 Open Access
 Total Downloads : 8690
 Authors : Rasik A Upadhye, Dr. Ishwar P Keswani,
 Paper ID : IJERTV1IS7476
 Volume & Issue : Volume 01, Issue 07 (September 2012)
 Published (First Online): 25092012
 ISSN (Online) : 22780181
 Publisher Name : IJERT
 License: This work is licensed under a Creative Commons Attribution 4.0 International License
Optimization of Sand Casting Process Parameter Using Taguchi Method in Foundry
Rasik A Upadhye
Department of Industrial Engineering,
Shri. Ramdeobaba College of Engineering and Management, Nagpur 440013 India
Dr. Ishwar P Keswani
Professor, Department of Industrial Engineering,
Shri. Ramdeobaba College of Engineering and Management, Nagpur 440013 India
Abstract
The purpose of this paper is to optimize 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. A Taguchi approach is used to capture the effects of signal to noise ratio of the experiments depending upon the orthogonal arrays used, an analysis of variance and optimum conditions are found. This paper demonstrates a robust method for formulating a strategy to find optimum factors of process and interactions with a small number of experiments. The process parameters considered are moisture, sand particle size, green compression strength, mould 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.
The improvement expected in reduction of casting defects is found to be 37.66 percent.
Keyword(s): Casting defects; Taguchi method; Iron foundry; Control factors; Sand casting

Introduction
Quality improvement in foundry industry have been carried over by researchers and foundry engineers for robust product at the customers end by applying various optimization methods to the sand casting process parameters: the gradient search method, the finite element method (FEM) based neural network method and the Taguchi method [1, 2]. Taguchi [3, 4] has introduced several statistical tools and concepts of quality improvement that depend heavily on the statistical theory of experimental design. Some applications of
Taguchis method in the foundry industry have shown that the variation in casting quality caused by uncontrollable process variables can be minimized [5, 6].
The concept of robust projects for the process and the product introduced by G. Taguchi [7, 8] seeks to make both the process and the product insensitive to disturbing factors that occasionally or systematically affect the variability of the process that may lead to imperfections in the products.

Literature Review
Sand casting is one of the most common production technique used for manufacturing ferrous castings. Cupolas are solely used by iron foundries for continuous production of molten iron. Der Ho Wu et al [9] applied the Taguchi method to optimize the process parameters for the die casting of thinwalled magnesium alloy parts in computer, communications and consumer electronics industries. The results confirmed the effectiveness of robust design methodology. Sushil Kumar et al [10] have carried out an optimization technique for process parameters of green sand casting of a cast iron differential housing cover based on
Taguchi parameter design which indicated in determining the best casting parameters for differential housing cover. Pradeep Kumar et al [2] applied the Taguchis approach to the vacuum sealed process to obtain an optimal setting of the control factors that yielded the optimum surface roughness of the Al 11 per cent Si alloy castings. Muzammil et al [11] made a study for optimization of Gear Blank Casting Process by Using Taguchis Robust Design Technique. In this study they demonstrated that casting process involve a large number of parameters affecting the various casting quality features of the product. The reduction in the weight of the casting as compared to the target weight was taken to be proportional to the casting defects. B.
H. Kim et al [12] in their study, the relationship between casting process parameters and mechanical properties in a 14.5% Si containing corrosion resistant cast iron was statistically investigated using Taguchi method. The effects of casting process parameters on mechanical properties and corrosion resistance were further confirmed by combined analysis of fractography, hydrogen content determination, microscopic test and acid resistance test. Ballal Yuvraj P. et al [13] describes that in order to produce any product with desired quality proper selection of process parameters is essential. This paper describes use and steps of Taguchi design of experiments and orthogonal array to find specific range and combinations of turning parameters. Quality achieved by means of process optimization is found to be cost effective in gaining and maintaining a competitive position in the world market. Dr. M. Arasu [14] the approach taken in this paper expects the foundries to use a standard classification system to describe undesirable casting artifacts for more effective failure analysis. It deals the various aspects of a systematic approach to understanding and development of quality cost system in cast iron foundries. A. Noorul Haq.et al [15] in their study demonstrates optimization of CO2 casting process parameters by using Taguchis design of experiments method. The effect of the selected process parameters on casting defects and subsequent setting of the parameters are accomplished by using Taguchis parameter design approach.

