Ferry M. Direct Stripcasting of Metals and Alloys: Processing, Microstructure and Properties
Подождите немного. Документ загружается.
Secondary
processing
and
fabrication
233
Strip-cast
aluminium
alloys
Most
fcc metals
produced
by
cold rolling
and
annealing
have
a
low
r -value
«0.8)
and
a
high
M -value (-0.6)
which
results
in
limited
deep
drawability
and
problems
with
earing
(Humphreys
and
Hatherly
2004). To enhance the
drawability, either a reduction
in
{100}<001>
or
an
increase
in
{11I} <
uvw
> is
desirable but, unfortunately,
the
latter
does
not
form easily
in
these materials.
There
have
been
several
attempts
at
improving
the
formability
parameters
in
strip-cast
aluminium
alloys
and
the
following discussion outlines a few of the
more
recent.
Uppperdie
Lower die
Figure
6.31.
Schematic illustration of the continuous strip shearing process based on
DCAP, after Han
et
al.
(2002)
(with kind permission of Elsevier Limited).
Table
6.4.
Texture evolution and calculated formability parameters as a function of
processing route for strip-cast
AA1050,
adapted from Han
et
al.
(2002).
Process Effect
Major texture
r-value
M-value
Strip
casting
<110>//ND
i
<110>//ND
Moderate
(0.9)
Very low
(0.05)
DCAP
<ll1>//ND
i
<111>//ND
{lOO}<Ol1>
i
{lOO}<Ol1>
High (1.1)
High
(0.6)
<110>//ND!
DCAP + annealing
{lOO}<Ol1>
!
<111>//ND
High
(LOS)
Very low
(0.05)
{lOO}<OOl>
i
{lOO}<Ol1>
{lOO}<OO1>
Han
et
al.
(2002)
have
developed a novel technique for
improving
the
strength
of
the
{11I} <
uvw
> texture
in
strip-cast (1.55
mm
gauge) AA1050
Al
alloys
by
incorporating dissimilar channel
angular
pressing
(DeAP)
as a secondary
processing step, Figure 6.31.
In
this process,
the
as-cast
strip
is subjected to
234
Direct
strip
casting
of
metals
and
alloys
shear deformation through the roll bite of a rolling mill which tends to generate
a range of important texture components after annealing (Table
6.4).
DCAP
generates
<ll1>//ND
and {l00}<011> texture components through the shearing
action of the die with the sheet which is similar to the surface shear texture that
is produced
in
twin roll cast Al alloy strip (see
e.g.
Figures
5.35
to
5.37).
On
annealing, the major texture components include a combination of
<l1l>//ND,
{100}<011>
and
{l00}<001> and generates sheet material
with
a high r -value
and very low
M-value; this equates to an overall improvement
in
formability.
It
is pertinent to note, however, that plastic straining
by
DCAP will have a
deleterious effect
on
the surface condition of the strip as considerable contact
occurs between the die wall and the strip surface.
GallerneauIt and Lloyd
(2002) investigated the effect of secondary processing of
twin roll cast
AA5754
on mechanical behaviour (see Table
6.3)
and compared their
data with material produced by conventional
DCC!I'MP. Figure
6.32
shows the
effect of processing variables on formability of both types of material; formability
is assessed as a limiting dome height before splitting using a lubricated bulge
tester with a
100
mm
diameter die. Despite the elaborate secondary processing
schedules, the strip cast material exhibits slightly inferior formability compared
with DCC material; this result was tentatively attributed to differences in the
distribution of intermeta1lic phases in the alloy and not considered to be a result of
differences
in
texture.
50,----------------------------------,
~~----~--------~--------~------~
1.0 2.0 3.0 4.0
Sheet thickness (mm)
Figure
6.32.
Comparison of limiting dome height for conventionally-produced
and
10
mm
gauge
twin
roll cast AA5754 alloy processed further
by
the following routes, after
Gallemeault and Lloyd
2002
(with kind permission of Trans Tech Publishers):
Process 1 - Warm roll to 6 mm, cold roll to 3.5, 2.0
and
1.0
mm
and
recrystallize.
Process 2 - Warm roll to 3.5, 2.0
and
1.0
mm
and
recrystallize.
Process 3 - Warm roll to 6 mm, interanneal, cold roll to 3.5, 2.0
and
1.0
mm
and
recrystallize.
Process 4 - Cold roll to 3.5, 2.0
and
1.0
mm
and
recrystallize.
Secondary
processing
and
fabrication
235
6.7.2
Surface treatments and coatings
Sheet metal products are often surface treated
or
coated for decorative purposes
and/or protection against wear, oxidation
and
corrosion. A range of surface
treatments which may be classified as either metallic or non-metallic coatings
are shown in Figure
6.33.
