FIELD OF THE INVENTION
The invention relates to control the mold behavior based on measured billet
rhomboidity and incorporating mathematical model of dynamic cooling and
casting speed of mold. More particularly, the invention relates to a system for
real-time measurement of rhomboidity of hot billets.
BACKGROUND OF THE INVENTION
Quality problems, such as cracks, rhomboidity and breakouts are often observed
in continuously cast steel billets. One of the problems frequently encountered in
casting is distortion of the cross-sectional profile of the casting strand. Such
distortion results from the fact that the cast strand, during its transit through the
mold and for a real time thereafter, is not a completely solidified condition.
Rather, the strand has a tendency towards cross-sectional deformation during its
movement. While this deformation can occur with any cast shape, strands which
are rectangular in cross-section are particularly susceptible.
DD (diagonal-difference) is measure of rhomboidity - shape related defect in
square cross section billet. Billet Rhomboidity starts with non-uniform shell
solidification in the mold which is due to inconsistent mold cooling causing

irregular heat transfer. The higher diagonal difference greatly impacts the quality
of billets to be rolled at various mills. Rhomboidity at or over 4% lead to billet
twisting in the roughing stands of the rolling mill.
In a conventional method, DD is measured manually at the end of casting
operation by using calipers. By the time measurement is known and corrective
action is initiated, huge quantity of defective product is already produced. Also,
there is always safety concern in manual method as measurement is done on hot
billets (around 50-60 degree C).
It has also been observed that loss of production at customer end due to rolling
of defective billets leading to breakdown in mills. Therefore a real-time, non
contact system is required to measure rhomboidity of hot billets during
manufacturing process only to prevent further production of defective billets as
soon as rhomboidity detected.
Inconsistent mold cooling due to an uneven gap between the water jacket and
the mold outside surface can cause irregular heat transfer. Poor mold water
quality may cause 'plating' of non-metallic compounds on the outside of the
copper tube. This will greatly impede heat transfer. Other mold related factors
include mold taper and wear. Steel chemistry, starting with the peritecti grades
up through the high carbons, can also greatly contribute to billet rhomboidity.

The higher diagonal difference greatly impacts the quality o billets to be rolled
at various mills of Long Product. High DD causes internal cracks (from slight to
massive depending on severity of DD) in billet and can cause cobble in the mill.
If rhomboidity starts suddenly during a cast, it is usually due to a blocked or
misaligned spray nozzle(s) just below the mold. Some operators correct this
problem by selectively increasing the water flow on the sides of the billet,
however this is usually effective only for a short period of time.
To determine % billet rhomboidity, measure and record the two Diagonal
distances, dl and d2.
% Billet Rhomboidity = ((d2 - dl -)/((d2+dl)/2)) x 100
Following are the drawbacks of conventional methods.
1. Due to very high temperature (900-1000 degree C) of billets,
manualmeasurement is taken after 4-5 hours of billet production when
the temperature of billet lowers to 70-80 degree C. By the time
measurement is known and corrective action is taken huge quantity of
defective is generated.
2. There is always safety concern during measurement as measurement
is done on hot billets (around 50-60 degree C).

3. Segregation of these defective billets is cumbersome process adversely
impacting the delivery compliance to customer.
4. Loss of production at customer end due to rolling of defective billets
leading to breakdown in mills.
SUMMARY OF THE INVENTION
Accordingly, there is provided an improved system and method of real-time
rhomboidity measurement of billet. More specifically this a machine vision
system wherein a CCD camera continuously monitors the cross-section of
billet. As billet rests on turn over cooling bed (TOCB), the images are
captured and in turn fed into an embedded hardware unit hosting the
algorithm wherein the captured frames are analyzed. It employs edge
detection and other image processing method on each captured image frame
to determine the billet edges. Once all four edges have been detected,
Rhomboidity can be measured by calculating diagonals of billets by meazns of
Edges.
Since, the aforesaid imaging system is installed two meters away from billet
face resting on turn over cooling bed, the ambient temperature remains
consistently around 80 °C. Thus the camera needs to be properly cooled to

get working in such a high temperature environment. A camera enclosure is
used to cool entire camera from high temperature effects. Camera enclosure
with an purging and water cooling provides the facility to prevent the passage
of dust inside the enclosure and to clean up the dust accumulated onto the
glass in front of the lens as well as cooling of entire system.
The major advantages of this system are :
1. This system reduces the rejection and defective billet generation.
2. This system measures billet rhomboidity only 10-15 minutes after
production online when billet comes on turn oven cooling bed. The
corrective action can be taken quickly and the result of the corrective
action can be seen in next 5-10 min.
3. This will also reduce the inspection cost and loss of production at
customer end by reducing the break down caused due to the rolling of
defective billets.
4. Man dependency will be completely eliminated.
5. The defective generations were at the cost of scarce prime natural
resource. The technology will conserve the natural recourses. This is step
towards Tata Steel commitment for Green earth.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS
Fig.#1 Schematic diagram of Real-time rhomboidity measurement system.
Fig.#2 Cross section of billet with detected edges using image Processing
algorithm.
Fig.#3 Installed machine vision system diagram is caster of long product.
Fig.#4 Trend diagram of rhomboidity for billets of each six strand.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
Figure 1 illustrates the schematic diagram of rhomboidity measurement system.
In a continuous billet caster, as torch cutting machine cut the continuous casting
billet in particular length, it comes over caster bed passing through different
rollers. Then these billets are transferred to turn over cooling bed by pusher
machine.
As shown in figure 3, the field unit of proposed machine vision rhomboidity
measurement system 30 installed around 2 meter away from billet face 303
while it is on turn over cooling bed 304. imaging System, surrounded by cameras
enclosure 310 is mounted on a mechanical stand 301 continuously monitors the
advancement of the billets 302 on turn-over cooling bed where the billets from

six stands are pushed. The V shaped notch 311 of cooling bed 304 collect billets
one by one. The light rays 303 from the cross section of billet resting on TOCB
304 goes inside the assembly 310 through plane toughened glass 306 of
diameter of 100 mm. Glass 306 protects the camera optics and electronics to get
warm from high ambient temperature. An air purging outlet 309 is provided to
prevent dust entering the housing 310. The light from billet face is gathered,
paralleled to Camera lens assembly by lens 307 and focused on CCD (Image
. Plane) of Camera 308. As shown in figure 1 Camera is driven by 12V DC power
provided by Adapter which converts 220V AC to 12V DC. The images captured by
camera is sent to an embedded hardware Online RMS placed in control room,
herein known as the 'the host' through a physical Ethernet link via Media
converted.
The camera continuously grabs the image of cross section of red hot billet. As
shown in Figure 3, as Billet rests on notch of turn over cooling bed, system
(Figure 1) receives the "billet rest" signal on notch with "Strand no." by
Progammable logic controller (PLC) through as Advantech USB module (Figure 1)
connected with RMS system. Image is captured by the camera and sent to
Online RMS system. The Control room unit hosts the sophisticated image
processing algorithms to detect the edges of billet face so that billet diagonals
can be computed to measure rhomboidity. The Hough transform is used to
determine the billet face edges provides a basis for robust extraction of shapes.
Properties of HT are robust to impulsive noise, insensitivity to partial occlusion of
patterns which made the system working even in presence of many uncertainties
like Scales, Burrs as shown in Figure 2. once all four edges have been detected,
rhomboidity can be measured by calculating diagonals of billets.

Sometimes low intensity, high scale on billet and billet entanglement may result
in false and erroneous measurements of rhomboidity of particular billet as shown
in Figure 4 so, statistical techniques have also been incorporated which rectifies
the effect of false and erroneous measurement so it can be used as rhomboidity
indicator in particular strand as shown in Figure. The output in the form of trend
chart and histogram of strand wise rhomboidity as shown in Figure 5 is provided
in control room for operator to monitor and take corrective action.
Patent References

Other References
1. WORLD STEELUNIVERSITY WEBSITE, Hot Rolling Link:
htt://www.steeluniversity.org/content/html/eng/default.asp?catid=30
&pageid-2081272061.

2. M Atiquzzaman, Multiresolution Hough transforms - An efficient method
of detecting patterns in images, IEEE Trans. Pattern Anal. Mach.
Intell. 14(11), 1992, 1090-1095.
3. T.M.Breuel, Finding lines under bounder error, Pattern Recog.29(l),
1996, 1090-1095.
4. SD Shapiro, Properties of transforms for the detection of curves in noisy
pictures, Computer Graphics Image Process, 8, 1978, 219-236.
5. Tsai, R.Y. HUANG, T.S. 1984 Uniqueness and estimation of three-
dimensional motion parameters of rigid objects with curved surfaces,
IEEE PAMI 6, 1, 13 27.
6. ZOLTOWSKt, M.D. 1987 Signal processing applications of the method of
total least squares. In IEEE 21st Annual Aszlomar Conference on Signals,
Systems, and Computers (Pacific Grove, CA. Nov), 290 296).
7. http://www.dresseltech.com/billet%20rhomboidity.pdf

WE CLAIM :
1. A system for real-time measurement of rhomboidity of hot billet
comprising a high resolution camera unit, an embedded hardware unit
which hosts an analysis software, a telephoto lens to focus billet cross
section image onto the CCD surface, a camera enclosure arrangement
for imaging system which makes it to work in very harsh environment, a
clean air purging unit which delivers clean air, and a flow of the clean
air which is delivered from said clean air unit, wherein the system is
operative to billets cross section edge detection to determine the billet
rhomboidity, and wherein the known Hough Transform is used to
determine the billet face edges which provides a basis for robust
extraction of shapes, the properties of the Hough Transform being
robust to impulsive noise, insensitivity to partial occlusion of patterns,
the system is enabled to work even in presence of uncertainties like
Scales, Burns.
2. The system as claimed in claim 1, wherein the system is disposed
nearer to the billet face and over the cooling bed.

3. The system as claimed in claim 1, wherein the CCD camera continuously
monitors advancement of the billet.

ABSTRACT

The invention relates to a system for real-time measurement of rhomboidity
of hot billet comprising a high resolution camera unit, an embedded hardware
unit which hosts an analysis software, a telephoto lens to focus billet cross
section image onto the CCD surface, a camera enclosure arrangement for
imaging system which makes it to work in very harsh environment, a clean air
purging unit which delivers clean air, and a flow of the clean air which is
delivered from said clean air unit, wherein the system is operative to billets
cross section edge detection to determine the billet rhomboidity, and wherein
the known Hough Transform is used to determine the billet face edges which
provides a basis for robust extraction of shapes, the properties of the Hough
Transform being robust to impulsive noise, insensitivity to partial occlusion of
patterns, the system is enabled to work even in presence of uncertainties like
Scales, Burns.