|
International Journal of ChemTech Research CODEN (USA): IJCRGG, ISSN: 0974-4290, ISSN(Online):2455-9555 Vol.9, No.06 pp 260-269, 2016
|
Optimization on Stir Casting Process Parameters of Aluminium Alloy (AL6061) Nickel Coated Graphite (NCG) Metal Matrix Composite using Taguchi Based Rsm
Lakshmikanthan.P1 , Prabu. B2
Department of Mechanical Engineering Pondicherry Engineering College, Pondicherry, India
Abstract : The investigation presents the results of experimental investigations on mechanical property and optimization of stir casting process parameters of aluminium 6061 nickel coated graphite metal matrix composite using Taguchi based Response surface methodology (RSM). The investigation has been made the effect of casting parameters such as percentage of reinforcement, stirring time, stirring speed and casting temperature of Al6061-NCG metal matrix composite manufactured by pellet method. The responses considered for the analysis are Hardness and tensile strength. An empirical model has been developed for predicting the Hardness and tensile strength of aluminium 6061 nickel coated graphite metal matrix composite. Response surface model and analysis of variance (ANOVA) are used in order to study the effects of casting parameters. Optimum result in maximizing the Hardness and tensile strength are determined using desirability function approach. The influences of different casting parameters of Al6061-NCGparticulate composite have been analyzed in detail.
Keywords : Al6061, Nickel coated graphite, stir casting, Anova, RSM.
1. Introduction
The Metal matrix composites have potential advantages than monolithic alloys and this has activated considerable attention in the past years. These composites give the solution of increasing problems in service requirements for various Engineering and structural applications by adding reinforcement of metal matrix by metallic or non-metallic materials. Generally Metal matrix composites (MMCs) are reinforced with low density, high strength and high modulus ceramic phases in the form of fibre, whiskers or particulates. The adding of ceramic reinforcement to metal matrix improves strength and stiffness while ductility is compromised1. Considerable research in the material science has been directed toward the development of new engineering materials and structural application which possessing high strength, high specific strength, good creep, and fatigue and wear resistance.
Properties of composites are mainly depends on the properties of their basic materials, their distribution and the interaction between the material. Apart from the nature of the parent materials, the geometry of the reinforcement such as size, shape, concentration, distribution and orientation of the reinforcement influences the properties of the composite. The most commonly used reinforcements are Silicon Carbide (SiC), Graphite, Zirconia and Aluminium Oxide (Al2O3). Graphite, in the form of fibers or particulates, has long been recognized as a reinforcement material because of its low co-efficient of friction, low specific gravity, good conductor of heat and electricity2. The aluminium alloy and its composites have been used in many engineering applications (for example automotive components, bushes and bearings) because of its low density, more
specific strength, corrosion resistance, etc.3. The previous attempts made of introduction of nickel-coated graphite, alumina and SiC in aluminium and Zn based alloys, by using stream of nitrogen gas reinforcement particles are injecting into the melt. A later method vortex is used to produce the composite, here stirring is carried out by using impeller and then reinforcement material such as copper coated graphite, nickel coated graphite nickel-coated alumina particles etc. Introducing to Centre of the vortex produced by stirring4.
Hsiao YehChu5 Investigated nickel coated metal matrix composite, from the experimental study, nickel coating on graphite particles allows contact between matrix and reinforcement phase (metal to metal contact), thus producing a large surface-to hardness ratio. Material bond can protect the specimen from severe wear.Lakshmikanthan.P6 also investigated characterisation of nickel coated graphite aluminium composite, from the analysis at the lowest percentage of reinforcement having better hardness and tensile strength due to lubricant properties of graphite.
Response surface methodology (RSM) is used to find the relationships between several explanatory variables and one or more response variables, it is a well-known approach on optimization of the input parameters models based on physical, simulation and experimental observations. These models need to be calculated statistically for their suitability, and then they can be utilised for an optimisation of the initial model. It is also measures the relationships between the controllable input parameters and the obtained response surfaces. The input parameters are sometimes called independent variables, and the performance measure is the response. By using the results of a numerical experiment in the points of orthogonal experimental design, response surface methodology analysis is computationally much less expensive than a solution using the traditional method. By using the analytical model, the objective function problem can be easily solved and time in the calculation can be saved7.
In the present work, an attempt is made to study experimental investigation on mechanical property and optimization of stir casting process parameters of aluminium 6061 nickel coated graphite metal matrix composite using taguchi and Response surface methodology (RSM), metal matrix composite are synthesized using stir casting technique by “PELLET METHOD”. By using these techniques of producing nickel coated graphite aluminium alloy composites are more attractive and controls the composition using this technique is better than using either method. The cast composites were tested for hardness, tensile strength and microstructure8.
