Tribological and machining characteristics of AA7075 hybrid composites and optimizing utilizing modified PROMETHEE approach

In this research work an attempt was made to reinforce AA7075 composites with B4C and SiC particles through stir casting route. SEM with EDS mapping revealed that the reinforcement were uniformly over the composites and hardness reduces with the addition of SiC particles owing to the inverse hall petch effect. The results revealed that wear rate reduces with addition of SiC particles owing to the formation of mechanically mixed layer and protective oxide layer confirmed through SEM with EDAX mapping. Three distinct cracks were formed, when slides at different temperature as confirmed through worn surface morphology, pits cracks and plasticization of material were the other features observed. The used motor oil properties were analyzed and results divulged that the oil suitable for dielectric fluid. Increase in Material Removal Rate (MRR), reduction in Tool Wear Ratio (TWR) and Surface Roughness (Ra) was observed with the incorporation of powder particles owing to the bridging effect. Black spots, craters, micro pits, globules and micro crack are the distinct observed on the machined surface topography. The modified Preference Ranking Organization METHod for Enrichment of Evaluations (PROMETHEE) optimization technique was utilized to find the optimal parametric combination.


Introduction
Aluminium Matrix Composites (AMC) were in great demand in the aerospace sector because of their high strength-to-weight ratio, improved tribological and mechanical properties [1]. The reinforcements were in the form of particulates, fibres and whiskers; powder metallurgy, casting, in situ fabrication and laminates are the distinct method utilized for the production of composites [2]. Attaining uniform distribution of reinforcement over the matrix material was the challenging task [3]. Preheating and addition of flux increases the wettability of composites. Stir casting, two step stir casting, ultrasonic stir casting, and electromagnetic stir casting are some of the stir casting technologies used in composite manufacturing [4][5][6][7]. In the stir casting technique, the mixture was heated to the molten state before being stirred by the mechanical stirrer [8]. In two step stir casting process, the material was heated above the liquidus temperature, and to keep the slurry semi-solid, it was gradually cooled below the liquidus temperature. Manual mixing was performed after the preheated particles were introduced. After reheated the composite slurry to a liquid state, mechanical mixing was performed for the prescribed time [9]. In the process of compo casting, the charge was heated to the alloy's melting point before being cooled to a semi-solid condition (0%-33% in solid fraction), using a paddle-like blade, semi-solid composites were whirled [10]. The electromagnetic field was produced by AC induction motor which was used for stirring of matrix and reinforcement hence uniform dispersion achieved [11]. An ultrasonic aided probe was employed for mixing in ultrasonic vibration followed by conventional stirring [12].
One of the most common industrial problems is wear, where the primary factors influencing the wear rate are velocity, temperature, distance and operating load. Composite pins wear less when lubricated, compared to unlubricated and abrasive surfaces [13]. When the particulates were added, it enhances the wear resistance because of the formation of Mechanically Mixed Layer (MML). Wear rate transfer from mild to severe when this melt and stirred at the speed of 1000 rpm. The 1wt% of magnesium powder was added as the flux to improve the wettability and the molten mixture was bottom poured into the preheated mould made of die steel of geometry 105 mm × 12 mm (l × f). The composites were machined to the dimension of (100 mm × 10 mm) to eliminate its surface defects. The same technique was used to produce a hybrid composite with different SiC weight percentages (0, 2.5, 5, 7.5, and 10) while keeping the B 4 C weight percentage constant at 5%.

Mechanical and tribological properties of composite
The uniform distribution of composites was analysed through SEM with EDS mapping. The Rockwell hardness were measured according to the ASTM standard ASTM E18-22, liquid immersion technique was utilized to find experimental density and it was compared with the theoretical density to compute void fraction. The wear experiments were conducted as per the ASTM standard G-99 and runs were designed using the Taguchi orthogonal array as depicted in the table 2. The parameters %reinforcement, applied load, sliding velocity, distance and temperature were varied for five distinct levels whereas wear, coefficient of friction and worn surface hardness were measured as the responses. According to the equation (1), the wear rate was calculated and it was measured in unit mm 3 /min. The hardness were measured at 10 distinct spots using Rockwell hardness testing machine as per the standard ASTM E 18-07 and average value was taken as the worn surface hardness [40]. The worn surface morphology was measured using the SEM.

