Optimization of Turning Process during Machining of Al-SiCp Using Genetic Algorithm

Transcription

1 Optimization of Turning Process during Machining of Al-SiCp Using Genetic Algorithm P. G. Karad 1 and D. S. Khedekar 2 1 Post Graduate Student, Mechanical Engineering, JNEC, Aurangabad, Maharashtra, India 2 Mechanical Engineering, JNEC, Aurangabad, Maharashtra, India Abstract Genetic algorithm has been recognized as one of the most popular multi-objective optimization techniques. In this work genetic algorithm has been used to optimize the CNC turning process parameters. Objective of the present study was to develop empirical models for predicting material removal rate and surface roughness in terms of speed, feed rate and depth of cut using multiple regressions modeling method. Experiments were carried out on CNC machine tool by taking Al-SiCp (10%) MMC as workpiece material and CBN inserts as cutting tool. The set of Pareto-optimal front provides flexibility to the manufacturing industries to choose the best setting depending on applications Keywords Aluminium metal matrix composites, Turning, Genetic Algorithm. I. INTRODUCTION Manufacturing industry needs to produce a large number of products within relatively lesser time. However it was observed that reduction in manufacturing time may cause severe quality loss. To deal with such situation of these two conflicting criteria it is necessary to check quality level of the item either on-line or off-line. Turning is a machining process used to obtain the desired dimension of round metal. The main objective in present industrial era is to produce low cost quality product with required dimensions in an optimum time. Therefore the optimum cutting parameters are to be recognized first. The composite material under consideration, which is cast A356 aluminum alloy reinforced with 10% volume fraction of SiC particulates, is cited as Al-SiCp (10p). These metal matrix composites (MMCs) using A356 aluminum as the matrix material with SiC particles reinforced in it (A356/SiCp), have found vast applications in automotive, aerospace, Electronics, Medical, Recreation and other allied fields, which have aggressive environments [1,2]. The combination of ceramic particles in Al alloy increases both mechanical strength and wear resistance of the composite. Machining becomes difficult with the hard abrasive SiC particles in Al SiC composite. Thus the machinability of particulate MMCs can be improved by reinforcing soft particles like graphite along with hard ceramic particles [3]. There is need of advance materials which can fulfill the requirement of automotive and aerospace industries, Metal matrix composites (MMCs) are the new age materials that are being preferred for their improved properties. Compared to conventional metals and alloys, these materials gives higher strength to weight ratio, hardness, stiffness, wear resistance etc. as. However, these very properties make these material difficult to machine [4]. There are several multi-objective optimization techniques for the same like goal programming, simulated annealing (SA), grey relation, and genetic algorithms (GA). GA is very different from most of the traditional optimization methods. GA finds applicability in the field of conventional machining processes. It works with a random population of solution points and a set of Pareto-optimal solutions is obtained for the best performance measures. Suresh et al. [5] developed a mathematical model for predicting value of surface roughness while machining mild steel using response surface methodology and optimized the developed model using genetic algorithm, in order to attain the required surface quality All rights Reserved 239

2 II. EXPERIMENTAL SET-UP Experiments are carried on CNC lathe of Hass automation USA. Experimental set-up is shown in figure 1. Al-SiCp MMC was used as work piece material of dimension ϕ 40 mm x 105 mm long and CBN insert as cutting tool. Figure 1. Turning set-up In this study, spindle speed, feed rate and depth of cut were considered as machining parameters and turning was carried out. Experiments were designed using L9 Taguchi orthogonal array. Table 1 shows the machining parameters and their levels. Table 1. Machining parameters with their levels Speed (rpm) Feed (mm/rev) Depth of cut (mm) The weight of specimen is taken before and after the machining process using a digital weighing machine. (Model-SC400, Max wt.: 7000 gm; Min wt.: 1 gm DIGITAL reading electrically operated). To measure the surface roughness of machined work piece the surface tester of Taylor Hobson surtronic 3 series was used. It can evaluate 36 kinds of roughness parameters conforming to the latest ISO, DIN, and ANSI standards, as well as to JIS standards (1994/1982). III. RESULT AND DISSCUSSION 3.1 Multiple regression models: The turning experiments were conducted by using the parametric approach of the Taguchi s method. Regression analysis has been performed to find out the relationship between input factors and responses using Minitab 16 statistical software. During regression analysis it was assumed that the factors and the responses are linearly related to each other. General first order model was developed to predict the material removal rate over the experimental region (equation 1). MRR = SPEED FEED DOC.. All rights Reserved 240

