Heat and Mass Transfer (2019) 55:1959–1967

https://doi.org/10.1007/s00231-018-2447-7

ORIGINAL

Simulation of casting process: case study on the gating and feeding

design for outlet diaphragms of iron ore ball mill
Mojtaba Javahery 1 & Majid Abbasi 2

Received: 9 February 2018 / Accepted: 6 August 2018 / Published online: 24 September 2018
# Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract
In this paper, the sand casting process was evaluated using ProCAST simulating software and practical experience
for manufacturing of the high chromium stainless steel outlet diaphragms used in iron ore ball mill. In the simu-
lation practice, the effects of feeding and gating designs on the liquid flow, heat transfer and solidification phenom-
enon were evaluated to minimize the liquid surface turbulence during filling of the mold cavity and the shrinkage
porosity formation during the liquid solidification in the mold. The effects of different gating ratios and number of
feeders were numerically studied using different indicators in the software such as linear velocity, niyama criteria,
fraction solid and shrinkage porosity. Further, the optimal design obtained by simulating process was empirically
examined. The simulation results showed that the 1:3:4 gating ratio is the proper gating system that it obtains an
unpressurized fluid flow without any surface turbulent of the liquid steel in the filling step. In addition, the
simulation of solidification phenomenon was indicated that four feeders are required to eliminate of the shrinkage
porosities. The empirical casting examination shown the optimal design by the simulation is proper for manufactur-
ing of the steel part without critical defects.

Nomenclature 1 Introduction
Ac Chock (cm2)
m Weight of Steel Part (gr) Casting is the exclusive process for manufacturing of the
t Filling Time (s) complex steel parts. Contrary to the simple appearance,
he Effective Height of Part (cm) designing of the process is extremely complex for the
g 981 (cm
s2 ) production of high performance components. Many of
H Height of Sprue (cm) casting parameters have significantly effects on the ap-
hc Height of Part (cm) pearance of different defects and variations of mechanical
S Thickness of Part (cm) properties. The chemical composition, size and shape of
G Weight of Steel Part (gr) cast parts, it’s molding and melting technologies are key
μ Casting Factor for Steel (0.3–0.4) parameters for casting design. Nevertheless, all casting
ρ Density (pmgrffiffiffiffiffiffiffiffi
3 ) process have concluded two main experiences or con-
Ny Niyama K:Sec cm: cepts: filling and feeding [1, 2]. In the filling practice
k
G Temperature Gradient (cm ) (according to fluid dynamics of liquid metals), the gating
L Cooling Rate system composed of sprue, runner and gates, is designed
to guide the liquid metal for whole filling of mold cavity
with a proper velocity, lowest surface turbulent and in a
short period [3]. On the other hand, in the feeding practice
* Majid Abbasi (according to the heat transfer and mold dynamics), the
[email protected] feeder (or riser) is designed for compensation of the con-
traction of the liquid metal in mold cavity that takes place
1
Babol Noshirvani University of Technology, Babol, Iran on freezing and solidification in order to prevention of the
2
Department of Materials Engineering, Babol Noshirvani University shrinkage defects. The feeding is the long and slow pro-
of Technology, Babol, Iran cess that takes minutes or hours depending on the size of
1960 Heat Mass Transfer (2019) 55:1959–1967

Fig. 1 Image of a used outlet

diaphragm and b) inner image
from outlet diaphragms in ball
mill during PM

(a) used outlet diaphragm in position X and W (b)

