from meniscus (m)
'/
-.y·
l.OUO 1.500 2.000
Figure 13 : Relation entre le temps de coulée
et l'épaisseur solidifiée S ou la distance I e
largeur brame (mm) entre la surface de brame et les criques
Figure 12 : Influence de la largeur de la internes observées par autoradiographie
brame sur les criques internes (d'après [4])
(d'après [5])
A = Usual casting
Β = Casting under pressure
x a x i s = Slab width
125
Casting speed
Vitesse de coulée
(m/min)
Casting time
ü 10 .»U 30 W Sot 60 TU »J to ίου temps de coulée (min)
Stoppage arrêt
Figure 14a : Diagramme de vitesse pendant un essai
d'arrêt de brame
Χ
ι|ι ΙΙ.ΤΤΤ 4' · Γ "¡¡Til
M.
Μ Ίι
I II τ—Γ
Ό O O 5Γ Q~ S Ö ® σ
= 36 =37 ? 38 = 39 \ * *° = 41 = 42 =43 =44
This mark indicates a driven roll.
t ^ j , | ι. ,1 11 If .y,...
V r ^ι- ^-^- r - y - · i|l|l|i|i[|i||iy""-i"'i|i||i"·' -min ■!■■"" *.*^IÚÍ
II ,,|
"Ο σ Ό O ö &
I5O0
1450
1400
1350
1300
%C
1) Hydraulic ram
2) Liquid steel A criques Cracks
3) Shell Δ pas de criques No c r a c k s
4) Steel mould 2
o acier 40 Kg/mm steel
5) Internal cracks
2
0 acier 50 Kg/mm steel
Ingot
lingot
2.0 0
1) / 250*300. I
V r n A
h Jt ι· ' *l70O ^ t e / S45C °
.hydraulique CN
C
1 5 V\
O \cV * o
■H « \ o
u Ε ι.ο
5) μ O °oS \
" ^ o
XI <9
0.5
o
4)
uj
. t . t . 1
ε vitesse de déformation (s l i
R a t e of s t r a i n
1.5
• With cracks
o Without cracks
Έ 10
C c
i V 0.5
1 2 3 4
Square root of solidification time (min"3)
•
without bulge
o
(Before modification) < with bulge
CJ
Dangerous zone
1.6
for inner crack ( S S 4 1 ) )
1.4 1
/200mm t h i c k n e s s \ "■3
\ hard c o o l i n g / "
a.
8
ï 1-2
■
2
10 c
■ » « s r ··# 4.*J ■ ·
o o
í * Ό
H 16 18 20X10-* (%) 1 » ^i> Sy,
Carbon content
06 °of
80 ΙΟΟ 120 140
Distance trom slab surface (mm)
o550*P R
»280.» Ρ R
3.0 t-
2.0
1.0
/
/
/ · · — · <
/
il
H ca
«
: : ·m ··
/
O) A /
■H c? ? + Í-
s. « /
tri rj
&.| / ·· · /
Q) ^ y
·· · /
••—XX-
· /
/
Ì3 1
Gonflement (mm)
Bulging
1. B rame
2. Rouleau d'appui
3. B ras de mesure
4. Capteur de déplacement
5. Plaque refroidie par l'eau
Nuance d'acier X 60
Rayon de courbure de l'installation : 3,9 m
Largeur des brames supérieure à 1300 mm
Vitesse d'extraction : 0,85 m/min
Epaisseur de manchon solidifié : 79 mm
distance au ménisque
lu.um
K.5m
17 " Τ Ι !
Ξ Rollenobstond 860 mm
Ξ :o • mit Zusatzkúhlunq Espacement des rouleaux 860 mm
o oh ie Zusotzkúhlun g Strongurerte Stahlsorte
c I o , > 1300 mm Χ 60 • Avec refroidissement supplémentaire
Ol Ι „
^ c
I J o Sans refroidissement supplémentaire
H b ι
w
O <ι
Ι · .« !* o _ ^ i ; t ; n r nm St 1 7 Courbe I : Acier X 60 brames de plus de
<t
Is.—f 1300 mm de largeur
!0 I
■J ι I
OL.
