6340(Print), ISSN (Online) Volume 4, Issue 2, March - April (2013) IAEME AND TECHNOLOGY (IJMET)

Transcription

1 INTERNATIONAL International Journal of Mechanical JOURNAL Engineering OF MECHANICAL and Technology (IJMET), ENGINEERING ISSN 0976 AND TECHNOLOGY (IJMET) ISSN (Print) ISSN (Online) Volume 4, Issue 2, March - April (2013), pp IAEME: Journal Impact Factor (2013): (Calculated by GISI) IJMET I A E M E PROCESS PARAMETERS OPTIMIZATION IN SLS PROCESS USING DESIGN OF EXPERIMENTS Dr. T. Nancharaiah Professor and Head, Department of Mechanical Engineering DMSSVH College of Engineering, Machilipatnam A.P., INDIA. M. Nagabhushanam Associate Professor, Department of Mechanical Engineering DMSSVH College of Engineering, Machilipatnam A.P., INDIA. B. Amar Nagendram Associate Professor, Department of Mechanical Engineering DMSSVH College of Engineering, Machilipatnam A.P., INDIA. ABSTRACT Rapid prototyping (RP) is an additive manufacturing process which builds the parts directly from CAD data sources. RP Techniques are increasingly being used to manufacture complex precision parts for the automotive, aerospace and medical industries. One of the popular RP processes is the Selective Laser Sintering (SLS) process which manufactures parts by sintering metallic, polyoneric and ceramic powder under the effect of laser power. This paper presents the effect of slice thickness and part orientation on total area of sintering (TAS) and the laser energy. Using design of experiments (DOE) L 9 orthogonal array was selected and experiments were conducted. The experimental results were statistically analyzed using ANOVA analysis, S/N ratio to find the contribution of each parameter and to optimize the process parameters. The significance of each process parameter is further strengthened by the correlation analysis. Finally confirmation tests were conducted for optimum process parameters in SLS process. 162

2 Keywords: Rapid Prototyping, additive manufacturing, selective laser sintering, laser energy, ANOVA. 1. INTRODUCTION Rapid prototyping (RP) is the most common name given to host of relation technologies that are used to fabricate physical objects directly from three dimensions CAD model. These methods are unique in that they add and bond materials in layers to form objects. Such systems are also known by the general names free from fabrication (FFF), solid free from fabrication (SFF) and layered manufacturing. The materials used in rapid prototyping are numerous plastics, ceramics, metals ranging from stainless steel to titanium and wood like paper. There are different types of RP processes and are classified based on their layer generation methods. Among these, most commonly used processes are Stereo lithography (SLA), fused deposition modeling (FDM), Selective laser sintering (SLS), laminated object manufacturing (LOM) and 3 D printings (3DP). Among the different RP processes the SLS process has gained traction in the manufacturing industry due to its capability to produce complex parts of any geometry without the need for special tooling and support structures. SLS also able to manufacture parts from materials such as metal and nylon which are difficult to fabricate using traditional methods. RP processes offer several advantages but have limitations like low productivity (large build time), low part quality (dimensional accuracy) and low yield. A need thus exist to carry out research and development on RP process to enable it to produce functional parts of good quality with reduced production time and cost. The present work primarily focuses on determining optimum slice thickness and part orientation for minimal process energy consumption in selective layer sintering (SLS) process. Selective laser sintering (SLS) is an additive manufacturing technique that uses a high power laser (for example, a carbon dioxide laser) to fuse small particles of plastic, metal (direct metal laser sintering), ceramic, or glass powders into a mass that has a desired three-dimensional shape. The laser selectively fuses powdered material by scanning cross-sections generated from a 3-D digital description of the part on the surface of a powder bed. After each cross-section is scanned, the powder bed is lowered by one layer thickness, a new layer of material is applied on top, and the process is repeated until the part is completed. Figure (1) shows the basic process of SLS. Compared with other methods of additive manufacturing, SLS can produce parts from a relatively wide range of commercially available powder materials. These include polymers such as nylon (glass-filled, or with other fillers) or polystyrene, metals including steel, titanium, alloy mixtures, and composites and green sand. The physical process can be full melting, partial melting, or liquid-phase sintering. Depending on the material, up to 100% density can be achieved with material properties comparable to those from conventional manufacturing methods. In many cases large number of parts can be packed within the powder bed, allowing very high productivity. 163