Some examples relevant to
DSC
include
electrochemical coatings such as anodising of aluminium
and
magnesium,
conventional coatings such as tin plating
and
galvanising of low carbon steel
and polymeric or ceramic coatings
on
both aluminium
and
low carbon steel.
,---------
-----------1
i Anodising, Phosphate, Chromate, Oxide J
,------------------------------------
Figure
6.33.
Possible coatings for sheet metal products.
The process of anodising is important
in
the aluminium sheet metal industry
and
involves the generation of a coherent and pore-free oxide layer
on
the metal
substrate that makes the underlying material impervious to chemical attack.
The mechanism of anodising is a chemical conversion of the surface of the
aluminium to produce a
hard
oxide layer during immersion of the material in
an acidic solution. These anodic layers have varying degrees of porosity and, to
improve corrosion resistance, a sealing operation is carried
out
by
exposure to
steam. During this stage, dyes can be used to produce a corrosion and wear
resistant coating in a range of colours for decorative applications.
There are difficulties
in
generating a high quality surface in strip-cast Al alloys
which often lead to a poor finish after anodising
and
preclude the production of
components where surface quality is paramount. For example, surface
segregation
on
strip-cast aluminium alloys results
in
a grey surface appearance
after anodising (Emley
1976).
For thicker gauge strip produced
by
belt casting
(Figure
3.9),
there is scope for substantially altering the cast structure by hot
and cold rolling which favours the ability to etch and anodise the final product.
The ability to produce a satisfactory finish
in
DSC Al alloys is expected to
extend the market range of this material.
236
Direct
strip
casting
of
metals
and
alloys
The processes of
hot
dip
galvanising and tin plating are cost effective means of
protecting mild steel against corrosion
and
involve the application of a layer of
zinc
or
its alloys and tin, respectively, to the steel surface. For sheet products,
these processes are carried
out
as a continuous process (Llewellyn 1994). The
application of paint to a strip surface (that may
be
galvanised) is similar to the
previous processes
and
is generally carried
out
as a batch
or
continuous process
(Llewellyn 1994).
In
galvanised steel, the zinc alloy layer acts as
both
a physical
barrier
and
sacrificial anode and thus, cathodically protects the underlying steel
if there is any surface damage.
In
contrast, tin plating
and
painting generates
an
impervious surface layer more resistant to chemical attack than the
underlying steel.
Figure 6.34 is a schematic diagram showing a possible
plant
setup for the
continuous processing of galvanised steel sheet products from a strip caster;
similar configurations are likely for tin plating
and
painting. Strip-cast low
carbon steel is highly amenable to surface modification
by
these processes and
will make them important additions to any DSC process since possible
applications of strip-cast steel sheet include structural decking, light gauge steel
framing, tubular goods, office furniture, appliances, food packaging and
exposed automotive panels.
Twin roll caster
Inert atmosphere
Single/multi-strand
Rolling
/
Heat treatment
furnace
\
Coiler
Zinc bath
Figure
6.34.
Schematic diagram
of
a possible configuration
of
a twin roll casting,
cold rolling and galvanising plant.
In summary, the final quality of any surface treated product is strongly
dependent
on
the structure
and
properties of the sheet material to
be
treated.
The various types of surface defects described in §4.6.2 need to
be
minimised or,
preferably, eliminated
in
order to generate high quality coated products
produced
by
strip casting.
Appendix A
Simulation
and
Modelling
of
Direct
Strip
Casting
A.I Introduction
Direct strip casting is a complex process involving solidification of liquid metal
directly into thin-gauge strip.
To
control both the process and the microstructure
and properties of the strip, an understanding of various phenomena such as heat
transfer, fluid flow and thermomechanical behaviour is important. A more
.
complete understanding of these (and many other) aspects of
DSC
can be
achieved using casting simulators and/or mathematical modelling. To be
effective, however, the outcome of these methods must compare well with actual
plant data. As it is impossible to depict the overall strip casting process by a
single model, the process is generally broken
up
into component parts. Examples
include the prediction of molten metal flow, interfacial heat transfer or the
temperature profile of the strip during casting. This appendix describes some
useful laboratory-based techniques for simulating strip casting together with
relevant mathematical techniques for modelling heat transfer, fluid flow and
microstructural evolution etc. The appendix precludes a detailed discussion of
the very wide range of mathematical
and
computer-based techniques available for
modelling
DSC and does not cover models and simulations associated with
machine dynamics and process automation.
