10 I January 2022

https://doi.org/10.22214/ijraset.2022.39852
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

Minimizing Shrinkage Porosity by Optimizing

Some Parameters of Al Alloy (6061T6) Using DOE
Greeshma Arya1, Pratesh Jayaswal2, M. K Gaur3
1, 2, 3
Department of Mechanical Engineering, Madhav Institute of Technology & Science Gwalior, M.P -474006

Abstract: Response surface methodology is an efficient and powerful tool which is widely applied for casting optimization. In
this research aluminum alloy wheel hub casting is done by using BOXBEHNKEN design, three level of each parameter were
taken. Solid modeling of casting and
gating system is done by CAD. Simulation of Aluminium Alloy (6061 T6) casting were perform in PRO-cast (2009.1) the
simulation result indicates that selected parameters significantly affect the quality of casting. ANOVA is employed to examine
the relationship between the factors. Input parameter namely flow rate, pouring temperature and runner size were taken to
reduce the volume of shrinkage porosity. Experimental Design consist 15 experimental trials and output data obtained from
simulation will be optimized through minitab-18. Result indicates that selected independent variables are significantly influence
the response. ANOVA gives the optimized value of selected factors which reduces the porosity volume up to 30cm³.
Keywords: Sand casting, Shrinkage porosity, Simulation, DOE, Response surface method.

I. INTRODUCTION
From casting it has been observed that melting and pouring conditions directly or indirectly affects such mechanical properties of
cast materials as: hardness, percentage elongation, percentage reduction in diameter, toughness and so on. There is no doubt that
casting as a process involves so many parameters such as charge melting temperature, mould charge, pouring speed, pouring
temperature, composition, feeder design, core design, microstructure, size of casting, runner size, composition of the alloy and
solidification time just to mention but a few have successfully carried out studies on the varying effects of casting process
parameters on the mechanical properties of casted metals and their alloys. These all above factors or parameters affect the properties
both physical and mechanical. Unfortunate mechanical properties leads to defect like blow hole and cold shuts etc. Mainly ductility
and brittleness of a casted material is depends upon its grain size. Larger grain size makes the material ductile and smaller makes
them brittle these grain sizes are influenced by different factors during casting of alloy. To improve the mechanical properties of
these alloys either grain refining is to be done by adding grain refining element or by using cast technology depending upon
particular alloy. The wide range of the application of aluminum alloys is very obvious. The wheel hub (also abbreviated as wheel or
hub) is a critical safety component of transportation equipment such as vehicles like automobiles and trains, and it has close relation
with many performances of vehicles, for example motion properties, comfort ability and dependability.
Three parameters pouring temperature, Flow rate and runner size are taken into consideration. It has been stated that when pouring
temperature is lower than optimum the mould cavity will not fill the gate or riser will solidify too rapidly and intercept directional
solidification. On the other hand, higher pouring temperature causes shrinkage of the casting and mould wrapping. Many casting
defects result because the optimum casting condition was not used during casting process. Similarly optimum value of permeability
number is required because lower value of permeability number leads to poor porosity property of moulding sand this is because of
gap between sand particles and additives which doesn’t allows the gases to pass through the mould and these entrapped gases will
be the reason for blow holes defect, whereas with higher permeability number gap between sand particles and additives is more so
to increase permeability sand particle size should be more but this may leads to surface roughness and strength will be low.
Increasing the runner size will decrease the solidification time & other mechanical properties the decrease in mechanical properties
may be result of non-uniform crystal structure formed because of decrease in solidification time. Since the velocity of molten metal
going in the mould cavity is decrease, the filling of the mould cavity is delayed resulting to non-uniform solidification of molten
metal and this can result to non-uniform microstructure formation within the casting. (i.e. Coarse and fine grains within the casting)
but decreasing the runner size at great instead will may result sand erosion but keeping the runner size at certain level will leads to
increase in solidification time, ultimate strength, and elongation, therefore it is necessary. If we talk about our third factor i.e. Flow
rate and Pouring speed, slow pouring speed reduces the flow rate of molten metal hence metal is incompletely filled in thin areas,
fast pouring causes air entrapment into the mould cavity which leads to blow holes.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 454
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

