Title of Invention

LOW COST SLAG DETECTION SYSTEM FOR MEASURING SLAG CARRY OVER TO THE MOLTEN STEEL DURING TAPPING

Abstract The present invention relates to a low cost slag detection system for measuring slag carries over to the molten steel during tapping comprising a direct vision spectroscope (8) enclosed in a CCD camera (7) by which images from light source (4) are captured and processed through hardware card interface routines, a host PC (5) which generates intensity patterns of spectrum of steel and slag generated by said direct vision spectroscope, the classification of intensity distribution for steel and slag being carried out by the detection system via interfacing with the host PC (5) characterized in that the system can differentiate slag from steel during hot metal tapping from BOF and the identification is done in the visible region of the spectrum.
Full Text

Field of Invention:
The present invention relates to a development of low cost slag detection system during
tapping to avoid slag carry over to the molten steel during tapping in steel manufacturing
processes specifically in converter processes.
Background of the invention:
During primary refining in the BOF/EAF (Basic Oxygen Furnace/Electric ARC Furnace)
converter, additions are made to form a slag to remove impurities such as sulphur and
phosphorus from the steel. During converter tapping, in addition to the crude steel some
converter slag is carried over into the steel-making ladle and impurities from the slag
revert to the steel in the ladle. Thus me slag carried over into the ladle during tapping
produce a negative influence on the quality of the melt. Iron oxide contained in the slag
makes de-oxidation of the steel difficult and cause re-phosphorisation of the steel.
The increasing demand for high quality standards, increased yield and higher production
in steel making has necessitated the development of products and systems to provide a
competitive advantage to steel business. For production of higher steel grades, secondary
metallurgy is higher and growth in this area calls for more reliable methods avoiding slag
carry over during tapping. It is in this area of development that slag-containing impurities
from primary refining of steel should be desirably kept at Zero level of carry over to the
steel in the ladle as variable levels of carryover slag cause undesirable variation in the
ladle top slag composition which necessitates control of both the consistency and level of
slag cany over.
Presence of excess converter slag in ladle could raise the following problems:
• Lower recoveries of alloys and conditioners (and thus increased alloying costs) as
additions are hindered by presence of a slag layer.
• Higher levels of FeO and MuO (slag constituents) result in high Oxygen content
of steel leading to increased processing time and treatment costs.
• Phosphorus and silicon reversal occurs.
• Ladle desulpherization is hamped
• High inclusion formation,

• Increased risk of caster nozzle clogging.
• High wear of Ladle refractory.
• Increased consumption of Aluminium.
• Lower yield
To meet these disadvantages of slag carry over to the steel the present invention has proposed to develop
a system to minimize the outflow of converter slag into a ladle and during hot metal transfer and thereby
maximize the amount of steel removed from the vessel during further treatment thereby improving the
quality of the steel and ensuring efficient process treatments of steel.
A prior art patent CN1426146 discloses an invention "Signal transmission cable connector of slag
detection system", the abstract of which recites that this invention discloses a slag detection system signal
transmission cable connecting device including wire clipping components composed of clipping slots
latches and quick connector base seat at right/left sides of a box. A T-shape clipping latch mounted in the
clipping slot to fasten metal sheathed wire transducer wire at the left of the box, a wire connecting screw
insulation base on the box bottom with four sets of screws, each two sets of which is connected with
conducting plates and dust proof cover mounted outside of the quick connector base at the right of the box
with the advantages of high stability of fine anti-mechanical, electric interface and small volume.
Another prior patent GB1509746 discloses an invention entitled "Slag-metal interface detection system
with a bottom pouring ladle", the abstract of which recites that a bottom pouring ladle 1 having a pair of
axially aliened trunions 2 for suspending the ladle in a normally vertical position has a slag metal
interface detector consisting of a pair of probes 13, 14 mounted in the wall of the ladle in the vertical
plane passing through the trunions. The probes are connected by plug and socket connections to a circuit
adapted to indicate when the metal surface lies between the two probes, this indication being unaffected
by small angles of tilt.
The cited prior specification discloses the inventions which is a direct method for slag detection in liquid
steel.
The present invention proposes a system development for detection of slag carry over in liquid steel while
tapping from BOF vessel to ladle in an indirect method.
In the past it was a common practice in the steel melting ships to visually detect slag carry over during
tapping. Towards the end of tapping, experienced workers observe the stream and subjectively determine
the time when the converter vessel is to be titled back to its upright position. Due to different experience
personal skills of the melting personnel, the slag carried over always varies.
To overcome these problems presently different methods were applied like ball and dart arrangement,
electromagnetic coils etc. Presently all over the world the most favored way to detect slag carry over
during hot metal tapping is using Thermal Imaging Camera. This concept is based on the existence of
difference between the emissivities of hot metal (steel) and slag at higher wavelengths. The detection is
possible only in the Infrared region. The system uses an Infrared (IR) camera that captures the thermal
image of the hot metal stream and calculates the slag percentage. The same system is in use in four BOF
vessels of Tata Steel. The system uses a long wave Infrared camera (7-14 micron), Image grabber card,
Special Processing unit, and related accessories.
One of the main disadvantages associated with this system is the high cost. It takes generally two years to
recover the cost of investment.

