International Journal of Advance Research In Science And Engineering http://www.ijarse.

com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)

INFLUENCE OF SAND MOLDING PROCESS

PARAMETERS ON PRODUCT QUALITY OF AL-SI
ALLOY CASTING - AN ANOVA APPROACH
M. Viquar Mohiuddin1, A. Krishnaiah2, S. Ferhathullah Hussainy3
1, 3
Mechanical Engineering Department,
Muffakham Jah College of Engineering & Technology, (India)
2
Mechanical Engineering Department, Osmania University, (India)

ABSTRACT
Sand mold casting is an age old metal casting process. Even today, though there are many advanced techniques
present to produce metal castings, sand molding remains one of the most widely used casting process because of
its ability to produce low cost and wide variety of castings. Sand casting process involves many parameters
which affect the quality of the produced castings. In this study, an attempt has been made to optimize the sand
molding parameters, and to study the influence of these parameters on quality of Al-Si alloy castings. Taguchi’s
experimental design approach is adopted to achieve optimum level of parameters, and Analysis of Variance
(ANOVA) technique is used to identify the significant factors and their percentage contribution for improved
mechanical properties. Experimental results show that amount of clay and amount of moisture are the
parameters making significant influence on mechanical properties of these castings. Based on optimum level of
parameters, verification test is performed and the results are found to be in confidence level. Experimental
results confirmed the validity of selected approach in enhancing and optimizing the sand mold casting process
for aluminium alloys.

Keywords: Al-Si alloy, Analysis of Variance, Product quality, Sand casting, Taguchi design

I. INTRODUCTION

Aluminum is the third most abundant element in the earth's crust. After iron and steel, the most popular metal
used is aluminium. It is widely used in transportation, defense, construction, aerospace, maritime, domestic, and
general engineering purpose. This is because of its unusual combinations of properties such as light weight (high
strength to weight ratio), resistance to corrosion, good cast-ability, good machinability, chemical resistance,
excellent conductor of heat and electricity, etc. Pure aluminium is soft having comparatively poor casting
features and little strength, that's why aluminium castings are prepared from aluminium alloys. The main
alloying elements are silicon, copper, magnesium, zinc, etc. Fluidity of aluminium increases with silicon
addition. Al-Si alloys have good casting and corrosion resistance properties. In transportation industry, more use
of these alloys decreases weight of vehicle, improve its performance and reduces fuel consumption.
Sand mold casting is an important process for casting aluminium alloys. It is used to create castings of any
shape and size. The process involves, melting a metal, pouring it into a previously prepared mold cavity formed
out of natural or synthetic sand, allowing liquid metal to solidify, and breaking sand mold to get the castings.
For every single casting new mold is to be prepared, quality of mold determines the quality of produced

1751 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
castings. For this reason aluminium foundries have to focus more on optimizing on sand molding process to
improve the quality of products and reduce cost.
According to Taguchi, the key element for achieving high quality and low cost product is parameter design.
Through parameter design optimal levels of process parameters are selected and these parameters should be
controlled to improve the quality of both casting process and the product [1]. A number of automotive suppliers
have achieved quality and cost improvement through Taguchi‟s robust design. These applications include
improvement in metal casting, injection molding of plastic parts, etc. [2]. Taguchi has introduced several
statistical tools and concepts of quality improvement that depend heavily on the statistical theory of
experimental design [3]. To overcome the problems of defects, flaws and imperfections in castings, and improve
the quality of products produced, optimization of process parameters for sand mold casting should be carried out
[4].
In this study, the experiments are designed by selecting four parameters (sand grain size, amount of clay,
amount of moisture and number of ramming) all at three levels. The required number of experiments under full
factorial method was 34 = 81. Using Taguchi‟s L9 orthogonal array, the number of experiments are reduced to
nine. All these nine experiments are performed to produce test castings which are then examined for improved
mechanical properties – tensile strength and hardness.
The signal to noise (S/N) ratio for each parametric setting is computed. The optimal level of process parameters
promotes highest S/N ratio [5]. Analysis of variance (ANOVA) on the collected data from the Taguchi design of
experiments is used to select new parameter values to optimize the performance characteristic. The data from
the arrays is analyzed by performing a visual analysis. ANOVA and Fisher's exact test (F-Test) is applied to find
significant factors and test the level of significance. The percentage contribution of significant factors is also
obtained.