Dr.B. Ravi, et al[1] discussed about the simulation in cating. His study describes the tangible and non tengible benefits of casting
simulation, bottlenecks ,study is based on survay of 150 foundary industries about the computer application during 2001-2006. He
suggested that casting simulation technique has become and nacessary tool for removing the casting defects and method
optimization.
Feng Liu et al. [2]in this paper, with the aid of parametric modeling technology of runner and riser are modeled parametrically. By
varying each parameter, it is easy to get different casting cad models. These models output data populate the orthogonal matrix,
which is used in the orthogonal array testing strategy to define the most suitable combinations of runners and risers parameters.
After inputting the completed orthogonal matrix data and all cad models into the simulation software the simulation result can be
obtained.
S.Shamasunder, et al.[3] applied ADSTEFAN simulation software for optimizing the gating and riseing system for valve body
castng to reduce the defects followed by number of itration. Nx V9(UG) CAD software is used to develop the solid model and
imported in to adstefan software. This software package perticularly helpful for foundry application to visualize and predict the
casting result so as to provide guidelines for improving product as well as mould design in order to achive desired quality.
T.R. Vijayaram, S.Sulaiman, et al.[4] had presented the objective, steps involved and significance of modelling and simulation
process, author also discussed about the importance of heat transfer in simulation process and details of computer simulation of
solidification of casting in metallurigical engineering foundary. Suggested that casting simulation can be employed not only to find
defective location. Beside it is use to find behaviour of different material under different conditions. It is use to determine the
cooling rate which affect the grain size of the casting. Solidification simulation provides time temperature which explains the effect
of under cooling of solidifying casting which reflects more on the inside microstructure responsible for material properties.
Dr.p.prsbhakara rao et al.[5] had presented their work on solidification simulation on flywheel casting made by gray cast iron to find
potential location of defects by using CLICK2CAST software package. The modification on existing riser and gating design were
changed to improve the casting quality.
The aim is to determine the optimum settings by using response surface methodology (BBD) at which these parameter produce good
quality casting. The analysis of variance (ANOVA) and regression analysis are employed to find the optimum level and to analyse
the effect of casting process parameter on ultimate tensile strength, hardness, impact strength. Therefore, this research presents
extensive research into solidification simulation to minimize shrinkage during the development of aluminum alloy wheel hub. The
material of alloy wheel hubs is aluminum alloy. The grade of the aluminum alloy is AL6061-T6; the temperature of environment is
30℃.

A. Sand Casting and Parameters

Basic elements of gating system
1) Pouring Basin: Its acting as a reservoir for supplying molten metal to the cavity. It also separates the impurity present in the
charge Shape and size of pouring basin does not affect the pouring time and filling time of molten metal in the casting.
2) Sprue: It is a connecting passage between pouring basin and runner it is always vertical with straight tapered circular cross
section the height of the sprue mainly responsible for producing velocity of molten metal in the gating system hence height of
the sprue is selected in such that velocity of molten metal in the gating system must ensure the laminar flow.

Vmax = 2 ℎt

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 455
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

Figure: 1 Dimensions of sprue [6]

Mass flow rate= ρAV= constant
Applying continuity equation between point 2 and 3 we get
₂ ₃
₃
= ₂
= =

₂
= ²
₃
Actual shape of sprue is parabola but to avoid manufacturing difficulty we use tapered cylinder shape

3) Runner: Connecting passage between bottom of the sprue and Ingate its always horizontal with uniform trapezoidal cross
section, trapezoidal cross section is preferred because it gives the high coefficient of discharge(Cd)
Cact= Cd.Q theoritical
Variation of Cd in only in the case of horizontal passage whereas in vertical passage flow of liquid is due the datum head
irrespective of cross section the value of Cd is remain constant hence whichever cross section is easier to manufacture the
corresponding cross section is use to making the runner.