3 General first order model was developed to predict the surface roughness over the experimental region (equation 2). Ra = SPEED FEED DOC.. (2) 3.2 Multiple objective optimization: To solve optimization problem using GA, fitness value is required. Fitness values, in fact, are the objective function values. In this work, multiple regressions modeling method has been employed to developed the mathematical model which establishes the relation between input and output. The developed mathematical model was converted into a MATLAB (R2009a) function. In the objective function f (1) and f (2) are MRR and Ra respectively. function f = simplemulti (x) f(1) = ( *x(1))+(7.43*x(2))+(2.02*x(3)); f(2) = 1/( ( *x(1))+(1.02*x(2))+(0.715*x(3)); This function was input to the GA Toolbox of MATLAB 2009a as the objective function. Upper and lower bounds were specified as per the levels of the machining parameters and the number of variables was set at 3. The objective function values are obtained for maximization of material removal rate and minimization of surface roughness. Here, an initial population size of 60 is taken and optimization is carried out by setting simple crossover and bitwise mutation with a crossover probability Pc = 0.8, migration interval of 20, migration fraction of 0.2 and Pareto fraction of Solver: ga-multiobjective optimization using Genetic Algorithm Fitness Number of variables : 03 Lower Bounds: [600, 0.1, 0.4] Upper Bounds: [1800, 0.2, 0.8] Iteration required: 153 According to the algorithm, ranking and sorting of solutions are done. The Pareto-optimal solutions (along with corresponding performance measure values) are reported in table 2. Table 2. Pareto optimal solutions INDEX MRR Ra Speed Feed DOC f1 f2 x1 x2 x All rights Reserved 241

4 Ra International Journal of Modern Trends in Engineering and Research (IJMTER) Fig. 2 shows the formation of Pareto-optimal front that consist of the final set of solutions. The shape of the Pareto optimal front is a consequence of the continuous nature of the optimization problem posed. The results reported in table 2 clearly show that in 15 pareto-optimal solutions, the whole given range of input parameters is reflected and no bias towards higher side or lower side of the parameters is seen Pareto front MRR Figure 2. Pareto optimal front 3.3 Confirmation experiments: From the Pareto-optimal solution, randomly run was chosen to verify the prediction of responses (MRR and Ra). It was found that validation experiments showed a good agreement with the predicted values of responses with an error less than 5% (table 3). Table 3 Validation experiment results based on multi-objective optimization Predicted Experimental % Error MRR (gram/sec) Ra (µm) IV. CONCLUSION In this study turning experiments were conducted by using the parametric approach of the Taguchi s method. Regression analysis has been performed to find out the relationship between input factors and responses using Minitab 16 statistical software. General first order model was developed to predict the material removal rate and surface roughness over the experimental region. Based on multi-objective optimization genetic algorithm the best material removal rate obtained was mm/gram and the best surface roughness value obtained was 0.27 μm. V. ACKNOWLEDGMENT I have taken efforts in this paper. However, it would not have been possible without the kind support and help of many individuals. I would like to extend my sincere thanks to all of them. I am highly indebted to Prof. D. S. Khedekar for their guidance and constant supervision as well as for providing necessary information regarding the paper. I would also like to express my deep gratitude to the Dr. M. S. Kadam for timely support and motivation for this All rights Reserved 242

5 REFERENCES [1] D.B. Miracle Metal matrix composites From science to technological significance Composites Science and Technology 65 (2005) [2] Quan Yanming, Zhou Zehua Tool wear and its mechanism for cutting SiC particle-reinforced aluminium matrix composites Journal of Materials Processing Technology 100 (2000) [3] Monaghan J, O Reilly P. Machinability of aluminum alloy:silicon carbide metal matrix composite. Process Adv Mater 1992; 2: [4] Tomac N, Tonnessen K. Machinability of particulate aluminium matrix composites. Ann CIRP 1992;41:55 8. [5] Suresh P.V.S., Rao P. V., and Deshmukh S.G., A Genetic Algorithmic Approach for Optimization of Surface [6] Roughness Prediction Model. International Journal of Machine Tool and Manufacture 42, [7] Saha A. and Mandal N.K., Optimization of machining parameters of turning operations based on multi performance criteria. International Journal of Industrial Engineering Computations 4, 51 All rights Reserved 243