II. LITERATURE REVIEW

Mekonnen Liben Nekere and Ajit Pal Singh [1] has examined optimal settings of two groups of aluminum blank
sand casting processes. Single aluminum blank sand casting and double aluminum blanks sand casting for
process robustness comparison. The results have shown that single aluminum blank sand casting process is more
robust than double aluminum blank sand casting process. The experimental results confirmed the validity of
used Taguchi robust design method for enhancing sand casting process and optimizing the sand casting
parameters in aluminum blank casting process.
Lakshmanan Singaram [6] has studied the analysis of green sand process parameters strength, moisture content,
permeability, mold hardness using Taguchi method and ANN Analysis. Outcome is optimized green sand
process parameters which lead to improved process performance, reduced process variability and thus minimum
casting defects.
Rasik A Upadhye and Dr. Ishwar P Keswani [7] has studied the sand casting process parameters of the castings
manufactured in iron foundry by maximizing the signal to noise ratios and minimizing the noise factors using
Taguchi method. The process parameters considered are moisture, sand particle size, green compression
strength, mold hardness, permeability, pouring temperature, pouring time and pressure test. The results indicated
that the selected process parameters significantly affect the casting defects in the foundry.
P. Senthil and K. S. Amirthagadeswaran [8] has investigated the influence of process parameters on mechanical
properties of the castings prepared through squeeze casting process using Taguchi method. Experimental results

1752 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
showed that squeeze pressure, die preheating temperature and compression holding time were the parameters
making the significant improvement in mechanical properties.
Jhon O. OJI et. al. [9] has investigated the effect of mold temperature and pouring temperatures on ultimate
tensile strength of aluminum alloy sand castings. The result shows that mold temperature is the significant
factors which influence the casting quality.
L. Ceschini et. al. [10] has investigated the relationships between ultimate tensile strength and micro structural
parameters for the sand cast A357 aluminum alloy. Starting from the micro structural parameters and taking in
to account the material hardness, a relationship able to predict the ultimate tensile strength of the alloy was
found.

III. EXPERIMENTAL APPROACH

To find the optimum conditions of the control factors, Taguchi‟s robust experimental design methodology is
used to design the experiments. This methodology is used by applying eight experimental steps that can be
grouped into three major categories as follows [1, 2]:
 Planning the experiment
1. Identify the main function of casting process.
2. Identify the quality characteristic to be observed and objective function to be optimized.
3. Identify control factors and their alternate levels.
4. Identify noise factors and testing conditions of the process.
5. Design matrix experiment and define data analysis procedure.
 Performing the experiment
6. Conduct the matrix experiment.
 Analyzing and verifying the experimental results
7. Analyze the data, determine the optimum levels, and predict the performance under these levels.
8. Conduct the verification test (also called confirmation experiment) and plan future actions.
Experimental data can also analyzed using Analysis of Variance (ANOVA) where the relative percentage
contribution of all factors is determined by comparing with the relative variance. In ANOVA calculations, the
degree of freedom for all factors is obtained first, the values of variance for all factors are then calculated and
finally F-ratio and percentage contributions for the factors are calculated. The intention is to determine the
significant control factors and their level of significance to optimize the process.

3.1 Sand Mold Casting Process and its Main Function

Initially, two solid square cross section wooden patterns are prepared with dimensions 3cm x 3cm x 16cm to
make mold cavities in which molten metal is poured to produce castings which are used to prepare test
specimens. Molding sand is prepared using silica sand as base sand, bentonite powder as clay or binder, and
water as moisture. Nine molds are prepared by varying grain fineness of silica sand, amount of clay, amount of
moisture and number of ramming, as per the experimental design. Molds are then dried in air to remove the
moisture completely. Aluminium alloy LM25 (AlSi7Mg) is melted for pouring into the molds. Melting is done
in lift out crucible type coke fired pit furnace.