4) Ingate: It is the last point of the gating system from where molten metal is entering in to casting cavity. Fig represents the
elements of gating system

Figure: 2 Elements of gating system [6]

II. RESEARCH METHODOLOGY

The entire study has been carried out in three sections, designing of casting and gating system in creo parametric 3.0, Numerical
simulation for percentage of shrinkage porosity by using PROcast, analysis for the significant factors and optimization has been
carried out through minitab18.1

A. Modeling of Wheel hub in 3D using CAD

The solid model information regarding the overall shape of casting, base features, local features and tooling is stored in geometry
objects using the symmetric data structure for boundary representation. This involves storing the bounding faces, edges and vertices
of the corresponding solid model, along with the relevant topological relationships (such as adjacent faces for an edge) and
geometric parameters (plane coefficients, direction cosines and Cartesian coordinates) The solid model information regarding the
overall shape of casting, base features, local features and tooling is stored in geometry objects using the symmetric data structure for
boundary representation. This involves storing the bounding faces, edges and vertices of the corresponding solid model, along with
the relevant topological relationships (such as adjacent faces for an edge) and geometric parameters (runner size.)

Figure: 3 Solid model of Wheel hub (Creo Parametric 3.0)

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 456
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

B. Meshing with mesh CAST

The work steps which you follow when using mesh CAST depend upon the following: the nature of your mission, the upcoming use
of the meshes generated by mesh CAST and the type and quality of CAD model you use as the initial input.

Figure: 4 Performing MESH cast (element size-7mm)

C. Solidification Simulation using Pro- CAST

Pro CAST is the world's most intuitive, incredibly fast, and truly integrated casting software for methods design, simulation and
optimization Pro CAST is fast and easy to use, even if you are new to computers. It is also the most economical casting methods
software available today, and is supported by a network of highly qualified and helpful consultants.

Figure: 5 Simulation of wheel hub by using PRO cast (2016)

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 457
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

D. Analysis of data using Response Surface Methodology in MINITAB 18.1

1) Create Response Surface Design: Response surface methods are used to inspect the connection between a response and a set of
measurable experimental variables or factors. These methods are often hired after you have identified a "vital few" manageable
factors and you want to find the factor settings that optimize the response. Designs of this type are usually chosen when you
suspect curvature in the response surface.

Figure: 6 creating experimental design using minitab18.1

III. SELECTION OF MATERIAL

The material of alloy wheel hubs is aluminum alloy. The grade of the aluminum alloy is AL6061-T6; the temperature of
environment is 20℃. The material composition including its properties are shown in Table 2.3 & 2.4

A. Composition (mass %)

Table 1. Material composition [7]

Al Cr Cu Fe Mg Mn Si Ti Zn
96.45 0.4 0.3 0.7 1.0 0.15 0.6 0.15 0.25

B. Solidus and Liquids Temperature

Tsol = 600ºС, Tliq = 642ºС

C. Density
ρ = 2705 Kg/m³
Table 2. Recommended values for the thermo physical properties of Al alloy - 6061-T6 [7]

Temperature (℃) Density Heat Capacity Enthalpy Thermal Viscosity

(Kg/m3) (KJ/KgK) (J/Kg) Conductivity (mPaS)
(W/mK)
25 2705 0.87 0 -
100 2695 0.95 69 195
200 2675 0.98 166 203
300 2655 1.02 266 211
400 2635 1.06 370 212
500 2610 1.15 370 225
600 2415 1.17 981 200
700 2400 1.17 1049 91 1.05
800 2372 1.17 1166 92 1

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 458
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

IV. RESULT AND DISCUSSION

A. Statistical Analyses (ANOVA)
The statistical foundation for design of experiment and the analysis of variance (ANOVA) was first introduced by Sir Ronald A.
Fisher the British biologist. ANOVA is a method of partitioning total variation in to accountable sources of variation in an
experiment.it is a statistical method used to interpret experimented data and make decision about the parameters under study the
basic equation of ANOVA is given by:
Total sum of square = sum of square due to factors + sum of square due to error
Here are some important terms which are important to interpret the result:
1) Coefficient of Determination (R2): Explains amount of variation explained by the model. Larger value of R2 does not necessarily
imply that model is better one by adding variable to the model always increase the value of R2 whether the variable is
significant or not. Hence we prefer R2(adj).

R2 = where SSR = sum of square of regression

SST= total sum of square

2) Adjusted coefficient of determination (R2): In general R2 (Adj) value does not increase as variable are added to model the
addition of unnecessary variable will decrease the value of R2(Adj).

R2 (Adj) = n= total number of runs

n-p = error degree of freedom

n-1 = total degree of freedom
Corresponding value of R2 (adj) tells amount of variation is responsible for response out of total variation in input factor.

3) Predicted Coefficient of Determination (R2) pred.: how well your model predict the response for the new observation, model
which have larger predicted R2 value have better predictive ability. (R2) pred. subsequently less than R2 which indicates model
is over fit (when you add terms in the model which are not important).it is calculated with observation that are not included in
the model calculation.
4) Value of S: Defines how well model describe the response. S is measured in units of response variable and represents how far
the data values fall from the fitted values. Lower the value of S, the better the model describe the response. However a lower S
value by itself does not indicate that model meets the model assumption.
5) Residuals: Adequacy of the model is also investigated by examination of residuals examine by normal probability plot of
residuals.

Now after getting the response for every combination of experiment click on analyse the response surface design AVOVA table
will be given by the software which will tell the selected factors are significant or not and it also test the hypothesis table 3 shows
ANOVA table for linear term, squared term , and interaction term.

Table 3. ANOVA table

Analysis of Variance
Source DF Adj SS Adj MS F-Value P-Value
Model 9 77.9594 8.6622 15.42 0.004
Linear 3 32.3734 10.7911 19.21 0.004
flow rate 1 5.6252 5.6252 10.01 0.025
pouring temperature 1 5.4197 5.4197 9.65 0.027
runner size 1 21.3284 21.3284 37.96 0.002
Square 3 3.2198 1.0733 1.91 0.246
flow rate*flow rate 1 0.3256 0.3256 0.58 0.481

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 459
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

pouring temperature*pouring temperature 1 2.5334 2.5334 4.51 0.087

runner size*runner size 1 0.2891 0.2891 0.51 0.505
2-Way Interaction 3 42.3662 14.1221 25.14 0.002
flow rate*pouring temperature 1 5.8430 5.8430 10.40 0.023
flow rate*runner size 1 8.2387 8.2387 14.66 0.012
pouring temperature*runner size 1 28.2845 28.2845 50.35 0.001
Error 5 2.8090 0.5618
Lack-of-Fit 3 2.8090 0.9363 1.66 0.28907
Pure Error 2 0.0000 0.0000
Total 14 80.7683

S R-sq R-sq(adj) R-sq(pred)

0.749529 96.52% 90.26% 44.36%

At 95% confidence interval, significance level-5%

 F value (fisher) used to check the significance; more value of F gives less value of p.
 F value of model is 15.42 and value of p is 0.004 which is less than 0.05 means model is highly significant.
 All terms in the model are significant accept squared terms since value of p is greater than 5%
Difference between R2 and R2 (Adj.) values are 6.26 it means generated model can predict quality within +/- 6.26 with 95%
confidence level.
So remove the Insignificant Terms until All the Values in Terms Becomes Less Then 0.05
After removing the unnecessary terms software will generate new ANOVA table

Table 4. Modified ANOVA table

Source DF Adj SS Adj MS F-Value P-Value
Model 6 74.7396 12.4566 16.53 0.000
Linear 3 32.3734 10.7911 14.32 0.001
flow rate(kg/s) 1 5.6252 5.6252 7.46 0.026
Pouring temperature(◦c) 1 5.4197 5.4197 7.19 0.028
Runner size(mm²) 1 21.3284 21.3284 28.30 0.001
2-Way Interaction 3 42.3662 14.1221 18.74 0.001
flow rate(kg/s)*Pouring temperature(◦c) 1 5.8430 5.8430 7.75 0.024
flow rate(kg/s)*Runner size(mm²) 1 8.2387 8.2387 10.93 0.011
Pouring temperature(◦c)*Runner size(mm²) 1 28.2845 28.2845 37.53 0.000
Error 8 6.0288 0.7536
Lack-of-Fit 6 6.0288 1.0048 1.34 0.341415
Pure Error 2 0.0000 0.0000
Total 14 80.7683

S R-sq R-sq(adj) R-sq(pred)

0.868100 92.54% 86.94% 63.40%

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 460
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

 Now All The Values Are Coming Out To Be Less Than 0.05 Which Means All Terms In The Model Is Significant.
 Value Of R² & R² (adj) Are Close To Each Other Which Represents Goodness Of Model.
 R² (pred) value (63.40) represents how well your model predict the response for new observation. model which have large value
of R² (pred.) value have better predictive ability
 S = 0.868100 value defines how well model describe the response lower the value of s, the better the model describe the
response.