EDM of composites
The next module of work was the EDM of AA7075 composites and experiments were performed on the die sink EDM machine in which UMO was used as dielectric fluid. The parameters current, gap distance, pulse on, powder concentration and polarity were varied as shown in table 3, in which MRR, TWR and Ra were recorded as response, runs were designed using the Taguchi mixed orthogonal array. the MRR and TWR were calculated according to the equations (2) and (3). Ra was calculated using the Mitutoyo roughness tester at 10 different places and average was recorded as the Ra value. The Machined surface morphology was analysed using the SEM. The parameters were optimized using the modified PROMTHEE approach Wb and Wa-Weight of the work piece before and after wear test Sb and Sa-Weight of the work piece before and after machining Tb and Tb-Weight of the electrode before and after machining ρ-density of the material t-Machined time in sec

Modified PROMETHEE
In the PROMETHEE approach, the performance of the alternatives was compared with each other and optimal solution were identified. The approach begins with formation of decision matrix, say the problem has n alternative and it has accessed in terms of m performance a decision matrix of (nxm) was formed. The next step was the making all response as the unit less value, which was referred as the normalisation, the normalisation matrix Bij is the ratio of the ith element of the response to the square root of the sum of all the square values of that specific response as shown in equation (4). The weighted normalized matrix is the product of weight and the normalized value as shown in equation (5).
The optimal best value for the beneficiary element is the maximum of the weighted normalised decision matrix, while the optimal worst value is the minimum of the weighted normalised decision matrix, and vice versa for the non-beneficiary attributes as shown in equations (6) and (7). For Beneficiaries The distance between the C ij element to the ideal best and worst value element was computed as the ideal best and ideal worst as depicted in equation (8). The ratio of the ideal worst to the sum of the ideal best and worst was taken as the assessment value as shown in the equation (9). The parametric combination with the highest assessment value was taken as the optimal solution. å = -+ - The next step phase of the optimization technique was the comparing the performance of the individual, it begins with the summation of the all the assessment value of that specific candidate. For example, if m no of alternatives were available, n criteria was utilized for evaluating performance and a surrogative decision matrix of (mXn) was formed as shown in equation (10). The difference between the alternatives viz (m 1 −m 2 ), (m 1 −m 3 ), (m 1 −m n ) were computed for all the responses and it was termed as the preference value. The next step was to compute the aggregate preference value; the preference value can be taken as it as for non-negative elements and 0 for negative components as depicted in the equation (11). After the formation of aggregate preference value, the next step was the calculation of entering and leaving flow as shown in equations (12) and (13). The entering flow is the summation of the rows whereas leaving flow is the summation of the columns of the aggregate preference matrix. The net flow is the difference between the leaving flow and entering flow as depicted in the equation (14). The alternative with the high net flow was considered as the optimal.

Module 1
Tribological behaviour of composites The microstructure of the composites revealed that the particles were dispersed consistently across the composites, as seen in figure 1. Each element was presented in a different colour, with reinforcements displayed in black and equally distributed over the matrix material. The hardness of the composites increases until the 2.5 SiC weight percentage and there after it declines attributed to the fact inverse hall petch effect and increase in void fraction as shown in figure 2. With the addition of flux, it refines the grain particles, hence hardness reduces, similar trend was observed by the distinct researchers. According to an inverse Hall-Petch, materials get softer when grain size is decreased below a threshold level [41].
