Note: Descriptions are shown in the official language in which they were submitted.
The present invention relates to a method for
determining the level of a melt in a continuous-casting
moldO More particularly, the invention is concerned with a
measurement procedure for measuring exactly the position of
the melt level in the mold.
In the continuous casting of steel to produce in a
single operation shapes equivalent in section to conventional
semifinished shapes, molten steel is poured from a tundish or
ladle into the top of a cooled mold having a throughgoing
passage. The steel cools and at least its surface hardens
in the mold, so that a cast strand can be pulled continu-
ously by pinch rollers from the bottom of the mold. The
mold is frequently reciprocated continuously vertically in
order to prevent the casting from adhering to it, as such
sticking could cause a brea~out of molten steel in or below
the mold.
One of the parameters which is absolutely critical
is the melt level, or the vertical level of the upper surface
of the liquid steel in the mold. This level must be maintained
above a lower limit so that the casting is sufficiently strong
as it exits from the mold to not break out, and below an upper
limit which would cause the mold to overflow or the casting
to be so rigid as it exits the mold that, for example, it
could not be bent, as the strand invariably exits down from
the rnold and usually must be bent to a horizontal position for
~urther treatment. The range hetween these two limits is
quite small. The melt level can be varied either by with-
drawing the casting at a greater speed to lower it or more
510wly to lift it, or by reducing the fill rate to lower it or
increasing the fill rate to lift it.
-- 1 --
In French Patent No~ 2,251,811, measurements are
made of the impedance of a coil, in the moving alternating
current field of which a conductor, e.g~ a melt, is moved.
This coil is juxtaposed with the top of the melt, which
obviously is of conductive material while whatever above it
is not, and can therefore roughly detect the melt level.
Two U-shaped inductor cores are ernployed which are parallel
to each other but whose coils are oppositely connected. Such
inductors must be aligned with an opening or nonmagnetic
window in a mold at the melt level since a standard conductive
mold would effectively shield the material it holds and make
this type of measuring device useless. It is unfortunately
impossible to provide such a window in a steel-casting mold
which is normally made of thick water cooled copper in which
eddy currents would form that would create a powerful secondary
field completely masking the secondary field created by the
eddy currents forrned in the melt inside it. Arranging the
system to hang down inside the mold has not worked out, as
the level meter gets in the way and is quickly destroyed by
the heat and corrosive chemicals generated by the processA
U~S. Patent No. 4,279,149 describes a system which
eliminates the influence of the mold~ According to this
patent, an alternating current field-forming primary coil
and two oppositely connected similar secondary coils as well
as the liquid metal form a system in which the position of the
melt level relative to the coils produces an induced voltage.
The primary and secorldary coils engage without contact
coaxially around the mold without touching it. The position
of the melt level determines the voltages induced in the
secondary coils as well as the electrical conductivity of the
melt. The detected voltage must be corrected for the
particular conductivity to determine the melt level. Such
an arrangement can only be used for small tubular molds due
to the necessity of providing annular coils surrounding the
mold and the difficulty of providing such arrangements on a
large scale.
European Patent Application No. 82O630~020 published
on Septer~er 22, 19~32 under Serial No. 0060800 describes a sys-
tem used with a vertlcally reciprocating continuous-casting
mold. A nonhomogeneous steady magnetic field is formed ex-
tending horizontally through the mold into the melt generally
at the melt level. This field is vertically reciprocated
jointly and synchronously with the mold and its field strength
is detected at a sensing location after the field passes
through the mold. This location is also vertically recipro-
cated jointly and synchronously with the mold. The melt level
is derived from the detected field strength, normally taking
into account the melt conductivity and relative displacement
rate of the mold and melt. Obviously, the readings taken will
vary regularly with the reciprocation of the mold, but such
variation can easily be compensated out by standard electronlc
techniq~tes.
Such use of a steady field, that is, a ~ield
produced by a permanent magnet or one produced by an electro-
magnet energized with a direct current or alternating current
of at most 10 Hz cmd preferably no more than 5 Hz, auto-
matically eliminates the eEfect of the mold. Since the field
cloes not move appreciably relative to the mold, it generates
virtually no eddy currents in it that would generate secondary
fields. Thus, the field simply passes through the normally
nonrnagnetic copper mold. The field will be affected by the
-- 3 --
13
ferrous melt however and this effect can be measured and from
it can be derived the melt level~ Several field sources can
be used and when these are electromagnetic they can be
connected in series. The magnetic permeability of the liquid
metal of the melt is normally similar to that of vacuum, air,
or a protective gas. Thus, the magnetic permeability of the
melt is largely irrelevant.
Such arrangements are adequate for measuring the
melt level within a limited range. Determining the melt
level on system startup, however, requires that several such
sets of coils be provided along the mold.