the casting part. Therefore many of the casting defects In this study, a casting process is carried out followed by
such as film oxides, non-metallic inclusion, gas porosities, through ProCAST simulation in combination with the experi-
shrinkage porosities and hot tears can be related to mental of outlet diaphragms used in iron ore ball mill. The main
pouring and solidification conditions of the liquid metal objectives of the present work are: a) to effect of position and
in the mold cavity [4]. number of feeder and evaluation of these designs in numerical
Nowadays, with the development of computer science simulation and experimental condition. b). to compare
technology, it is possible to solve and predict many prob- ProCAST results such as solid phase fraction percentage,
lems in the casting processes. These improvements save shrinkage porosity and niyama criterion to experimental results
time and money, as well as provide the optimal design for c) to analyze whether effect on melt flow behavior in different
casting. Generally, foundries usually optimize casting pro- gating system ratios were simulated by ProCAST software, d)
cess depending on experiences. However, it is difficult to to evaluate the new concept in gating design developed by Prof.
know the metal flow and temperature distribution during John Campbell (unpressurized gating system for controlling of
mold filling and solidification, which are undoubtedly im- surface turbulence of the liquid metal) [1, 2].
portant to obtaining high quality castings. For this reason,
to select a production method of a casting part had to test
in different methods and design for choosing an appropri-
ate production method. In fact, this continuous operation 2 Experimental
is called trial and error, in practical condition; it will cause
loss of money, materials, time and energy. But, by using 2.1 Materials and characterization
simulation techniques for optimization casting conditions,
it is possible for a specific casting part the simulations run Figure 1 shows a used outlet diaphragm and inner image
to be performed at the least amount of time, cost, and from outlet diaphragms network in iron ore ball mill (dur-
energy to select the optimal design [5–7]. ing productivity maintenance (PM) at Golgohar Mining
A number of casting simulation programs is available and Industrial Company in Iran). The low carbon-high
t o d ay, s u c h a s M A G M A , N o v a C A S T, P r o C A S T, chromium steel (called FMU 27) was used for production
SUTCAST and SolidCAST. The simulation ProCAST soft- of the outlet diaphragms. Table 1 presents the chemical
ware by solving equations based on finite element method composition of FMU 27.
(FEM), will be able to solve complex geometries and allows The steelmaking was performed in 1000 kg high frequency
the user to simulate heat transfer [8], fluid flow [9], stress induction furnace accordance. Molding and casting was per-
conditions [10], microstructure [11] and shrinkage porosity formed using the silica sand and silicate sodium as binder
[12, 13]. There have been researches in this field, Wang et (called CO2 sand). The pouring of the liquid steel was per-
al. [14], use the niyama criteria for Shrinkage porosity and formed at 1620 °C in 7 s.
optimized design of a large forging steel ingot, Li et al. [15],
prediction of micro porosity defect in sand cast magnesium
alloy by using ProCAST result. Sabau [16], evaluation the Table 1 Chemical composition of the steel (wt.%)
shrinkage porosities of step block in investment casting
C Si Mn P S Cr Mo
A356 alloy. Panchal et al. [17], studied on design of gating
system for casting of ball valves with numerical simulation 0.22 0.70 0.62 0.02 0.02 11.00 0.20
and could increase the casting yield by 8%.
Heat Mass Transfer (2019) 55:1959–1967 1961

Fig. 2 Schematic illustration of

the outlet casting parts

W1 W2

2.2 Feeder design relationships (1), (2) and (3) were used to calculate the dimen-
sions of the system [4].
In this study, the number and position of feeders in the casting m
process were selected for simulation of solidification and Ac ¼ pffiffiffiffiffiffiffiffiffiffi ð1Þ
ρ:t:μ 2ghe
predication of shrinkage cavities. Figure 2 shows the scheme
of outlet diaphragms of iron ore ball mill according to the hc
number and position of feeder (designs of W1 and W2). The he ¼ H− ð2Þ
2
first part designed with two feeders (W1) (preliminary design) pffiffiffiffi
and the second part designed with four feeders (W2). The t¼S G ð3Þ
dimensions and sizes of feeders were constant and equal in
both parts. Weight of each part (with feeders) in order for W1 2.4 Numerical simulation
was 164 kg and W2 was 201 kg.
The mold filling and solidification processes have been sim-
2.3 Gating system design ulated using the ProCAST 2016 software. The Details of the
numerical model are shown in Table 3. Input mesh data, step
In the another part of this research, the effect of different number and time step provided for W1 and W2 models. Step
gating systems ratios on the fluid dynamics of molten metal number Specifies the number of time steps to take in the cur-
rent run and ProCAST will terminate the run when it reaches
and the condition of filling the mold were investigated. Figure
3 shows the X part was designed in defferent gating system this limit. Time step consists of 3 steps, the initial time step
ratios. For this purpose, four different gating systems ratios size, the maximum time step size (which will be used during
the filling stage only) and the maximum time step size. In this
were used. The bottom gating system was selected for all of
designs. In the each design, the certain proportion between the research, their value was determined by the software default
cross sections of the components of a gating system (called 0.001, 0.1 and 1, respectively.
The main inputs including the part design, feeders, and
SRG ration) was used for the gating design. The smallest
cross-section in a gating system is called a Chock that deter- gating system, thermo-physical properties of material such
as; density, specific heat, and thermal conductivity, de-
mines the filling time. Table 2, shows the SRG ratios were
scription of boundary conditions and import of process
used in this research. For designing of gating system,
parameters such as pouring rate, time and temperature.
Figure 4 shows the geometry and boundary conditions