The impact of distinct process parameters on the wear rate of composites were shown in the figure 3. From the graph it was confirmed that wear resistance enhances when 2.5% of SiC particles were added to matrix material. With the inclusion of particles, it removes material from the contact surface while sliding, resulting in third body abrasion and the formation of Mechanically Mixed Layer (MML) [42], as confirmed by the EDAX analysis shown in figure 4. When the particle concentration exceeds 2.5%, the wear rate increases to a maximum of 0.0089 mm3/min for SiC (10wt%). From the trend line, it was evident that wear rate increases with raise in temperature and maximum wear rate of 0.0085 mm3/min was observed at the temperature of 250C. As it reaches plastic deformation at higher temperatures, aluminium loses its hardness [43]. At this point, the detached reinforced particles had ploughed material from the composites pin as a result of the third body abrasion. The wear rate increases steadily until the saddle point of 60N, there after it declines. At higher load, the  contacting surface exerts high pressure which results in the breakdown of MML leads to the direct metal to metal contact, hence wear rate increases. The transition of wear rate from the mild from the severe occurred at the load of 45N, which confirmed that the developed composites resolute until 45N. The sliding velocity followed the similar trend, as wear rate raises until 3.6 m s −1 beyond that it reduces. When tested at the temperature of above 150C, At higher load and sliding speed the temperature of the composite pin reaches 469C measured using the thermocouple, which results in the formation of tribo rich Oxidation Layer, which was confirmed through the presence of O in the EDAX analysis [44]. The temperature on the pin's surface rises quickly and the pin surface exhibits significant distortion at greater sliding distances. The MML was broken down by the distance increase, resulting in delamination wear.
The frictional behaviour of AA7075 hybrid composites was 0.201 when 2.5% SiC particles were added, 0.221 when 7.5% SiC particles were added, and 0.225 when unreinforced composites were added as shown in figure 5.  The decrease in COF with SiC addition was attributed to the formation of MML between the surface; the findings indicated that 2.5wt% produces optimal MML, and therefore these composites had the lowest wear rate, which was well correlated with the experimental data. In terms of load and sliding distance, the MML layer breaks at parametric values of 45N and 3000 m, respectively, implying that beyond these levels, the wear rate shifts from mild to severe. When sliding at high velocity and temperature, materials reach a plastic state, resulting in a decrease in hardness and a drop in COF. From the figure 6 it was evident that the temperature was the most influential factor which impacts the worn surface hardness [45]. The increase in worn surface hardness was attributed to the following facts (i) composites pin attains the temperature of 469C, hence recrystallization occurs, further investigation required (ii) because of transfer of Fe metals from the counter face, evident from EDAX mapping.
The worn surface topography of the pin machined at 150C was depicted in the figure 7(a). The cracks and pits were clearly visible on the worn surface. The cracks starts from distinct region and ends up in the pit and this type of crack was termed as Type 1 crack. The materials shifted towards the surface rather than being removed from it, indicating ploughing wear. Because of the high temperatures, some of the materials deformed and resolidified on the surface. At the magnification of 5000×, the plastic flow was evident which confirmed that the materials reaches it deformation temperature as shown in figure 7(b). The pits are scattered over the surface and its size ranges from 1 μm to 5 μm. When slides at the temperature of 50°C, micro crack of size 5-6 μm were observed, it has definite starting and ending point and was classified as type 2 crack. The resolidified materials and larger number of tiny pits were clearly visible on the surface as shown in the figure 8(a). At 5000×, nano pits of size 0.25 μm to 0.5 μm were clearly visible. The plasticized of materials occurred and mode of wear was observed as adhesive as shown in the figure 8(b). At the temperature of 200°C, the cracks initiated at different points and ends in a single point, classified as the type 3 cracks was clearly visible as depicted in the figure 9(a). Direct metal-to-metal contact results from the breakdown of the MML and protective oxide layers, which creates a conduit for delamination wear. Cracks, micro pits and material ploughing were clearly visible at the higher magnification as shown in the figure 9(b).
The next phase was finding the optimal solution from the available alternative, as 25 experiments were conducted and 3 three responses were measured, a decisive matrix of 25 × 3 was formed as depicted in the table 4. The responses were squared up, and the normalised value was calculated by dividing the response value by the sum of square roots of squared up values, which was then multiplied by the weight to form the Weighted Normalised Decision Matrix (WNDM). The ideal best was calculated as the square root of the difference between the WNDM and the maximum WNDM, and the ideal worst as the difference between the WNDM and the minimum WNDM. The composites are capable of withstanding a load of 60N, a sliding speed of 2.4 m s −1 , a sliding speed of 5000 m, and an operating temperature of 150C, according to the experimental run number 19, which also exhibited the highest net value as shown in the table 5.
