Ferry M. Direct Stripcasting of Metals and Alloys: Processing, Microstructure and Properties
Подождите немного. Документ загружается.
Continuous
casting
processes
and
the
emergence
of DSC
83
3.3
The emergence of direct strip casting
3.3.1 General evolution
of
the process
The earliest reported research
on
continuous strip casting dates back to
1846,
when Sir
Henry
Bessemer developed
and
patented a twin roll casting machine
to produce tin
and
lead strip products (Cramb
1988;
Wolf 2001). The caster is
depicted
in
Figure 3.17 which shows a pair of
300
mm
diameter water-cooled
rolls
in
the horizontal plane. One of the rolls was flanged to contain the molten
metal pool and was fixed while the other roll was weighted
and
able to move
laterally to relieve pressure
due
to excess molten metal between the rolls. Metal
was transferred from various holding furnaces to the mould
where
thin-gauge
strip
was
formed directly from the molten state. In 1856, Bessemer used the
design
shown
in
Figure
3.18
to produce 0.9-1.2 m lengths of steel strip (Cramb
1988). However, major problems were associated
with
steel strip casting such
as the control of fluid flow and metal distribution between the rolls, refractory
problems in the
"pour
box", physical properties of the as-cast strip, edge
containment
and
surface quality.
Cooling
rolls
(-------1-
Strip
guidance
'----------+--
Shear
I--------+-Work
rolls
Figure
3.17.
Schematic diagram
of
Bessemer's
twin
roll strip caster,
adapted
from
Hendricks
(1995).
The technology available
in
the 1850's was insufficient for Bessemer to
overcome the problems associated
with
DSC
and
several years elapsed before
another attempt was made
at
strip casting. In 1890, Edwin
Norton
of
USA
conducted experiments
on
twin roll casting using Bessemer's original concept
and
managed to strip cast
3-5
mm
thick sheets but,
at
the time, various
problems
>still
needed to be overcome to justify further investment in the
process.
Bessemer's ideas were eventually resurrected
in
the 1920's
when
the American
engineer, Clarence
W.
Hazelett, built a
number
of commercial twin-roll casting
84
Direct
strip
casting
of
metals
and
alloys
units to produce lead, brass and aluminium strip (Cramb 1988). These casters
were more rigid than Bessemer's original design
and
reduction of the strip was
also achieved during the casting process. Using
300
mm
diameter rolls, steel
strip of
width
600
mm
and
thickness
-2
mm
was
produced
at
a rate of
12
m/min. Due to problems with the twin roll design, Hazelett subsequently
developed the ring caster which has a large casting roll of diameter
up
to 7 m,
with a pair of rolls used to drive the casting roll (Cramb
1988).
Silicon steel
and
stainless steel strip of width
200
mm
and thickness < 1
mm
were produced
at
a
rate of
15
m/min. Again, variability in strip thickness, properties
and
edge
containment forced Hazelett to abandon both types of roll casting processes in
the 1940's.
A major
new
development in strip casting took place
in
1949
when
Hazelett
designed
and
patented a
twin
belt caster suitable for casting a wide range of
ferrous
and
non-ferrous alloys (see Figure
3.9).
This development has proven to
be highly successful and there are currently over sixty Hazelett casting
machines around the world with the number expected to expand rapidly; the
biggest area of growth is the production of aluminium sheet for car
body
panels
(Hazelett
2006).
The concept of twin roll casting was embraced
by
the aluminium industry in
1950's
with
FATA Hunter Inc.
in
USA
(formerly Hunter Engineering)
and
Pechiney (France) the first to commercially produce aluminium strip. The
original Hunter 'standard' casters generated a cast gauge of about
6-10
mm
with the as-cast product amenable to cold rolling without the need for a
hot
rolling mill (Vangala
et
al.
1992).
There were, however, limitations
in
the range
of casting alloys
and
low productivity caused
by
low casting speeds
and
large
as-cast gauge thicknesses. This inspired Hunter to develop
in
1970's a much
more robust
twin
roll caster called the
Supercaster™
which was capable of
casting a wider range of alloys
at
sheet widths
up
to
2000
mm. Further
engineering developments
in
1990's generated the
SpeedcasterTM
capable of
casting wide
(>
2000
mm), thin-gauge strip « 3 mm)
at
high casting speeds.
The world's first operational
Supercaster™
in
USA produced edge-trimmed cast
strip at a width of
2134
mm
and gauge thicknesses as low as 0.635
mm
at
a
throughput of
38
m/min. In parallel with FATA Hunter's development
programs, Pechiney Aluminium Engineering developed
in
late 1970's a
new
generation of
twin
roll caster called JUMBO
3CTM
with further developments in
1990's producing the JUMBO
3CMTM
caster. While there are notable differences
in
the design and operation of the twin roll casters between these organisations,
the casters appear to be comparable in both performance
and
productivity.