Another disadvantage associated with this system is the non-availability and
maintainability of the hardware, specially me Infrared camera. As the infrared (IR)
cameras are considered sensitive for defense applications it requires special Government
sanctions to procure such cameras. Again whenever the camera stops functioning it has to
be sent to OEM (Original Equipment Manufacturer) for repair as the camera is a
proprietary item and can be repaired by OEM only.
According to the main objective of the present invention it is proposed a system to
differentiate slag from steel during hot metal tapping from BOF vessel to ladle in die
visible region of the spectrum.
According to another objective of the present invention it is proposed an automated
assessment and decision making system based on a spectral imaging principle which will
be of very low cost, in-house maintainable, and operator friendly.
According to a further objective of the present invention there is proposed a low cost area
scan CCD camera based system to view the spectrum.
According to a still further objective of the invention it is proposed to generate intensity
distribution pattern for slag and steel and identify slag phase by providing a system units
for spectrum generation, image capturing and processing, camera enclosure with water
cooler and system cabling and accessories.
DESCRIPTION FO THE INVENTION:
Spectroscopy is the use of absorption, emission or scattering of electromagnetic
radiations by atoms or molecules (or atomic or molecular ions) to qualitatively or
quantitatively study the atoms or molecules, or to study physical processes. The present
invention in this case depends on spectroscopy. In this process atoms or molecules
excited to high energy levels, decay to lower levels by emitting radiation (emission or
luminescence). For atoms excited by a high temperature energy source, this light
emission is commonly called atomic or optical emission.

When a matter is being heated (or excited) to 1700 Degree Celsius it luminescence's in
the visible range. The spectrum so generated varies characteristically for materials of
different atomic composition.
Though the composition of steel is grade dependent but mainly the constituents of steel
are Carbon, Manganese, Phosphorus, Sulphur, Silica, Nickel, and Chromium While steel
slag mainly comprises of oxides of different materials like Ca0, Si02, Fe0, Mn0, Mg0,
Al203, P205, Sulphur and Metallic Fe. Thus steel and slag generate different spectra when
heated to about 1700°C.
The crux of the system development lies in to register the spectral constituents of
incoming light onto a normal CCD (charged coupled device) plane with the help of a
direct view spectroscope and men to determine the presence of the slag by its specific
signature.
The system operates as a computer assisted visual decision system, using efficient Image
Processing techniques. High-resolution areas scan CCD camera captures the spectrum
generated by the direct vision spectroscope and behaves like intelligent eyes of the
system. The captured images are sent to the computers in the control room for fiirther
processing. The processing is done with Image Processing Hardware and by supporting
software. The software generates an array of pixel intensities along a line within the
image of the spectrum. The intensity distribution of the spectrum generated by the steel
and slag are different By using the thermal imaging based system as reference,
classification was done for the intensity pattern of steel as well as slag. Different
classification algorithms were tried like Support Vector Machines (SVM), K-Nearest
Neighbors (KNN) etc and most suitable one was found to be K Nearest Neighbors. After
the classification, when this algorithm is used to test the real and mixed data pattern of
steel and slag, the test accuracy was found to be 96% repeatedly. The algoritlim testing
was done using MATLAB.