The following phase is a performance comparison of the produced composites, in which produced composites performance (0, 2.5, 5, 7.5, and 10wt%) were compared to each other. The process starts with the creation of a surrogate decision matrix, which is the sum of all the assessment values calculated for that specific composite as shown in table 6. The difference between the alternatives that is (m1-m2) (0.67001-0.52198) were calculated, negative values were taken as zero and positive as it is, and the formed matrix were named as aggregate preference function as depicted in table 7. The assessment value was taken to be the maximum value of aggregate performance, and each assessment value of that particular composite was denoted in the leaving and entering flow matrix as shown in table 8. The net flow was calculated as the difference between the leaving flow  table 9, and the alternate with the highest net flow was chosen as the best. According to the results composites AA7075/5B 4 C/2.5SiC was the best alternative and it can withstand the load upto 60N and it can operate until the temperature of 150°C.

Module 2
Electric discharge machining of composites In this work an attempt was made to utilize UMO as the dielectric fluid with the objective of attaining wealth from waste. To access the feasibility of the UMO as the dielectric fluid its characteristics were compared with the conventional dielectric fluid. The flash point of UMO is 221°C higher than Hydro Carbon Oil (HCO) which ensures enhanced worker safety [46]. The density of UMO is 900 Kg/m 3 , hence better flushing occurs and break down strength is 63 KV results in larger machining cycle. The carbon content ranges from 91%-94% which may leave black spots on the machined surface. The kinematic viscosity and thermal conductivity are 2.3 mm 2 /s and 0.136 W/m.k, respectively, ensuring higher heat transfer and dissipation. The characteristics of UMO was depicted in the table 10.
Machining composites is a challenging task because particles in the matrix cause excessive wear, hinders surface quality. EDM machining was accomplished by melting and vaporization, using heat generated by the spark and no direct contact with materials [47]. The trend line in figure 10 demonstrated that when connected to the positive polarity, the composites proffered 28.5% higher MRR. Because the preponderance of the heat   produced during the spark cycle was absorbed by the materials connected at the cathode, numerous researchers reported comparable results. The results showed that at unmixed UMO medium a MRR of 0.20 mm3/min was and it was increased to 0.28 mm3/min, when 4 g l −1 of Al 2 O 3 particles were added to the dielectric medium. The rise in MRR was ascribed to the fact that when particles enter the spark gap, they migrate in a zigzag pattern between tool and electrode, resulting in the bridging effect, which causes an increase in heat intensity and hence an increase in MRR. Densification of machined debris occurred with 6 g l −1 of Al2O3 particles were added to the dielectric fluid, resulting in a reduction of MRR. The higher current produces a high-intensity spark, which generates more heat and hence removes more materials from the workpiece, as evidenced by the experimental results. At higher parametric level of current, MRR reduces owing to the widening of plasma channel. The generated spark was kept inside the machined gap for the prescribed duration of time referred to as the Ton, higher Ton significantly improves MRR [48]. When the parametric level exceeds 45 μs, the MRR decreases due to inadequate flushing since a substantial volume of materials were removed from the surface. The span between the electrode increases with raise in gap distance hence lower heat intensity and reduction in MRR.