The quest for thinner gauge aluminium strip requires a caster that operates both
at
high casting speeds and high rolling loads (in the range 1.25-1.35 tonnes/mm
of strip width, Thomas
2003).
The high loads are needed to reduce buckling
defects
in
the strip which is considered to be the limiting factor
in
improving
Continuous
casting
processes
and
the
emergence
of
DSC
85
productivity (§3.4.2). To accommodate these high loads, the high-speed casters
developed
at
FAT A
Hunter
and
Pechiney use steel rolls
with
diameters in
excess of
1100
mm.
It
was argued by Thomas
(2003)
that large diameter rolls:
(i)
cause problems
with
setup
and
starting the caster;
(ii)
result in casting defects
and
buckling; (iii) increase the cost of strip casting since they are a consumable
item and, (iv) require larger scale infrastructure since the entire roll assembly
would
exceed
40
tonnes. To alleviate such problems,
an
alternative high-speed
twin roll caster was developed in
1990's
by
Kvaerner ASA (formally Davy
International) capable of producing aluminium strip in the thickness range
1-6
mm. Figure 3.18 is a schematic diagram of the caster showing a set of small
diameter casting rolls
and
a set of larger diameter backup rolls needed for the
high rolling loads associated with high casting speeds. This casting
configuration is therefore very similar to a conventional 4-high
hot
rolling mill.
It
also permits a
wide
range of casting conditions
and
allows a variety of alloys
and
strip widths to be produced (Thomas 2003).
Load
Load
Ne~latl\le
bend
Back
N"'"''''.''' bend
Backup roll
Figure 3.18. Schematic diagram of the 4-Hi strip caster introduced
in
1995
at
Granges
Eurofoil, Germany for the high-speed production of aluminium strip, after Thomas
(2003)
(with kind permission of Institute of Physics, UK).
3.3.2 The revival
of
strip casting
of
iron alloys
Unlike TRC of aluminium sheet, the development of this process in the steel
industry was terminated in the
1940's
due
to problems
with
roll wear, low
productivity,
poor
as-cast strip quality,
and
variable solidification structure
and
mechanical properties (Cramb
1988,
1995;
Manohar
et
al.
2000;
Zapuskalov
2003). Research
and
development was subsequently focused in the area of
oscillating moulds that eventually led to the development of continuous casting
of large section sizes such as slabs, blooms
and
billets (§3.2.1.2). Factors
86
Direct
strip
casting
of
metals
and
alloys
responsible for the delay
in
commercialising TRC of steel include the high
melting point
and
density, the complexity of the
Fe-Fe3C
phase diagram (Figure
1.4)
and a low thermal conductivity compared with aluminium (Table 1.1).
Consequently, significant engineering advances in
DSC were needed before the
vision of Bessemer could be translated into commercial reality.
It
is difficult to ascertain exactly the 'who', 'when',
and
'where' of the
commencement of
modem
steel strip casting research. However, it is known
that research was being carried
out
in several countries
by
the mid-1980's
(Shibuya
and
Ozawa
1991;
Cramb
1995;
Manohar
et
al.
2000;
Zapuskalov
2003)
and
the global competition amongst steel companies to produce the first
commercial-scale direct strip caster was intense. While the commissioning of
pilot plants has been reported
in
the literature, it is likely that these were
preceded
by
laboratory-scale experiments and/or computer modelling. The
details of these preliminary investigations, as well as some pilot-scale casters,
have been kept confidential for obvious commercial reasons
but
Table II
by
Manohar
et
al.
(2000)
shows
TRC
to be the most promising strip casting process.
The following example
ifaken from the open literature provides insight into the
extent of research
and
development required to produce a commercially-viable
steel strip casting plant.
The
development
ofTRC
for
steel
BHP Steel of Australia (now Bluescope Steel) and Ishikawajima-Harima Heavy
Industries (IHI) of Japan began a collaborative effort
in
1988
to gauge the
commercial feasibility of strip casting steel (Castrip
LLC
2006).
This joint
venture commenced with the pilot-scale production of twin roll cast
AISI
304
stainless steel
in
800
mm
wide coils of
up
to 5 tonnes with commercially
acceptable grades produced
by
1991.
