Engineering, Environment
Abrasive wear behaviour of AlCuMg/palm kernel shell ash particulate composite
Gambo Anthony VICTOR
Department of Mechanical Engineering, Ahmadu Bello University, Zaria, Nigeria
Email: gambo.anthony@yahoo.com;
Corresponding author, phone: +2348038467245
Received: June 20, 2017 / Accepted: December 02, 2017/ Published: December 30, 2017
Abstract
This paper presents a systematic approach to develop a wear model of AlCuMg/Palm kernel shell ash particulate composites (PKSAp) produced by double stircasting method. Four factors, five levels, central composite, rotatable design matrix was used to optimize the number of experiments. The factors considered were sliding velocity, sliding distance, normal load and mass fraction of PKSA reinforcement in the matrix. Response surface methodology (RSM) was employed to develop the mathematical model. The developed regression model was validated by statistical software MINITAB and statistical tool such as analysis of variance (ANOVA). It was found that the developed regression model could be effectively used to predict the wear rate at 95% confidence level. The regression model indicated that the wear rate of cast AlCuMg/PKSAp composite decreased with an increase in the mass fraction of PKSA and increased with an increase of the sliding velocity, sliding distance and normal load acting on the composite specimen.
Keywords
AlCuMg alloy; Analysis of variance; Wear rate; Response Surface Methodology; Palm kernel shell ash particles
Introduction
Industrial technology is growing at a very rapid rate and consequently there is an increasing demand and need for new materials [1]. To meet such demands, particulate reinforced composites constitute a large portion of these new advanced materials. The choice of the processing method depends on the property requirements, cost factor consideration and future applications prospects [2].
Incorporation of hard second phase particles in the alloy matrices to produce MMCs has also been reported to be more beneficial and economical [1, 3] due to its high specific strength and corrosion resistance properties. In the past, various studies have been carried out on metal matrix composites. SiC, TiC, TaC, WC, B4C [4] are the most commonly used particulates to reinforce in the metal or in the alloy matrix or in the matrices like aluminium or iron.
Recently, there has been an increasing interest in composites containing low density and low cost reinforcements [5, 6]. The availability of natural fibers has tempted researchers to try locally available fibers and to what extent they satisfy the required specifications as good reinforcement for tribological applications. Natural fibers such as banana, cotton, coir, sisal and jute have attracted the attention of scientists and technologists for application in consumer goods, low cost housing and other civil structures. It has been found that these natural fiber composites possess better electrical resistance, good thermal and acoustic insulating properties and higher resistance to fracture. Natural fibers have many advantages compared to synthetic fibers, for example low weight, low density; low cost, acceptable specific properties and they are recyclable and biodegradable. They are also renewable and have relatively high strength and stiffness and cause no skin irritations [7]. However, there are also some disadvantages, for example moisture uptake, quality variations and low thermal stability. Many investigations have been made on the potential of the natural fibers as reinforcements for composites and in several cases the result have shown that the natural fiber composites own good stiffness, but the composites do not reach the same level of strength as the glass fiber composite [8].
Among various discontinuous dispersions used palm kernel shell ash (PKSA) has been found to be one of the most inexpensive and low density reinforcement available in large quantities as solid waste from coconut processing industries [9]. Hence, composites with palm kernel shell ash as reinforcement are likely to overcome the cost barrier for wide spread applications in automotive and small engine applications. It is therefore expected that the incorporation of palm kernel shell ash particles (PKSAp) in aluminium alloy will promote yet another use of this lowcost waste byproduct and, at the same time, have the potential for conserving energyintensive aluminium and thereby, reducing the cost of aluminium products [10].
There have been few dry sliding wear behaviour studies based on various reinforcements like SiC, Al_{2}O_{3}, fly ash and Zircon [11]. The principle tribological parameters that control (load, sliding velocity, sliding distance, counterpart material, weight % of reinforcement, shape, and size) specific wear rate and coefficient of friction were analysed. From the literature, it is understood that the relationship between the parameters in dry sliding wear is complex and independent, selection of the optimal parameter of combination is important to reduce specific wear rate and coefficient of friction. Design of experiment, Genetic algorithm and response surface method is widely used to optimize the dry sliding parameters. There has been experimental investigation using Taguchi and ANOVA to identify the significant factors while testing with Al 2219 SiC and Al 2219 SiC graphite material shows that the sliding distance, sliding velocity and load are having significant effect [11]. From these discussions it is clear that though lot of work has been done on MMCs, as per the information of the author, no work has been done on the use of Response Surface Methodology (RSM) technique to predict the tribological performance of AlCuMg/ PKSAp composite. Therefore, this work aims at adopting RSM technique to obtain an empirical model of wear loss (response) as a function of amount of reinforcement, applied load, sliding velocity and sliding distance (input factors).
Materials and methods
Preparation of palm kernel shell ash
Palm kernel shell was crushed and grounded to form palm kernel shell powder; the powder was packed in a graphite crucible and fired in electric resistance furnace to a temperature of 1300^{0}C for 1 hour to form palm kernel shell ash (PKSA).
Particle size analysis of the palm kernel shell ash particles was carried out in accordance with BS1377:1990 [4]. About 150g of the ash particles were placed into a set of sieves arranged in descending order of fineness and shaken for 15 minutes which is the recommended time to achieve complete classification, the particles that were retained in the BS 75µm was used in this study.
Chemical composition of the palm kernel shell ash particles is presented in Table 1.
Table 1. Chemical composition of palm kernel shell ash
Element 
Al_{2}O_{3} 
SiO_{2} 
CaO 
Fe_{2}O_{3} 
MgO 
K_{2}O 
Na_{2}O 
LOI 
% by wt 
13.17 
65.56 
6.56 
4.79 
1.24 
5.56 
1.20 
0.90 
Fabrication of composites
A high purity aluminium electrical wire obtained from Northern Cable Company (NOCACO) Kaduna, Nigeria, was used as the matrix. Synthesis of the metal matrix composite was by double stircasting method at the foundry shop of the National Metallurgical Development Centre (NMDC), Jos, Nigeria. The specimens were produced by keeping the percentage of copper and magnesium constant and different volume fraction of PKSAp 5, 10, 15 and 20% with particles size of 75µm were added in the mix. All the melting was carried out in a claygraphite crucible in a resistance furnace. Al4%Cu0.8%Mg alloy was preheated at 450^{0}C for 2 hours before melting, and before mixing the palm kernel shell ash particles, it was preheated at 1000^{0}C for 1 hour to make their surfaces oxidized[12]. The furnace temperature was first raised above the liquids temperature of aluminium (720^{0}C) to melt the alloy completely and then cooled down just below the liquids to keep the slurry in a semisolid state. At this stage, the preheated palm kernel shell ash particles was added and mixed manually. Manual mixing was used because it was very difficult to mix using automatic device when the alloy was in a semisolid state.
After sufficient manual mixing, the composite slurry was reheated to a fully liquid state and then automatic mechanical mixing was carried out for about 20 minutes at an average stirring rate of 150 rpm. In the final mixing processes, the furnace temperature was controlled between 730 and 740^{0}C and 0.01%NaClKCl was added as a covering flux. The pouring temperature was controlled to be about 720^{0}C [13].
Specimen of 6mm × 6mm ×50mm were cut from the cast composite, the end of the specimens were polished with abrasive paper of grades 600 followed by grade 1000. Dry sliding tests were carried out as per ASTM G9995 test standards on pinondisc equipment [14], the disc of which is of EN31 steel with surface roughness, Ra 0.1µm. The sample pins were cleaned with acetone and weighed before and after testing using an electronic balance to an accuracy of 0.0001g to determine the amount of wear. The sliding end of the pin and disc surfaces were cleaned with acetone before testing. The specific wear rate (Ws) was calculated using the following Eq. (1) [11]:
Ws = (1)
Where: Ws is the specific wear rate, is the mass loss, is the density, is the normal load, and is the sliding distance. After the test, the sliding surface of selected test samples was observed by scanning electron microscopy (SEM) as shown in figure 1.
Figure 1. Pin on disc equipment [11]
Identifying the important process parameter
Based on preliminary trials, the independent process parameters affecting the tribological behaviour of a composite were identified as: sliding velocity A(m/s), sliding distance B(m), applied load C(N), and fraction of reinforcement D (wt%).
Finding the limits of control variable
Many trial experiments were conducted on the aluminium matrix composite (AMC) specimens to find out the feasible limits of the process in such a way that the lower limit of each parameter was fixed to yield a noticeable wear. The upper limit was selected when wear was not severe. In order to have an easy interpretation of results and to understand the effect of each parameter on the response, the lower and upper levels of the parameters are coded as 2 and +2 respectively. The coded values for any intermediate range were calculated using the following relationship, Eq. (2):
= 2[2X – ( + )]/ () (2)
Where: X_{i }is the required coded value of a variable, X is any value of the variable from X_{max} to X_{min}; X_{max} is the upper limit of the variable; X_{min}_{ }is the lower limit of the variable.
Table 2 below shows the factors and their level employed in the experiments. Wear in terms of specific wear rate is the measured response used to evaluate the tribological behaviour.
Table 2. Factors and their levels in central composite design experimental plan
Factors 
Level 

2 
1 
0 
1 
2 

Sliding velocity A (m/s) 
0.4 
0.8 
1.2 
1.6 
2.0 
Sliding distance B(m) 
400 
800 
1200 
1600 
2000 
Applied load C (N) 
10 
20 
30 
40 
50 
Reinforcements D (wt %) 
0 
5 
10 
15 
20 
Design of experiment (DOE)
In the present investigation, experiments were designed on the basis of the Design of Experiments (DOE) technique proposed by Box and Hunter [16, 17]. A 2^{k }factorial, where k is the number of variables, with central composite secondorder rotatable design was used to improve the reliability of results and to reduce the size of experimentation without loss of accuracy.
The fourfactor, five level central composite rotatable design requires 31 experiments with 16 factorial points. The next 8 experimental runs comprised a combination of each process variable at either their lowest (2) or highest (+2) level with the other three variables kept at the intermediate levels (0) constituting the stars point and seven centre points for replication to estimate the experimental error. The experiment has been carried out according to the run order in the experiment design matrix. At the end of each run, settings for all four parameters were changed and reset for the next run. This was essential to introduce variability caused by errors in experimental settings [17].
Development of wear model
The wear loss (W) of the AlCuMg/PKSAp composite is a function of sliding velocity, sliding distance, normal load and mass fraction of PKSAp reinforcement in the aluminium alloy matrix. It can be expressed as, Eq. (3):
W = f (A, B, C, D) (3)
Where: W – Response; A  Sliding velocity; B  Sliding distance; C  Normal load; D  Fraction of CSAp reinforcement.
For the fourfactors, the selected polynomial (regression) could be expressed as Eq. (4):
W = b_{0} + b_{1}A + b_{2}B + b_{3}C + b_{4}D + b_{11}A^{2} + b_{22}B^{2} + b_{33}C^{2} + b_{44}D^{2} + b_{12}AB + b_{13}AC + b_{14}AD + b_{23}BC + b_{24}BD + b_{34}CD (4)
Where: b_{0} is the free term of the regression equation, the coefficients b_{1}, b_{2}, and b_{3} are linear terms, the coefficients b_{11}, b_{22}, and b_{33}, are quadratic terms, and the coefficients, b_{12}, b_{13}, and b_{23}, are interaction terms. The values of the coefficients were calculated with the help of MINITAB a statistical analysis software, which is widely used in many fields of engineering research. The values of the coefficients for the wear model is presented in Table 3:
Table 3. The value of the coefficients model
Coefficients 
Value 
Coefficients 
Value 
b0 
126 
b14 
– 5.830 
b1 
– 2.303 
b22 
–2.111 
b2 
– 3.056 
b23 
– 0.167 
b3 
2.028 
b24 
– 4.292 
b4 
4.806 
b33 
– 4.861 
b11 
– 0.611 
b34 
– 0.417 
b12 
– 3.170 
b44 
– 2.264 
b13 
– 1.295 


Results and discussion
Verification of the adequacy of the developed model
Analysis of Variance (ANOVA) and the Fratio test was performed to check the adequacy of the model as well as the significance of the individual model coefficients. The ANOVA was carried out for a confidence limit of 95% or Pvalue of 0.05. This implies any factor with Pvalue equal to or less than 0.05 is significant. From the analysis of the results obtained in table 3, it is clear that sliding velocity (A), sliding distance (B), load (C) and fraction of PKSAp reinforcement (D).
Are significant along with the interactions BD and CD and the quadratic terms of A, B and D as the P value of these terms is less than 0.05. Other model terms can be said not to be significant. These insignificant model terms can be removed and may result in an improved model [11].
Another criterion that is commonly used to illustrate the adequacy of a fitted regression model is the coefficient of determination (R2). For the models developed, the calculated R2 value and adjusted R2 value are 97.84% and 93.62% respectively.
Table 3. ANOVA for specific wear rate of AlCuMg/PKSAp
Source 
SS 
DF 
MS 
FValue 
PValue 
Constant 
2.02436 
14 
0.96569 
53.86 
0.0021 
A 
0.003263 
1 
0.003263 
17.82 
0.0046 
B 
0.054209 
1 
0.054209 
89.09 
0.0075 
C 
0.009741 
1 
0.009741 
7.37 
0.0224 
D 
0.000448 
1 
0.000448 
0.074 
0.0016 
AB 
0.001627 
1 
0.001627 
15.66 
0.0576 
AC 
0.054825 
1 
0.054825 
2.89 
0.0682 
AD 
0.007683 
1 
0.007683 
3.26 
0.0976 
BC 
0.009569 
1 
0.009569 
1.52 
0.2179 
BD 
0.000321 
1 
0.000321 
0.96 
0.0013 
CD 
0.008346 
1 
0.008346 
2.21 
0.0363 
A^{2} 
0.002771 
1 
0.002771 
2.73 
0.0044 
B^{2} 
0.002681 
1 
0.002681 
2.46 
0.0268 
C^{2} 
0.001762 
1 
0.001762 
1.08 
0.8743 
D^{2} 
0.059178 
1 
0.059178 
3.79 
0.0466 
Residual error 
0.734415 
6 
0.037769 
 
 
Lack of fit 
0.008382 
10 
0.008382 


Fratio as per table (14, 6, 0.05) = 3.96 
These values indicate that the regression model is quite adequate. The adequacy of the model was further confirmed by a scatter diagram. Figure 2 shows a typical scatter diagram for the model (of wear of AlCuMg/PKSAp). The observed values and predicted values of the responses are scattered close to the 45^{0}C line, indicating an almost perfect fit of the developed model.
Figure 2. Normal probability plot of residuals
Conformity tests
Five experimental data which were never used in the modelling process were used to test the performance of the model.
Table 4. Results of conformity test
Input variables 
Specific Wear rate m3/Nm×1013 

A 
B 
C 
D 
Measured 
Predicted 
% Error 
0.5 
300 
10 
3 
8.998 
8.714 
3.259 
0.9 
500 
20 
6 
9.444 
9.106 
3.712 
1.5 
1000 
35 
9 
5.717 
6.071 
5.831 
1.5 
1400 
35 
12 
7.320 
6.801 
7.631 
3.0 
1900 
65 
12 
6.112 
5.905 
3.506 
The difference in the experimental values of wear corresponding to a set of input parameters and the predicted values were taken as error of prediction and are calculated as per Eq. (5) reported as % error in Table 4 along with other results. It is observed from this table that the results are within acceptable range and the average deviation is 2.455%.
% = (5)
Effects of the different process parameter
The effects of the different process parameter on the wear behaviour of AlCuMg/PKSAp composite are predicted from the developed mathematical model by varying one parameter value from its maximum level to maximum level while keeping the other three parameters values at their centre levels. The experimental results are plotted and presented in figures (47) as a function of wear. The general trends between cause and effect are discussed below.
Effect of sliding velocity
The effect of sliding velocity on the specific wear rate of AlCuMg/PKSAp composite is shown in Figure 4a; it is obvious from the figures that wear rate increase with sliding velocity. Sliding velocity influences the frictional heat developed in the area of contact between the test pin and counter surface.
Figure 4a. Effect of sliding velocity on the wear rate
More frictional heat is developed in the contact area when sliding velocity is increased [19]. Thus micro thermal softening of matrix material may take place which lowers the bonding strength of PKSAp with aluminium matrix alloy [20].
Figure 4 (b) shows SEM micrographs of worn surface of the cast AlCuMg/20%CSAp composite at sliding velocity of 2m/s with normal load of 50 KN and sliding distance of 2000 m.
Figure 4b. SEM image of worn
The figure shows that the extended PKSAp can be easily pulled out from the matrix as a result of micro thermal softening of matrix and higher shearing force developed on the contact surface. Those pulledout PKSAp particles may act as wear debris between test pin and counter face and form the third body abrasive wear mechanism [21].
Effect of sliding distance
Figure 5a shows the variation of wear rate with sliding distance of AlCuMg/PKSAp composite. It is evident from the figure that sliding distance increases linearly with the wear rate.
Figure 5a. Effect of sliding distance on the wear rate
When the sliding distance increases, frictional heat on the contact surface also increases. As the raised temperature in the contact surface decreases, the resistance offered by the matrix against the shear force, the rate of deformation as well as pullout of PKSAp from the matrix are increased. This leads to subsurface cracks which nucleate at the interfaces between PKSAp and aluminium alloy matrix as depicted in Figure 5b (worn surface of AlCuMg/20%PKSAp composite tested at sliding distance of 2000 m).
Figure 5b. SEM image of worn surface
The worn surface of the composite reveals continuous deep grooves. The edges at the groves are plastically deformed, due to the generation of higher frictional heat [21].
Effect of normal load
Figure 6 a show the effect of applied load on the wear rate of AlCuMg/PKSAp composite.
Figure 6a. Effect of applied load on the wear rate
From Figure 6 a, it is observed that wear rate of the composite increases while increasing the applied load. This is because at higher load, frictional thrust increases, which results in increased deboning and fracture. A similar effect of normal load on wear rate has been observed by [22]. Figure 6b shows SEM micrograph of the worn surface of cast AlCuMg/ 20%PKSAp composite at normal load of 50 N with sliding distance of 2000 m and sliding velocity of 2 m/s.
Figure 6b. Worn surface tested at applied load of 50N
The worn surface shows asperities of broken soft aluminium matrix due to shear force on the worn surface. Due to the high applied pressure, cracks nucleate as a result of ploughing of hard asperities.
Effect of fraction of reinforcement
Figure 7a shows the wear rate of AlCuMg/PKSAp composite as a function of mass fraction of PKSA in the matrix. It is obvious from the figure that the wear rate decreases with increase in the mass fraction of PKSA by keeping other wear process parameters constant, this result is in agreement with similar works by [21].
Figure 7a. Effect of fraction of reinforcement on the wear rate
Figure 7b shows the uniform distribution of PKSAp in the AlCuMg matrix and Figure 7c shows the SEM image of worn surface of AlCuMg/20%PKSAp composite.
Figure 7b. SEM image of worn surface 
Figure 7c. SEM image of worn surface 
Conclusions
The relationships between process parameters for wear behaviour of AlCuMg/PKSAp have been established. Response surface methodology was adopted to develop the regression models, which were checked for their adequacy using ANOVA test, scatter diagrams are found to be satisfactory.
Confirmation experiments showed the developed models are reasonably accurate. The accuracy of the developed model can be improved by including more number of parameters and levels. Wear coefficient tends to decrease with increasing particle volume content. It also indicates that coconut shell ash addition is beneficial in reducing wear of the AlCuMg/PKSAp composite.
References
1. Hamouda A.M.S., Suliaman S., Vijayaram T.R., Sayuti M., Ahmad M.H.M., Processing and characterization of particulate reinforced aluminium silicon matrix composite, Journal of achievement in materials and manufacturing engineering, 2007, 25(2), p. 1116.
2. American Society for Testing and Material, Annual Book of ASTM Standards, USA, 1999.
3. Burr A., Yang J.Y., Levi C.G., and Leckie F.A., The strength of metal matrix composite joints, Journal Pergamon, 1995, 43(9), p. 33613373.
4. Kaczmar J.W., Pietrzak K., Wlosinski W., The production and application of metal matrix composite materials, Journal of Materials Processing Technology, 2000, 106, p. 5867.
5. Bienia J., Walczak M., Surowska B., Sobczaka J., Microstructure and corrosion behaviour of aluminium Fly Ash Composites, Journal of Optoelectronics and Advanced Materials, 2003, 5(2), p. 493 – 502.
6. Aigbodion V.S., Hassan S.B., Dauda E.T., Mohammed R.A., Experimental study of ageing behaviour of AlCuMg/bagasse ash particulate composites, Tribology in industry, 2011, 33(1,), p. 456469.
7. Oksman K. M., and Selin J. F., Natural fibers as reinforcement in polylactic acid (PLA) composites, J. Comp. S. Skrivars ci. Technol., 2003, 63, p. 13171324.
8. Hussain S. A., Pandurangadu V., Palanikuamr K., Mechanical properties of green coconut fiber reinforced HDPE polymer composite, International Journal of Engineering Science and Technology (IJEST), 2011, 3(11), p. 79437952.
9. Butler R.A. (online), “Why is coconut replacing tropical rainforest”? Why are biofuels fueling deforestation”? Available at: http://news.mongabay.com/2006/(0425ai) palm.html (accessed 11/11/2016)
10. Hassan S. B., and Aigbodion V. S., The study of the microstructure and interfacial reaction of AlCuMg/bagasse ash particulate composite, Journal of Alloy & Compounds, 2010, 491, p. 571–574.
11. Rajesh S., Rajakarunakaran S., Sudhakara R., Modelling and optimization of sliding specific wear and coefficient of friction of aluminium based red mud metal matrix composite using Taguchi method and response surface methodology, Journal of Materials Physics and Mechanics, 2012, 15, p. 150166.
12. Aigbodion V. S., and Hassan S. B., Effects of silicon carbide reinforcement on microstructure and properties of cast AlSiFe/Sic particulate composites, Journal of Materials Science and Engineering, 2007, 447, p. 355 – 360.
13. Aigbodion V. S., Development of AlSiFe/rice husk ash particulate composites synthesis by double stir casting method, Usak University Journal of Material Sciences, 2012, 2, p. 187 – 197.
14. Suresha S., and Sridha B.K., Parametric studies on the tribological behaviour of aluminium matrix hybrid composites, International Journal of Design and Manufacturing Technologies, 2010, 4 (1), p.1520.
15. Cochran G., and Cox M., Experimental designs, Newyork, John Wiley & sons, 1987.
16. Velmurugan C., Subramanian R., Thirugnanam S., Ananadavel B., Experimental investigations on machining characteristics of Al 6061 hybrid metal matrix composites processed by electrical discharge machining, International Journal of Engineering, Science and Technology, 2011, 3 (8), p. 87101.
17. Harris P., and Smith B. L., Factorial technique for weld quality prediction, Met Construct, 1983, 15, p. 661666.
18. Douglas C. M., Design and Analysis of Experiments, New Delhi, Wiley India (P) Ltd, 2007.
19. Rao R. N., and Das S., Effect of SiC content and sliding speed on the wear behaviour of aluminium matrix composites, Journal of Materials and Design, 2011, 32, p. 1066−1071.
20. Kumar S., Balasubramanian V., Developing a mathematical model to evaluate wear rate of AA7075/SiCp powder metallurgy composites, [J]. Wear, 2008, 264, p. 1026−1034.
21. Kumar B. A., Murugan N., and Dinaharan I., Dry sliding wear behaviour of stir cast AA6061T6/AlNp composite, Trans. Nonferrous Met. Soc. China, 2014, 24, p. 2785−2795.
22. Alpas A. T., Embury J. D., Sliding and Abrasive Wear Behaviour of an AluminiumSiC Particle Reinforced Composite, J. Scripta Metallurgica et Materialia, 1990, 24, p. 931−935.
23. Cirino M., Pipes R. B., and Friedrich K., The abrasive wear behaviour of continuous fibre polymer composites, Journal of Material Science, 1987, 22, p. 2481.