5th International DAAAM Baltic Conference
"INDUSTRIAL ENGINEERING - ADDING INNOVATION CAPACITY OF LABOUR
FORCE AND ENTREPRENEURS"
20-22 April 2006, Tallinn, Estonia
SIMULATION METHOD OF DUCTILE IRON SOLIDIFICATION
Čiuplys, A.; Bočkus, S. & Žaldarys, G.
Abstract: The purpose of this work was
to compare computer program of the
mathematical model created by authors for
calculation of heat exchange in a
solidifying casting, which helps to
calculate volume changes in the solidifying
casting with well known simulation
software – Novacast’s NovaFlow&Solid.
After research of technological ductile iron
casting factors, the program can rather
precisely, without test castings, find the
best dimensions of risers for ductile iron
castings. Correlation of simulation results
received by authors program and
Novacast’s NovaFlow&Solid software, and
experimental
results
showed
high
reliability of this mathematical model.
Key words: Simulation, solidification,
ductile iron, riser.
casting. Higher metal consumption makes
a casting more expensive, therefore, the
dimensions of a riser must be optimal, i.e. a
riser must do its function, and a shrinkage
cavity can not penetrate into a casting.
The dimensions of risers can be calculated
in different empiric methods [1]. When
temperature-dependent volume shrinkage
rate of an alloy is known, the general size
of a shrinkage cavity is usually calculated.
After that, volume of a riser is found
according to its type (open or closed).
These dimensions of risers are not precise,
and later on, they are corrected in the
process of production. Such creation of
technology takes a lot of time.
Risers can be calculated according to the
heat balance of a casting solidification [2]
too, i.e. a casting is divided into separate
simple geometric elements, and heat
change is calculated for each of them in a
definite period of time when heat is
transferred from metal into mould walls, as
well when heat exchanges among different
elements of the casting. Different
dimensions of a riser are tried, and the best
dimensions of a riser are found when metal
solidify in it last. Other authors simulate
the same heat balance with respect to
material balance [3]. Such simulation is
more precise because it estimates the
change of metal amount in a riser, but it
does not estimate the formation of a solid
metal crust.
Disadvantage of the most mathematical
models, used for calculation of heat
exchange, is that the temperature of liquid
metal poured into the mould is considered
as instant. In fact, mould pouring lasts from
ten seconds up to some tens of seconds.
1. INTRODUCTION
Most castings of different alloys need
risers. Risers are not necessary in some
cases, i.e. when castings are thin-walled or
they do not have hot spots. Risers are not
necessary sometimes due to specific alloy
properties. Castings of alloys, which form
a concentrated shrinkage cavity during
solidification, need risers. Under some
technological casting conditions and
dispersion, the alloy which has the same
metallurgical quality can form a shrinkage
cavity of different size. The formation of a
shrinkage cavity is complicated, and it
depends on many factors. The size of a
shrinkage cavity during crystallization of a
casting determines the dimensions of a
casting riser. The size of a riser has a big
influence to metal consumption for a
251
Therefore,
metal
looses
part
of
superheating energy and shrinks partially,
and after mould pouring, a temperature
gradient in metal appears.
One of the way which helps crystallization
of liquid metal act at the last moment of
solidification, is warming of a riser with
exothermic mixture. Heat flow at such riser
is less than in other part of a casting,
consequently heat balance of a riser and
riser itself can be smaller than in moulds
with homogeneous mixture. Therefore,
such method makes moulding technology
complicated and expensive [4].
On the base of Fourier heat conductivity
equation for a rectangle parallelepiped
surface and for a cylindrical surface under
marginal conditions and initial conditions a
digital calculation method of solidifying
metal was created [5]. Calculation was
made using definite differences in the
method of a sieve. A calculation method
for solidifying metal, flowing in the mould
cavity, was created as well. According to
this model, it was possible to calculate
solidification of metal for castings that
have elements of different cross-section,
and different speed of metal flow is
estimated in each element. After mould
pouring, a temperature gradient and initial
conditions were found for the further
calculation of solidification in a casting. A
program was created for volumes changes
of a solidifying casting. The whole volume
change of a casting in time was
transformed into volume change in a riser,
i.e. it is calculated in what size the level of
liquid metal decreases. Moreover, when the
level of liquid metal changes in a riser, the
thickness of a solid metal crust was
estimated as well. On the base of this
mathematical model, a computer program
was created for calculation of heat
exchange in a solidifying casting and
volume shrinkage of metal.
Nowadays, one of the most popular
software for riser simulation in the world is
NovaCast’s NovaFlow&Solid [6]. It makes
a virtual view of a casting solidification in
the method of definite elements. When the
formation of thermal centres is observed in
a casting, it is possible to predict not only
the places of thermal stresses but also the
places for risers in a casting. When
different
casting
technologies
and
parameters are tried, it is possible to direct
the crystallization of a casting towards its
riser. Afterwards, by changing the
dimensions of a riser, it is supervised that
crystallization in a casting would happen
last.
2. EXPERIMENTAL
Density of melted ductile iron and
temperature-dependent volume shrinkage
rate were found in the method of
hydrostatic weighing when a graphite buoy
was submerged into liquid metal, and
density was found according to the lifting
force of metal [7].
Experimental cast iron was melted in the
induction metal melting furnace. The
charge consisted of steel scrap, waste of
production, carburizer, and ferro-silicon.
Spheroidizing inoculation of cast iron was
made using magnesium ferrosilicon alloy
in the method of overpour.