2. Materials used
The matrix material used in this study is Al 6061, obtained from PMC Corporation, Chennai, as blanks. The reinforcement material used is Nickel coated graphite NCG obtained from Alfa Aesar Ltd, USA in powder form (Powder size: < 150 μm).
3. Experimental Work
All the pellets are compacted to a density of 2.6gm/cm3. The compaction process is carried out at Tenneco Automotive Pvt.ltd, Pondicherry. Pellets are made from the mixture of 65% of nickel coatedgraphite powder and remaining of pure aluminium powder in weight percentage. In order to produce the composites different parameters are selected such as melting temperature, stirring time, stirring speed and percentage of reinforcement. Initially weight percentage of reinforcement such as 2.5%, 5%, 10% produced in form of pellet is pre heated at the temperature of 450oC, aluminium alloy was melted at defined selected temperature of 700 oC,750 oC,800 oC stirring is carried out for 10 minutes (constant) for all different level of experiment at different selected speed say 400rpm, 450rpm, 500rpm. After adding pellets in the molten aluminium alloy stirring was carriedout at 10, 15, 20 minutes and the resultant slurry was then cast into the metallic mould to form the composite specimens. Fig 3 shows photos of the prepared specimens.
Fig.1:Prepared specimen, clockwise Al6061, Al-NCG 2.5%, Al-NCG 10%,Al-NCG 5%.
Brinell hardness test was performed on the composites with 10 mm ball indenter at a load of 500kgf.for each sample average of the hardness reading was considered. Tensile test was carried out on the prepared specimens as per ASTM standard E8 in order to evaluate the tensile strength of the composites.
Fig. 2: SEM image at 1500x magnification
The Fig 2. shows the SEM image of thecomposite. From the SEM analysis it is confirmed that the presence of reinforcement in the composite and the particles were distributed uniformly, there is no large pore and particles clustering existed.
3.1 Plan of Experiments
The Taguchi‟s orthogonal array,which helps in reducing the number of experiments, according to 3-level experiments L27 orthogonal array are conducted. The casting parameters considered for the present investigations are: (1) % weight fraction of NCG (2) stirring speed (3) stirring time (4) stirring temperature. The interactions in the casting parameters may also play constructive role in deciding the Hardness and tensile strength of the composite. After careful investigation,only square effects and two factor are selected ,from this outcome effects are not linearly related, so it is decided to use 3 level tests for each factor9.
Table 1 Process Parameter and Level
Parameter |
Temperature (oC) |
Time (sec) |
Stirring speed(rpm) |
% of Reinforcement |
Level 1 |
700 |
10 |
400 |
2.5 |
Level 2 |
750 |
15 |
450 |
5 |
Level 3 |
800 |
20 |
500 |
10 |
3.2 Design of experiment based on Taguchi method and selection of the casting process parameter
Taguchi's design of experiments (DoE) is used for experimentation. Taguchi's approach is a systematic and efficient method provides parameter design to the design engineer for determining optimum design parameters. This method can dramatically reduce the number of experiments required to gather necessary data. In Taguchi‟s orthogonal array, the most suitable array is L27, which needs 27 runs and has (n-1) 26 degrees of freedoms (DOF)10. It can conduct three levels of parameters, for the three level tests, 10 DOFs (5X2) are taken
for 5 main factors and the remaining DOFs are taken by interactions. The 3 level L27 orthogonal array is shown in Table 2,