Table 2 The Indicators characteristics of gating system in different ratio

Gating System Ratio: Gate Cross Section of Flow Type in

(SRG Ratio)* Numbers Sprue (cm2) Gating System

1: 0.9: 0.7 2 16 Pressurized

1: 2: 1.5 2 16 Unpressurized
-pressurized
1: 3: 4 4 16 unpressurized
5: 3: 1 2 80 pressurized

Fig. 3 Schematic illustration of X part was design for gating system * S: Sprue; R: Runner and G: Gate
1962 Heat Mass Transfer (2019) 55:1959–1967

Table 3 the detail of numerical models Table 4 Information about the parameters used in the simulation

part No. of No. of Solid No. of Step No. Row Index Amount
Nodes (Tetrahedral Material (50000)
elements) 1 Pouring Temperature (°C) 1620
2 Mold Temperature (°C) 25
W1 200,541 991,311 26 6460
3 End Simulation Temperature (°C) 1410
W2 339,047 1,718,099 28 7800
4 Liquidus Temperature (°C) 1490
5 Solidus Temperature (°C) 1420
6 Mass Flow Rate (kg/s) 25
of the model were used in this study. The finite element 7 [18] Heat transfer coefficient between 500
molten metal and mold (W/m2K)
mesh of the mold (30 mm) and part (10 mm) and feeders 8 [18] Heat transfer coefficient between 50
and gating systems (15 mm) were selected. The mesh was molten metal and sleeve (W/m2K)
selected based on several mesh refinements. The casting 9 [18] Heat transfer coefficient between 500
parameters were the same as those in the simulation. As it molten metal and cores (W/m2K)
10 [18] Heat transfer coefficient between 200
can be seen, sleeves were assigned over the molten metal mold and cores (W/m2K)
surface at the hot top. Other information and simulation 11 [18] Film Coefficient (W/m2K) 10
related parameters are shown in Table 4. 12 [18] Emissivity 0.5
In ProCAST user manual, to analyze the porosity in a
casting, several options are available: Temperature field
with the cut-off option, fraction of solid field with the cut-  
off option, shrinkage porosity field, niyama criteria, hot spot  T upper −T lower 

L¼  ð6Þ
and Specific RGL criteria. The RGL function is able to t upper −t lower 
calculate the solidification rate (R), the temperature gradient
(G) and the cooling rate (L). In ProCAST, there are ways to
For calculation of the niyama criterion, suggested values
combine the R, G and L variables together so as to obtain a
are [18]:
combined criterion function (called Mapping factor) as fol-
lows. The values of the coefficients a = 1, b = 0, c = 1 and
d = −0.5 were considered [14, 19]: LUpper Temp ¼ T liquidus þ 2 ð7Þ
M ¼ aR G L b c d
ð4Þ
LLower Temp ¼ T solidus ð8Þ
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
s   2  2ffi
∂T 2 ∂T ∂T  
G¼ þ þ ð5Þ
∂x ∂y ∂z R; G Temp ¼ T solidus þ 0:1  T liquidus −T solidus ð9Þ

Fig. 4 The geometry and boundary conditions

Heat Mass Transfer (2019) 55:1959–1967 1963

Fig. 5 Distribution of simulated solid fraction during the filling and solidification of the W1 part with section cut and time simulation (sec)

