Tribological Behavoiur of Silicon Carbide/ Rice Husk ash Reinforced Al7075 Hybrid Composites Using Optimization Techniques

In this paper, aluminium alloy Al7075 was reinforced with silicon carbide (SiC) particles and rice husk ash to improve hardness and wear resistance. The SiC particulate and rice husk ash composition in composite were taken in five different proportions such as 3%, 6%, 9%, 12% and 15% for % of SiC particulates and 2%, 4%, 6%, 8% and 10% for % of rice husk ash. Nine different compositions to cater the above proportions were fabricated using stir casting technique. Microstructure examination was conducted on the prepared specimen using optical microscope to determine the homogeneity of reinforced particle distribution. The tribological properties were examined by measuring the wear on the prepared specimen using pin on disc apparatus. The input parameters considered were materials factor such as % of SiC and % of rice husk ash and mechanical wear factors such as load applied, speed of rotation and sliding distance. The design of experiment technique, central composite design (CCD) was employed for conducting experiments with different combination of parameters. Empirical relationship was devised to predict wear of the composite specimen in terms of parameters such as % SiC, % of rice husk ash, load applied, speed of rotation and sliding distance. The empirical relations adequacy was tested using Analysis of variance (ANOVA). The direct and interaction effect of input parameters on wear were analyzed which helps to determine the significance of these parameters. Microstructure examinations were conducted to analyze assimilation and uniform distribution of reinforcement particles in prepared composites. Optimization of process parameter was obtained for minimum wear using genetic algorithm.