3 Fig: (1) Selective Laser Sintering Process 2. LITERATURE REVIEW RP in general and SLS in particular have recently gained popularity as a main stream of manufacturing process for producing functional parts in bulk quantities. However, there has been very little reported research on understanding the relation between the part shape, process parameters and energy consumption in SLS process. Traditionally, researches have concentrated on the physical, chemical and mechanical changes involved in the creation of slices in different RP process. Phatak and pande (1) designed a modular system and implemented to find the optimum orientation of the CAD part quality model in RP process using generic algorithm technique for improvements in manufacture. Canellidis et al (2) proposed methodology using genetic algorithm, to get optimum part orientation using a multi criteria objective function comprising of the estimated build time, post processing time and the average surface roughness of the part. Nelson et al (3) analyzed the SLS process and developed a one-dimensional thermal model to predict the laser energy required for the complete sintering of bisphenol A polycarbonate powder. They studied the effect of laser scan speed, laser power, and powder size and powder bed temperature among various other parameters on the development of the sintered layers. Eho et al(4) used a genetic algorithm (GA) method to optimize the SLA process by analyzing the dimensional errors in SLA parts and correlating them to the laser power used for creating the slices. LW et al (5) investigated the environmental effects of three RP process: SLA, SLS and FDM and calculated the life cycle energy utilization in these processes. No systematic work has however, been directed towards findings as optimal process parameters for minimum laser energy consumption in SLS process. The work reported in this paper is an attempt in this direction. 3. METHODOLOGY This section presents a methodology to analyze the energy utilization in the SLS process by modeling the virtual manufactures of a part and correlating the energy to the slice thickness and part orientation. The overall methodology for calculations the laser energy for manufactures a part in SLS is shoes in fig (2). The part is first modeled in a CAD system and the CAD model is exported to the STL file format. The part orientation and slice thickness are selected and the STL file is sliced. For each slice, the sintering area is calculated and the 164

4 total area of sintering (TAS) is calculated by adding up the sintering areas for all the slices. The total laser energy is than calculated from the TAS and correlated to the orientation and slice thickness. STL file as the part Fix build orientation and slice thickness For each slice, calculate sintering area using connect hull approach Sum sintering areas for all slices to calculate total area of sintering (TAS) Calculate laser energy from TAS Repeat for different values of build orientation and slice thickness 4. DESIGN OF EXPERIMENTS Fig (2): Over all Methodology to calculate laser energy Design of experiments and analysis of results are engaging the attention of the Research Scholars and also practicing engineers. Many statistical tools are being used in the recent past. Present day competition in the industry is pushing for more and more emphasis on quality. Improved quality and enhancement in the market share can be achieved through preventive action rather than inspection and process control techniques. Design of experiments is one such quality improvement process which builds quality into products and process as that eliminates expensive controls and inspection. It is a valuable tool to optimize product and process design, to accelerate development cycle and to reduce development cost. This will also improve easy transition of products from R & D stage to manufacture. 4.1 Selection of Process Parameters in SLS Process When preparing to build SLS parts, many fabrication parameters are needed in the software. To achieve optimum quality, these parameters are set differently according to requirements of applications. Therefore, the first step in the experiment was to identify the process control parameters that are likely to affect the laser energy in SLS process. The two process parameters are selected at three different levels as shown in table (1). 165

