An Investigation of Effect of Dressing Parameters for Minimum Surface Roughness using CNC Cylindrical Grinding Machine. Dadaso D.

Transcription

1 An Investigation of Effect of Dressing Parameters for Minimum Surface Roughness using CNC Cylindrical Grinding Machine Dadaso D. Mohite 1, PG Scholar, Pune University, NBN Sinhgad School of Engineering, Pune, India. S.M. Jadhav 2 Asst. Prof. / Dept. of Mechanical Engineering, NBN Sinhgad School of Engineering, Ambegaon (Bk.), Pune-041 ABSTRACT The ability of a grinding wheel to perform is significantly affected by the way in which the wheel is dressed. The four important parameters of blade type multi point diamond dressing operation are, dressing depth of cut, dressing cross feed rate, drag angle of dresser and number of passes. The effect of these parameters is measured in terms of surface of work piece during subsequent grinding operation. In this work, empirical models are developed for surface roughness by considering selected four parameters of dressing as control factors using Taguchi design methodology. The mathematical models in terms of dressing parameters were developed for surface roughness on the basis of experimental results. The model selected for the optimization has been validated by F-test and Chisquare test. In this paper, Taguchi design methodology has been applied to determine the optimum dressing parameters leading to minimum surface roughness using CNC cylindrical grinding machine on EN19 Steel bar. In the present work, experimental results were used to find the analysis of variance (ANOVA) which explains the significance of the parameters on the responses and predicted that dressing cross feed rate is most dominating parameters for surface roughness in cylindrical grinding. Keywords: CNC cylindrical grinding, Dressing, Taguchi Design, ANOVA, Modeling and Optimization. 1. Introduction The grinding wheel performance is significantly affected by the way is dressed. Basically, dressing is a grinding wheel conditioning process, in which suitable wheel topography is generated over the grinding wheel surface [Malkin,(1989), Pande and Lal, (1979)]. The four important parameters of blade type multipoint diamond dressing are dressing depth of cut, dressing cross feed rate, drag angle of the dresser and number of passes. The effects of these four parameters on the grinding wheel topography are measured in terms of the surface finish which is produced on the work piece during subsequent grinding operation. Experimental evaluation of volume of percentage of grain fracture & bond fracture of grinding wheel and to predict the surface finish of work piece and MRR is based on the same [Malkin,(1989), Chen and Rowe, (1996)]. In industries, the prime objective in grinding process is to get a better surface finish or to get high material removal rate (MRR) of the work piece. Better surface finish can be obtained by using fine grained grinding wheel whereas higher MRR can be obtained from the coarse grained grinding - editorijrim@gmail.com, 59

2 wheel. Fine grained topography is obtained by providing a lower dressing depth and dressing the wheel for a short amount of time while coarse grained topography is obtained by providing a greater dressing depth and dressing the same wheel for more time duration (Pande and Lal, (1979)). Pattinson and Chiholm (1972) studied the effect of dressing cross feed rate on radial wheel wear and number of active cutting edges. They have shown that the initial were rate of the wheel is greatly influenced by the dressing feed rate during diamond dressing, but the stable regime this wear rate is independent of dressing severity. Vickerstaff, (1976) has analyzed the effect of the wheel dressing condition on surface roughness of the work piece and on the distribution of metal removal rate over the wheel surface. A new model of dressing method was proposed which claims to have advantages of both fine dressing and coarse dressing and also proved that the radial wheel wear is significantly affected by the grinding condition and topography of the wheel surface as determined by the dressing operation. M. Janardhan et al. (2011) has developed model to predict and optimize the process parameters using Taguchi method, ANOVA and regression analysis. The main objective of this work is to minimize the surface roughness (Ra). The design of experiments is conducted on CNC cylindrical grinding machine. The empirical models are developed using design of experiments and surface roughness methodology. Regression analysis used to determine the relationship of the input variables which affect the output responses. In this study, four important parameter of grinding wheel dressing (i.e., dressing depth of cut, cross feed rate, drag angle of dresser and number of passes) were selected as variable parameters and other grinding process parameters were fixed. Taguchi design methodology has been applied to determine the optimum dressing parameters leading to minimum surface roughness and maximum MRR in CNC angular grinding machine on EN19 Steel bar. Also, empirical models are developed for surface roughness and MRR by considering selected four parameters of dressing as control factors using Taguchi design methodology. The second order mathematical models in terms of dressing parameters were developed for surface roughness and MRR on the basis of experimental results. In the present work, experimental results were used to find the analysis of variance (ANOVA) which explains the significance of the parameters on the responses. Conformation experiment was conducted at the optimum level to verify the effectiveness of the Taguchi approach. 2. Proposed Methodology In this present paper, effort is made to find the effect of dressing parameters on the grinding wheel topography and it is measured in terms of minimum surface roughness (R a) during the cylindrical grinding for EN19 steel bar. For this purpose, experiments at various levels of multi point diamond dressing parameters such as, dressing depth of cut, dressing cross feed rate, drag angle of dresser and number of passes, were conducted in order to explore the effect of the dressing conditions. An experimental design based in L9 orthogonal array is used to check the interactions between the factors. In the present work, experimental results were used to find the analysis of variance (ANOVA) which explains the significance of the parameters on the responses. To establish the suitable correlation between the input dressing parameters and the response surface roughness a regression model is develop. Finally, the predicted value is validated and compared with experimental result. - editorijrim@gmail.com, 60