Attention was focused
on
the feasibility of
TRC
of carbon steel with a full-scale casting facility being completed
in
1995
at
Port Kembla, Australia. This facility produced -30 tonne coils of low carbon
steel of
1350
mm
width
and
2.5
mm
thickness with
an
eventual reduction
in
thickness to below
2.0
mm
improving the productivity dramatically (Castrip
LLC
2006).
During these early stages,
an
extensive research
and
development
program was carried
out
to:
(i)
generate a better understanding of early
solidification;
(ii) develop procedures for uniform delivery of molten metal; (iii)
devise methods for containing the melt pool edge; (iv) control roll distortion,
and
(v)
understand the interactions between molten steel and refractory
materials.
In March
2000,
the
US
steel company, Nucor, joined BHP
and
IHI to establish
Castrip Limited Liability Company (Castrip
LLC)
with Nucor becoming the
first licensee of the
TRC
technology. Nucor completed construction
and
commissioning of a commercial-scale
TRC
plant
in
Crawfordsville, Indiana,
USA and started production
in
2002.
A schematic diagram of the plant setup is
shown
in
Figure
3.19
with major plant specifications given
in
Table 3.2. The
Continuous
casting
processes
and
the
emergence
of
DSC
87
initial thickness of sheet produced
by
CastripTM
was 1.7-1.8
mm
but
there
appears to be the capability of casting sub-millimetre gauge strip. The process
in Figure 3.19 also shows an in-line rolling mill which can generate a further
reduction of 15-30
% in strip thickness. In
2005,
CastripTM
has the capability of
producing over
500,000 tonnes p.
a.
of strip-cast steel in a range of grades for
various markets.
Ladle
In-line hot rolling
1'
-
---
1-
--"
~
,----
_____
"'"
Solidified strip
Co
ilers
Figure 3.19. Schematic diagram of the
twin
roll strip casting process
at
CastripTM,
Crawfordsville, Indiana, USA, adapted from Castrip
LLC
(2006).
Table 3.
2.
Specification of twin roll casting at
CastripTM,
Crawfordsville, Indiana, USA,
adapted from Castrip
LLC
(2006).
Unit
Ladle size
Caster type
Roll diameter
Casting speed
Sheet thickness
Sheet
width
Coil size
In-line mill
Rolling force
Coiler size
Coiler
mandrel
diameter
Annual capacity
Specifications
60
-110
tons
Twin
roll
500mm
80
- 150 m/min
0.7-2.0mm
2000
mm
(maximum)
25
tons
4-high single strand
30 MN (maximum)
2
x
40
ton coilers
760mm
500,000 tons/year
88
Direct
strip
casting
of
metals
and
alloys
3.3.3 Current status
of
strip casting
Iron
alloys
The commercial viability of
DSC
of steel remains
under
scrutiny
but
it is highly
likely that stainless steel, carbon steel and silicon iron will
be
mass produced
by
this process. Figure 3.20 shows the status of steel strip casting throughout the
world in
2005.
The figure also shows the development of strip casting since the
early
1980's from preliminary stages through to industrial-scale casters. This
figure can be viewed in conjunction with the data compiled by Cramb
(1995)
and Manohar
et
al.
(2000)
to show the substantial reduction in the number of
organisations involved
in
steel strip casting either through natural attrition
or
industrial collaboration, cooperation and networking, which reflects the high
level of financial and research commitment required to develop the technology.
Nevertheless, Figure
3.20
shows that
TRC
has already made its industrial debut
for the production of both stainless steel and carbon steel strip.
Armco
Voest
Alpine
Allegheny
Us
i
nornRSID
ThyssenlRwrH
CSM/AST
Krupp/AST
Nippon
Steel
Corp
/
Mitsubishi
BHPIIHI
Hitachi
Bri
ti
sh
Steel
Bessemer
C/IMI
POSCO/RIST
_; r
op
1
_~~
~==
~r=~Stop
Carbo
n
Stainless
-
+
Clecim!
Stainless
Stainless
Carbon
Stainless
Stainless
.....,
: !
..
1 Eu
ros
trip
....
;
======~r~==
~i
==~
+~
~
oe~st
..
I::
:.
t~"·
i:
~"""
~.:
-=
=
=t===::J
1
1:.
Stop
1 Ni
ppon
-
Stainless
..
+
Nucor
~
Cast
rip
Carbon
Stainless
~
+ Pacific
m~tals
••
_
-===~CI
==~==
=:;::4~
?
Stainless
Carbon
Stainless
!"'r-'---
-
--
_
~
l
-~+-:-:-v:-tc-h
-&
A~
·
S-S
OC
-
i
at-eS-~~i-St-op---'1
?
~
und
er
con
s
tru
ctio
n
1982
..