5 Table (1) Selection of Process parameters and their levels Process parameters Level 1 Level 2 Level 3 Slice thickness in mm (A) Part orientation in degrees (B) Orthogonal Arrays (OA) An experiment in which all possible combination of factor levels are used is called full factorial experiments. If an experiment consists of n number of factors and each factor at X levels. Number of trials possible (treatment combinations) = Xn. As the number of factors considered at multi levels increases, it becomes increasingly difficult to conduct the experiment with all treatment combinations. In this situation, orthogonal arrays are at our rescue (which are highly fractionalized factorial layouts), becomes useful in reducing the number of trials. 4.3 Selection of Orthogonal Array The first step in selecting the correct standard OA involves counting the total degrees of freedom (dof) in the study. This count fixes the minimum number of experiments that must be run to study the factors involved. In counting the total dof, the investigator commits 1 dof to the overall mean of the response under study. This begins the dof count as 1. The number of dof associated with each factor under study equals one less than the number of treatment levels available for that factor. One determines the total dof in the study as follows: if n A and n B represent the number of treatments available for two factors A and B respectively, Then 1 = dof to be used by the overall mean n A 1 = dof for A n B 1 = dof for B An example will illustrate this procedure. If a design study involves three 3-level factors (A and B), then the dof would be as follows: Source of dof required dof Overall mean 1 A,B 2(3 1) = 4 Hence, Total dof = = 5 Therefore, in this example, one must conduct at least 5 experiments to be able to estimate the desired main effects. The corresponding OA must therefore have at least 5 rows. Therefore L9 orthogonal array is selected for experimentation and shown in table (2). 166

6 Table (2) L 9 orthogonal array Expt. No. Columns EXPERIMENTAL RESULTS A trial run was performed in which a series of samples were built on the SLS machine. Totally 9 samples were produced by SLS according to the L9 array. The dimensions of the sample specimen shown in figure (3) Fig (3): CAD model of the Part 5.1 Results The study involved 9 sample components produced by SLS machine. Experimental results for laser energy were shown in the table (3) and in figure (4). From graphs it was found that the slice thickness increases as laser energy required sintering the part decreases and part orientation effects moderately. 167

7 Table (3) Experimental results for Laser Energy Expt. No. TAS in mm 2 Laser Energy in Kilo Joules Laser energy Slice thickness Vs Laser energy Slice thickness in mm Las er energy Part orientation Vs Laser energy Part orientation Fig (4): Variation of Laser Energy with respect to Slice thickness and Part orientation 5.2 Analysis A. Signal to noise (S/N) ratio The signal to noise ratio measures the sensitivity of the quality characteristic being investigated to those uncontrollable external factors. To minimize the problem, the governing relationships for the S/N ratio in terms of the experimentally measured values of laser energy, i.e., S/N ratio = -10 log 10 MSD Where MSD = (yi - ) / y)2 /n, y the target value that is to be achieved, the number of samples. The S/N ratio values obtained for the trials are listed in Table (4). From the results optimum laser energy value obtained at level 3 of slice thickness and level 1 of part orientation. 168

8 Table (4) S/N ratio for optimization of Laser energy Factor Level 1 Level 2 Level 3 Max. Min. Slice thickness Part build orientation B. ANOVA analysis ANOVA analysis provides significance rating of the various factors analyzed in this study. Based on the above rating, factors, which influence the objective function significantly, could be identified and proper control measures adopted. In a similar way, those factors with minimum influence could be suitably modified to suit economic considerations. The ANOVA computations are carried out based on procedure out lined in ref (10) and listed table in (5). A variable possessing the maximum value of variance is said to have the most significant effect on the process under consideration. When the contribution of any factor is small, then the sum of squares, (SS) for that factor is combined with the error (SSe). This process of disregarding the contribution of a selected factor and subsequently adjusting the contributions of the other factors is known as pooling. Table (5) ANOVA analysis Factor Sum of squares (SS) Degree of freedom (dof) Variance Percentage of contribution F test f-table value Slice thickness Part orientation Error Total C. Correlation analysis In process control, the aim is to control the characteristics of the output of the process by controlling a process parameter. One succeeds if the parameters are chosen correctly. The choice is usually based on judgment and knowledge of the concerned technology. A correlation is assumed between a variable product characteristic and a variable process parameter. 169