3 3. Experimentation Details A set of Experiments were conducted on the CNC angular grinding machine (AHG 60X300 CNC) on EN19 steel bar to determine the effect of the dressing parameters on Surface roughness. The grinding wheel is used for the experiment work has Aluminum oxide (Al2O3) abrasives and vitrified bond. A blade type multi-point diamond dresser was used with a water soluble coolant. Fig. 1 shows the grinding operation on CNC angular grinding machine and Fig. 2 shows the dressing process. Figure 1: Experimental setup for CNC grinding process In this present work, three levels at four factors have been employed to predict the optimal values as shown in Table 1. Ranges of dressing parameters have been established based on review of literature and by performing the pilot experiments using one factor at a time (OFAT) approach. The number of experiments to be conducted can be reduced by using orthogonal array method of Taguchi optimization technique. Table 1: controllable factor and their levels Factors Parameters Unite Symbol Levels L1 L2 L3 A Dressing Depth of Cut μm B Dressing cross feed rate mm/min C Drag Angle of dresser B Number of passes Experimentation condition Table 4.4 lists the experimental condition to obtain the surface roughness. Grinding wheel of Aluminum oxide is used. Specimen is EN19. The grinding wheel speed was constant as 1000 rpm and the spindle speed was also constant as 100 rpm. A depth of cut for grinding was constant 100 µm. Coolant was continuously on during grinding operation. Grinding was plunge type. Multi pointed dresser was used for dressing. Dressing depth of cut was µm. 3.2 Work piece material In this work, EN19 steel bar of 40 mm diameter and 120 mm length has been selected as a work piece material. EN19 is a high quality carbon alloy steel which offers a good ductility and shock resistance. EN19 steel is widely used for manufacturing Axles, Drive Shafts, Crankshafts, Connecting Rods, High Tensile Bolts, Studs, Propeller Shaft Joints, Rifle Barrels, Breech mechanisms for Small - editorijrim@gmail.com, 61

4 Arms, Induction Hardened Pins. It is a good quality steel for wear resisting machine parts and widely used for power transmitting gears, pinions, spindles etc. The chemical composition for the work piece (EN19) is listed in Table 2. Table 2 Chemical composition of work piece (EN19) Constituent C Mn Si S P Cr % Composit ion The turning operation is done on the universal central lathe machine by holding the work piece between the two centers of the lathe. After the turning operation, heat treatment is done at C so as to increase the hardness of the material. The hardness of each job was maintained to 45HRC. After heat treatment, the specimens were ready for experiments. The close up view of the single job is shown in the figure Mathematical formulation Figure 2: Heat treated specimen The statistical tools such as Multiple Regression Analysis, RSM and Taguchi method are widely used for development of conventional predictive modeling. Taguchi is a collection of mathematical and statistical techniques for empirical model building. It is used for the problems in which output parameters are influenced by several input parameters and the objective is to optimize the output response. In this work, the first order and second order mathematical model were developed using multiple regression analysis in order to investigate the influence of dressing parameter on grinding wheel topography which is measured in terms of output response Surface roughness. The predicted surface roughness of dressing process can be expressed in terms of the investigation independent variable, Where Q is the grinding/dressing response, ф is the response function and and are dressing variables. Expressed in non-linear form, Eq. (1) becomes, (2) (1) - editorijrim@gmail.com, 62

5 In this work, following mathematical models are formulated: Surface Roughness model, (3) To determine the above constants and exponents these mathematical models are converted from non-linear to linear form by performing a logarithm transformation. The above function can be represented in linear mathematical form as follows: Equation (5) and (6) are re-expressed into generalized model for Surface roughness and MRR respectively as: (4) Where y is true (measured) response value on logarithmic scale, is dummy variable and its value is equal to 1, and,, and are the logarithmic transformation of input variable.,,, are the parameters to be estimated for Surface roughness model. The constants terms from above equations are estimated by the method of least square. The validity of the mathematical model will be tested using F-test, Chi-square test before going for the optimizations. The error between experimental values and predicted values from the mathematical model can be calculated by the method of least square. Equation (5) gives the least square error between observed values and computed values by model. 5. Result and Discussion The objective of this work is to investigate the effect of the dressing parameters on grinding wheel surface topography measured in terms of output response Surface roughness for EN19 steel. Total 9 experiments were conducted using Taguchi experimental design methodology. Experimental results for surface roughness are given in Table 3. In the present study all the designs, plots and analysis have been carried out using Minitab statistical software. 5.1 Surface Roughness Measurement Surface roughness values are obtained from MITUTOYO Surf test SJ-210 surface roughness tester for each experiment. The obtained values are used for the Taguchi optimization process. (6) - editorijrim@gmail.com, 63

6 Table 3: Experimental result of Surface Roughness Input Parameter Response Expt. No Surface Roughness (μm) a) Main Effects Plots The main effects plots for the experiments have been given in Fig. 5. Figure 5: Main Effects plot for Mean of Surface Roughness Fig.5 shows that the surface roughness increases with increase in the dressing depth of cut and dressing cross feed rate. Also, Surface roughness decreases with increase in drag angle of dresser and number of passes to some extent and then it increases with further increase in the drag angle and number of passes. It is observe that, The optimal set of dressing parameters obtained for surface roughness using Taguchi design of experiment methodology is dressing depth of cut: 20 μm, dressing cross feed rate: 80 mm/min, drag angle: 50 0 and number of passes : 4. b) Interaction Plots The Figure 6 shows the interaction plot between all parameters such as dressing depth of cut, dressing cross feed rate, drag angle of dresser and number of passes. It is observed that there is no strong relation between dressing parameters. - editorijrim@gmail.com, 64

7 Dressing DOC IJREAS VOLUME 6, ISSUE 6 (June, 2016) (ISSN ) c) Contour Plots Figure 6: Interaction plots of Dressing Parameters Contour plots are used to explore the relationship between three variables. Generally, there are two predictors and one response variables. Contour plots are useful for establishing desirable response values and operating conditions. Minitab plots the values for the x and y factors (predictors) on the x and y axes, while contour line and colored bands represents the values for the z-factors (response). In Fig.7 contour plot shows that the minimum surface roughness occurs below 22 and 85 mm/min dressing feed rate dressing DOC Contour Plot of PMEAN1 vs Dressing DOC, Dreesing Feed Rate PMEAN1 < > Dreesing Feed Rate Figure 7: Contour plot of Surface roughness Vs Dressing depth of cut and Dressing Cross feed rate 5.2 Analysis of Variance The experimental results were analysis of variance (ANOVA) by using MINITAB 17 statistical software. The ANOVA results for the response are shown in Table editorijrim@gmail.com, 65

8 Source Table 4: Analysis of Variance of Surface Roughness DF Sum of Square Variance F-Ratio P (%) Error ANOVA results determine the significance of factors. This significance of the factors is calculated by using F-ratio. It contains percent (%) P-value, defined as the significance rate of the process parameters on the surface roughness. From ANOVA Table 4, the dressing feed rate is most affecting factor on Surface Roughness. 5.3 Mathematical Modeling of response Surface Roughness (Ra) model: By the method of multiple regressions analysis equation (5.11) solved by using MATLAB software and found values of constants C1, a1, a2, a3 and a4. Hence, while developing the model for roughness, only the individual variables, and are considered and the non-linear fit between response and dressing variables is given by: ( ) (6.15) It is observed from the mathematical model of surface roughness that the roughness is increases with increase in dressing depth of cut, dressing cross feed rate and drag and angle of the dresser. But, roughness is decreases with increase in number of passes. Table 6.9.Comparison of Experimental Values and Predicted Values Expt. No. Surface Roughness (μm) Experimental value Predictive Value Error Least Square Error (% ) Table 6.9 gives the comparison between the experimental values and predicted values. The least square error is obtained by equation (5.13). - editorijrim@gmail.com, 66