3 Result and discussion One of the parameters for evaluating the porosity loca-
tion is the percentage of solid phase fraction. The fraction of
3.1 Solidification evaluation solid and liquid metal in the mushy zone is a function of
time and temperature [20]. In Fig. 6, this criterion represents
Figure 5 shows the filling and solidification sequences of the percentage of solid phase formation with decreasing
the W1 part with the melt flow rate of 25 kg/s. The pictures temperature. As shown in this figure, the area shown with
demonstrate the filling and solidification process. In the the black arrow is related to the presence of feeders, and the
thermal simulation, solidification starts from the edges and areas shown with the blue strip in the W1 are related to the
sharp corners and moves towards the center of the part, and formation of hot areas. As seen, hot areas are formed in
finally the solidification ends in the feeding. Use of insula- areas that are far from the feeders and then the feeding op-
tion around the feeders and the addition of exothermic ma- eration in those areas is not well done. It is anticipated, leads
terials at the top of the feeder were considered for the for- to a delay in solidification and creating shrinkage porosity at
mation of directed solidification. The simulation time for these points [21]. In W1 part it is expected that these prob-
W1 5666 s. and W2 6197 s. were obtained. lems will be overcome by adding feeders.

Fig. 6 The results of solid

fraction calculations and
prediction of hot forming place in
same time simulation (100 s)

W1 W2
1964 Heat Mass Transfer (2019) 55:1959–1967

Fig. 7 Results of a shrinkage W1 W2 Character

porosity, b niyama and c [%]
temperature gradient index for W1
and W2 models

(a)
√

(b)

(c)

Figure 7 shows the results of the shrinkage porosity, indicates that the probable hot areas are prone sources for
niyama and temperature gradient criterions for models. In forming shrinkage porosity [1]. The niyama criterion
shrinkage porosity criterion pictures, the probable areas with according to Eq. (10) is defined as the local thermal gradient
their probability of formation are shown. As shown in the divided by the square root of the local cooling rate [14, 22]:
figure, the formation areas of porosity with probability 3%
are specified. The results of the porosity criterion are equal G
Ny ¼ pffiffiffi < Constant Temperature ð10Þ
to the results of the solid phase fraction percentage and L
Heat Mass Transfer (2019) 55:1959–1967 1965

resulting niyama values. When the constant temperature is

low, the heat transfer or gradient temperature is decreases
and so probability of formation the shrinkage porosity is in-
creases. The constant temperature is dependent on the material
and size of the castings.
Figure 8 shows the shrinkage porosity in parts in the prac-
tical conditions for W1 part. As shown in the simulation re-
sults, in areas where feeders are far away or feeding are not
well performed, the shrinkage porosities defects are formed at
the surface and below the surface. Therefore, it is expected
that these defects will be eliminated by increasing the number
of feeders in these regions, which is shown in simulation re-
sults that this defect is not observed in W2 part and the prob-
ability of formation of shrinkage porosities defects decreases
from 40% to less than 2%.

3.2 Gating system evaluation

In the present case, we compare the metal flow simulation

results of the parts at different gating systems. One of the
requirements for good operation of a gating system is that,
Fig. 8 Shrinkage porosity in W1 part in practical conditions
the cross section of the sprue, runner and gates should have
an appropriate ratio among themselves [2]. This ratio is called
The niyama criterion is measured near the end of solidifi- gating system ratio and it is expressed as AS: AR: AG. AS is
cation, when shrinkage porosities form. The choice of niyama cross section of sprue, AR is cross section of runner and AG is
evaluation temperature can significantly influence the summed cross sections of the gates [23, 24].

Fig. 9 The images obtained from ProCAST software and the analysis of the speed and molten flow into the mold for different ratios of the gate systems
1966 Heat Mass Transfer (2019) 55:1959–1967

1: 2: 1.5 1 : 0.9: 0.7

5: 3: 1 1: 3: 4
Fig. 10 Disturbance of melt flow with white vector in the mold for all ratios after 40% filled