9 In the present study, a relationship is assumed between the slice thickness (process parameter) and laser energy, part orientation and laser energy. Slice thickness is the property which significantly affects the quality of the prototypes in SLS process. This is proved by the contribution at 99% level of significance. The correlation coefficient (r) obtained from the results is for slice thickness. The range of values for (r) lies between 1 and -1. The experimental value indicates a reasonably strong negative (indirect) relation. Therefore, as slice thickness increases, the laser energy decreases. The correlation coefficient (r) obtained is for part orientation. The experimental value indicates moderate positive (direct) relation. Therefore, as part orientation increases, the laser energy value increases moderately. 6 CONFORMATION TESTS Once the optimal level of design parameters has been selected, the confirmation tests were conducted. The experimental results confirm the prior design and analyses for optimizing he process parameters. Table (6) Results of the confirmation experiments for laser energy Optimal process parameters Prediction Experimental Level A 3 B 1 A 3 B 1 Laser energy in KJ CONCLUSIONS This paper presents the effect of slice thickness and part orientation on laser energy of a part manufactured in the SLS process. Nine sample parts were virtually manufactured and their laser energies were calculated for different sets of slice thickness and part orientations and the results are presented. From the results, it can be concluded that the slice thickness in inversely proportional to the total laser energy building the part, and the effect of part orientation on laser ene4rgy is dependant upon the geometry of the part. In this study the design of experiments and S/N ratio provides a systematic and efficient methodology for the design optimization of the process parameters. From the ANOVA analysis it was found that the effect of slice thickness on laser energy is 97.79% and part orientation is 1.538%. This significance is further strengthened by correlation analysis. The confirmation experiments were conducted to verify the optimal process parameters. 7.1 Scope of future work In this paper, only the SLS laser energy has been analyzed while other energy components such as platform energy, energy for heating the bed etc., have been neglected. This work can be extended considering other energies. This work can also be extending to other process parameter incorporating in addition to these two process parameter for optimization. 170

10 8 REFERENCES 1. Amar M.Phatak & S.S.Pande, Optimum part orientation M Rapid phototyping using genetic algorithm, Journal of manufacturing systems 31 (2012) Canellidis V. Giannalisis J, Dedoussis V., Genetic algorithm based Multi Objective optimization of he build orientation in stereolithograthy, international journal of Advanced Manufacturing Technology 2009, 45: Nelson J C, Xue S.Barlow JW, Beaman J J, Marcus H L, Boureil DL. Model of he selective laser sintering of bisphenol a polycarbonate, industrial and Engineering Chemistry Research 1993; 32 (10) Cho H S, park WS, Choi B W, Leu Mc, Determining optimal parameters for stereolithography processes Via genetic algorithm, Journal of Manufacturing systems 2000; 19(1) Luo YC, Ji ZM, Leu MC, Candill R., Environmental performance analysis of solid free form fabrication processes. In proceedings of the 1999 EEE international symposium on electronics and the Environment, ISEE 1999, P Diane A. Schaub, Kou-Rey Chu, Douglas C. Montgomery, Optimizing Stereolithography Throughput, Journal of Manufacturing Systems, Vol. 16/No.4, (1997). 7. Diane A. Schaub, Douglas C. Montgomery, Using Experimental Design to Optimize The Stereolithography Process, Quality Engineering, 9(4), pp , (1997). 8. A. Armillotta, G.F. Biggioggero, M. Carnevale, M. Monno, Optimization of Rapid Prototypes with Surface Finish Constraints: A Study on The FDM Technique, Proceedings 3rd International Conference on Management of Innovative Technologies, Piran, Slovenija, 28-29, (October 1999). 9. D.Karalekas and D.Rapti, Investigation of the processing dependence of SL solidification residual stresses, Rapid prototyping journal, vol. 8, number 4, pp , (2002). 10. Douglas C. Montgomery, Design and Analysis of Experiments, 3rd edition John Wiley & Sons, (1991). 11. Tapan.P. Bagchi, Tagichi Methods Explained Practical Steps to Robust design Prentices Hall of India Pvt Ltd, New Delhi, (1993). 12. U. D. Gulhane, A. B. Dixit, P. V. Bane and G. S. Salvi, Optimization of Process Parameters for 316L Stainless Steel Using Taguchi Method And Anova International Journal of Mechanical Engineering & Technology (IJMET), Volume 3, Issue 2, 2012, pp , ISSN Print: ,, ISSN Online: Raju B S, Chandra Sekhar U and Drakshayani D N, Web Based E- Manufacturing of Prototypes By Using Rapid Prototyping Technology, International Journal of Mechanical Engineering & Technology (IJMET), Volume 4, Issue 2, 2013, pp , ISSN Print: ,, ISSN Online: