Experimental and optimised data set for hot extrusion of B4C/Al 6061 composite using Taguchi coupled GRA technique

Modern aluminium composites reinforced with variety of hard particles become more usage in Industrial environment. But manufacturing of composites with homogenously distributed reinforcement becomes the challenge. To overcome this challenge, most of the Aluminium composite are undergone a secondary extrusion process. The data presented here are related to hot extrusion of round geometry to hexagonal section Al/B4C composite. Availability of data is extended to expose the optimal parameters of the process over the extrusion load and tensile strength of the extrudate. Ram speed, geometry of die profile, billet temperature and friction within the die and billet interface have been considered as chief process parameters which influence the extrusion load and strength of the product. Totally, nine experiments were conducted as per Taguchi's L9 orthogonal array to reach optimal parameters. Most influencing parameters with ranking significance have been arrived through ANOVA, MRPI and grey grade. Optimal parameters were compared with confirmation experiments and predicted one to justify the investigation


a b s t r a c t
Modern aluminium composites reinforced with variety of hard particles become more usage in Industrial environment. But manufacturing of composites with homogenously distributed reinforcement becomes the challenge. To overcome this challenge, most of the Aluminium composite are undergone a secondary extrusion process. The data presented here are related to hot extrusion of round geometry to hexagonal section Al/B 4 C composite. Availability of data is extended to expose the optimal parameters of the process over the extrusion load and tensile strength of the extrudate. Ram speed, geometry of die profile, billet temperature and friction within the die and billet interface have been considered as chief process parameters which influence the extrusion load and strength of the product. Totally, nine experiments were conducted as per Taguchi's L9 orthogonal array to reach optimal parameters. Most influencing parameters with ranking significance have been arrived through ANOVA, MRPI and grey grade. Optimal parameters were compared with confirmation experiments and predicted one to justify the investigation © 2020 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license.
( http://creativecommons.org/licenses/by/4.0/ ) Value of the data • The data finds a pivotal role to demonstrate the optimum process conditions of hot extrusion not only for extruding Al/B 4 C composite but also for all materials. • The data are found useful to fellow researchers and industrialist to know the intense effect of three chief extrusion process parameters over the extrusion load. • The data can be useful to compare further optimisation of processes through any of the conventional or unconventional way of optimisation technique for the hot extrusion process. • The data presented here are much more beneficial for the design and development of extrusion dies.

Data description
The data presented in this article are both raw and analysed in natures which are related to hot extrusion process of Aluminium 6061 alloy reinforced with boron carbide particles as shown in Fig. 1 . Table 1 shows the specification details of Scanning Electron Microscope employed for the experiment. The chemical composition of Aluminium 6061 and Boron carbide are tabulated    in Table 2 and Table 4 , respectively. Mechanical properties of Aluminium 6061 and Boron carbide are shown in Table 3 and Table 5 , respectively. Experimental condition followed for each trial has been given in Table 6 . Selection of process parameters and their levels with corresponding values are given in Table 7 . Observed values of responses (extrusion load and tensile strength) through nine experiments under the principle of Taguchi's L9 orthogonal array have been tabulated as a raw data in Table 8 . Table 9 depicts the manipulated S/N ratios and normalised S/N ratios. Computed grey coefficients and grade has given in Table 10 . The ranking effect of each parameter over the response can be known from Multi Response Performance Index which has tabulated in Table 11 . The percentage of contribution over the impact of response of each parameter has been given as ANOVA Table 12 . Table 13 interprets the comparison of grey grade between random and optimal parameters. Fig. 2 depicts the stir casting set up and the casted specimen is shown in Fig. 3 . The schematic view of three different profiles used for experiment

Materials
Boron carbide particles with particle size of 35 μm observed through Scanning Electron Microscope whose specification is shown in Table 1 , were used to reinforce the proposed composite. Microscopic view of the particles is shown in Fig. 1 . Aluminium 6061 alloy was utilised as matrix material.

Methods
Stir casting set up as shown in Fig. 2 was employed to fabricate the composite. Initially Aluminium alloy 6061 whose chemical composition and mechanical properties has been given in Tables 2 and 3 respectively was melted at 750 °C in a crucible, made out of graphite. Then, B 4 C particles whose chemical composition and mechanical properties has been given in Tables 4  and 5 respectively were preheated up to 500 °C and mixed with molten bath. In order to achieve a homogenous mixture of composite, stirring process with nearly 400 rpm was maintained for about 45 minutes. The mixture was poured in to the cylindrical mould having with dimensions of 12 mm diameter and 25 mm long. Fig. 3 shows the casted composite. Experimental extrusion     was carried out by following the conditions given in Table 6 , with the help of UTM shown in Fig. 5 (a). Three extrusion dies with different die profile namely, fillet radius, conical and cosine curve, illustrated in Fig. 4 were manufactured through Die sink and wire cut EDM and one of the die manufactured with cosine profile is shown in Fig. 5 (b) and typical cosine curve has been shown in Fig. 6 . Initially, the casted specimen was heated to a given temperature within muffle furnace. The hot round billet was loaded in the die and then the punch was used to press the

. Selection of process variable and their levels
Extrusion is one of the important metal forming operation, which is highly influenced by geometry of the die profile, initial billet temperature, friction condition between the die and billet interface and also ram speed. These four parameters were considered to be more significant and their levels were also decided to cover low, medium and higher region of magnitudes so as to accomplish optimal parameter set, Whenever frictional factor is a matter of concern during the experiments, die surface was fully lubricated by applying a graphite powder as higher level, partially lubricated as medium and zero lubrication or dry extrusion as low level.