Another disadvantage of the known systems is that
they are relatively inaccurate when the melt level lies close
to the desired location. The induced voltages form a bell
curve (voltage, normally in mV, plotted against melt height,
normally in mm) which peaks at the middle of the range,
making accurate measurement of it difficult or impossible~
The field density is greatest at the top of the melt so that
as it comes level with a coil, the induced voltage in this
coil reaches a maximum, then drops off as the top of the melt
moves past the coil, forming a perfectly symmetrical bell
curve if the relative displacement speed is constant. The
relationship of induced voltage to mold position, which is
fairly uniform in khe two flanks of this bell curve, changes
drastically at the peak thereof such that the accuracy of
measurements suf~ers greatly in this region.
It i8 therefore an object of the present invention
to provicle an improved method for measuring the melt level in
a continuous-casting mold, which overcomes the aforesaid
drawbacks.
-- 4 --
~:~9~
It is another object of the invention to provide a
method which gives accurate measurements along its entire
measurement range and which can readily be adapted to a very
wide range, so it can also be used at startup when the mold
is filled.
In accordance with the present invention, there is
thus provided a method of detecting the level of a melt in a
continuous-casting mold by means of a primary coil and at
least three vertically spaced secondary coils surrounding the
mold. According to the invention, a primary magnetic field
is formed with the primary coil, which extends horizontally
through the mold into the melt generally at the melt level.
This primary field generates eddy currents in the melt that
in turn generate secondary fields that induce in the secondary
coils voltages that peak as the melt level comes level with
the secondary coils. Each of the induced voltages is combined
with each of the other induced voltages to provide a plurality
of output voltage curves each corresponding to the combined
outputs of the respective two induced voltages and each peaking
when the melt level lies at a respective location intermediate
the respective induced voltages. The melt level position
relative to the coils is derived from these output voltage
curves.
The invention is based on the principle that the
individual voltage curves of the secondary coils that are
connected opposite to one another have a shape that is either
generally linear or made of straight portions. rrhus, with
one primary and t:hree secondary coils, there are according to
the invention thr.ee curves, that of the first coil minus
the second, the first minus the third, and the second minus
-- 5 --
:~19~3293
the third. of course, with a larger number n of coils a
number of output voltages will be derived equal to the sum
of all the whole numbers between 1 and (n-l). All of these
curves are employed and the linear or straight portions are
compared with the nonlinear or nonstraight portions and the
dif-ferences are recorded. It is a relatively easy procedure
to ignore readings corresponding to the inaccurate peaks of
the curves, which pea~s are according to this invention
positioned level with straight portions of the other curves so
accurate readings can be made along the entire measurement
range. In other words each of the curves has at least one
fairly straiyht portions and the secondary coils are spaced
such that the straight portions cover substantially the entire
vertical measurement range.
In practice, the voltages in the three curves are
followed and compared almost simultaneously with a linear
curve derived from the immediate surroundings to correct for
local magnetism. With this system, it is therefore possible
to determine the melt`level within about 1 mm. Such accuracy
has been hitherto unattainable.
Accordiny to a preferred embodiment of the invention,
the output voltages are compared to derive the melt level.
Preferably the melt level position derived principally by
analysis of the output voltage corresponding to the location
to which the level is closest. The other output voltages
can be compared with respective set points to determine the
approximate position of the melt level.
When the position of the melt level is derived by
analysis of the output voltage corresponding to the location
to which the level is closest, the direction of change of the
-- 6 --
other output voltages may be detected to determine the
direction of displacement of the melt level.
It is also possible to detect the position of the
melt level by comparing the change of at least one of the
output voltages with respect to time.
In addition, the displacement of the melt may be
varied in response to the position o-f the melt level. Means
of doing so are described in detail in the afore-mentioned
European patent application ~o. 82.630.020.
Further features and advantages of the invention will
become more readily apparent from the following description of
preferred er~bodiments, reference being rnade to the accompanying
drawings in which:
Fig. 1 is a diagrammatic representation of a system
for carrying out a method according to the invention;
Fig. 2 shows another system using two coil sets
each having one primary and three secondary coils; and
Fig. 3 shows an arrangement using one secondary and
three unevenly spaced secondary coils.
As shown in Fig. 1, a coil arrangement SPl coaxially
surrounds a rnold K in which a ferrous melt M centered on a
vertical axis A has a li~uid level L. The arrangement SPl
includes a single primary coil Pl and n secondary coils Sl,
S2, S3 .~. Sn 1 and Sn in which respective voltages Vl, V2,
V3 ... Vn 1 and Vn are induced from the eddy currents induced
in the melt M by the field of the primary coil. As is well
known in the art, the melt is continuously fa]ling and is
replenished frorn a ladle or tundish from above through a
valve V (Fig. 3) so that it is moving relative to the steady
primary magnetic field of the coil Pl. It is this relative
~8~
rnotion that generates eddy currents in the melt M that in
turn general secondary magnetic fields.
When the mold K is empty, the total voltage induced
in the secondary coils Sl -~ Sn is equal to zero~ As the level
L rises, at first the induced voltage in the lowest coil Sn
decreases so that the level of the resultant voltages
/~Vn 1) - Vn ~ or (Vl - Vn) increases. At maximum asymmetry,
that is when level L lies equidistant between the two like
secondary coils whose induced voltages are combined for a
given output voltage, this output voltage is at a maximum.