Preliminary
phase
1986
D
Hot model
1990
1994
D
Pilot caster
1998
2002
2004
.. ..
Industrial-scale Sale
of
steel cast on
caster industrial-scale caster
Figure 3.20. The development of steel strip casting technology since 1982 showing the
type of
R&D
equipment used -
hot
model, pilot caster
and
industrial-scale caster
(organisations in italics are machine suppliers
or
engineering firms),
adapted from Luiten and Blok
(2001).
Continuous
casting
processes
and
the
emergence
of DSC 89
Aluminium
alloys
Strip casting of a
wide
variety of aluminium alloys
has
been
a commercial
operation for five decades using a range of casting techniques
and
several
hundred
casting facilities are currently
in
operation
throughout
the
world
with
annual production of several million tonnes. A major current development
in
twin roll casting of aluminium (see
e.g.
§3.3.1) is aimed
at
decreasing the gauge
thickness. This is expected to improve productivity, reduce capital
and
operating costs, enhance properties, develop
new
alloys
and
expand the range
of
end
products (Menet
et
al.
1997,2001; Beals
et
al.
1998;
Thomas 2003).
The current status of DSC of aluminium alloys is
shown
in
Figure
3.21
which
lists the alloys
and
end
products that are currently achievable. Aluminium
foilstock
and
building products have been available for
many
years with more
advanced sheet products for the automotive, aerospace
and
beverage industries
under
development.
It
is pertinent to note, however,
that
some of the
applications shown
in
Figure
3.21
may
not be achievable
due
to problems with
generating the desired properties
and
strip quality (see
e.g.
§1.3.1.3).
Aerospace
sheet
Automotive
sheet
Can body
stock
Can end and
tab stock
Foilstock
Common
alloys
Building
sheet
Proven
Proven
Proven
I
Today
I
Figure 3.21.
Current
status
and
expected
trends
in
direct strip casting
of
aluminium
alloys
and
potential
end
products, after Thomas
(20D3)
(witll
kind
permission of Institute of Physics, UK).
90
Direct
strip
casting
of
metals
and
alloys
Magnesium
alloys
Strip casting of magnesium is currently at the pilot plant stage with various
organisations investing heavily in the process, particularly
TRC
by
CSIRO
in
Australia. In
2005,
CSIRO
signed an exclusive option
with
Magnesium
International Limited to licence their
TRC
technology. Despite this positive
development,
DSC
of magnesium lags behind the aforementioned alloys
and
the challenges outlined
in
Table
3.3
need to be met to take the process to the
commercial stage. For example, magnesium has a high oxidation potential
but
this may
not
be problematic
if
the procedures adapted for magnesium die
casting are used, together with oxidation control of the as-cast strip. The large
freezing range of Mg-Al alloys and high solidification rate of magnesium also
make DSC difficult. However, these problems may be alleviated
by
changes
in
composition to generate a new range of alloy grades
and
by
controlling the heat
transfer rate
at
the melt/substrate interface. CSIRO have successfully twin roll
cast
AZ31, AZ61,
AZ91
and AM60 alloys and these are expected to be produced
commercially
in
the near future.
Table
3.3.
Challenges associated with
twin
roll casting of magnesium alloys,
adapted from Liang
and
Cowley
(2004)
Challenge
Oxidation
potential
Solidification
rate
Freezing
range
Costs
and
performance
Comments (Liang
and
Cowley
2004)
Molten Mg oxidises readily
and
ignites without good protective
atmosphere systems and
control.
Molten Mg freezes faster than
wrought aluminium alloys
due
to a lower specific heat and
latent heat of fusion. This
makes it more difficult to
achieve uniform solidification
across the strip.
Mg-base alloys have a larger
freezing range than aluminium.
This may lead to formation of
cracking
and
segregation
defects in the as-cast strip.
Continuous requirement from
industry to reduce the
operational cost of
manufacturing processes.
Magnesium alloy sheet
must
compete with other materials
on
the basis of both cost
and
performance.
Possible Solution
The development of safe
but
simple protection for
both
molten magnesium
and
containment materials is crucial
for achieving high quality strip
and
preventing fire ignition.
Control of interfacial heat·
transfer during casting
(§4.5)
Careful alloy design
and
the
development of
new
alloy
compositions.
Like aluminium, TRC of
magnesium is most cost-
effective for small tonnages
« 50,000 tonnes p.a.)
Continuous
casting
processes
and
the
emergence
of
DSC
91
Copper
alloys
A wide range of strip-cast copper alloys have been produced commercially for
several years
with
Cu
(>99
.