Table 5 Percentage contribution of each linear and interaction terms

Factors Degree of Sum of Percentage Contribution
freedom Square
Flow rate 1 5.6252 6.97%
Pouring temperature 1 5.4198 6.72%
Runner size 1 21.3282 26.42%
Flow rate* pouring temperature 1 5.8430 7.234%
Flow rate* runner size 1 8.2387 10.20%
Pouring temperature* runner size 1 28.2854 35.01%
Lack of fit 6 6.0288 7.4643%
Pure error 2 00000 00%
TOTAL 14 80.7683 100%

From ANOVA table and observing p value it was observed that “runner size” is the most significant factor in linear term,
percentage on contribution by runner size is 26.42% while pouring temperature and runner size together are the most significant
factor in the regression model, corresponding p-value Is 0.0001 percentage of contribution is 35.01%.

6) Lack of Fit: A regression model exhibits lack of fit when it fails to adequately describe the functional relationship between the
experimental factor and response variable, lack of fit can occur if important terms in the model such as interaction or quadratic
terms are not included it can also occur of several unusually large residual from fitting the model.
7) Pure Error: Occur from repeated value of dependent variable(Y) for a fixed value of independent variable.

Figure: 7 Residual plots

8) Normality: The population normally checked from normal probability plot of residual if the distribution of residual is normal
the plot will be reassemble in straight line. In this design the probability plot (fig-5.1a) shows strongly linear pattern this is
verified by the correlation coefficient of 0.8694 of the line fit to the probability plot. The fact that the points in lower and

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 461
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue I Jan 2022- Available at www.ijraset.com

extremes of the plot do not deviate significantly from their straight pattern indicates that there are not any insignificant outliers
(relative to normal distribution). In this case it conclude that the normal distribution an excellent model for the data.
9) Constant Variance: The variance of the observation in each treatment should be equal. A constant variance assumption can be
checked with residual versus fit plot. This plot shows (figure: 7) a random pattern of residual on both side of zero and not
showing any recognizable patterns.
V. CONCLUSION
Conventional optimization studies are time consuming and expensive. To overcome these problems, a BOX BEHNKEN design
(BBD) was used for the optimization of process conditions. From the present study, it is evident that the use of statistical process
condition optimization approach, response surface methodology has helped to locate the most significant conditions with minimum
effort and time. In addition, it has also proved to be useful for predict less shrinkage porosity volume only 15 experiments were
necessary and the obtained model was adequate (p < 0.00041). By solving regression equation, the optimum process condition was
determined. The value of R² and adjusted R² represent the proportion of variation in the response. In this study the value or R² is
92.54 for wheel hub which describes the amount of variation in the observed responses and it is more significant.

REFERENCES
[1] Dr. B.Ravi “casting simulation & optimization, benefits, bottlenecks and best practice” Technical paper for Indian foundry journal (jan-2008)
[2] Feng Liu, Worcester polytechnic institute, December 2008
[3] Saad Chowdhary, Santhosh Kumar, S. Shamasundar “optimization of gating system and risering system for casting of valve body using ADSTEFAN
simulation software”IJESC 2016
[4] T.R. Vijayaram, S. Suleiman, A.M.S. Hamouda, M.H.M. Ahmad, numerical simulation of casting solidification in permanent metallic molds, journal of
materials processing technology 178 (2006) 29–33
[5] B. Shradesh Kumar, Dr. P.Prabhakararao Rao Predicting the Casting Defects through Simulation International Journal of Engineering Technology Science and
Research (Dec-2017)2394 – 3386
[6] https://www.slideshare.net/93Bhardwaj/gating-system
[7] Kenneth C Mills Recommended Values of Thermo physical Properties for Selected Commercial Alloys National Physical Laboratory the Materials Information
Society WOODHEAD P U B L I S H I N G L I M I T E D Cambridge England

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 462