As the level L moves up past this middle point, asymmetry
decreases along with this combined output voltage. Two
induced voltages which each increase and decrease when the
melt level comes up to and passes the respective secondary
coils are combined to form an output voltage which peaks
instead when the level reaches a point midway between the
two secondary coils. This generally straight flanks of this
output voltage can be interleaved with those of other
secondary coil pairs so that the entire vertical measurement
range is covered.
Fig. 2 shows an arrangement having an upper coil
set SPl' and a lower coil set SP2' having respective primary
coils Pl' and P2' and secondary coils Sl 1' Sl 2' Sl 3, and
S2 1~ S2 2~ S2 3. The number of coil sets needed can be
increased to follow the melt level during mo]d filling. In
any case, there are at least three secondary coils for each
primary coilO
In Fig~ 3, the system is identical to that of Fig.
1, but the coils Sl and S2 are spaced more than the coils
S2 and S3. This offset subdivides the entire measurement
range into three nearly equal zones.
-- 8 --
The individual curves shown to the right in Fign 3
resemble Gaussian curves, none of which can be used to
determine the melt level L. The vertical line from +Z -to -Z
represents the position of the level L, normally represented
in millimeters, and the horizontal line the voltage, normally
in millivolts. The curve Vl - V3 only is generally straight
between the points 1 and 2 and between the points 3 and 4,
but is curved between the points 2 and 3. More particularly,
there is a~airly linear relationship in these straight portions
between the distance from the vertical coordinate line and the
vertical position of the level L along this line. In the
curved peak of the output-voltage curve, this relationship
does not hold. The same is true for the curves Vl - V2 and
V2 ~ V3.
When the level L is in the indicated position, it
would be impossible to accurately determine its position with
the curve Vl - V3, at the point X3~ as one would be using the
curved central portion where there is little relationship
between voltage and actual position. In fact, if the level L
rises to the center of this curve there is virtually no
voltage change corresponding to a change of position. It is
however possible to determine at the points X1 and X2 on the
other two output-voltage curves V1 - V3 and V2 - V3 the
exact position of the level L, since these points lie on
approximately straight portions of these curves.
Obviously the coils are arranged so that any given
melt height in the desired measurement range lies on only a
single straight stretch of the curve of a single output
voltage. Nonetheless, the use of two overlapping stretches
as sh~wn in Fig. 3 adds to the accuracy of the measurement.
_ g _
93
Taking two such measurements allows outside factors to be
canceled out, here by determining the difference between the
points Xl and X2~
Determining that a given value of Vl - V2 is greater
than the value at the points 2 and 3 establlshes that the level
L is in a given region. This allows the readings V2 - V3 to
be pinpointed.
Fig. 3 also shows how a controller C is connected to
a valve V that controls the filling rate for the mold K, and
through an unillustrated connection to a traction roller R
that pulls the strand out of the mold ~. The coils Sl, S2,
and S3 are connected to three combiners or adders Al, A2, and
A3 to produce the output signals shown to the right in FigO 3.
The controller C is also connected to a transponder T
constituted by a coil that generates an output signal that is
subtracted from all the signals to eliminate ambient magnetismO
The controller C is an electronic microprocessor
which also compares the incoming signals to set points to
determine the approximate position of the melt level L so that
the signal from the curve that would give the most accurate
reading for that portion of the range is employed. This is
most easily done by establishing set points that represent
crossovers between deriving the melt level from one output
voltage and to a mode deriving it from another. Thus when,
for instance, the level is being derived from the curve
(Vl - V3) between the levels equal to points 3 and ~, a set
point at the point 3 triggers the contro~ler C to take over
the reading from the curve (Vl - V2) which is more linear
above this level. In addition this controller C, in accordance
with standard electronic practice, monitors the direction of
-- 10 --
293
change of any of the incoming signals to determine whether
the level L is rising or falling.
The method according to this invention allows one
to continuously monitor the system in that according to Fig. 3
an increase in the value Xl must be accompanied by a decrease
in the value X2 and must indicate a ~all in the level L. In
general, not only the individual values but also their changes
with time can be used for monitoring the system and determining
melt level. Since one normally only measures the effective
value of the differential voltages, negative differences are
not shown in the bottom half, but instead are flipped up
over the abscissa. This creates as a rule bends in the curves
whose tips lie on the abscissa. Angles to both side of the
point touching the abscissa are identical. These inter-
sections with the abscissa only occur at the outer ends of the
bell curve and can be avoided by increasing or decreasing the
measurement voltage.
Of course the various measurements are handled by a
microprocessor which in turn controls the continuous-casting
operation. When the voltages on the upper spools Sl and S2
are affected the drawing-off speed is increased and when the
lower spools S3 or Sn are affected the speed is decreased by
means of the roller R. Instead of changing the speed, it is
also possible to vary the input rate for metal into the top of
the mold K by adjusting the valve V.
,~