ι Courbe 2 : Acier st 37 brames de 550 mm
550 900 950 1000 1050 1100 1150 1200 1250 de lareetr
StrongoDertlöchentemDerctur in°C
Key to Fig. 22 to 24
Key to Fig. 25 to 27
1 1 1
Stahlsorte St 37
Rollenabstand 860 mm Nuance d'acier St 37
Strangbreite 550 mm
E Abstand vom Gieflspiegel 10 m
Ecartement des rouleaux : 860 mm
E
c Largeur des brames : 550 mm
<υ
■w
c Distance ménisque : 10 m
υ ·**
•
c
o
O
54 56 58 60 62 64 66 68 70
Schalendicke in mm
Epaisseur de peau solidifiée, en mm
Stahlsorte St 37
Rollenabstand 860 mm
Nuance d'acier : St 37 Strangbreite 550 mm
Abstand vom GíeDspiegel 10 m
Ecartement des rouleaux : 860 mm
Largeur des brames : 550 mm
II'
y— °*
513
Distance au ménisque : 10 m
<υ= 2
OJ
c
o
O 0
I 1.4 1.5 1.6 1.7, 1.8 1,9 2.0 2.1
Gieilgescnwindigkeit in m/min
1 1 1 1 1—
Stahlsorte Χ 60
Anlagenradius 3.9 m
Rollenobstand L 860 mm
Nuance d'acier x 60
Strang bre ite ^ 1300 mm Rayon de courbure de l'installation : 3,9 m
Gieilgescnwindigkeit 0.85m/min
Strangschalendicke 79 mm
Espacement des rouleaux : 860 mm
Oberflachentemperatur~ 1030 °C Largeur des brames supérieure à 1300 mm
Vitesse d'extraction : 0,85 m/min
Epaisseur du manchon solidifié : 79 mm
c
Température de surface de la brame A/ 1030°C
V
Ξ
V
c
o
Ό
3.1 3.2 3.3 0.4 0.5 3.6 0.7 3.8 0.9
Loge zwischen zwei Soilen x / l
Position x/L entre deux rouleaux
Figure 27 : Déformée de gonflement mesurée en deux
rouleaux d'appui (d'après [29])
133
Pasty zone
zone pâteuse
Solid solide J Ι ' M * Liquid
solide Solid
Roll rouleau
domaine A D o r n a i n A
domaine Β D o m a i n Β
Roll rouleau
Axe de la brame
Slab axis
Figure 28 : Représentation schématique de la zone de gonflement
pour la modélisation de la ségrégation axiale (d'après [34])
Figure 29 a figure 29 b
■ 2 mm
\ . ' » 1ι ι ι ι ι
\ ■ ' » 1 l . 7,87 m
* » 1 M 11 * * 1 r —
x .S
H" — TTTTTT"'
•
• / » ' ' |— ■ * · * \ \ \ f \
2 1 10 domaine A
< ,. , * χ μ ' ' ' » \ \ "^c^ Domain A
ai • > / / / L*
._» « ♦ ♦ t t
• · »· \\\ j \
•S I * . ν \ V \ » ...,,, ¡
\ ' ' ' ' / / t
^ •KK'»^'^ t
5 a 5» » x v ·»* * . t
• ' ' ' // / \ Domain
domaine B
B
•H *a) test » » % % ^ t fest \
o
to „ 40
• ' ' ' ι / / ι
DJ
« (D
c
\ 'ΓΤΤΤ"·,.^ 1
\ ' · « » ι ι ι ?r 0.02 cm s'
5« \ 1 fl '<Λ
CL 9 10 8 9
Abstand von der Srammenoberflãche x in cm
Distance x
Distance x à la surface de la brame en Cm"*" from slab sorface in cm
Çs · densité de la phase solide Density of s o l i d phase
¿I : densité de la phase liquide D e n s i t y o f l i q u i d Dhase
¿max : gonflement maximal Maximum b u l g i n g
ZB : distance des rouleaux au ménisque D i s t a n c e o f r o l l s from m e n i s c u s
Slab a x i s
axe de la axe de la.
bra/ne brame
Manganése |
mm
2 \ : gonflement maximal
"max
3
Maximum bulging
■^1
9 io a 9 10
Abstand von der Brommenoberflõcrie χ in cm » D i s t a n c e f r o m s l a b surface X i n cm
Distance au c e n t r e de la brame en cm
Ol Elastic
élasticité élasto plasticité Elastoplastic
•»E
A 1
Γ C
i
\
— — — ■
' B Ρ
ι ι ι ι ■ ι ι l M I M I t I !
Ferrosunc Pressure a I Eletnenl:4Nodes Plsir. Slrain
P / 2 199.5 B r Total 2X Elements
] Ι Toi·! ?3! Nodes
0 0.01 0.02 0.03 0.CM 0.05 0.06 0.07 0.C
F e r r o s u u r Pressure I t g / W
/1JMC
10
■
* IJ00C
— / s i
/ 1
1 ii»«.
L^^ ^
2
IIO0C
Figure 33 : Définition des différents domaines 10O0C
de fluage
Ό IO a Kl SO 60 '0 w »ιοα
I : fluage primaire ou transitoire t (>«c)
II : fluage secondaire ou stationnaire
III : fluage tertiaire Figure 34 : Essais de fluage (influence
de la température) acier calmé AlSi
T = 1 kg/mm 2 (d'après [19])
C Mn Si Ρ S Al
0,15 0,68 0,25 0,012 0,014 0,023
40
measured by
Wünnenberg
—calculated
u>T42cms'
ho ui=280kPa
z d =10m io "I I 1 1 1 T-
1 = 8 6 cm measured by Wünnenberg·
u=142cms~'
ζ
O I8 ZcplOm üj=280kPa
υ 2C
UJ
UI
Q
Σ
D
Ι io
χ
<
Ί. 04 06
Z'( = z/I)
σι
4 6x10-
- 138 -
y
II
J
Gl
H—h b.1
OJ
Thickness measurement a t
e x i t from t h e m a c h i n e
mesure d'épaisseur de brame
en sortie de machine
EXTRADOS
Outer r a d i u s
mesures de
ΘΘΘΘΘΟ©
ι» itu, 10i ι·6 Ho I ια
gonflement
Bulging COKttT If.
measurement
- 140
USINOR Dunkerque
Roll nos.