Methodology
The objective of this paper is on optimizing the process parameters of sand casting process including optimum levels and the case study is done in a job foundry in central India. The Taguchi method can be applied by using eight experimental steps that can be grouped into three major categories as follows [8]. :

Planning the experiment: (1) Identify the main function of casting process. (2) Identify the quality characteristic to be observed and the objective function to be optimized. (3) Identify the control factors and their alternate levels. (4) Identify noise factors and the testing conditions of the process. (5) Design the matrix experiment and define the data analysis procedure.

Performing the experiment: (6) conduct the matrix experiment.

Analyzing and verifying the experimental results: (7) Analyzing the data, determining the
Optimum levels for the control factors, and predicting performance under these levels: (8) Conducting the verification (also called confirmation) experiment and planning future actions.
The basic steps for achieving the above target are summarized below [16, 17]

To select the most significant parameters that causes variations in the quality characteristics.

Casting defects have been selected as the most representative quality characteristics in the green sand casting process, as it is related to many internal defects (sand blow holes, pinholes, scabs, metal penetration, mold shift, mold crack, sand drop.). The target of the green sand casting process is to achieve lower casting defects while minimizing the effect of uncontrollable parameters.

Make the green sand casting process under the experimental conditions dictated by the chosen orthogonal array and parameter levels. Based on the experimental conditions, collect the data.

An analysis of variance (ANOVA) table is generated to determine the statistical significance of the parameters. Response graphs are plotted to determine the preferred level for each parameter.

Beside the optimum settings of the control parameters and predict the results of each of the parameters at their new optimum levels

Verify the optimum settings result in the predicted reduction in the casting defects.
Sand casting is used to manufacture complex shapes of various sizes depending upon the customer requirements. The basic requirements casting are pattern making, preparing a mold, pouring a molten metal, cooling of mold, shakeout, fettling. The main causes of rejection in castings are due to improper pattern, improper gating system, improper control of sand parameters, improper molten metal composition. The process parameters of the sand casting can be listed as follows:

Sand particle size

Moisture percentage in sand

Green compression strength

Mould hardness

Permeability

Pouring temperature of molten metal

Pouring time of molten metal in mold

Pressure test

For each process parameter two/three levels are selected which define the experimental region. The levels selected are based on the standards acceptable and foundry men experience in this organization for engine castings and fittings. Significant interactions within control parameters are also considered. The parameters, along with their ranges are given in Table 1.
Table 1.Control factors of process parameters and their levels
Factor Control factors Range Level 1 Level 2 Level 3
designation
A 
Moisture (%) 
3.5 4 
3.5 
4 
— 
B 
Sand particle size (AFS) 
50 55 
50 
53 
55 
C 
GCS (g/cm^2) 
900 1200 
900 
1100 
1200 
D 
Mould hardness (nu) 
50 80 
50 
70 
80 
E 
Permeability (nu) 
150 220 
150 
185 
220 
F 
Pouring temperature (deg c) 
1300 1420 
1300 
1390 
1420 
G 
Pouring time (sec) 
20 28 
20 
24 
28 
H 
Pressure test (MPa) 
1.5 2.5 
1.5 
2 
2.5 
3.1 Quality Characteristics
Casting defects was selected as a quality characteristic to be measured. The most common defects occurring in the foundry were monitored and recorded. The smaller the better number of casting defect implies better process performance. Here the objective function to be maximized is:
= 10 log
/ ( ) = 10log(yi^2)/n
Maximizing leads to minimization of quality loss due to defects. Where S/N ratio is used for measuring sensitivity to noise factors, n is the number of experiments orthogonal array, and yi the ith value measured.