9 ACKNOWLEDGMENT First of all I would like to express my deepest gratitude and sincere thanks to my guide, Prof. S. M. Jadhav, Assistant Professor, Mechanical Department, NBN Sinhgad School of Engineering, Ambegaon, Pune his valuable time and keep interest in my project research work. His intellectual advice has helping me in every step of my research work. I also thank thanks to our beloved Head of Department, Prof. (Dr.) S.Y. Gajjal who has contributed towards the completion of my project. I am indebted towards Prof. S. S. Patil, Prof. V.S. Jadhav, Prof. (Dr.) R. N. Patil, Prof. (Dr.) S.S. Kadam, and Prof. (Dr.) V. P. Shukla for given valuable suggestions and motivate this effort. CONCLUSION On the basis of this investigation, the following conclusion can be drawn. 1. It was found that, for Surface Roughness (Ra), out of four dressing parameters dressing cross feed rate is the most influencing factor for EN19 work material followed by dressing depth of cut which have a lower influence while drag angle of dresser and number of passes has relatively weak effect on surface roughness. 2. The interaction between dressing depth of cut and cross feed rate is dominating factors for surface roughness. The percentage contributions of both parameters on response are 75.52% and 15.27% respectively by analysis of variance. 3. The mathematical model is generated for Surface roughness by using multiple regression analysis which shows that the surface roughness increases with increase in dressing depth of cut, dressing cross feed rate and drag and angle of the dresser, and decreases with increase in number of passes. As dressing cross feed rate is the dominating factor, it should be low while dressing the grinding wheel. 4. The optimal set of dressing parameters obtained for surface roughness using Taguchi design of experiment methodology is dressing depth of cut: 20 μm, dressing cross feed rate: 80 mm/min, drag angle: 50 0 and number of passes : 4. REFERENCES [1] Malkin S. and Guo C. (2008), Grinding Technology: Theory and Applications of Machining with Abrasives, Second edition, Industrial Press, Inc. 989, New York, NY [2] W. Brian Rowe (2009), Principle of Modern Grinding Technology, William Andrew, Second Edition, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK. [3] Douglas C. Montgomery, (1991), Design and Analysis of Experiments, 7 th edition, Wiley, Singapur. [4] T. J. Vickerstaff (1970), The Influence of Wheel Dressing on the Surface Generated in the Grinding Process, International Journal of Machine Tools Design, Res. Vol. 16, pp [5] Pacitti, V. and Rubenstein, C. (1972), The influence of the dressing depth of cut on the performance of a single point diamond dressed alumina grinding wheel, International Journal of Machine Tool Design and Research, Vol. 12, pp [6] P. Ramchandran and S. Vaidyanathan (1976), Statistical Evaluation of Parameters in Grinding, Wear, Vol. 36, pp [7] S.J. Pande and G.K. Lal, (1979), Effect of Dressing on Grinding Wheel Performance, International Journal of Machine Tools Design, Res. Vol. 19, pp editorijrim@gmail.com, 67

10 [8] N.P. Fletcher (1978), Single Point Diamond Dressing of Aluminium Oxide Grinding Wheel and its Influence in Cylindrical Transverse Grinding, International Journal of Machine Tools Design, Res. Vol. 20, pp [9] Buttery, T.C., Statham, A., Percival, J. B. and Hamed, M. S. (1979), Some Effects of Dressing on Grinding Performance, Wear, Vol. 55, pp [10] Andrzej Koziasrski and Andrze Golabczak (1985), The Assessment of the Grinding Wheel Cutting Surface Condition after Dressing with Single Point Diamond Dresser, International Journal of Machine Tools Design, Res. Vol. 25, No.4, pp [11] Xun Chen, W. Brian Rowe (1996), Analysis and Simulation of Grinding Process. Part I: Generation of Grinding Wheel Surface, International Journal of Machine Tools & Manufacture, vol.36, No. 8, pp [12] B.Linke (2008), Dressing Process Model for Vitrified Bonded Grinding Wheel, CIRP Annals - Manufacturing Technology, Vol. 57, pp [13] Thiagarajan, R. Sivarama krishnan and S. Somasundaram (2011), Modeling and Optimization of Cylindrical grinding of Al/SiC composites Using Genetic Algorithms, ABCM Journal of Braz. Soc. Of Mechanical Sciences and Engineering, Vol. 34, No. 1, [14] M. Janardhan and Dr. A. Gopala Krishna (2011), Determination and optimization of cylindrical grinding process parameters using Taguchi method and regression analysis, International Journal of Engineering Science and Technology, Vol. 3, No. 7, pp **************** - editorijrim@gmail.com, 68