Introduction
Metal matrix composites, MMCs are compounded with at least two constituent parts. One part is a metal and other part constitute of other metal/ ceramic/ organic materials. If in such combinations two or more materials reinforces with the matrix metals, then the composites are known as hybrid composite. MMCs are fabricated to have enhanced mechanical properties such as good wear and corrosive resistance, high strength, low density, good thermal conductivity, low coefficient of expansion, etc. Owing to its unique characteristics and properties, MMCs find wide application in automobile parts, aerospace structures, deck panels, etc. Particulate reinforced MMCs contain reinforcement materials as either particles or platelets that range in size from 0.5 to 100 µm. The MMCs are fabricated by the accretion of reinforcement material in the matrix metal. Some suitable methods for the fabrication of MMCs are stir casting [1,2,3], squeeze casting [4], spray atomization and co-deposition [5], plasma spraying [6] and powder metallurgy technique [7]. The need for hybrid MMCs had been increasing in the recent years for the various product developments with high stiffness and strength. The most common matrix material used was aluminium and its alloy due to its commercial standpoint which is employed for structural application in aerospace, automotive and sports equipment sectors. The reinforcement materials are implanted into a matrix metal material to enhance its mechanical properties such as wear resistance and hardness. Non oxide ceramic material SiC exhibits good thermal conductivity and low density is commonly used reinforcement material. Rice husk ash is an agricultural by product available in abundance contains amorphous silica and carbon used as reinforcement materials. Sliding wear is perceived as a relative motion between two surfaces in contact under load which is inevitable in many industrial applications. The composite materials fabricated should be highly wear resistant. Aluminium MMCs manufactured with reinforcement of both SiC particulates and rice husk ash resulted in enhanced wear resistance compared to the reinforcement with either SiC alone. In this study the MMCs fabricated is of particulate type, where Al 7075 matrix composite was reinforced with different percentage of SiC particulate and rice husk ash using stir casting technique. Further the analysis was conducted to understand the effect of independent parameters on the wear resistance of the composites fabricated.
Different researchers have been carried out investigations on the mechanical properties of hybrid composites and shows an enhancement. The inclusions of mica and SiC particles reinforcement and its positive effects was reported by [8]. It has been concluded that increase in mica and SiC mass fraction resulted in better strength, wear resistance and hardness. Wear mechanism based on the pattern and stability of oxide layer was described by [9] in aluminum composites impregnated with SiC particles. Baradeswaran et al [10] fabricated MMCs with aluminium alloy (Al 7075) as matrix materials and reinforcement with various proportions of boron carbide. It was concluded that the wear resistance improves with the increasing content of boron carbide particles. The reinforcement of % of particulate boron carbide and rice husk in aluminium alloy were investigated by [11]. It was concluded that the hardness and strength increases with an increase in % of reinforcement. The effect of heat treatment in Al alloy reinforce with 6% rice ash husk was described by [12]. Three different aging processes is employed at temperature 135°C, 175°C and 225°C. Improvement in hardness and reduction in wear was observed at aging temperature 175°C in composite development. More importance are given for hybrid composites by adding two different reinforcements in matrix alloy resulted in better tribological characteristics. Tribological properties of hybrid composites with Al as matrix alloy and SiC and graphite as reinforcement as particulate of were investigated by [13,14]. Results have shown that the SiC and graphite particles positively influences the tribological behavoiur of fabricated composites. Friction and wear behavoiur of hybrid composites composed of magnesium reinforced with SiC and grapheme oxide using solvent based powder metallurgy method was studied by [15]. Predictive model was developed and optimal combination of process parameter for specific wear rate was obtained. Tribological studies on wear measurement of alumina reinforced MMCs was conducted by [16] for varying load conditions and sliding speed. It was concluded that wear rate increases with an increase in load. Wear characteristics of A356/SiC/MoS2 [17], Al6061/SiC/Al2O3 [18], Al6061/SiC/B4C/talc [19], AA1100/SiC/B4C [20], and Al7075/SiC/graphite [21] hybrid composite fabricated using friction stir processing were analyzed and found an improvement in wear resistance. Ravindran et al [22] studied the tribological properties of Al/SiC/graphite hybrid MMCs using powder metallurgy technique.
They concluded that the addition of both a hard reinforcement (e.g., SiC) and soft reinforcement (e.g., graphite) significantly improved the wear resistance of hybrid MMCs. Thus hybrid MMCs consist of matrix metal alloy with strongly reinforced multiple particulates enhance the wear resistance properties. Dey et al [23] studied the inclusion of SiC particulates in Al2024 alloy and investigated the effect of SiC content on mechanical properties and tribological behaviour of the prepared composites.
The design of experiment techniques intends in describing information on variation collected under certain conditions that are assumed to impersonate the variation. Central composite response surface methodology was employed by [24] to study the direct and interaction effect of volume fraction, matrix particle size and reinforcement particle size on the wear resistance of Al/SiCp composite. Taguchi technique was used by [25] for optimization of controlling parameters such as sliding distance and applied in wear test of Al/Al2O3 composites. Poria et al. [26] studied wear behavior of Al-TiB2 composites using Taguchi based L27 orthogonal array design matrix. Wear and friction behaviour of pure aluminium, Al alloy (Al-112Si) reinforced with ZrC with various percentage combinations was studied by [27] using full factorial experimental design.
The optimization of composite materials for enhanced properties and its utility to determine best combination parameters has been subjected to exhaustive research. Chidozie and Chizoba [28] applied Non Dominated Sorting Genetic Algorithm II for multi objective optimization of Al 6351/ Egg Shell Reinforced composite. Chidozie et al [29] used genetic algorithm with artificial neural network as fitness function for multi objective optimization of A356 alloy/ cow horn particulate composite. . Sudarsana et al [30] employed genetic algorithm based neural network model to determine computationally fast and efficient analysis of interface strength of ceramic matrix composites. Gangadhara et al [31] used genetic algorithm to obtain optimal values of the input process parameters related to the powder metallurgy-based process of producing Al/SiC composites.
The literature review revels that SiC, graphite, B4C, TiO2, ZrC and rice husk ash can be adequately mix up with aluminium alloy to form a hybrid MMCs. The low density and high thermal conductivity nature of SiC and heat resistant capability of rice husk ash makes the fabricated hybrid MMCs to exhibit superior mechanical properties. In this work, % reinforcement of SiC and rice husk ash particles varied individually and independently which has not been considered in the previous work. The main objective of the present research is to establish empirical regression relations in terms of parameters such as % SiC, % rice husk ash, speed of rotation of disc, load and sliding distance to predict wear in the fabricated hybrid composite specimen. The direct and interaction effects of these parameters on wear, to understand the effect of these parameters on wear resistance were studied. The optimal values of considered process parameter on wear was determined using genetic algorithm (GA).