Selection of Orthogonal Array
Selection of an orthogonal array depends upon the number of control factors and interaction of interest. It also depends upon number of levels for the control factors of interest. Therefore with one control factor moisture percentage of two levels and other control factors sand particle size, green compression strength (GCS), mould hardness number, permeability number, pouring temperature of molten metal, pouring time of molten metal in mould and pressure test for casting leakage with three levels are considered, L18 orthogonal array is selected with 18 experimental runs and eight columns. Taguchi has provided in the assignment of factors and interaction to arrays. The tools are: (1) the linear graph and (2) triangular tables. Linear graphs indicate various columns to which factors may be assigned and the columns subsequently evaluate the interactions of those factors. The assigned L18 orthogonal array is shown in Table 2 and the experimental orthogonal array having their levels are assigned to columns are shown in Table 3.
Table 2.Orthogonal array L18 (control factors assigned)
Trial no. A B C D E F G H
1
1
1
1
1
1
1
1
1
2
1
1
2
2
2
2
2
2
3
1
1
3
3
3
3
3
3
4
1
2
1
1
2
2
3
3
5
1
2
2
2
3
3
1
1
6
1
2
3
3
1
1
2
2
7
1
3
1
2
1
3
2
3
8
1
3
2
3
2
1
3
1
9
1
3
3
1
3
2
1
2
10
2
1
1
3
3
2
2
1
11
2
1
2
1
1
3
3
2
12
2
1
3
2
2
1
1
3
13
2
2
1
2
3
1
3
2
14
2
2
2
3
1
2
1
3
15
2
2
3
1
2
3
2
1
16
2
3
1
3
2
3
1
2
17
2
3
2
1
3
1
2
3
18 2 3 3 2 1 2 3 1
Table 3.Experimental orthogonal array
Trial no.
A
B
C
D
E
F
G
H
Moisture
Sand
Green
Mould
Permeability
Pouring
Pouring
Pressure
content
particle
Strength
hardness
(nu)
temperature
time
test
(%)
Size
(g/cm^2)
(nu)
(deg c)
(sec)
(MPa)
(AFS)
1
3.5
50
900
50
150
1300
20
1.5
2
3.5
50
1100
70
185
1390
24
2
3
3.5
50
1200
80
220
1420
28
2.5
4
3.5
53
900
50
185
1390
28
2.5
5
3.5
53
1100
70
220
1420
20
1.5
6
3.5
53
1200
80
150
1300
24
2
7
3.5
55
900
70
150
1420
24
2.5
8
3.5
55
1100
80
185
1300
28
1.5
9
3.5
55
1200
50
220
1390
20
2
10
4
50
900
80
220
1390
24
1.5
11
4
50
1100
50
150
1420
28
2
12
4
50
1200
70
185
1300
20
2.5
13
4
53
900
70
220
1300
28
2
14
4
53
1100
80
150
1390
20
2.5
15
4
53
1200
50
185
1420
24
1.5
16
4
55
900
80
185
1420
20
2
17
4
55
1100
80
220
1300
24
2.5
18
4
55
1200
70
150
1390
28
1.5

Experiment Results and S/N Ratios
The experiments were conducted thrice for the same set of parameters using a singlerepetition randomization technique [18]. The casting defects that occur in each trial conditions were found and recorded. The average of the casting defects was determined for each trial condition as shown in Table 4. The casting defects are the lower the better type of quality characteristics. Lower the better S/N ratios were computed for each of the 18 trials and the values are given in Table 4:
Lower is better: S/N = 10 log [(y^2i) / n]
For example the S/N ratio for trial number 1 is: = 10 log [(y^2i)/3]
S/N ratio = 10 log [(5.25^2 + 5.7^2 + 4.21^2)/3]
= 10 log [(27.5625 + 32.49 + 17.7241)/3]
= 10 log [(77.7766)/3]
= 10 log [25.925]
= 14.138