1753 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
After melting degassing is done using Hexa-chloro-ethane (C2Cl6) tablets which are added to the molten metal
which liberates stable chlorine gas to flush out the dissolved hydrogen gas by creating partial pressure in molten
metal. To protect the molten charge from getting oxidized, coveral-36 is used as a flux powder to cover the
molten metal from atmospheric gases. Thermocouples with digital temperature indicator are used to measure the
temperature of molten metal. All nine molds are poured to produce test castings. After solidification the molds
are broken to obtain the castings. Finally machining operations are performed on CNC machine to prepare
standard test specimen as per ASTM A370 standard, 12.5 mm round tension test specimen with 50 mm gauge
length, for measuring tensile strength. Universal Testing Machine is used for measuring tensile strength and
Dynamic Hardness Tester is used for measurement of Brinell Hardness.

3.2 Quality Characteristics and Objective Functions

To determine the effect each parameter has on the response, the signal-to-noise ratio, or the SN ratio, needs to
be calculated for each experiment conducted. Taguchi defines it as three categories of quality characteristics
which are: Lower-the-better, Larger-the-better and Nominal-the-best. As in this study, tensile strength and
hardness are selected as response; both are larger-the-better type of quality characteristics. The objective

function is to be maximized by using the S/N ratio: ------------ (1)

Where „n‟ is trial number, „y‟ is response and „i' is experiment number. After calculating the S/N ratio for each
experiment, the average SN value is calculated for each factor and level.

3.3 Control Factors and Their Levels

In general, sand casting process involves many parameters such as type of sand, size of sand grain, clay type and
amount, percentage of moisture, green compressive strength, shear strength, permeability, number of ramming,
shatter index, mold type, mold hardness, pouring temperature, pouring time, pouring height, metal composition,
metal fluidity, running and gating, risering and feeding, casting design, alloy type, and many more [4,7]. As this
study aims to optimize sand molding process, four key sand molding parameters – sand grain size, amount of
clay, amount of moisture and number of ramming are selected at three different levels as given in the Table 1,
keeping other parameters constant.
Table 1: Control Factors with Levels
Control factor Levels
Control factors
designation 1 2 3
X1 Sand grain size 40 55 70
X2 Amount of clay (%) 6 9 12
X3 Amount of moisture (%) 7 10 13
X4 Number of ramming 2 3 4

3.4 Noise Factors and Testing Conditions

Number of noise factors affecting the casting process was identified from literature for sand mold casting
process. Some of these factors are variation in humidity, pouring temperature, metal flow rate, ambient
temperature and so forth. For this experiment the important noise factors considered were: pouring temperature,

1754 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
metal flow rate and humidity. To capture the effects of noise factors during the casting process, different sand
mold castings were prepared and examined.

3.5 Matrix Experiment Design and Data Analysis Plan

The experiments are conducted by keeping four parameters at three levels; the number of experiments required
was 34 = 81 under full factorial method. Using Taguchi‟s orthogonal array table, for four parameters at three
levels, L9 orthogonal array is selected. With this number of experiments is reduced to nine. These nine
experiments are conducted in order to optimize the process parameters to achieve castings with less defects and
improved mechanical properties.
The standard table of combinations for experiment for L9 orthogonal array is as given below.
Table 2: Standard L9 orthogonal array table
Experiment Control Factors
Run Order X1 X2 X3 X4
1 1 1 1 1
2 1 2 2 2
3 1 3 3 3
4 2 1 2 3
5 2 2 3 1
6 2 3 1 2
7 3 1 3 2
8 3 2 1 3
9 3 3 2 1

3.6 Conducting Matrix Experiment

As per the nine rows of L9 standard orthogonal array table, nine experiments are conducted. To prevent the
translation error, the table is translated incorporating level table to create experimental design sheet as shown in
the Table 3. The responses of each experiment are tabulated in adjacent columns.
Table 3: Experimental Design Table and Responses
Control factors (or) Process parameters Responses
Experiment Sand Grain Amount of Amount of Number of Tensile Hardness
Run Order Size Clay (%) Moisture (%) Ramming Strength (MPa) (BHN)
X1 X2 X3 X4 YTS YH
1 40 6 7 2 104.5 41
2 40 9 10 3 106.8 41
3 40 12 13 4 125.7 51
4 55 6 10 4 102.9 40
5 55 9 13 2 119.2 46
6 55 12 7 3 117.7 44
7 70 6 13 3 111.1 42

1755 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
8 70 9 7 4 108.9 42
9 70 12 10 2 114.8 42

3.7 Analyzing Data, Determining Optimum Levels And Predicting Performance Under These
Levels
For analysis of results obtained from the experiment, the SN ratios are to be calculated [2]. Test samples are
prepared for each experiment, tensile strength and hardness tests are performed on each sample. The results
obtained for tensile strength and hardness are listed under responses in Table 3. Based on these responses S/N
ratio calculations and ANOVA calculations are carried out.

3.7.1 S/N Ratio Calculations [4, 11]

S/N ratio for tensile strength and hardness are calculated using „larger-the-better‟ quality characteristics for each
experiment run order by using expression (1) of Para 3.2, the calculated values are tabulated in the Table 4 as
shown below. The Table 5 gives the calculated average S/N values for each factor and each level for both tensile
strength and hardness.
Table 4: S/N ratio calculations
Experiment S/N Ratio
Run Order Tensile strength Hardness
1 40.38 32.25
2 40.57 32.25
3 41.98 34.15
4 40.24 32.04
5 41.52 33.25
6 41.41 32.86
7 40.91 32.46
8 40.74 32.46
9 41.19 32.86

Table 5: S/N ratio response table

Tensile Strength Hardness
LEVEL X1 X2 X3 X4 X1 X2 X3 X4
1 40.97 40.51 40.99 41.03 32.88 32.25 32.52 32.78
2 41.05 40.94 40.64 40.96 32.71 32.65 32.38 32.52
3 40.94 41.52 41.47 40.98 32.59 33.29 33.28 32.88
Max-Min 0.11 1.01 0.83 0.07 0.29 1.04 0.90 0.36
Rank 3 1 2 4 4 1 2 3
Optimum
X1,2 X2,3 X3,3 X4,1 X1,1 X2,3 X3,3 X4,3
level

1756 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
3.7.2 ANOVA Calculations
In order to find out the level of significance of control factors and their percentage contribution towards tensile
strength and hardness in sand mold casting process, Analysis of Variance approach is followed and ANOVA
calculations are made.

3.7.3 ANOVA Calculations for Tensile strength

ANOVA calculations are made using the experimental results data from Table 3. Sample calculation is
presented below.
Number of experiments, N = 9
Total Degrees of Freedom (DOF) = N-1 = 9-1 = 8
Sum of the response, T = 1011.6
Correction Factor, C.F. = T2/N = 113703.84
Total of response for factor X1 at level 1, X1,1 = 337
Similarly total response for all four factors at all three levels are calculated for both tensile strength and hardness
and are tabulated in Table 6.
Table 6: Level totals table
Control factor Level Total for Tensile Strength Level Total for Hardness
Control factors
designation 1 2 3 1 2 3
X1 Sand grain size 337 339.8 334.8 133 130 128
X2 Amount of clay 318.5 334.9 358.2 123 129 139
X3 Amount of moisture 331.1 324.5 356 127 125 139
X4 Number of ramming 338.5 335.6 337.5 131 127 133

Where „i‟ is experiment run order number.

Total sum of square, ST = 454.94
Number or repetitions of factor X1 at level 1, =3

Sum of squares of X1, = 4.18

Similarly SX2, SX3 and SX4 are calculated.