Design of experiment
Taguchi's method of experimental design can be found very much useful in solving the complex engineering problems with lean data. Implementation of orthogonal array paves the way to decrease the number of experiments drastically. Each column in OA represents important parameters which influence the responses. The degree of freedom and number of trials were decided by the number of parameters and their levels. The order of parameter level for each trial has mentioned in Table 8 .
Whenever, the extrusion process is considered, load needed to extrude the billet becomes a significant response because it decides the press capacity [1][2][3] . Moreover the importance of secondary extrusion process for the processing of composite lies with homogenous distribution of reinforcement in order to enhance the strength of the composite, hence tensile strength of the composite being another objective [4] . It is most important to know the degree of influence of the extrusion parameters over these objectives with optimised way. However, inclusion of two responses changes the problem into multivariable approach. As per the Taguchi's robust design    of experimental approach, there shall be nine experiments conducted based on L9 OA [5] . For each experimental trial, the extrudate shown in Fig. 5 (C) was made into observation of extrusion load and tensile strength using UTM.

Grey relational analysis
GRA can be applicable to evaluate the problem with more than one objective. The multi response optimisation can be changed in to single objective problem. The data observed through experiments were analysed and normalised between zero to one so as to generate the grey relational coefficients. Initially the response data recorded in experiment trials were transformed into S/N ratio. The effect of response in terms of larger or smaller was arrived. For the present case, one of the objective extrusion load to be as lower as possible and another objective tensile strength should be as higher as possible are preferred. S/N ratio were calculated as per smaller the better type and larger the better type approaches as follows, For smaller the better type S N ratio = −10 log 1 n n i =1 For larger the better type The purpose of normalisation is to express the analysed data in to single decimal ranging from 0 to 1. The following relation was employed to execute the normalisation under smaller the better type approach. Z ij = normalised value for the i th experiment/trial for the j th response From the normalised S/N ratio, the grey relational coefficient can be manipulated by using the following relation, Where GC ij = Grey Relational Coefficients for the i th experiment/trial for the j th response = Absolute difference between Y oj and Y ij which is a deviation from the target value and can be treated as quality loss Y oj = Ideal normalised value of the j th response Y ij = the i th normalised value of the j th response min = Minimum value of max = Maximun value of λ = Distinguishing coefficient defined in the range 0 ≤ λ ≥ 1 The Grey Relational Grade (G i ) can be determined with the help of following relation, The manipulated Grey Relational grade now can be equated with Multi Response Performance Index (MRPI), so as to convert the multi objective problem into single objective. The optimal parameters required to extrude the composite with minimum extrusion load and maximum tensile strength can be earned through MRPI data.
It is quite clear from the MRPI analysis that, the effect of die profile geometry is more significant than the rest of other three parameters followed by the effect of friction. The higher values of MRPI are taken into consideration for arriving the optimal parameters. It is quite clear from the grey grade graph presented in Fig. 7 which confirms the optimal level of parameters for the extrusion process as A 3 B 3 C 3 D 1.
The main intention of constructing ANOVA table is to evaluate the quantum of significance of each parameter over the responses. The rank and percentage of contribution of each parameter can be known through ANOVA. From Fig. 8 , it is very much clear that the die profile has got the most influencing characteristic over the responses by achieving highest contribution of 50% followed by friction factor as second and billet temperature as third influencing parameters. It is evident that ram speed becomes the least significant parameter in the study. The physical reason behind this influence has observed that, the nature of profile with uniform curvature of cosine profile ensures the homogeneity in plastic deformation and material flow by preventing the chances of agglomeration of B 4 C particles near the die entrance [6] . Development of more frictional effect within die and billet region makes the peripheral layer of the work material to deform much slower than the central zone. This imbalanced strain behaviour during plastic deformation obviously requires more extrusion load, which in turn shows the importance of friction factor over the extrusion process. At higher temperatures, the molecular bonding between adjacent molecules becomes diffused, which in turn causes quick deformation by absorbing minimum extrusion load than at low temperature.

Verification of optimal parameter through confirmation experiment
Confirmation experiment was carried out by following the optimised process parameters A 3 B 3 C 3 D 1 . The experiment was conducted with a ram speed of 12 mm/min, with cosine profiled die, applied with full lubrication and with a billet temperature of 400 °C. Where α predicted is the grey relation grade for the predicted parametrrs. α m is the mean average of the grey relational grades. α o is the average grey relational grade of the optimal level of the fluid parameters (A 3 B 3 C 3 D 1 ) and ' n' is the number of significant factors considered from the response table. The computed predicted grey relational grade was 0.689.
Comparison table confirms that, the difference between predicted and confirmation experiment is within the allowable value confident interval. Hence this type of statistical investigation through minimum number experiment can be very much useful in solving complex problem in extrusion Industries to arrive for the optimal process parameters over the quality and product cost.