Roll 0
Distance between: Bearers (? Traverses)
Inner radius rolls
Inner radius fulcrums
Outer radius rolls
Outer radius fulcrums
Name of the zone
141 -
(29)
M e a s u r e m e n t s b y Wünnenberg
I Mesures de Wünnenberg [29] 1
h
= 860 mm
= 79 mm
conditions aux limites n° 1 V = 0,85 m/min
Boundary conditions no.l
T = 1030°C
conditions aux limites n° 2
Boundary conditions no.2 ρ = 2,8 bars
conditions aux limites n° 3
Boundary conditions no.3
•
•*
••
.· ••
••
τ
••
•
•
••
•
ε
-2
''•y
' ¿I
E
c Ni \ •
•
•
I \ v-
o*-
■rH C ••
4J 11 •
υE
OJ OJ
* ·/«
OJ o
y
I
Ό ao
o«
•H oj
tn.c
Η υ
3-aj l·-
«IT
-10
-M,
20 40 60 Θ0 100
>>
>X
S
-2 ~ \ y
\
il
II
1j
C
\
1
\
\
\ ti
c
o
OJ \ \ 1
•H
+-·
c \ \ / /
+J
O
OJ \ \ / /
E
OJ
\ N / /
I
rH
ω co / /
Ό Só \ / /
Cn OJ \ / /
J
L \ / /
3 -OJ
\ \ /
CQ —
' /
\ \ /
y
-10
-14
20 40 60 80 100
Position y entre deux rouleaux (%)
P o s i t i o n x / 1 between2> arioils (%)
143
M e a s u r e m e n t s by Wünnenberg
I Mesures de Wünnenberg
.acier 18M 5 s t e e l 1 8M 5
1 = 860 mm
1 = 79 mm
V = 0,85 m/min
T= 1030°C
, _ a c i e r 18M 5Nb S t e e l 18M 5 Nb ρ = 2,8 bars
/ y"
/ / \ \
li \
c 11 \
It \\
h
■H
C
h \\
o
ω
•H
Ο oj
+J
s h
·-* c
Ή o
OJ OC
Ό
OJ
Di'0
C OJ
'ri.-c
tn υ
r-l-4)
m u. -1
ρ; '
il
\\ II
\\ il
\ \ il
il
\ 11
V s • /
\ V -*
20 40 60 80 100
1)
2)
I Mesures de Wünnenberg
/ /
•• / /
•f
-2 i
: i 11
S '
E
E
w
c c -4 % / /
■ri,
OJ
G
O «--
■H c
+J
U
OJ : /
E
m
01
_0J
4-1
C -6
·. \
( ï .· /
••
\ 1
OJ
o
Ό
00 •• • /
OJ *
• /
Di
C T: ••
•H • /
Oi \ •
rH υ
Ρ -OJ
-8
• ι
\
*. \
•
10
...···
···
12
20 40 60 80 100
Position r entre deux rouleaux (mm)
h= 89 mm
V = 0,85 m/min
T= 1030°C
Ρ= 2,8 bars
012.7
_J
Ι
(
curve 6
υ
ι
c . courbe 5 fig 36 [47]
•H 3 - curve 5
c
o
•H O V modèle IRSID
P viscoplastique
l
υ
0) IRSID v i s c o
p l a s t i c model
S S»
OJ
D> Ό
Ç OJ
'il-G
D> (J
Ή -OJ
mE / /
1 / /
1/
Α
/
y
V
oL 20 ¿0
ÍÍ'
¿Mi
60 30 100 120
0 140
Steel
«___ acier 18M5 (14)
S
a a e r 18M5Nb(15) en
T = 1000°C
I E
π
c
1** X
I* "*» x
■Η
0J
C ,-.
o
δι / ι \ \ χ \
s *■ \
■Η
ί v
Il
-Ρ
υ ο
11
rH c \ χ
HH o s χ
UJ
TI
DO
Ν χ
m
11
Ό ^***-- ƒ ' ν χ
c *— —J / Ν Χ
•H
0J
r / Ν Χ
ΓΡ o Ν ^
r H -OJ Ν.
3 / / Ν.