In this case, since there are no empty columns or repetitions, experimental and reputational error cannot be
calculated. Pooling the smallest variance factors X1 and X4 would provide some information of error so that we
could proceed with the construction of ANOVA Table.
Sum of square of pooled error, Se = SX1 + SX4 = 5.62
DOF of pooled error = DOF of X1 + DOF of X4 = 2 + 2 = 4
DOF of X3 = 2

Variance of X3, VX3 = = 132.66

Similarly VX2, VX1 and VX4 are calculated.

1757 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
Variance of error, Ve = = 1.40

Pure variation of X2, = ) = 262.52

Similarly is calculated.
Pure variation of pooled error, ] = 11.22

Percentage contribution of factor X2, = = 57.70%

Percentage contribution of pooled error, = = 2.46

Similarly ANOVA for hardness is calculated.

Table 13: ANOVA table for tensile strength

Source Pool S DOF Variance F-ratio S' P%
X1 yes 4.18 2 2.09 - - -
X2 265.32 2 132.66 94.75 262.52 57.70
X3 183.98 2 91.99 65.70 181.18 39.82
X4 yes 1.44 2 0.72 - - -
Pooled error 5.62 4 1.40 1.00 11.22 2.46
Total 454.92 454.92 100.00
From the ANOVA table for tensile strength, F-ratio for amount of clay and amount of moisture is
greater than F critical value, which is 18 at 0.01 level of significance, this makes them the significant factors.
Table 14: ANOVA table for hardness
Source Pool S DOF Variance F-ratio S' P%
X1 yes 4.23 2 2.11 - - -
X2 43.56 2 21.78 8.34 38.34 41.56
X3 38.23 2 19.11 7.32 33.01 35.79
X4 yes 6.23 2 3.11 - - -
Pooled error 10.46 4 2.61 1.00 20.9 22.66
Total 92.25 92.25 100.00

From the ANOVA table for hardness, F-ratio for amount of clay and amount of moisture is greater than F
critical value, which is 6.94 at 0.05 level of significance, this also makes them the significant factors.

3.8 Verification Test

Based on optimum level of process parameters, verification test is performed; the results obtained are:
Tensile strength: 130 MPa, and Hardness: 53 BHN.

IV. RESULTS AND DISCUSSIONS

From the above experimentation, it is observed that amount of clay and amount of moisture are the significant
factors which are affecting tensile strength and hardness of aluminum alloy sand mold castings.
Increase in the amount of clay decreases permeability, which increases porosity in castings hence reducing
mechanical properties. If amount of clay is decreased, mold hardness, shatter index and green compressive

1758 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
strength will also decrease which decreases the quality of the sand mold. Therefore optimum amount of clay for
making a sand mold is to be selected.
Increase in the amount of moisture will also decrease permeability which affects the quality of produced
castings. If more amount of moisture present in sand mold, it creates more amount gases during pouring which
in turn results in casting defects like blowholes, open holes, porosities, etc. Hence optimum amount of moisture
is to be used while making a sand mold.
From S/N ratio response table, Table 5, it is found that the optimum level of significant parameters is 12%
amount of clay and 13% amount of moisture.
Verification test conducted based on optimum level of process parameters shows that the values of tensile
strength and hardness are 130 M Pa and 53 BHN respectively. These values were found to be in confidence
level.
From ANOVA tables, it is observed that amount of clay (X2) and amount of moisture (X3) are the factors
contributing 57.70% and 39.82% for tensile strength, 41.56% and 35.79% for hardness respectively; these are
represented in the chart below.