The gating systems ratios are divided into two categories: the walls, which can lead to corrosion, abrasion. With this
pressure and unpressurized system. If the total level of the parameter, melt motion can be optimized in different parts
cross sections of the gates is less than the cross section of area. In the unpressurized system 1: 3: 4, the molten metal is
the sprue, the system called pressure system (such as 1: 3: 4 entering into the mold by four channels, which reduces the
ratio) and if the opposite of the above mode is present, the vector density in this section, but it is not seen in other sys-
system called unpressurized. So the cross sections of the gates tems, especially in pressure systems.
are high in unpressurized systems [9]. In Table 5, the other results obtained from the simulation
In Fig. 9, shown the results obtained from ProCAST soft- of different gating system ratios are shown. The filling time
ware consist of analyses of the melt speed flow into the mold is slightly longer for the casting with the unpressurized ratio
in Y vector for all ratios. As shown in the images, as the cross gating system. It means that the filling velocity is slower
sections of the gates are increases, the speeds of the melt are than that of the pressure ratio. The faster filling velocity of
decreases. Also, with increases in cross sections of the gates, the molten metal will increase the gas involvement and
the filling state of the molten metal is stable when compared to nonstationarity and result in porosity defects and surface
the other gating systems [25]. Figure 10 shows the disturbance enters the oxide [2, 24]. Also, in the pressure gating system,
of melt flow with white vector in the mold. The higher density there is always a pressure behind the melt, but in unpressur-
of the vectors shown the more pressure of the molten metal to ized gating system, the pressure of the melt is taken over at

Table 5 The results obtained

from the simulation of different Gating Time Filling Melt Pressure Melt speed
gating system ratios System Ratio to Solidus (sec) Time (sec) in Gate (bar) in Gate (ms )

1: 0.9: 0.7 5320 6.6 1.1 2

1: 2: 1.5 5230 6.9 1 0.9
1: 3: 4 5266 7 0.9 0.6
5: 3: 1 5477 5.4 1.1 1.5
Heat Mass Transfer (2019) 55:1959–1967 1967