C3 Lu Ν.
y—->. ƒ 1
10 15 20
Longueur métallurgique en m
"'"tallurgical length in m
147
. PÍo^chiudiT * losjc)
. Profil froid,(Τ décir
C o l d p r o f i l e <680°<
(Τ s t r a i g h t e n i n g )
ai OJ
680 °C
Η zi
ω o
Ό ω
01
-π
■Η υ
tn .c
)
¿0,5
10 15
Longueur métallurgique en m
Metallurgical length in m
Figure 49 : Influence du type d'acier sur le profil de gonflement
entre deux rouleaux (cas statique)
Steel
__ acier 18M5 (14)
Steel
_. acier 18M5 Nb (15)
.. Mesures de Wünnenberg [29]
Wünnenberg's measurements
20
iE
•H C .fc.
co
ss
υ t
(DÛ
Η Ή
αϊ o
DTJ
c
•Η ^
-*-
r^OlO
3'JH * Ν»
COU-
"ν
Ni
y
y
y ·Ν«ν *< \
/ y
r..'f
^¿¿* ^ ^
capteur
capteur
transversal longitudinal
M
O
ω
d
QJ
co
ω
to
u
Q)
>
to
O
10 SH r H
C to 0 ω
Q) αϊ CO tn
co -P C to
ro
rH
0) Q)
UI
>
ft <JJ
0)
rH
Ï7
rd
C
•H
C υ to
Íá 0
•H
•rH U
Ρ ω ÍH
Di Q) >
f) ■Ρ to r H OJ
C IS Fi ra rrj r H P
O O d) ■Ρ rH o rd
PI A « co M S
ft Rgure 50 : Schéma de principe du système de soutien de la plasticine
Ή CM m ·<* m trj.r~ et du réglage des rouleaux de la maquette
y/ 72 y
Travel of vessel
défilement du bac
100 -
Stress
contrainteö (kg/cm 2 )
1
30°C, E = 78 kg/cm 2 Un
1
40°C, E = 36. kg/cm 2
50°C, E = 25 kg/cm 2
Profil initial O r i g i n a l p r o f i l
de la plaque o f t h e plate
t = 0
flèche instantanée
Instantaneous deflection
t = 1'
υπ
ro
t r 2'
0,1 mm
Figure 56 : Exemple d'enregistrement lors des essais cinématiques
(Influence de la hauteur d'eau dans le bac)
Plate thickness épaisseur de plaque : 10 mm
R a t e of t r a v e l vitesse de défilement = 0,6 m/mn
V ( d i r e c t i o n of t r a v e l of t h e v e s s e l ) H e i g h t of w a t e r hauteur d'eau = 80 mm
V (sens de defilement du bac)
h = 20 mm
h = 40 mm
(_n
h = 60 mm
h = 80 mm
h = 100 mm
j 1
1200
1Ω00
800
b - 6,5
600
^
400
y
300 y'
v'
1
b* - 6,3
200
y
150
y y m
y
y y
100
S i ! .0
10) 10
V 1 m/min Ecartement des rouleaux (mm)
10,5 cm R o l l s p a c i n g i n mm
P
h 13,5 mm
T 20°C
- 155 -
1000
500
a function of
speed
Height of water ρ
V (m /mn)
156
Viscoplastic flèche
deflection I viscoplastique
0,04
& Ν
0,02
\
\
0,01
10 15 4
~20
épaisseur de plaque (mm)
Plate thickness
Figure 60 : I nfluence de la temperature de surface externe de plaque Ts
sur le logarithme de la flèche viscoplastique mesurée
* : f2 vp
* ! 1/2 [fivp + f2vp]
αϊ
Τ : °K
ρ : mm
1 : mm
h : mm
f = 2.67.10"\exp (0,0617.T) a l i a l i with t : s
V
P hd avec a = 2,1
b = 6,07
c = 0,23
d = 2,38
In
fvp t
158
Figure 61 : Comparaison entre les flèches totales mesurées et les flèches calculées
par modèle statique dans les mêmes conditions (cas statique)
Bulging d e f l e c t i o n
Flèche de gonflement (mm)
0.7
Mesures M e a s u r e m e n t s
l.h = 10 mm P = 110 mm
2.h = 10 mm P = 70 mm
3.h = 10 mm P = 40 mm
4.h = 25 mm P = 125 mm
1 = 1 1 1 mm T = 20°C
0,1
4 Creep time
temps de fluage
t (s)
900 1830 2700 3600
159 -
l'
Figure 62 : Coefficients de régression de la formule f = K
1.0 1.2
.^ A Log h (mm)
1 0.9 _^ χ Log p (mm)
—I
flèche de gonflement
viscoplastique mesurée (mm)
Measured viscoplastic
bulging deflection (mm)
Casting d i r e c t i o n
direction de coulée
CTS
O
No load p o s i t i o n of r o l l s
Inner r a d i u s
Intrados
Extrados
Outer r a d i u s
Figure 65 : Exemple d'enregistrement brut, de la vitesse d ' e x t r a c t i o n ,
des positions de rouleaux et de la brame
(Acier pour tôles fortes avec Nb - épaisseur brame : 250 mm)
P o s i t i o n of r o l l s and s l a b
Position rouleaux et brame (mm)
Ό c
0) o
<D -
ft υ
to ra
rd
OJ
rî 1.
rd Χ}
y en 0_ σι
rC M
+J
•H .. .8
S >
.6
.4
.2
Γ •nr
ΙΟ 15
Temps écoulé à partir du passage de la brame
au droit du capteur de brame (5)
Time elapsed from passage of the slab to
the vertical from the slab sensor (s)
1/4 turn of Position of the outer sur
rol 1 face of the slab relative to
h sum of the positions of a fixed position
1/4 tour de
rolls 76 & 78 rouleau position de la surface externe de la
4 brame par rapport a une position fixe
+2
c
T3E Fleche de B ulging
gonflement d e f l e c t i o n
to OJ
rdCJ
1.5
ω !=,
ε LO cn
OJ
C cc
oc
'1 J.0
■p o
•H
to
ri ^
. .5
1 .
,_0
lïnrs ( t) T i m e
» «. · » ■ ■ · · ■ . . , , „ ¿r^ir^i1 ί.'j^ 1 ¡¿.-χ.·;¿·ίί·^.·a^a,·-«ya.- ;ί , • m m IM t u χ. X4 ut
m»·.; (*)Time
· ■ < · «a < · aa i a η a i ra·'m¡TíuV ÜT '
Figue 67a : Profil du signal de flèche de gonflement filtré a 0,028 Hz Figure 67c : Profil du signal de flèche de gonflement fUtré à 0,028 Hz en
Spectrum
Spectrum
SPECTRE
S t e e l f o r heavy p l a t e Extrasoft steel
Aciers pour tôles fortes SPECTRE
v
_ Λ = 0rS6 m/mn Aciers extradoux
V = 1,0 m/mn
s,
ι Hj(«iaoo) ¿3^ \ *
■ I II l.lli.ri H¿(*4ooo)
i'l' ■ · ■ I 'I ■ ■ I · ■ M. ι h . . . , , | i | . M I . I ||,,,..|. , ren^çEs,,,,
ι I idi ι Ιι·' 'I 1"' m u l l í . . . i . n . u i l i l . i l l , i.illi ι ,,,, FREQUENCES
Frequencies » _,
ngçreJTb : Spectre de fréc.jenœ di signa, de flèche de gentanent non filtré F ^ e 67d : Spectre de fréquence du sigral de fieche de gonflement non f i l t r é " ^ " " "
Figure 67
Fia. 67 a/c: Profile of bulging deflection signal filtered at 0.028 Hz
Fig. 67 b/d: Unfiltered frequency spectrum of bulging deflection signal
- 165 -
Key to Fig. 68
1 Roll deflection
6 R 78 (Idler roll)
7 R 76 (Driven roll)
8 Withdrawal stopped
1
Flèche rouleau (nm)
Décharge du rouleau et
suppression des contraintes Position des capteurs du
themniques après le passage signal délivré en régime permanent de rouleau avant l'arrivée
Refroidissement et contraction du rouleau de la brame coulée (rouleaux en charge) de la brame
en
arrêt d'extraction
en
Bulging d e f l e c t i o n
Fleche de gonflement (mm)
ι i
■ ■
■■ ^y
20 .
■ ^y
ιI
m •
m
1.5 Tiickness
• Epaisseur
• 210 mm •
•
• 250 rrm ■
• 300 mm 4
•
ΙΛ S* •
m •
• «
OJ ,
B u l g i n g d e f l e c t i o n (mm)
Flèche de gonflement (mm)
i ι ιι r
2.0 ■
■^••^
■
• y*^ ι ' 1"
• ··
•
^*y**^ *1
1.5
y*· m
m
• ^ * * •·
1.0 t
• • Thicknesi
• Epaisseur
■ 250 mm
• 210rrm
0.5
f ( i n mm) f(tnmm)
Ρ (in m) Pt (*"ƒ*)
1 ι^
i-3
Al
At.
AS
A}>
H
o.
-1
temps (s)
Time
250 mm
O
1.5 ·
•
•
•
• *
o
% •
1.0
oo
o J
o
Steels for sheet
O Aciers pour tôles fines
Aciers pour tôles fortes
1 avec Nb
O. )
S t e e l s w i t h Nb," i u i heavy
plate
0
I
1.0 1.5 2.0 2.5 3.0
Faux rond du rouleau 76 ou 78
(qui a l'amplitude maximum) (mm)
Influence du plus grand fauxrond des rouleaux 76 ou 78 Eccentricity of roll 76 or 78
Figure 73
sur l'amplitude de la flèche de gonflement (whichever has maximum amplitude
(mm)
171
1.5 Λ Ál
Withdrawal
speed
liai i f f I iëimMàhi r\ UM
Vitesse
1.0 ■ -V—? ii
d'extraction (m/]rnr
Jl
1.2
0.5
1.0
0.8
0.6
0.4
S t e e l w i t h Nb, f o r h e a v y p l a t e
Acier pour tôles fortes avec Nb V = 0,86 m/mn
Coulée n° 2
Bulging d e f l e c t i o n Cast no.2
Flèche de gonflement (mm)
2.0.
1
! \
1.5 ■■ I 1
ï 1 Λ I 11
J !
.Λ ¡Λ Ί
1 ί ί1
\f Ν
1.0
/
»ι II η f ï ι ij I j ij ; !j M fl il (j ;
0.5 ■·
4000 ol
4500 E
5000 5500 6000
2.0
214-
213
Figure 75 : Evolution de la flèche de gonflement et de l'épaisseur (sortie machine) pendant une coulée
à régime permanent V = 0,86 m/mn (acier pour tôles fortes avec Nb)
décharge du rouleau 78
Profondeui du puits
(m)
- 15
14
13.75
12
^J
- 11
Flèche des
3
rouleaux et de la
brame (mm)
A000
Temps (s)
10
Figure 76 : Détection du passage du puits liquide au niveau du rouleau 78
désaligné de + 4 mm vers l'extrados
174 -
Key to Fig.76 :
6 Roll 76 (fx)
10 Time (s)
- 175
Appearance of segregation
----.---.--*-„- 5
10
at i t a t e d
^J
σ
4 --
1.0 1.1
Vitesse d'extraction (m/mn)
Figure 79 : Influence du brassage électromagnétique sur la flèche de gonflement w i t h d r a w a l s p e e d (m/min)
178
Slab thickness
Epaisseur de brame
210 mm
Slab thickness
Epaisseur de brame
250 mm
60 60
40 40
20 20
►o V = 1,6 m/min
Bulging s t r a i n Roll spacing
déforrnaticn de ■· V = 13 m/min Ecarterrent
gonflement £ Q 6
(9Ö ^¿8S
10 15 , 20
Distance t o meniscus (m) distance au ménisque en m
1,6 m/min
] fleche de gonflernent
1,3 m/min / B u l g i n g d e f l e c t i o n
Deflections Ecartement R o l l
fleches (rrm)
s rouleaux s p a c i n g
'°0 (mm)
15 . 20
distance au ménisque en m
Distance to meniscus (m)
- 180
/
v
3,5 - ^ ^ \
!
\ \ V = 1,0 m/min
V = 1,3 m/min )
10
I
1/2 épaisseur de
—-"" brame
h thickness
of s l a b
;'
15
V = 1,6 m/mi\ \ N
S t r a i g h t e n Lng \ \
20
L _
50 100 110 150
épaisseur solide en mm
S o l i d t h i c k n e s s i n mm
181
§ ε
~ε
CO £
*(DH <Ûi
Φ g
S*
'0 iß
H Jg
H
ε
o m
«a
Distance au ménisque en m
Distance to meniscus (m)
182
V = 1,6 m/min
T = 1000°C
1,5 mm
0,5%
1 mm
0,3%
0,5 mm
£ = 0,03 %
15 20
Distance au ménisque en m
Distance to meniscus (m)
F^jre 85 : Profil de flèche de gonflement le long d'une machine de coulée
continue (R = 12,2 m) induisant une déformation de gonflement (£) constante
V = 1,6 m/min
c t = 1000°C = cíe
■H
Ό E
ε
O b
υ ^
tu c
ι«
0)
■α
1 4
C' DO
c CO
■Η
Oi
.Η
U_
e =i %
^ ^ _ _ _ £ = 0,5 %
b = 0,3 %
£ = 0,03 %
10 15 20
Distance au ménisque en m
Distance to meniscus in m
184 -
600
Distance au ménisque en m
Distance to meniscus in m
185
T = ÎOOCPC
Bulging deflection(δ) in mm τ = 900°C
or s t r a i n (ε) i n % 0 flèche de gonflement = B u l g i n g d e f l e c t i o n
Flèche de gonglement (ò ) en mm déformation de gonflement £ = B u l g i n g s t r a i n
#
α ι iHéformaticri ( £ ) en %
1.0
0,8
0.6
0.4
0.2
15 20
Distance au ménisque en m
Distance t o meniscus i n m
186
L = 2150 mm
o
1 0 = 1600 mm
V = 1,6 m/min
T = 1000°C
machine courbe de rayon 12,2 m
Bow r a d i u s 1 2 . 2 m
c
■H
H
ω
-p
0)
■r-i
O
Κ
Distance au ménisque en m
Distance to meniscus i n m
187
x χ T = 900°C, E = 0 , 5 %
f
o r = 1 mm, L 0 = 2800 mm, 1 0 = 1600 mm
T = 1000°C, £= 0,5 %
500
Ξ
Ξ
Γ.
Ξ υ
h χ
Ί
C Λ
■H D
u η
0 1-
4J 71
0 Î1
- "Ο
■H υ
Ό 1
—*
5)
H
ε
0
2
Κ Q
10 15 20
Distance au ménisque en m
Distance t o meniscus in m
Figure 90 : Composantes de la déformation totale au front de solidification \£ T )
(évolution le long de la machine n° 1)
Total s t r a i n
, # déformation totale By . —— · déformation de gonflement ί = Bulging s t r a i n
0,2
Distance au ménisque en m
Distance t o meniscus i n m
Figure 91 : I nfluence des déformations distinctes du gonflement dues à la géométrie de la machine
sur le profil optimal de diamètres de rouleaux
500
400
co
a t four points
a t two p o i n t s
- 190 -
LIST OF APPENDICES
APPENDIX 1
where
σ = the stress
ε = the strain
έ = the rate of strain
Τ = temperature.
193 -
APPENDIX 2
In the case of the static model with a beam at uniform temperature, the
following assumptions have been made:
4) o(y) + o(-y) = 0
5) when t = 0, C(χ) = O yχ
6) Π t u(0) = u/(l/2) = 0
In this case, the neutral axis is confounded with the mean axis, there
fore y = 0
o
ε = -Cy (Al)
1/2
C(x) dx = 0 (A3)
/o
which indicates that the mean axis of the slab is tangent horizontally
to the upper generatrix of the second roll (assumption 6)
- 194 -
(A4)
, η -m n+m+1 , , _.