Experimental castings had two elements
(140×44×140 mm; 170×160×65 mm) and a
riser (Ø90×130 mm) [5]. The temperatures
of the solidifying casting: in the riser and
in different elements of the casting were
measure by three thermocouples of
platinum and platinum rhodium. In the
experiment, metal pouring temperature was
1300 °C.
3. RESULTS AND DISCUSSION
Temperatures of the solidifying casting are
given in Figure 1 [5]. The results show that
a big temperature gradient appeared in the
casting after pouring into the mould. When
the casting was solidifying, temperature
decreased the most rapidly in its first
element, and the most slowly in the riser
because the first element has the smallest
relevant wall thickness, and the riser has
the biggest one.
252
1300
Temperature, C
1250
o
o
Temperature, C
1300
1200
1150
1250
1200
1150
1100
1100
0
0.5
1
1.5
2
Time, min
1
2
2.5
0
3
0.5
1
1.5
2
Time, min
1
3
2
2.5
3
3
Fig. 1. Temperature of solidifying casting
during the experiment: 1 – in the first
element of the casting; 2 – in the second
element of the casting; 3 – in the riser.
Fig. 3. Calculated temperature change
(simulation with NowaFlow&Solid) in the
different elements of cooling casting.
Designation of curves is presents in Fig. 1.
Temperatures of different elements of the
casting, calculated according to the
mathematical model [5], are given in
Figure 2. The calculated temperatures in
this Figure is given as funkcion of
solidification time but not as function of
time steps we used in our previous work[5].
That let to compare the given results with
the experimental results and the results of
other works miles easer. Figure 2 shows
that the calculated solidification path is
essentially similar to the solidification path
determinated by the experiments.
Figure 3 presents the cooling curves of
experimental casting calculated by
NovaFlow&Solid software. The similarity
of these curves to experimental curves is
better fractionally than the curves showed
in the Figure 2.
Correspondences of the mathematical
model to a real process and to
NovaFlow&Solid software were tested
using software “Statistica”. The change of
metal temperature in different elements of
the casting during solidification was
chosen
for
statistical
evaluation.
Correlation between the experimental
change of metal temperature in time period
and the change of average temperatures in
each element of the casting, calculated
according to the mathematical model and
NovaFlow&Solid software were estimated.
With a reliable interval of 95 %, there is a
good correlation between experimentally
measured temperatures and calculated
average temperatures. A correlation rate
for different elements of the casting
changes from 0.906 to 0.982 (Fig. 4).
A correlation rate between experimentally
measured temperatures and calculated
average temperatures by NovaFlow&Solid
software for different elements of the
casting changes from 0.976 to 0.989
(Fig. 5).
o
Temperature, C
1320
1280
1240
1200
1160
1120
0
0.5
1
1.5
2
Time, min
1
2
2.5
3
3
Fig. 2. Calculated temperature change in
the different elements of solidifying
casting. Designation of curves is presents
in Fig. 1.
253
o
Calculated temperature, C
1230
of shrinkage lets to eliminate test castings
and to find optimal dimensions of a riser
for the lowest metal consumption.
1210
5. ACKNOWLEDGEMENTS
1190
Simulations
with
NovaFlow&Solid
software were made in CAD/CAE
Laboratory of Materials Technology at
Poznan University of Technology as a part
of WP4 project “Virtual prototyping in
foundry. Application of simulations codes
in concurrent engineering”.
1250
1170
1140
1170
1200
1230
1260
o
Experimental temperature, C
o
Calculated temperature, C
Fig. 4. Correlation between experimental
temperatures and temperatures simulated
by mathematical model.
6. REFERENCES
1.
1250
1230
2.
1210
1190
1170
1160 1180 1200 1220 1240 1260
Experimental temperature, oC
3.
Fig. 5. Correlation between experimental
temperatures and temperatures simulated
with NovaFlow&Solid software.
4.
5.
Simulation results of our mathematical
model very less differs from the results of
NovaCast’s NovaFlow&Solid software,
accordingly created mathematical model
can be used for solidification simulation of
a casting.
6.
7.
4. CONCLUSIONS
A program for calculation of solidification
and volume changes in a casting was
compared with well known simulation
program NovaFlow&Solid. The results
showed high reliability of created program.
This
program
helps
to
change
technological conditions of casting, and to
simulate solidification of a separate casting
and volume shrinkage of metal. Simulation
Karsay, S. J. Ductile Iron Production
Practices. Am. Foundrym. Soc. Inc.
Des Plaines, Ill., 1994.
Swaminathan, C. R., Voller, V. R.
Towards a general numerical scheme
for analysis of solidification systems.
International Journal of Heat and
Mass Transfer, 1997, 40, 2859-2868.
Denisov, V. A., Kosteneckijj, S. V.,
Oshurkov, A. T., etc. Calculation of
risers of steel castings using ECM.
Foundry, 1988, 10, 17-18 (in Russian).
Ignaszak, Z. Virtual Prototyping in
Foundry. Wydawnictvo Politechniki
Poznanskiej, Poznan, 2002 (in Polish).
Bočkus, S., Venckūnas, A., Žaldarys,
G. Simulation of a shrinkage cavity in
the risers of ductile iron castings.
Materials Science (Medžiagotyra),
2005, 11, 19-23.
www.novacast.se
Bočkus, S., Dagys, V., Venckūnas, A.,
Žaldarys, G. Density and linear
contraction of cooling ductile iron.
Materials Science (Medžiagotyra),
2001, 7, 252-255.
8. CORRESPONDING ADDRESS
Dr. Antanas Čiuplys
Kaunas University of Technology
Kęstučio 27, LT-44025 Kaunas, Lithuania
E-mail: antanas.ciuplys@ktu.lt
254