Main effects of Factors and Analysis of Variance
The average effect of factors is shown in Table 5. After the experiments are conducted, the ANOVA is used to analyze the results of the experiments. The significant factors and/or their interactions are identified, for various trial conditions and the parameters which significantly influence the casting defects. However, some more information is
required to conclude with an optimum setting of parameters [19]. In applying ANOVA technique, certain assumptions must be checked through analysis of residuals before interpreting and concluding the results. It is highly recommended to examine these residuals for normality, independence, and constant variance, when using ANOVA [16]. In this paper, F ratio test is employed to check constancy of residual variance. If the F ratio test statistic is equal to or less than its corresponding critical value, the residuals have constant variance. The Fratio value can be found using the ratio of mean square of a factor to variance of error. It can seen from the Fratio value result that the significant factors are the control factors in the order of C (green compression strength), A (moisture), F (pouring temperature) and E (permeability). The other control factors are pooled due to less significance and percent contribution. The expected amount of sum of squares (SS) for each factor is computed by using variance. The percent contribution (P) for each factor is calculated by using expected amount of sum of squares (SS) in Table
6. The ANOVA table after pooling until the DOF of the error term is approximately half the total DOF of the experiment is shown in Table 6.
Table 4.Casting defects values and signaltonoise ratio against experimental trial numbers
Trial no. 
Percentage 
defects in experiment 
Average 
S/N ratio 

1 
2 
3 

Total 

1 
5.25 
5.7 
4.21 15.16 
5.053 
14.138 
2 
4.48 
5.58 
5.56 15.62 
5.206 
14.374 
3 
6.47 
7.13 
5.12 18.72 
6.24 
15.982 
4 
3.93 
4.16 
5.04 13.13 
4.376 
12.875 
5 
6.24 
6.68 
4.68 17.6 
5.866 
15.46 
6 
7.29 
7.17 
6.79 21.25 
7.083 
17.009 
7 
2.64 
3.05 
3.87 9.56 
3.186 
10.178 
8 
3.75 
5.54 
5.13 14.42 
4.806 
13.746 
9 
7.26 
8.15 
6.34 21.75 
7.25 
17.252 
10 
6.34 
4.83 
7.24 18.41 
6.136 
15.872 
11 
5.67 
7.1 
7.53 20.3 
6.766 
16.668 
12 
9.1 
9.53 
9.29 27.92 
9.306 
19.378 
13 
7.34 
7.76 
6.75 21.85 
7.283 
17.261 
14 
6.67 
5.2 
6.33 18.2 
6.066 
15.706 
15 
7.17 
6.87 
5.54 19.5 
6.526 
16.345 
16 
3.09 
5.89 
4.78 13.76 
4.586 
13.496 
17 
7.16 
6.74 
6.17 20.07 
6.69 
16.525 
18 
7.29 
7.64 
7.64 22.57 
7.523 
17.531 
Average casting defects = 6.108 Standard deviation= 1.548
Table 5 Main effects
Columns 
Factors 
Level 1 
Level 2 
Level 3 
L2 L1 
1 
Moisture (%) 
14.557 
16.531 
0 
1.974 
2 
Sand particle size 
16.068 
15.776 
14.788 
.292 
3 
GCS 
13.97 
15.413 
17.249 
1.443 
4 
Mould hardness 
15.634 
15.697 
15.302 
.063 
5 
Permeability 
15.205 
15.036 
16.392 
.169 
6 
Pouring temp. 
16.343 
15.602 
14.688 
.74 
7 
Pouring time 
15.905 
15.05 
15.677 
.854 
8 
Pressure test 
15.515 
16.01 
15.107 
.496 
Table 6.Analysis of Variance Results
Factors 
DOF 
Sums of Squares 
Variance 
FRatio 
Pure Sum 
Percent 