85%),
Cu-(10-40%}Zn, Cu-(1.5-8%}Sn, Cu-(0.
9-
25%}Ni
and
Cu-(1.7-
2.4%}Fe
alloys produced to thicknesses of
up
to
25
mm
(Carlsson
et
al
. 1990; Muller
2003;
Kim
et
al.
2005; Hazelett
2006)
. Copper alloys
are produced in moderate tonnages using a Hazelett belt caster
with
in-line
rolling operations which results in a very compact processing line. Figure 3.22
is a schematic diagram of the world's first continuous copper strip casting
and
hot
rolling line of Lamitref Industries
NV,
Germany (Conti
MTM)
which can
produce
up
to 100,000 tonnes p.a. of copper sheet. In the process, 10
mm
gauge
strip of
width
1250
mm
can be produced using a Hazelett caster
and
hot
direct
rolling.
Melting furnace Equalizing
/ furnace .
Holding furnace / Cooling tunnel
t:!
T~"
/belt caster
HjOlling
/ Surface mill
I~---'I
/
I
----
I
I
I
(not shown)
° °
(not shown) 1/
Coilers
Figure 3.
22
. Schematic diagram of the Conti
MTM
copper strip casting
plant
showing
the caster, in-line rolling mills
and
coiling setup,
adapted
from Muller (2003).
Other
alloys
DSC of titanium appears to be both technically
and
economically feasible,
but
developments are slow
and
restricted to laboratory-based research (Gaspar
et
al.
1991, 1994; Weaver
and
Garmestani 1998). The slow progress is probably
due
to the high melting
point
and low thermal conductivity of titanium
and
problems associated with oxygen pickup
at
elevated temperatures.
Nevertheless, commercial purity titanium has been cast into thin-gauge
(-0.5
mm) strip using a single-roll casting technique termed melt overflow rapid
solidification technology
(MORST)
(Gaspar
et
al
. 1991).
A range of strip-cast lead
and
zinc alloys have also been produced
commercially for several years. These alloys can be produced
by
TRC,
but
are
more often produced in a Hazelett belt caster
with
in-line rolling operations
which result in a very compact processing line.
92
Direct
strip
casting
of
metals
and
alloys
3.4 Comparison between DSC and other continuous
casting processes
3.4.1 Process characteristics and advantages
of
strip casting
The foregoing sections have shown that a wide range of metals
and
alloys can
be
cast continuously into large-sized ingots, near-net-shape slab
and
thin-gauge
strip
or
net-shape ribbons.
It
is clear that the scale of the overall
casting/downstream processing operation is downsized considerably as the
as-
cast
gauge is reduced. This is illustrated
in
Figures 3.23
and
3.24 showing
process flow diagrams for aluminium and low carbon steel.
Table
3.4.
Comparison
of
typical continuous casting characteristics
for
steel, adapted
from
Legrand
et
al.
(1997),
Blejde
et
al.
(2000a),
Zapuskalov
(2003).
Process Variables
Conventional Thin
Slab
Direct Strip
Continuous Casting Casting
Casting
(TSC)
(DSC)
(CCC)
Product Thickness
(mm)
150
-
300
20-60
1-4
Total Solidification
Time
(s)
600
-1100
40-60
0.15
-<
1.0
Casting Speed (m/min)
1.0
-
2.5
4-6
30-
90*
Average Heat
Flux
in Mould
(MW
/m2)
1-3
2-3
6-15
Weight
of
Melt
in Caster
(Kg)
>5000
-900
<400
Average Shell Cooling
Rate
(OC/s)
-12
-50
-1700
Scale
Loss
(kglm2)
7.8
<0.2
Plant
Capacity
(Mt
p.a.)
4-10
2
0.5
* Faster speeds are possible at Castrip
and
Eurostrip (Castrip
LLC
2006).
The CCC,
TSC
and DSC processes have different
mould
geometries, heat
transfer characteristics, cooling rates and casting speeds
and
produce widely
varying microstructures and properties. Typical process characteristics of slab-,
thin-slab-
and
strip-cast steel are given in Table 3.4 which shows the substantial
differences
in
casting parameters. For
DSC,
the casting speed is at least 30x
faster than conventional slab casting with solidification occuring
in
a fraction of
a second. However, to match the tonnage
output
of DSC
with
that
of other
casting routes, the operating speeds need to be higher than those given
in
Table
3.4 simply because DSC produces a much thinner as-cast product.
A key aspect of many of the DSC configurations that separate them from the
other casting processes is the direct contact of the liquid metal
with
the mould
(Figure 3.25a). In the production of steel using CCC
and
stationary mould
TSC