Figure 1: Bar chart of percentage contribution

V. CONCLUSIONS

On the basis of experiments conducted and analysis carried out in this study, many conclusions can be drawn;
few of them are mentioned as follows:
 Study shows that this approach can be effectively applied in foundry shops for deciding optimum setting of
process parameters to improve the quality of the castings produced.
 Results indicated that the selected process parameters are significantly affecting the selected mechanical
properties.
 It is observed that amount of clay and amount of moisture are the significant control factors from among the
selected factors.
 Amount of clay is the major contributing factor for the improvement of tensile strength and hardness of
LM25 alloy casting.
 It is also observed that amount of clay and amount of moisture are the factors contributing 57.70% and
39.82% for tensile strength, 41.56% and 35.79% for hardness respectively.
 Verification test conducted based on optimum level of process parameters shows the increase in values of
both tensile strength and hardness. Work is in progress to adopt ANOVA approach to study the influence of

1759 | P a g e
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.4, Special Issue (01), March 2015 ISSN-2319-8354(E)
other parameters such as melting and gating design and include other mechanical properties to optimize the
process for improving the quality of Al-alloy castings.

REFERENCES

[1] Mekonnen Liben Nekere, Ajit Pal Singh, “Optimization of aluminium blanks sand casting for process by
using Taguchi Robust Design Method”, International Journal for Quality research Vol.6, No.1, 2012, 81-
97.
[2] Phadke, M.S., “Quality engineering using robust design”. Prentice Hall International, Inc., 1989.
Englewood Cliffs, New Jersey, NJ.
[3] Kiyoung KIM and Kyongwhoan LEE, “Effect of process parameters on porosity in aluminum lost foam
process”. J. Mater. Sci. Technol., Vol.21 No.5, 2005.
[4] M. Viquar Mohiuddin et. al., “Effect of Composition of Sand Mold on Mechanical Properties and Density
of Al-Alloy Casting Using Taguchi Design Approach”, International Journal of Engineering Research and
Application, Vol. 5, Issue 3, March 2015, pp.37-41.
[5] G.O. Verran et. al. “DOE applied to optimization of aluminum alloy die castings”. Journal of materials
processing technology 200(2008) 120–125.
[6] Lakshmanan Singaram “Improving quality of sand casting using Taguchi Method and ANN Analysis”
International Journal on design and manufacturing technologies. Vol.4 No.1, January 2010.
[7] Rasik A Upadhye and Dr. Ishwar P Keswani, “Optimization of Sand Casting Process Parameter Using
Taguchi Method in Foundry”, International Journal of Engineering Research & Technology (IJERT) Vol.
1 Issue 7, September – 2012.
[8] P. Senthil and K. S. Amirthagadeswaran, “Optimization of squeeze casting parameters for non symmetrical
AC2A aluminium alloy casting through Taguchi Method”, Journal of Mechanical science & Technology
26(4) (2012).1141-1147.
[9] John O. OJI et. al., “Effect of Mould and Pouring Temperatures on Ultimate Tensile strength of
Aluminium Alloy Sand Castings - An ANOVA Approach”, Leonardo Electronic Journal of Practices and
Technologies, Issue 19, July-December 2011, p. 97-108.
[10] L. Ceschini et. al., “Correlation between Ultimate Tensile Strength and Solidification Microstructure for
the Sand Cast A357 Aluminium Alloy”, Material and Design 30(2009) 4525-4531.
[11] M. Viquar Mohiuddin et. al., “Experimental Study of Sand Mold Process Parameters on Al-Alloy Sand
Casting using DoE”, ISOR Journal of Mechanical and Civil Engineering Volume 11, Issue 6, Dec. 2014,
PP 01-06.
[12] Zillur Rahman and Faisal Talib, “A Study of Optimization of Process by Using Taguchi‟s Parameter
Design Approach”, The Icfai University Journal of Operations Management, Vol. VII, No. 3, 2008.
[13] Datta, G.L., Sand and mould related casting defects, Indian Foundry Journal, 44(9), 1998, 148-154.
[14] Johnston R.E., Design of experiments: Taguchi in the Foundry, AFS Trans 82, 1989, 415–418.
[15] Taguchi G., Introduction to quality engineering: design quality into products and process, (Asian
Productivity Organization, Tokio, 1986).

1760 | P a g e