AC C y dy (A5)
By putting
r = m/(m + η) (A6)
s = l/(m + η) (A7)
Cncm = (m+n+l)M(x)/(2A(h/2)m+n+1)
When C(χ) and Μ(χ) are any signs, one writes the relationship:
From relation (AIO) the slab deflection f(χ) can be calculated by two
successive integrations :
In the case of the static model in which the beam has a heat gradient
in its thickness, the following assumptions are used:
σ = σ (Τ = 1300°C)((1500 - T)/200)
ε = -C(y - y o ) (A13)
If t is the time during which the strain acts, the law of behaviour
of the steel, σ = λε έ may b e written:
with
or
σ = sign Cc(y - yQ) 1 L(t) |c(y - yQ) Γ + η (A18)
One then looks for the relation between moment and curvature after
time t. For this, for several values of the curvature C, one suc
cessively resolves the following equations:
+h/2
σ dy = 0 (A19)
/-h/2
with
Γ n
σ = ση (Τ) = sign Cc(y - y )]L(t)|c(y - y ) Γ for T < 1300 °
1 o o
One sees that with the couples (M(x), C(x)) obtained, the relationship
moment-curvature may be put, as in the isothermal case, into the form:
C(x) dx = 0 (A3)
(All)
APPENDIX 3
f (o) = 0 (Bl)
df(o)/dx = 0 (B2)
ε = -C (y - yo) (B4)
expressed by:
calculated.
199 -
(B6)
(B9)
with
/ A ( T ) | y | n + m + 1 dy (Bli:
200
n 1 1_
m m
|cl dc/dx = (l/(V x I ) ) | M ( X ) Γ M(X) (B12)
ii c I I
S- + 1 2- + 1 ì /Χ2
Cc(x2)]m - Cc(x1)]m = - 1/(V Icm) / M(x)|M(x)|m dx (B14)
In the general case, to resolve equation (B12) the three parameters C(ο),
M , and F (respectively the curvature in χ = o, the restraining moment,
and the separating force) must be known. They are determined so as to
verify three boundary conditions of the problem that we must impose in such
a way as to best describe the actual phenomenon.
Key to APPENDIX 4
Key to boxes:
6) Readout of f'(o)
A
7) Readout of C (o)
11) Calculation of f(χ) by double integration of C(x) with Mo, C(o), f'(o)
202
Introduction données
1) conditions initiales, bornes de dichotomie
3) Calcul de la relation
moment courbure
L
calcul de
9) P., (o) - f ' , (l) pour C (o)
f' 2 (o) - f ' 2 (l) pour C (o)
C A (o) =cA(o)+CB(o)
C(o) = C A ( o )
calcul de
f-| (o) - ι - , ί Ι ) pour f'A(o) fAfn^f'A(0)-rf'B(0)
10)
f 2 ( o ) - f 2 ( l ) pour f' B (o)
[f ( o ) - f ( I ) ] [f ( o ) - f ( l ) ] < 0 fn ( o ) - f 1 ( l ) < É3
1 1 2 2
No Yes
f'(o) =f'A(o)
&
End
- 203 -
APPENDIX 5
The introduction, into the equations defining the dynamic viscoplastic model,
of dimensionless reduced variables allows, by selection of symmetrical
boundary conditions over two consecutive rolls, a formula to be obtained
that gives the maximum bulging deflection as a function of roll spacing,
ferrostatic pressure, and speed of passage of the slab. We detail here the
determination of the regression coefficients of equation (33) with ρ = pH,
where ρ = the ferrostatic pressure and Η = the metallurgical height.
To do this, we start from the general expression for the curvature C(χ) (see
Appendix 3, formula (B14)).
1 χ 1 m
m ι im
C(χ) = sgnj I sgn M|M| dx sgn MΙMI dx m+n
with ε = -Cy, Oy aligned in the direction from the external face to the
internal face of the shell of the slab.
.1 1
sgn M|M|m = l(pl2/2)m 0(x')
ïï(l+2/m) i
.m+n m+n
c(x) = ± - £ —
m
m+n 1
m+n
v
cz
On the other hand
b = l/(m + η)
c = m/(m + η)
d = l/(m + η)
e = (m + η + 2)/(m + η)
APPENDIX 6
h = k/(L/v)
f - (K(T)/k 5 - 26 ) { (H 2 - 25 l 6 - 10 )/L 2 - 63 }v 2 · 06
APPENDIX 7
Q ( x \ t') = K Q(x, t)
ri
K
Κα = ο-'/σ Kp = F'/F K p = p'/P Κ γ = γ'/γ = V t ~
207 -
-2
(Ko/K ) difa + Κρ·ρ(Κ «F - Κ κ γ) = 0 (3)
χ F χ t
(Kp κ κ )/κσ = (κ + 2 κ - 2
κρ)/κσ = 1 (4)
F Χ Χ t
which reduces to
κ/ = KX/KF (5)
The two materials being in the same gravity field, one has
Kp = 1 (5a)
Therefore
Kt = /Κχ (6)
n .m
σ = α ε ε for steel
η' m'
α' = σ' ε' έ' for plasticine (8)
Κ ε = ε'/ε Κσ,
1 = σ',/σ,
1 1 Κη = η'/η Κm = m'/m
Κ. = έ'/έ = Κ ·Κ " 1
ε ε t
208
Κ ·η Κ *m Κ ·η Κ *m
η m η m
Κ ·σ = Κ ·σΊ Κ κ. ε έ (9)
α σ1 1 ε ε
In the case where K = 1, one puts :
Κ. = Κ ~
ε t
K = Κ ·Κ ~ m (10)
ε α t
η = η' (11)
m = m' (12)
These ratios depend on the grade of steel that one wishes to simulate
and on the position in the machine. We will make the calculations for
the case of steel 18 M 5 at 1000 °C.
Geometric ratio:
with D in mm.
_0 _. 0.24 .0.21
α = 38.01 ε ε
with
σ in N/mm2
ε in %
ε. in
. % s -ι
- 209
It transforms to:
_, p 0.24 .0.21
α = 027.5 ε ε
with
α in kg/mm2
ε dimensionless
. . -1
ε in s
- Ί„ 0.281 . 0.107
α = 0.12 ε ε
νρ νρ
with
α in kg/mm2
ε dimensionless
νρ
-1
ε m s
νρ
Time scale
1/2
One has seen that Κ = Κ
t χ
One has
Ρ = P1 g h (13)
In addition, with
ρ = density of liquid
h = height of liquid
3
Κσ = 4.4-10
Kp = 1/7 = 0.14
^ = 0.031 Κ χ -°·° 75
Κ = /κ = 0.476
t χ
Κα = Κσ ·Κ " m = 4.9·10~3
κ = Κσ/Κρ = 0. 0350
This height of water can be envisaged, if the sides of the vessel are
given rigid supports to prevent their collapse. But we have not taken
into account the equality to 1 of the two ratios of the relationship (4).
Therefore this result does not characterise a rigorous similarity.
Χ J_ L· J. Χ
whence
Κχ= (Κσ^Κρ)2/^
- 211
K = (κ σ ι /Κρ) 2 / ( 2 + ΐ η ) = 0.0207
Κ = /Κ = 0.143
t χ
Ι^ =(Κσ1·Κχ~Πΐ/2)/Κρ1 = 0.0421
The influence of the intrinsic weight of the plasticine has only been con
sidered when making an analysis of the results of the tests.
-m
Κα = Κα, ·Κ if one assumes Κ = 1 .
I t ε
The test parameters are in fact completely defined if, in addition, the
geometric scale is fixed.
- 213 -
APPENDIX 8
1) Principle
The equation giving the compressive force F has been established by the
slip lines method, giving:
Theory has shown the similarities between dual punching and rolling and
established the equivalence of these two methods of inducing strain.
έ = (2//3) «V «l/h
o
where
V = rate of descent of the punches
o
2) Execution
For each test, the position of the punches and the force, as a function of
time, are recorded by a microcomputer and output to a chart recorder. The
accuracy of the measurements is of the order of 1 kg for the force and of
a micron for the punch positions, but taking into account of other uncer
tainties introduced by the analysis.
The tests were made at constant strain ε, that is, with constant ratio
V /h, the velocity V being defined by the d.c. voltage feeding the d.c
motor that controls the descent of the support of the upper cylinder.
It has been decided to define Young's modulus as the gradient at the origin
of the curve (α,ε) obtained during these tests.
3) Results
o
Temperature C 20 30 40 50
f o r ε = 10 s
E (kg/cm2 ) 78 49 10 9
for έ = I O - 1 s " 1
It is apparent that Young's modulus is the lower, the higher the rate of
strain.
Even taking the experimental conditions used into account, the error in the
determination of these values is still substantial.
For sample A of plasticine, similar compression tests have been carried out
at a temperature of 20 C.
(63)
The following values have been obtained:
E (kg/cm 2 ) 44 60 45 52
APPENDIX 9
σ = λ (Τ) ε η έ m (37)
νρ νρ
We derived a law of viscoplastic behaviour for two lots of plasticine ' '
from relaxation tests that allow the coverage, by the minimum number of
tests, of a substantial range of rates of strain.
a) Principle
ε = ε - σ/Ε (38)
vp
έ = ε - σ/Ε (39)
νρ
217
b) Results
For plasticine B, the tests have been made using strains ε and tempera
tures Τ such that:
The four variables (σ, ε, έ, Τ) were made the subject of a linear regression
calculation, using a computer, to verify the law of the type
,_/m. n .m
σ = A exp(Β/Τ) ε ε
Τ in Kelvin
-1
έ in s
vp
ε is dimensionless
vp
^ 0.281 . 0.107 , 4,
σ = n0.12 ε ε 41
vp vp
218
with
α in kg/mm2 (sic)
ε dimensionless
vp
-1
ε m s
vp
CDNC08963ENC