1 
Moisture % 
1 
17.537 
17.537 
33.563 
17.015 
22.226 
2 
Sand particle size 
(2) 
(5.403) 
POOLED 
(CL = 84.14%) 
0.000 

3 
GCS 
2 
32.422 
16.211 
31.024 
31.377 
40.987 
4 
Mould Hardness nu 
(2) 
(.541) 
POOLED 
(CL = *NC*) 
0.000 

5 
Permeability 
2 
6.555 
3.277 
6.273 
5.510 
7.198 
6 
Pouring temperature 
2 
8.245 
4.122 
7.890 
7.200 
9.405 
7 
Pouring time 
(2) 
(2.349) 
POOLED (CL = 81.71%) 
0.000 

8 
Pressure test 
(2) 
(2.453) 
POLLED (CL = 84.58%) 
0.000 

Other/ Error 
10 
11.790 
1.179 
20.184 

Total 
17 
76.553 
100.000% 

Expected Cost Savings at Optimum Conditions
The performance of the expected optimum conditions is estimated considering the factors and interactions. The significant contributions made by factors A1, C1, E2 and F3 have a total contribution of 3.924 decibels. As the current grand average of performance is 15.544 decibels, the expected result at the optimum condition is computed at 11.619 decibels. Therefore the improvement expected in reduction of defects is 37.66 per cent. Estimating the performance of all factors at an arbitrary condition the expected result at optimum condition of S/N ratio is 9.689. In order to calculate the expected cost savings, Taguchis loss function has been used [20] using equation.
S S
L = 1 Ã—
10 N 1 Ã— N 2
10 Ã— 100% of L1
= 73.42 percent/$ loss Here (S/N) 1 = 15.444 and (S/N) 2 = 9.689
Table 7 Estimate of performance at any arbitrary condition
Column & Factor
Level description
Level
Contribution
1. Moisture
3.5
1
.987
2. Sand particle size
55
3
.756
3. GCS
900
1
1.574
4. Mould hardness
80
3
.242
5. Permeability
185
2
.508
6. Pouring temp.
1420
3
.856
7. Pouring time
24
2
.493
8. Pressure test
2.5
3
.436
Total contribution from all factors
5.855
Current grand average of performance
15.444
Expected result at optimum condition S/N ratio
9.689

Variation Reduction Data and Savings
The objective of the paper is to reduce the variation and getting closer to the target. After implementation of Taguchi method in the foundry a reduction of 37.66 % was observed. This was due to the improved signal to noise ratio from
15.544 to 9.689. The improved standard deviation was 0.788 from the current standard deviation 1.547.The optimum condition and performance factors are shown in Table 8, moisture percentage 3.5, GCS 900 g/cm^2, permeability number 185 and pouring temperature 1420 Celsius. The total contribution of all the factors is 3.924.
Table 8.Optimum condition and performance
Factors
Level description
Level
Contribution
1
Moisture %
3.5
1
0.987
2
GCS g/cm^2
900
1
1.574
3
Permeability
185
2
0.508
4
Pouring temperature
1420
3
0.856
Total contribution from all factors
3.924
Current grand average of performance
15.544
Expected result at optimum condition
11.619
Improvement expected
37.66%

Conclusions
The optimum conditions for the factors computed are: Moisture (%) Level 1 Minimum 3.5
Green compression strength (g/cm^2) Level 1 Minimum 900 Permeability Level 2 Minimum 185
Pouring temperature (deg. Celsius) Level 3 Maximum 1420
The improvement expected in minimizing the variation is 37.66 percent which means reduction of casting defects from present 6.1 percent to 3.84 percent of the total castings produced in the foundry. This also reflects that by
using Taguchi method the factor levels when optimized will result in reduction of casting defects and increase the yield percentage of the accepted castings without any additional investments. A usage of quality tools like Pareto chart is useful for finding the major defects in the daily operations of foundry. Quality of castings can be improved by aesthetic look, dimensional accuracy, better understanding of noise factors and the interaction between variables, quality cost system based on individual product, scrap reduction, reworking of castings and process control.