Stir casting technique
Stir casting technique is an appropriate economical processing method to manufacture Al based MMCs. The experimental setup consist of furnace, reinforcement feeder, stirrer and a permanent mold. A pit type furnace is used to melt the metal and the temperature is measured using J type thermocouple. A reinforcement feeder attached along with furnace, through which the reinforcement particles and wetting agent particles poured in to the melt. A mechanical stirrer is used to form a vortex in the melt to mix the reinforcement particles and also to stir the melt to reduce oxidation. The stirrer is composed of stirrer rod, impeller blade, a variable speed motor and a regulator to govern the speed of motor. A mold is used for pouring mixed slurry and composites ingots are casted. The procedure for stir casting process are as follows. Before starting the process of melting in the furnace, it is preheated to 100 0 C. The matrix materials was placed inside a graphite crucible and then the crucible is placed inside the furnace for melting.
The temperature of the furnace was increased slowly in a time period of 2-3 hrs. and maintained at that temperature for about 2 hrs. Once melting of matrix material is completed, the mechanical stirrer begins to stir to form a vortex in the molten metal to avoid oxidation and to facilitate reinforcement particles to add upon. Meanwhile the reinforcement particles are preheated in a different furnace. The preheated reinforcement particles are poured into the molten metal in the crucibles through feeder and the stirring process continuous. The stirring disperse the reinforcement particles in the matrix molten melt and maintain them in suspension.
The melt slurry was thoroughly stirred to ensure uniform of mixture of dispersed reinforcement particles. The resulting mixture was cast into a steel mould of 20 mm diameter and final MMCs cast ingots is fabricated.

MMCs fabrication
Aluminium alloy Al7075 was used as a matrix materials and powder particles SiC and rice husk ash were included as reinforcement materials. Al 7075 possess excellent castability, light weight, good oxidation resistance, better mechanical properties, etc., and used in the fabrication of structural elements in automotive and aerospace applications. SiC also called as carborundum was chosen as one of the reinforcement material owing to the fact as one of the most widely used particle material for aluminium alloy based MMCs. Since SiC is a ceramic materials, its inclusion in Al alloys increase hardness and wear resistance. Rice husk ash, a form of carbon is another material, which was chosen for reinforcement. It is easily available, a light weight material, economically lesser cost materials which is obtained from agricultural residue, can be used for effective reinforcement. Rice hush ash chemical composition varies with crop year and geographical regions, but predominately chemical proportions consist of SiO2 and Al2O3. Reinforcing rice husk ash enhances wear resistance and hardness of the MMCs.
The fabrication of hybrid MMCs is done using stir casting method, in which a molten aluminium matrix metal is mixed with a powder of dispersed of reinforcement particle, followed by casting and solidifying to a solid state. The matrix metal selected was Al 7075, which was received in the form of 2 kg ingot cylindrical rod. The chemical composition of Al The received Al7075 ingots were cut into smaller pieces using bandsaw cutting machine of about 50 to 100 grams. The smaller size will facilitate the metal to melt faster in the furnace.
The ingots after cutting into smaller pieces were washed with water and then dried. Metals of suitable amount, about 200 to 220 grams is placed into a graphite cylindrical crucible for each one experiments. The amount of SiC particles of size 20µm was varied as 3%, 6%, 9%, 12% and 15% relative to the volume of Al7075, each based on the combination given in the experimental design matrix. Dry rice husk was bought and it is burned in the furnace to make powdered ash. The amount of rice husk ash was varied as 2%, 4%, 6%, 8% and 10% relative to the volume of Al7075, each based on the combination given in the experimental design matrix.
The silicon carbide particle due to high density remains at the bottom and rice husk ash due to is low density tend to float on the top of the molten metal, even at high stirring speed. This is due to poor wettability of these two substance with Al molten metal. To improve the wettability, about 1% of Mg particles are added as a wetting agent. Al 7075 ingots, SiC particles, rice husk ash and Mg particles are weighted using a digital weighing machine with an accuracy of 0.01 grams. Pit type electrical furnace was employed for conducting experiments. The furnace was preheated initially before to start to melt. Al7075 collected in graphite crucible are kept steel chamber inside the furnace for melting. The temperature is then increased slowly to 850 0 and maintained at that temperature. The reinforcement particles to be dispersed in the melt is  employed for developing predictive statistical equations for the response measure of the process and to study the influence of process variable within its range [33]. When the input independent process parameter were controllable variable, then response of the process can be represented in terms of process parameters as given in the equation (1) as Where j = 1, 2, ……., N represents the N observation in the factorial experiment, xij = the level of the i th factor in the j th observation and ej = residual measures error of the j th observation. The function ψ is called the response surface. The mathematical formulation of response ψ was approximated within the experimental region using polynomials in terms of process variables.
The benefit of using polynomial response surface is easy to fit. The appropriate choice of design, a quadratic polynomial function in 5 variables which contains 21 coefficients is not an intricate task. The formulated polynomial surface has to be considered as an approximation. A polynomial surface should be considered as an estimation of function ψ within the range of input process parameters. The formulated quadratic polynomials which represent the relationship between the response and input parameter is (2) given below [34] yu = b0 + + + The above equation contains a linear terms in xiu, a quadratic terms in xiu 2 and an interactive terms xiuxju. b0 represents constant terms, bi, bii and bij represents the coefficients of the above equations for constant, linear terms, quadratic (squared) terms and cross product (interactive) terms. The computation of the constant and coefficient can be made easier based on the criterion of rotatability. According to the criterion of rotatability, the standard error of the response 'yu' at any point on the fitted surface can be computed from the results of the experiment. Evaluated standard error will be in terms of the coordinate's xiu of the point. The standard errors are equal and are at same distance ρ from the center of the region for all the points. ie., [35] The rotatable design is obtained by conducting experiments at number of points equally spaced around the circumference of a sphere with center as origin, plus experiments in the replicated center points. The central composite rotatable design is a 2 k factorial design with added 2K star points and replicated central points. The star points are considered with one parameter at lower or higher limits, while the other parameters are set at the middle level. The center points are replicated to provide degrees of freedom to compute experimental error and to determine the precision of the responses. The total number of experiment combinations is (2 k factorial points + 2K star points + replicated center points). Rotatable design for any number k of x-variable can be formed from these three components. The value of β (the distance of centre point to circumference of spherical space) must be 2 k/4 for full replicate and 2 (k-1)/4 for ½ replicate [36].
In this study, the tribological properties were examined by measuring the wear on the prepared specimen using pin on disc apparatus. The input parameters such as % of SiC, % of rice husk ash, load applied, speed of rotation and sliding distance were considered as process parameters  factorial points comprises of code value of -1 or 1 of each parameter.  The range of upper limit and lower limit of the each parameters are coded as +2 and -2 respectively, and for in-between were evaluated from the equation (5) given below [31].
Where Xi is the coded values of the process variable X, X is the value of the process variable inbetween the range of Xmin to Xmax, Xmin is the lower limit of the variable and Xmax is the upper limit if the variable. The in-between values are coded as -1, 0 and 1.

Wear test
The tribological behaviour of the prepared specimens was analyzed by conducting dry sliding  value of F-ratio of the model is less than the standard value and the estimated value of R-ratio is more than the standard value for a desired 95% level of confidence. The error between the observed values from experiments and predicted values from the equation 6 is found to be less than 5%. This is evident that the equation derived is capable of prediction to the near accurate value.

Results and discussion
The developed statistical model was used to envisage wear by substituting the particular values of the input process parameters. The effect of these process parameters on wear of MMCs was studied using the developed empirical model.

Direct effect of process parameters on wear
The direct effect of input parameters such as % SiC, % rice husk ash, speed of rotation of disc, load and sliding distance on wear were studied using the statistical empirical model. The direct effect was studied by varying levels of one process parameter whose effect need to be studied and the remaining parameter are kept at the middle level. The direct effect of input process parameter on wear were manipulated and plotted are shown in the figure 3. The cause and effect of each process parameter on wear were analyzed and discussed below. The effect of % rice husk ash particles on the wear loss of the composite is depicted in the figure   3 (b). The graph shows that the loss of material due to sliding wear decreases with an increase in % of rice husk ash in MMCs from 2% to 6%. From 6% to 8%, the wear loss don't show any significant change as it is in the same level. A slight increase in wear is observed when the % of rice husk ash increase from 8% to 10%. Addition of rice husk ash reduces wear rate due to its inherent wear resistance properties. This can be exhibited from the graph the wear loss decreases with an increase in % of rice husk ash. It can be concluded that the wear loss is minimum when the % of rice husk ash in the MMCs is between 6% -8%.

Interaction of process parameters on wear
The interaction effect aims in studying the change in influence of one process parameter on the response, while the second process parameter is changed from one level to another and the remaining process parameters are kept at the middle level. This analysis helps to understand the  to 8% resulted in decrease of wear and at 10% Rice husk ash, slight decrease in wear exhibited.
The similar trend is observed at last two levels of load, an increase in %Rice husk ash, for 2%, 4% and 6% resulted in decreased wear and an increase in %Rice husk ash, for 8% and 10% resulted in increased wear.

Optimization of process parameter
Genetic algorithm (GA) is a nondeterministic influential technique used in search and optimization of problems defining process with discrete independent variables [37,38]. Create the primary population at random using coding to represent process parameters, a crossover operator with probability and a mutation probability with probability. The maximum allowable generation number to succeed is 100.
ii   augmented with better string from the earlier generation after reproduction. A crossover operator in used to generate better strings in the population. In GA, crossover probability select two parent strings at random form the mating pool and combined them by exchanging some portion of the strings to form a two new offspring's. Point crossover and uniform crossover are two general crossover operator used for integer strings. In this paper uniform crossover with probability of 0.9 was used to populate new offspring's. A mutation operator is usually based on probability distribution such as normal, uniform and Cauchy distribution is a bit flipping operation for integer strings. In mutation, thorough bit flipping operation flips a bit from 1 to 0 and 0 to 1based on mutation probability. In this paper bit flipping mutation with probability of 0.1 was used to mutate a new offspring's.
After deploying GA operators, a fresh set of population is generated. This entire process completes one generation of GA. Such iterations are continued till the termination criterion is achieved. The iteration is repeated for a population size of 100 for 100 generations. Then, they are decoded and objective function values are calculated. The above process is simulated by a computer program developed using C language for computations. Figure 8

Conclusion
The experiments were conducted on hybrid MMCs to predict wear and to study the influence of process parameters such as % of SiC, % of rice husk ash, load applied, speed of rotation and sliding distance. The microstructure of fabricated composite materials were observed and analyzed. The following conclusions were obtained after detailed analysis and exploration. 5. The adequacy of the empirical equations developed was checked using ANOVA technique, validity of the model is checked by conducting conformity test.
6. The empirical model developed for prediction of wear can be further improved and used to optimize the process parameters using evolutionary algorithm.
7. The microstructure of fabricated MMCs reveals a uniform distribution of reinforcement particulate in matrix material which ensure better quality and enhancement in mechanical properties.
8. The genetic algorithm has been employed to optimize the process parameters to obtain the minimum wear in MMCs. The optimal combination of process for minimum wear of 34.5µm was found to be 12 0 , 7 0 , 1.9 m/s, 11.5 N and 715 mm for % of SiC, % of rice husk ash, sliding speed, load and sliding distance, respectively

Ethics approval and consent to participate
The authors of this paper declare that all procedures followed were in accordance with the ethical standards.
This paper is our original unpublished work and assures that it has not been submitted to any other journal for reviews.

 Disclosure of potential conflicts of interest
There is no conflict of interest in this work.