3.3 Experimental Results
Table 2:Experimental Results
Experimental Run |
Temperature (oC) |
Time (sec) |
Stirring speed(rpm) |
% of Reinforcement |
Hardness (BHN) |
Tensile (MPa) |
1 |
700 |
10 |
400 |
2.5 |
62.16 |
155.62 |
2 |
700 |
10 |
450 |
5 |
55.3 |
124.2 |
3 |
700 |
10 |
500 |
10 |
50.8 |
114.73 |
4 |
700 |
15 |
400 |
5 |
54.1 |
123.24 |
5 |
700 |
15 |
450 |
10 |
50.6 |
118.43 |
6 |
700 |
15 |
500 |
2.5 |
62.01 |
142.12 |
7 |
700 |
20 |
400 |
10 |
50.45 |
122.55 |
8 |
700 |
20 |
450 |
2.5 |
61.95 |
141.13 |
9 |
700 |
20 |
500 |
5 |
55.8 |
123.02 |
10 |
750 |
10 |
400 |
2.5 |
62.18 |
163.62 |
11 |
750 |
10 |
450 |
5 |
56.4 |
134.20 |
12 |
750 |
10 |
500 |
10 |
51.12 |
124.73 |
13 |
750 |
15 |
400 |
5 |
55.62 |
133.24 |
14 |
750 |
15 |
450 |
10 |
51.01 |
122.43 |
15 |
750 |
15 |
500 |
2.5 |
62.1 |
154.12 |
16 |
750 |
20 |
400 |
10 |
51.1 |
124.55 |
17 |
750 |
20 |
450 |
2.5 |
63.12 |
148.13 |
18 |
750 |
20 |
500 |
5 |
55.15 |
130.12 |
19 |
800 |
10 |
400 |
2.5 |
61.51 |
143.48 |
20 |
800 |
10 |
450 |
5 |
55.01 |
120.52 |
21 |
800 |
10 |
500 |
10 |
50.18 |
116.43 |
22 |
800 |
15 |
400 |
5 |
54.94 |
124.23 |
23 |
800 |
15 |
450 |
10 |
50.1 |
110.48 |
24 |
800 |
15 |
500 |
2.5 |
61.3 |
135.12 |
25 |
800 |
20 |
400 |
10 |
50.1 |
113.21 |
26 |
800 |
20 |
450 |
2.5 |
62 |
137.12 |
27 |
800 |
20 |
500 |
5 |
54.01 |
112.12 |
3.4 Modelling and Optimization of Machining Parameters Using Response Surface Methodology
Response surface methodology (RSM) method is used for determining the cause-and-effect relationship between input control variables and true mean responses influencing the responses as a two- or three-dimensional function11. Many researchers have applied response surface methodology for modelling and analysis of process parameters in manufacturing. Response surface methodology is one of the simple and easy method to use. In this present work, mathematical model is developed for correlating the interactive and second-order influences of various process parameters on Hardness and tensile strength of Al6061-NCG-MMC composites.
(1)
The „a‟ coefficients is determined by Higher order model which is obtained by the least square method and controllable parameters for product (process) x values result is obtained using response surface methodology can be satisfying several requirements to find that results in optimization of response. The second-order response surface representing the Hardness (H; BHN) and tensile strength can be expressed as a function of casting parameters such as (1) %volume fraction of NCG (2) stirring speed, (3) stirring time (4) temperature.
ANOVA tables are shown in Tables 3. & 4. for hardness and tensile strength. From the below table , analysis of varience for Hradness and tensile strength model, F value of 318.11 for hardness and F value of 29.88 for Tensile strengthin, it indicates that both model are Significan.The values of probability less than 0.05 imply that models are significant.
Table 3:Analysis of Variance for Hardness
Source |
DF |
Adj SS |
Adj MS |
F-Value |
P-Value |
Model |
13 |
602.440 |
46.342 |
318.11 |
0.000 |
Linear |
4 |
336.677 |
84.169 |
577.77 |
0.000 |
Square |
4 |
52.295 |
13.074 |
89.74 |
0.000 |
Interaction |
5 |
0.878 |
0.176 |
1.21 |
0.360 |
Residual Error |
13 |
1.894 |
0.146 |
|
|
Total |
26 |
604.334 |
|
|
|
Table 4:Analysis of Variance for Tensile strength
Source |
DF |
Adj SS |
Adj MS |
F-Value |
P-Value |
Model |
13 |
5079.21 |
390.71 |
38.92 |
0.000 |
Linear |
4 |
2223.64 |
555.91 |
55.38 |
0.000 |
Square |
4 |
1550.11 |
387.53 |
38.60 |
0.000 |
Interaction |
5 |
28.11 |
5.62 |
0.56 |