In the present investigation copper was used as the tool material, lowering the TWR and attaining High MRR improves the productivity of EDM sector. Low TWR was recorded when the electrodes were linked in the positive polarity, indicating that the majority of the produced heat was transferred to the work piece as shown in figure 11. TWR decreases until a concentration of 4 g l −1 , owing to the fact that the spark distance decreases due to the bridging effect, hence a minor expansion in the spark gap occurs to maintain the spark distance, resulting in decreased heat intensity and lowered TWR. In terms of current and Ton, TWR increases when either value increases due to the formation of a high-intensity plasma channel. The TWR reduces with upsurge in the gap distance up to the saddle point of 3 mm thereafter it raises [49]. Figure 11. Impact of distinct process parameters on the TWR of hybrid composites. Positive and negative polarity offer Ra of 2.93 μm and 2.97 μm, respectively, hence the polarity has minimum impact on the Ra of the composites as shown in figure 12. With an increase in current, the surface quality degrades because a high current produces a very intense spark that forms craters and tiny pits on the surface. Ra has no impact with the addition of 2 g l −1 of Al 2 O 3 particles, but when the concentration was increased to 4 g l −1 , the lowest Ra of 2.75 μm is achieved. The improved surface quality was engendered by the complete flushing of machined debris caused by the bridging effect. When the Ton was tuned at the parametric  level of 30 μs, an average Ra of 2.41 μm was obtained; however, when the Ton was increased to 60 μs, the average Ra dramatically increased to 3.75 μm due to the densification of the plasma channel, which prompted the formation of a remelted layer on the surface. The surface quality improves with increase in gap distance as the machined debris were completely flushed away which eliminates the formation of globules and remelted layer [50].
The machined surface topography of AA7075 hybrid composites machined with pure UMO were depicted in the figure 13(a). The topography revealed black spots all over the surface, indicating that the carbon content present in the dielectric fluid deposit over the surface. Craters formed on the surface as a result of the excessive heat generated by the spark. The eczema layer was clearly visible on the surface topography due to the uneven distribution of heat. Resolidified materials were clearly visible at a higher magnification of 2000X, as shown in the figure 13(b), confirming that inadequate flushing of machined debris. When machined with 6 g l −1 Al 2 O 3 incorporated UMO dielectric medium, black spots and crust were clearly visible on the surface as depicted in the figure 14(a). The crust and trough were clearly visible which revealed that generated heat was unevenly distributed over the surface. At higher magnification as in figure 14(b), globules ranging in size from 5 μm to 10 μm were clearly visible, as shown in the figure, exacerbated by insufficient flushing of machined debris. The other features observed on the surface were micro pits and micro cracks owing to which Ra deteriorates. At 4 g l −1 Al 2 O 3 particles incorporated dielectric medium, micro pits and craters were visible on the surface as depicted in the figure 15(a). Owing to the higher kinematic viscosity, machined cycle occurred more frequently, hence    attaining uniform distribution was the challenging task. At higher magnification as in figure 15(b) reveals micro pits, micro cracks, remelted layer and globules on the topography. In this case a decision matrix of 32 × 4 was formed as shown in table 11, it was normalized and weighted normalised matrix and its assessment value were shown in the table 12. In this case only 4 alternatives were available, hence a surrogative decision matrix of 4 × 4 was formed as shown in the table 13. The machining performance under distinct powder concentration were compared with each other as portrayed in the table 14, finally leaving and entering flow matrix was computed as shown in the tables 15 and 16. The results revealed that when 4 g l −1 were added, connected to the positive polarity, machined with 28A, 60 μs and 4 mm proffers best machining performance.

Conclusion
The AA7075 hybrid composites were successfully fabricated through stir casting route and following conclusion were obtained (1) The reinforcements were uniformly distributed throughout the matrix material, confirmed through SEM with EDAX mapping. Owing to inverse Hall Petch effect hardness reduces with the addition of SiC.
(2) The wear rate rises as the temperature does, and the highest wear rate was noted at a temperature of 250C. Higher temperatures cause aluminium to lose its hardness as it approaches plastic deformation. Because of recrystallization and the transfer of Fe metals from the counter face, the hardness of worn surfaces rises.
(3) The characteristics of UMO indicated that it can be used as a dielectric fluid. The MRR of composites rises due to the bridging effect and Ra due to the thorough flushing of machining debris. Black patches were seen across the topography, which revealed that the dielectric fluid's carbon component had deposited on the surface.
(4) The modified PROMETHEE approach were successfully applied for optimizing the parameters. The composites AA7075/5B4C/2.5SiC, which can sustain loads up to 60 N and function at temperatures as high as 150°C, were found to be the best solution. Composites machined with 28A, 60 s, and 4 mm offers the highest machining performance when 4 g l −1 were added, coupled to the positive polarity.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).