1MAGE CAPTURING AND PROCESSING
The image Processing System consists of the following:
1. High- resolution area scan camera with environmental enclosure installed just
below the tapping pulpit.
2. High-speed frame grabber and Image processing card (PCVision) for real-time
processing.
3. High-End PC.
4. Sophisticated software, coded in Microsoft Visual C++ Ver. 6.0, using specific
Image processing routine of MVTOOLS ver. 6.0 and HEX routine for card
interface. Mvtools is a set of C type routines for building short and elegant yet
reasonably complex code in fewer liues. And 1TEX library is adopted for
accessing the PCVision frame grabber card functions. The library consists of
routines for image grabbing and acquisition.
5. The algorithm for classification and testing was done on MATLAB.
Each data is a vector of length 720. Through PCA (Principle Component Analysis) me
dimensionality is reduced to 10. Training set consisting of 1200 samples of steel
spectrums and 373 samples of slag (steel + slag) spectrums are recorded. The first test set
consisting of 580 samples of steel spectrum and 300 samples of slag (steel + slag)
spectrums are computed from the training set. The same kinds of tests were carried out
for different set of samples.


CAMERA ENCLOSURE:
The camera assembly is specially designed and packaged to face the harsh plant conditions, and has
protective enclosure for dust.
The camera and the direct vision spectroscope are mounted inside the enclosure, equipped with air-
cooling.
The complete enclosure assembly is mounted on a rigid structure below the tapping pulpit.
According to the invention there is provided a low cost slag detection system for measuring slag carry
over to the molten steel during tapping comprising a direct vision spectroscope enclosed in a CCD camera
by which images are captured and processed through hardware card interface routines consisting of
MVTools and ITEX, a host PC, the ITEX routine being adopted for accessing the PC vision card
functions, the whole system being supported by a software to generate intensity patterns of spectrum
generated by the direct vision spectroscope of steel and slag, the classification of intensity distribution for
steel and slag being measured through a algorithm MATLAB with set of rules specifying a sequence of
actions carried out by the detection system via interfacing with the PC to identify slag phase in the molten
steel during tapping.
DETAIL DESCRIPTION OF THE ACCOMPANYING DRAWINGS
The nature of the invention, its objective and further advantages residing in the same will be apparent
from the following description made with reference to non-limiting exemplary embodiments of the
invention represented in the accompanying drawing, in which:
Fig.1 : represents a schematic diagram of the Slag Detection System.
Fig.2 : represents a schematic diagram of the Direct Vision Spectroscope.
Fig.3 : represents a flow chart of slag detection via software.


Fig 4: represents the graphical representation of the experimental results on MATLAB
for support vector machines method
In Fig 1 spectrum is generated of the tapping stream (4) by direct vision spectroscope (8)
enclosed in a camera (7) through a software (1) developed for this purpose. The spectrum
is pointed directly at a light source (4). The captured images in a high resolution area
scan CCD camera are sent to a PC vision (2) for further processing of the said captured
images through image processing hardware interface such as ITEX interline transfer and
supporting software (1).
The software generates an array of pixel intensities along a line within the image of the
spectrum. Classification of intensity distribution of metal and slag are carried through
different classification algorithms such as SVM, (Support Vector Machines) KNN (K-
Nearest Neighbors) and MVT tools (9). Algorithm testiqg is carried through MATLAB
(3). The Host PC (5) save records of the image processing classification and generates
report and display through the in home built hardwares (6 and 9) and software (1). The
test accuracy after the classification is found to be 96% repeatedly.
In fig 2 the spectroscope is shown which is an hand held instrument consisting of a small
tube (10) which is pointed at the light, which passes through a slit (11), a collimating lens
(12), a prism train consisting of two crown glass prisms (13) and one flint glass prism
(14) being cemented together, and exiting through a plain glass cover (15). The
instrument consists of two concentrical cylinders (16 and 17). The collimator lens and
prism assembly is mounted in the inner cylinder (17) and the slit is in the outer cylinder
(16). The focus of the spectrum can be adjusted by moving these cylinders and sharpness
of (lie spectrum is adjusted by varying me slit width (not shown).
Fig 2 also shows the normal CCD plane (18) on which the spectral constituents of
incoming light fall.
Fig 3 in a flow chart illustrates the strategy used for detection of slag. Any error in
captured images in PCVision (2) is corrected through resending the images in PCVision
card and camera again. Train set of samples of sloel and slag spcctrurus are fontted
through die software (1) coded in Microsoft visual f>+ vcr 6.0 (9) and UEX roulinet- (6)
for card interface to form n File Maintained Phase which is transmitted to the re?:;j!tiric
accumulated phase. Hie said phase is then classified and tested i>y MAT'lAls ('/?i to
identify slag phase h ottt the linked dabi oisV-ru of steel and slag.


Fig 4 the graphical representation of classification of intensify pattern of steel and mixed
intensity pattern of steel and slag are shown. The said pattern is obtained through support
vector machines (SVM) (9) by using thermal imaging based system as reference via
support vector for first and second principle components.
The graph generated through the image processing system clearly distinguish slag phase
in the molten steel and the system of identification of slag phase thus enables to take
precautionary measure to minimise the outflow of the converter slag into the ladle
containing molten metal and thus enables to produce quality steel to maintain its physical
characteristics as specified for various industrial applications.
The invention as herein described and illustrated should not be read in a restrictive
manner as various adaptations, modifications and changes are possible as encampused
within die scope of appended claims:



WE CLAIM
1. A low cost slag detection system for measuring slag carry over to the molten steel during tapping
comprising a direct vision spectroscope (8) enclosed in a CCD camera (7) by which images from
light source (4) are captured and processed through hardware card interface routines, a host PC
(5) which generates intensity patterns of spectrum of steel and slag generated by said direct vision
spectroscope, the classification of intensity distribution for steel and slag being carried out by the
detection system via interfacing with the host PC (5),
characterized in that the system can differentiate slag from steel during hot metal tapping from
BOF and the identification is done in the visible region of the spectrum.
2. A detection system as claimed in claim 1, wherein an array of pixel intensities are generated
along a line within the image of the spectrum and the host PC (5) records image processing
classification and generates reports.
3. A detection system as claimed in claim 1, wherein each data of the pixel intensities along a line
within the image of the spectrum formed is a vector of length 720.
4. A detection system as claimed in claim 1, wherein the said vector is dimensionally reduced to 10
through PCA.
5. A detection system as claimed in the proceeding claims wherein the record saving and report
generation for intensity distribution for steel and slag are carried through the host PC (5) via
graphical representation through first and second principle components of support vectors.

6. A detection system as claimed in the proceeding claims wherein the spectra]
constituents of incoming light from the tapping stream (4) is registered onto a
CCD plane with the help of a direct view spectroscope and then the presence of
slag is determined by its specific characteristics.
7. A detection system as claimed in the proceeding claims wherein the high-
resolution area scan camera (7) with protective enclosure mounted on a rigid
structure is installed just below the tapping point
8. A detection system as claimed in claim 7 wherein the camera and the direct
vision spectroscope are provided with air-cooling arrangement
9. A detection system as claimed in the proceeding claims wherein the low cost
direct vision spectroscope (8) comprises of a small tube (10) consisting of two
concentric cylinders (16 and 17) pointed at the light source (4) which passes
through a slit (11), a collimating lens (12), a prism train consisting of two glass
prisms (13) and one flint glass prism (14) and exiting through a plain glass cover
(15).
10. A low cost slag detection system for measuring slag carry over to the molten steel
during tapping as herein described and illustrated with reference to the
accompanying drawings.


The present invention relates to a low cost slag detection system for measuring slag carries over to the
molten steel during tapping comprising a direct vision spectroscope (8) enclosed in a CCD camera (7) by
which images from light source (4) are captured and processed through hardware card interface routines,
a host PC (5) which generates intensity patterns of spectrum of steel and slag generated by said direct
vision spectroscope, the classification of intensity distribution for steel and slag being carried out by the
detection system via interfacing with the host PC (5) characterized in that the system can differentiate slag
from steel during hot metal tapping from BOF and the identification is done in the visible region of the
spectrum.

Documents:

00480-kol-2006-correspondence others-1.1.pdf

00480-kol-2006-form-9.pdf

00480-kol-2006.abstract.pdf

00480-kol-2006.claims.pdf

00480-kol-2006.correspondece others.pdf

00480-kol-2006.description (complete).pdf

00480-kol-2006.drawings.pdf

00480-kol-2006.form-1.pdf

00480-kol-2006.form-2.pdf

00480-kol-2006.form-3.pdf

00480-kol-2006.gpa.pdf

480-KOL-2006-ABSTRACT.pdf

480-KOL-2006-CANCELLED DOCUMENT.pdf

480-KOL-2006-CLAIMS.pdf

480-kol-2006-correspondence.pdf

480-KOL-2006-DESCRIPTION COMPLETE.pdf

480-KOL-2006-DRAWINGS.pdf

480-kol-2006-examination report.pdf

480-KOL-2006-FORM 1.pdf

480-kol-2006-form 13-1.1.pdf

480-KOL-2006-FORM 13.pdf

480-kol-2006-form 18.pdf

480-KOL-2006-FORM 2.pdf

480-kol-2006-form 3-1.1.pdf

480-KOL-2006-FORM 3.pdf

480-kol-2006-form 9.pdf

480-kol-2006-gpa-1.1.pdf

480-KOL-2006-GPA.pdf

480-kol-2006-granted-abstract.pdf

480-kol-2006-granted-claims.pdf

480-kol-2006-granted-description (complete).pdf

480-kol-2006-granted-drawings.pdf

480-kol-2006-granted-form 1.pdf

480-kol-2006-granted-form 2.pdf

480-kol-2006-granted-specification.pdf

480-kol-2006-reply to examination report-1.1.pdf

480-KOL-2006-REPLY TO EXAMINATION REPORT.pdf

abstract-00480-kol-2006.jpg


Patent Number 246130
Indian Patent Application Number 480/KOL/2006
PG Journal Number 07/2011
Publication Date 18-Feb-2011
Grant Date 15-Feb-2011
Date of Filing 22-May-2006
Name of Patentee TATA STEEL LIMITED
Applicant Address RESEARCH AND DEVELOPMENT AND SCIENTIFIC SERVICES DIVISION, JAMSHEDPUR 831 001, INDIA
Inventors:
# Inventor's Name Inventor's Address
1 S. SISTLA TATA STEEL LIMITED RESEARCH AND DEVELOPMENT AND SCIENTIFIC SERVICES DIVISION, JAMSHEDPUR 831 001
2 ASHISH TIWARI TATA STEEL LIMITED RESEARCH AND DEVELOPMENT AND SCIENTIFIC SERVICES DIVISION, JAMSHEDPUR 831 001
3 ANINDYA SARKAR TATA STEEL LIMITED RESEARCH AND DEVELOPMENT AND SCIENTIFIC SERVICES DIVISION, JAMSHEDPUR 831 001
4 PRABAL PATRA TATA STEEL LIMITED RESEARCH AND DEVELOPMENT AND SCIENTIFIC SERVICES DIVISION, JAMSHEDPUR 831 001
PCT International Classification Number B22D 2/00
PCT International Application Number N/A
PCT International Filing date
PCT Conventions:
# PCT Application Number Date of Convention Priority Country
1 NA