0.729 |
Residual Error |
13 |
130.50 |
10.04 |
|
|
Total |
26 |
5209.71 |
|
|
|
ANOVA and F-ratio test are performed to validate the goodness of fit in the developed mathematical models, the developed model can be used to validate the effects of casting parameters on hardness and tensile strength of composite,the t-ratio is used to find the significance value of the individual coefficient factor. The larger and absolute value of the more significant factor will be the t-ratio. The probability value represents the factor coefficient and smaller value are more significant12. In this analysis, the origin value selected is 0.05. If the probability value is greater than the selected level, the null hypothesis is accepted and the coefficient is judged not to be significant. Finally, the values of R2represent the regression confidence. The larger value of R2is always desirable
Regression (R) =a0 +a Temp1 + a2 Temp2+ a3 Temp3 + a4Time1- a5TimeE2 + a6 Time3+ a7 RPM1 + a8 RPM2+ a9 RPM3 + a10of Rein1+ a11 % of Rein2 + a12 % of Rein3 (2)
Table 5:Tests on the Five Factors, Square Effects and Their Interactions for Hardness
Term |
Effect |
Coef |
SE Coef |
T-Value |
Probability |
R2 |
Constant |
------ |
53.436 |
0.240 |
223.02 |
0.000 |
99.54% |
Temp |
-0.4501 |
-0.2251 |
0.0908 |
-2.48 |
0.028 |
|
Time |
-0.1102 |
-0.0551 |
0.0917 |
-0.60 |
0.558 |
|
RPM |
0.028 |
0.014 |
0.127 |
0.11 |
0.915 |
|
% of Rein |
-11.570 |
-5.785 |
0.122 |
-47.41 |
0.000 |
|
Temp*Temp |
-1.476 |
-0.738 |
0.156 |
-4.73 |
0.000 |
|
Time*Time |
0.531 |
0.266 |
0.156 |
1.70 |
0.112 |
|
RPM*RPM |
-0.702 |
-0.351 |
0.165 |
-2.13 |
0.053 |
|
% of Rein*% of Rein |
6.822 |
3.411 |
0.189 |
18.06 |
0.000 |
|
Temp*Time |
-0.088 |
-0.044 |
0.110 |
-0.40 |
0.695 |
|
Temp*RPM |
-0.493 |
-0.247 |
0.110 |
-2.24 |
0.043 |
|
Temp*% of Rein |
-0.031 |
-0.016 |
0.108 |
-0.14 |
0.888 |
|
Time*RPM |
-0.280 |
-0.140 |
0.165 |
-0.85 |
0.411 |
|
Time*% of Rein |
-0.012 |
-0.006 |
0.162 |
-0.04 |
0.971 |
Regression Equation
Hardness(BHN) = -155.5 + 0.4857 Temp + 0.057 Time + 0.2090 rpm - 4.508 % of reinf. -0.000295 Temp* Temp + 0.01062 Time *Time - 0.000140 rpm *rpm + 0.2426 % of reinf. *% of reinf. - 0.000177 Temp*Time - 0.000099 Temp*RPM -0.000083 Temp*% of rein. - 0.000561 Time *rpm - 0.00032 Time *% of reinf.
Table 6:Tests on the Five Factors, Square Effects and Their Interactions for Tensile strength
Term |
Effect |
Coef |
SE Coef |
T-Value |
Probability |
R2 |
Constant |
--- |
123.39 |
1.99 |
62.04 |
0.000 |
97.50% |
Temp |
-5.702 |
-2.851 |
0.753 |
-3.78 |
0.002 |
|
Time |
-4.895 |
-2.448 |
0.762 |
-3.21 |
0.007 |
|
RPM |
-4.85 |
-2.42 |
1.06 |
-2.29 |
0.039 |
|
% of Rein |
-26.96 |
-13.48 |
1.01 |
-13.31 |
0.000 |
|
Temp*Temp |
-21.39 |
-10.70 |
1.29 |
-8.27 |
0.000 |
|
Time*Time |
2.52 |
1.26 |
1.29 |
0.97 |
0.348 |
|
RPM*RPM |
4.27 |
2.13 |
1.37 |
1.56 |
0.143 |
|
% of Rein*% of Rein |
28.67 |
14.34 |
1.57 |
9.14 |
0.000 |
|
Temp*Time |
-1.688 |
-0.844 |
0.915 |
-0.92 |
0.373 |
|
Temp*RPM |
0.715 |
0.358 |
0.915 |
0.39 |
0.702 |
|
Temp*% of Rein |
1.009 |
0.504 |
0.898 |
0.56 |
0.584 |
|
Time*RPM |
2.28 |
1.14 |
1.37 |
0.83 |
0.420 |
|
Time*% of Rein |
1.52 |
0.76 |
1.34 |
0.57 |
0.581 |
Tensile Strength(MPa) = -1907 + 6.330 Temp - 1.77 Time- 0.992 (RPM)-18.96 % of Rein -0.004278 Temp * Temp + 0.0504 Time*Time+ 0.000853 (RPM)* (RPM)+ 1.019 % of Rein * % of Rein - 0.00338 Temp *Time+ 0.000143 Temp * (RPM) + 0.00269 Temp*% of Rein + 0.00456 Time* (RPM) + 0.0406 Time(sec)*% of Rein
The casting parameter has been optimization using response optimizer in MINITAB and the result are discussed below.
4 . Result and discussion
In this analysis the objective set for casting process is maximizing the hardness and tensile strength, for optimization MINITAB 17 software is used. The results obtained from the experimental runs carried out,according to the taughi orthogonal array shown in Table 6. The main effects analysis plot is used to analysis the effects of each factors of casting parameter such as weight % of NCG, stirring speed , stirring time & stirring temperature versus Hardness have been shownin Figs. According to main effect plots, average hardness per casting process
Fig 3: Main effect plot for Hardness(BHN) Fig 4: Main effect plot for Tensile strength(N/mm2)
Analysis of Variance (ANOVA) technique is used to check the validity of the developed model and Student‟s t-test was utilised to find the significance of factors Analysis of Variance (ANOVA) and F test were analysis to find the significance level of each factor(confidence level of more than 95%). If the F value of a factor exceeds the F0.05 the factor is insignificant. ANOVA tables for all the four are presented in Tables 3 & 4.
Fig. 5:Residual plot for Hardness(BHN) Fig.6: Residual plot for Tensile strength(N/mm2)
From ANOVA of Hardness and tensile in Table 3 & 4, it can be observed that the influence of % reinforecement on hardness as well as tensile strength is more when compared to the other parameters. Hence, of % reinforecement is the most significant input parameter affecting Hardness and tensile, followed by stirring Temperature, stirring time and stirring speed. Consequently, it can be concluded that Hardness and tensile strength of Al6061–NCG are found to be more sensitive to percentage of reinforcement and stirring time than other two parametrs. Fig. 5 & 6.Shows the normal probabilities plot of residuals between predicted values and experimental value, from the figure it is clear that the residuals are spread closer to the straight line for both hardness and as well as tensile strength the residuals are fitted accurately so this entailed that the errors are distributed normally with respect to the predicted values. The residuals are distributed both in positive and a negative direction. This implies that the model is adequate.
Fig 7: correlation graph for Hardness(BHN Fig 8: correlation graph for Tensile strength(N/mm2)
From Fig. 7&8. it is declared that the predicted results are very close to the experimental results and hence the response surface models are suitable for predicting Hardness and tensile strength of the aluminium nickel coated graphite composites.
Minitab‟s response optimizer provide optimize set of response by combination a set input variable that satisfying the requirements..
The optimization graph obtained, the parameters used, its goal set and other information are given in Figure 9. Using the above the global statement
Fig 9.Optimization plot for Hardness (BHN ) and tensile strength (MPa)
From the optimization graph obtained, using the above the global obtained for maximization of Hardness and tensile strength are below shown.
Table 7:Multiple Response Prediction for hardness and tensile strength
Variable Setting |
Prediction Optimized Values |
Temperature(OC) |
750 |
Time(sec) |
10 |
Stirring Speed(RPM) |
450 |
% of reinforcement |
2.5 |
The response surfaces are plotted for the model established. Figure 11&12. shown below the response surface of Hardness (H) with respect to the Weight fraction of NCG, stirring temperature and stirring time
Fig. 10: Surface Plot for Hardness (BHN ) Fig .11:Surface Plot for tensile strength (N/mm2)
The surface plot shows that the Hardness and tensile strength are maximum at medium stirring temperature, stirring time, stirring speed and lower percentage of reinforcement. It is clear that the change in temperature and speed plays an important role in achieving maximum Hardness. The decreasein percentage of reinforcement leads to an increase strength of the composite material, this is subjected to increase in hardness and tensile strength of the casting process parameter. As a result, the observed hardness becomes higher. The effects of weight fraction NCG, while keeping the other parameters at centre level, are shown in Figure the above fig.10&11. It is evident from the plot that the Hardness of Al– NCG composites increase with a decrease in weight fraction of NCG.This is due to the solid lubrication of Gr particles.
Fig .12:Validation result for tensile strength (MPa) Fig.13:Validation result for Hardness (BHN )
or confirmation of the second-order response surface model, verification tests are conducted at three selected conditions. Figure 12. shows the validation of tensile strength result and Figure 13. shows the validation of hardness result, from this the difference between the predicted Hardness by the second order response surface model and the experiment are very closer, hence the second order response surface surface model is very useful for predicting the Hardness and tensile strength of stir casted Al6061-NCG composites within the selected parameter settings.
5. Conclusions
The Hardness and tensile strength in the casting process parameter has been investigated according to the Taguchi L27 orthogonal array. Based on the orthogonal array experimental and analytical results, the following conclusions are made:
References:
*****