A.2 Experimental techniques
A.2.1 Splat quenching
Quenching a molten droplet onto a substrate is a simple
but
useful method for
understanding rapid solidification behaviour. The droplet may be levitated above
237
238
Direct
strip
casting
of
metals
and
alloys
the substrate or injected from a small nozzle containing the melt. Other rapid
quenching techniques are possible such as sandwiching a molten droplet between
two platens. A schematic diagram of the interaction of a droplet with a substrate
is given in Figure
At.
Using this technique,
it
is possible to study systematically
the effect of solidification variables (melt superheat, melt atmosphere and alloy
composition) on cast structure development
and
heat transfer. Furthermore,
changes
in
the type, surface condition, topography and temperature of the
substrate are easily made. Splat quenching is also useful for investigating the
cooling rate
and
undercooling associated with the early stages of solidification
and is a particularly valuable technique for investigating interfacial heat flux.
The latter is made possible
by
carefully inserting thermocouples
at
various
positions within the substrate (Figure
AI)
or using optical pyrometry to
measure the droplet surface temperature.
I
Droplet maintained
by levitation or ejected
from
a nozzle
Droplet
Embedded
Thermocouples
Figure
A.t.
Schematic diagram of the interaction of a molten droplet with a substrate.
(a typical method of temperature measurement is also shown).
A notable downside of splat quenching is the difficulty
in
comparing directly
the microstructure and heat flux data with that generated
by
DSC.
For
example, the microstructures are not the same as those produced
by
strip
casting since the latter involves nucleation and growth of solidifying grains that
are affected
by
molten metal flow and the movement of the substrate relative to
the melt (see §5.2.I).
A.2.2 Substrate immersion
Substrate immersion to produce as-cast metal coupons is another useful
technique for simulating the solidification processes associated with strip
casting. Figure
A2
is a typical example of the technique whereby a water-
Simulation
and
modelling
of
direct
strip
casting
239
cooled paddle containing one
or
more types of substrate is
dipped
at
a certain
velocity into a pool of molten metal which is then retracted from the melt to
produce a solidified coupon
on
the substrate.
Slider Assembly
"
compumotors"
D
Immersion Paddle
Inert
Gas
Atmosphere
"
cy
1
..
........
.
................
.
..
..
..
..
....
...
..
..
..
..
..
..
...
.
....
....
......
............
..
........
Copper Substrates
~Induction
Furnace
75kW - 3kHz
Figure A.2. Schematic diagram of a typical immersion apparatus showing the melt,
furnace
and
substrates. The inset shows the types of substrate surface
topography discussed
in
the book: smooth, ridged
and
pyramid,
adapted from Strezov
and
Herbertson
(1998).
Compared with splat quenching, substrate immersion is a better DSC analogue
since the melt/substrate contact during dipping closely approximates the
melt/roll contacting geometry of the meniscus region of a single-
and
twin-roll
caster
and
produces as-cast microstructures comparable to that produced
by
these casting processes (Strezov
and
Herbertson 1998). The technique is also
capable of investigating systematically the same casting variables as splat
quenching
but
is more versatile since casting velocity is
an
additional variable.
Furthermore, several different substrates can be used concurrently during
casting (see inset of Figure
A2) which is particularly useful as
it
allows for the
240
Direct
strip
casting
of
metals
and
alloys
unambiguous interpretation of the effect of substrate type on heat flow and
microstructural development. Substrate immersion can also be used in
conjunction with optical pyrometry, Figure
A.3.
This allows temperature
fluctuations to be monitored during solidification and subsequent cooling and can
be used to investigate phase transformations in the as-cast strip. This
is useful in
the study of solidification behaviour of iron and titanium alloys due to the various
types of phase transformation that occur (see chapter
2).
High
volume
gas
flow
valve
~
Inert
gas
Data
acquisition
~UPPIY
and
quench
------'IIVI
control
Pyrometer
--a,
o
o
/
Strip sample
Melt
Immersion I withdrawal
Substrate
o
o
Figure A.3. Schematic representation of
an
experimental apparatus
used
to study
microstructural evolution
in
strip casting, after Evans and Strezov (2000a) (with
kind permission of The Metals, Minerals and Materials Society,
USA).
A.2.3 Laboratory- and pilot-scale strip casters
While a scaled-down strip caster
is
more capital intensive and expensive than the
foregoing techniques, it is a powerful method for investigating a wide range of
important variables pertaining to strip casting. For example, a laboratory-scale
twin roll caster is useful for correlating casting performance with a wide range of
variables including alloy type, melt superheat, molten metal delivery, mould level
control, furnace atmosphere, casting speed and roll properties (material type,
surface structure, roll shape, roll cooling, roll separating force and edge
containment) etc. Considering the substantial benefits of a scaled-down strip
caster, a large number of organisations have been active in developing research
programs over the past few decades.