Acknowledgements
The authors would like to thank the company Tru Form Techno Products Limited, Nagpur and staff of the quality and production department for their involvement in the study. The authors also gratefully acknowledge the casting experiments supported by the company Tru Form Techno products limited and the Taguchi approach software Qualitek 4 (demo version) supported by R Roy, Nutek, incorporated, sited at www.rkroy.com.

References

Johnston RE (1989) Design of experiments: Taguchi in the foundry. AFS Trans 82:415418.

Kumar P, Gaindhar JL (1995) Offline quality control for V process castings. Qual Reliab Eng Int 11:175181.

Masters I, Khoei AP, Gethin DT (1999) the application of Taguchi methods to the aluminium recycling process. In: Proceedings of the 4th ASM conference and exhibition on the recycling of metals. Vienna, Austria. pp. 115124.

Taguchi G (1993) Taguchi on robust technology development bringing quality upstream by Genichi Taguchi, ASME, New York.

Ghani JA, Choudhury IA, Hassan HH (2002) Application of Taguchi method in the optimization of end milling parameters. J Mater Process Technol 145:8492.

Osma A (2009) a statistical approach to comparing wheel hub forging processes. Proc I Mech Eng Part D, J Automob Eng 223(12):15591576.

Taguchi, G. Introduction to quality engineering: design quality into products and process. 1986, Tokio. Asian Productivity Organization.

PHADKE, M. S. Quality Engineering Using Robust Design. Prentice Hall, Englewood Cliffs, NJ, (1989).

Der Ho Wu, Mao Sheng Chang, Mater. Sci. Eng. A 379 (2004) 366371.

Sushil Kumar, P. S. Satsangi, D. R. Prajapati: Optimization of green sand casting process parameters of a foundry using Taguchis method, Int J Adv. Manu f Technology (2011) 55:23 34.

Muzammil, M., Pal Singh, P., Talib, F.: Optimization of gear blank casting process by usingTaguchis robust design technique. J.Qual. Eng. 15, 351359 (2003).

B. H. Kim, J. S. Shin, S. M Lee, B. M. Moon: Improvement of tensile strength and corrosion resistance of high silicon cast irons by optimizing casting process parameters, J Mater Sci (2007) 42:109 117.

Ballal Yuvraj P., Dr. Inamdar K. H., Patil P. V.: Application of Taguchi method for design of experiments in turning gray cast iron. International Journal of Engineering Research and Applications (IJERA) Vol. 2, Issue 3, MayJun 2012, pp. 13911397.

Dr. M. Arasu: Development of quality cost system in cast iron foundries

A. Noorul Haq, S. Guharaja, K. M. Karuppannan: Parameter optimization of CO2 casting process by using Taguchi method. Int J Interact Des Manuf (2009) 3:4150

Enright TP, Prince, B (1983) Offline quality control parameter estimation and experimental design with the Taguchi method. AFS Transactions. pp. 393400

Roy RK (1990), A primer on Taguchi method. Van Nostrand Reinhold, New York

ASM International Committee (1990) Metals handbook for castings, vol 15, 10th edn. ASM International,
Materials Park, OH

Syrcos GP (2002) Die casting process optimization using Taguchi methods. J Mater Process Tech 135:6874

V. D Tsoukalas, St A Mavrommatis, N G Orfanoudakis and A K Baldoukas, A study of porosity formation in pressure die casting using the Taguchi approach, Proc. Instn Mech. Engrs. Vol. 218 Part B: J. Engineering Manufacture 2004.
International Journal of Engineering Research & Technology (IJERT)
ISSN: 22780181
Vol. 1 Issue 7, September – 2012