the sprue section. In this case, the gas will not be easily optimization. Metall Mater Trans B Process Metall Mater Process
Sci 43(3):609–626. https://doi.org/10.1007/s11663-011-9617-z
involved in the molten metal, so the porosity defects can
8. Zhang L, Tan W, Hu H (2016) Determination of the heat transfer
be avoided. However, it has been reported that at speeds coefficient at the metal–sand mold interface of lost foam casting
above 1 (m/s), surface corrosion is observed in the molten process. Heat Mass Transf 52(6):1131–1138. https://doi.org/10.
metal, which causes the surface enters the oxide to the mol- 1007/s00231-015-1632-1
ten metal. 9. Wenming J, Zitian F (2014) Gating system optimization of low pres-
sure casting A356 aluminum alloy intake manifold based on numer-
ical simulation. China Foundry Research & Development 11(2):119–
124 https://doaj.org/article/041f71a010144dd58d4a2f4214967235
4 Conclusion 10. Mingguang W, Yong P (2017) Numerical simulation and analysis
of hot cracking in the casting of fork. Rare Metal Mater Eng 46(4):
946–950. https://doi.org/10.1016/S1875-5372(17)30127-3
The different gating and feeding designs were evaluated for
11. Szeliga D, Kubiak K, Burbelko A, Motyka M, Sieniawski J (2014)
casting of the high chromium stainless steel outlet diaphragms Modeling of directional solidification of columnar grain structure in
using ProCAST simulating software. Then, the optimal design CMSX-4 nickel-based Superalloy castings. J Mater Eng Perform
was empirically examined. It can be concluded that: 23(3):1088–1095. https://doi.org/10.1007/s11665-013-0820-8
12. Kang M, Gao H, Wang J, Ling L, Sun B (2013) Prediction of
microporosity in complex Thin-Wall castings with the dimension-
1. ProCAST simulating software is useful for predication of less Niyama criterion. Materials 6(5):1789–1802. https://doi.org/
surface turbulence and shrinkage porosities at different 10.3390/ma6051789
gating and feeding designs. 13. Zhang C, Bao Y, Wang M, Zhang L (2016) Shrinkage porosity
2. The simulation results showed that the unpressurized fluid criterion and its application to a 5.5 ton steel ingot. Archives of
Foundry Engineering 16. https://doi.org/10.1515/afe-2016-0021
flow by 1:3:4 gating ratio is the proper gating system that
14. Wang J, Fu P, Liu H, Li D, Li Y (2012) Shrinkage porosity criteria
obtains mold filling without any surface turbulent. and optimized design of a 100-ton 30Cr2Ni4MoV forging ingot.
3. The simulation of solidification steps was indicated that Mater Des 35:446–456. https://doi.org/10.1016/j.matdes.2011.09.
four feeders are required to eliminate of the shrinkage 056
porosities. 15. Li J, Chen R, Ma Y, Ke W (2014) Characterization and prediction
of microporosity defect in sand cast WE54 alloy castings. J Mater
4. The empirically studies indicates that the selected gating Sci Technol 30(10):991–997. https://doi.org/10.1016/j.jmst.2014.
and feeding designs has good agreement by the simula- 03.011
tion results. 16. Sabau AS (2006) Alloy shrinkage factors for the investment casting
process. Metall Mater Trans B Process Metall Mater Process Sci
Acknowledgements This research is financial supported by Gol-Gohar 37(1):131–140. https://doi.org/10.1007/s11663-006-0092-x
Iron Ore and Steel Research Institute. In addition, the authors greatly 17. Panchal P, Joshi S, Ghetiya N (2015) Design and analysis of gating
appreciate Babol Noshirvani University of Technology and Foolad and Risering system for casting of ball valves. Nirma University
Tanbour Company (Foundry industry) for their assists. Journal of Engineering and Technology 4(1):1–5
18. ProCast 2016 User's Guide, ESI group. www.esi-group.com
Publisher’s Note Springer Nature remains neutral with regard to jurisdic- 19. Fu PX, Kang XH, Ma YC, Liu K, Li DZ, Li YY (2008) Centrifugal
tional claims in published maps and institutional affiliations. casting of TiAl exhaust valves. Intermetallics 16(2):130–138.
https://doi.org/10.1016/j.intermet.2007.08.007
20. Lakum PM, Gondaliya RD, Chauhan HN (2016) Shrinkage
Porosity & Temperature Distribution analysis of air cooled cylinder
References liner block of Grey cast iron - ASTM class 35. Internation Journal
Of Advance Research And Innovative Ideas In Education 2(3):
1. Campbell J (2003) Castings, 2nd ed. Butterworth-Heinemann, 569–577
Oxford 21. Tavakoli R (2014) On the prediction of shrinkage defects by ther-
2. Campbell J (2004) Castings practice: the ten rules of castings. mal criterion functions. Int J Adv Manuf Technol 74(1):569–579.
Butterworth-Heinemann [Imprint] Elsevier Science & Technology https://doi.org/10.1007/s00170-014-5995-0
Books Available via http://worldcat.org. http://public.eblib.com/ 22. Carlson KD, Beckermann C (2009) Prediction of shrinkage pore
choice/publicfullrecord.aspx?p=226698 volume fraction using a dimensionless Niyama criterion. Metall
3. Jezierski J, Dojka R, Janerka K (2018) Optimizing the gating sys- Mater Trans A 40(1):163–175. https://doi.org/10.1007/s11661-
tem for steel castings. Metals 8(4):266–279 008-9715-y
4. ASM International (2008) ASM Handbook Volume 15: Casting. ASM 23. Nimbulkar SL, Dalu RS (2016) Design optimization of gating and
International feeding system through simulation technique for sand casting of
5. Bahmani A, Baghani A, Mirbagheri MH, Davami P (2017) wear plate. Perspect Sci 8:39–42. https://doi.org/10.1016/j.pisc.
Simulation of casting processes. Amirkabir University of 2016.03.001
Technology, Tehran (in persian) 24. Masoumi M, Hu H, Hedjazi J, Boutorabi MA (2005) Effect of
6. Chudasama BJ (2013) Solidification analysis and optimization gating design on Mold filling. AFS Transactions 05-152(2):1–12
using pro-cast. International Journal of Research in Modern 25. Jezierski J, Dojka R, Kubiak K, Zurek W (2016) Experimental
Engineering and Emerging Technology 1(4):9–19 approach for optimization of gating system in castings. Paper pre-
7. Kotas P, Tutum CC, Thorborg J, Hattel JH (2012) Elimination of sented at the in proceedings of the 25th anniversary international
hot tears in steel castings by means of solidification pattern conference on metallurgy and materials, Brno