Li
(1995b)
has compiled a comprehensive
list of the types of laboratory- and pilot-scale
twin roll strip casters used for
process development as well as some casting parameters and alloys investigated.
For obvious reasons of confidentiality, the list is incomplete
but
nevertheless,
highlights the extent of the research throughout the world.
Simulation
and
modelling
of
direct
strip
casting
241
A.3
Mathematical Techniques
Mathematical modelling of DSC
has
been a major contributor
to
our
understanding of various phenomena involved
in
the process
and
is a useful
tool for
the
prediction of productivity
and
for process optimisation. While a
large
number
of
mathematical techniques are available for modelling strip
casting, this section is restricted to models discussed
in
the
book
for
understanding
phenomena
such as
heat
transfer, fluid flow
and
microstructural
development.
A.3.1 Modelling of heat transfer
The successful operation of a strip caster is highly dependent
on
the optimal
thermal management of the casting process. Depending
on
the information
required, various heat transfer models have been developed including one
dimensional (I-D) transient models
or
two dimensional (2-D) steady state
or
transient models. For I-D models, it is assumed that heat flow is normal to the
~
mould wall
and
negligible
in
other directions; such models are easiest
to
develop
and
provide a reasonably good description of DSC. 2-D models are more
powerful
and
can generate information such as the spatial temperature
distribution throughout the strip
and
mould during casting (Figure
4.18)
and
temperature hysteresis behaviour (Figure
4.21).
An important aspect of all
transient
heat transfer models is the ability to calculate
the
time variation of
interfacial
heat
flux
during
casting (see
e.g.
Chapter 4).
Transient
heat
flux
calculations
for
a
stationary
mould
The transient interfacial
heat
flux can
be
determined from a time modified
version
of
Eq.
4.7:
(A.l)
where
Tm(t)
and
TM(t) is the metal
and
mould
surface temperature,
respectively,
and
q(t)
the
interfacial
heat
flux
at
a given time
t.
Heat
flux associated
with
the initial contact of molten metal
with
a
mould
can
be
calculated directly only if the interfacial temperature can
be
measured.
There are obvious difficulties associated
with
such a
measurement
in
DSC
and
heat
flux is usually calculated
in
conjunction
with
the
more
simplified casting
procedures
of
splat quenching
and
substrate immersion. Here, temperature
changes can
be
measured
by:
(i)
thermocouples placed carefully
at
a various
depths
below the surface of the substrate (Strezov
and
Herbertson 1998)
and
(ii)
optical
pyrometry
at
the surface of the melt (Wang
and
Matthys 2002)
or
at
the
metal/substrate interface via a specially machined hole
in
the
substrate
(Mizukami
et
al.
1992).
242
Direct
strip
casting
of
metals
and
alloys
The transient heat flux in the initial stages of melt/substrate contact is strongly
time-dependent, so
it
is essential to solve Fourier's time-dependent heat flow
equation
(Eq.
2.47):
The calculation of interfacial heat flux is
not
trivial
but
solutions of the heat flow
equation are possible using inverse numerical techniques. These techniques
estimate the interfacial heat flux based
on
careful temperature measurements in
the interior of the mould. The most commonly used techniques are based on
the finite difference method proposed initially
by
Beck
and
co-workers (Beck
et
ai.
1985).
Here, the temperature
at
one or several locations below the surface of
the mould is measured accurately and iterations of
an
initial estimate of the
interfacial heat flux are made until the difference between the computed and
calculated temperatures
at
prescribed positions in the mould fall below a
prescribed value. This iterative process is carried
out
over a range of
temperatures to calculate the interfacial heat flux as a function of time.
It
is
pertinent to note that small errors in the measured
data
can lead to large
uncertainties in the calculated value of heat flux. For example,
if
a
thermocouple is placed too close to the mould surface, the heat flux calculated
from the thermocouple
data
corresponds to the local heat flux across the
interface rather than an overall value.
On
the other hand,
if
the thermocouple is
positioned too far below the interface, it is less sensitive to the rapid changes in
heat flux associated
with
the initial period of contact between the melt
and
mould surface. Errors can be minimised both
by
the correct choice of
thermocouple
and
optimising its position within the mould.
30.-----------------------,
OL-
__
-L
____
L-
__
~
____
L_
__
~
o
10
20
30
40
50
Time
(ms)
Figure A.4. Variation
in
interfacial
heat
flux
during
substrate immersion, after Strezov
and
Herbertson (1998) (with kind permission of
the
Iron
and
Steel Institute of Japan).