Modeling and Optimizing of CNC End Milling Operation Utilizing RSM Method

Transcription

1 I Vol-0, Issue-0, January 0 Modeling and Optimizing of CNC End Milling Operation Utilizing RSM Method Prof. Dr. M. M. Elkhabeery Department of Production Engineering & Mech. design University of Menoufia Shebin El-kom, Menoufia, Egypt mmkhabeery@yahoo.com Dr. M. H. Kazamel Department of Production Engineering & Mech. design University of Menoufia Shebin El-kom, Menoufia, Egypt hany_kazamel@yahoo.com M.M.Mansour Department of Production Engineering & Mech. design University of Menoufia Shebin El-kom, Menoufia, Egypt engmohamed9@yahoo.com Abstract - In the present investigation, a response surface methodology (RSM) method is proposed for modeling and optimizing some process parameters (cutting force, surface roughness and material removal rate) during end milling of 83- Alumimum alloy on a CNC machine tool. The considered machining parameters were spindle speed, feed rate, axial and radial depth of cut. Both cutting force and surface roughness were measured whereas material removal rate was calculated. The analysis of variance (ANOVA) was performed. The results show that the feed rate is the most significant factor affecting the process parameters. Also, RSM optimization procedure is efficient and sufficient to obtain the optimum machining parameters during end milling operation. Index terms- End milling, machining parameters, Response surface methodology, modelling, optimization. I. INTRODUCTION Most of complex shapes can be machined with very accurate dimensions by using end milling operations. Recently, computerized numerical control (CNC) machine tools have been implemented to achieve full automation in milling process since they provide high production rate, increase the machined parts quality and require less operators []. The optimization of process parameters will cause a significant improvement in the end milling process and a reduction in the manufacturing costs. The response surface methodology (RSM) is one of several optimization methods that used by many researchers []. Tammineni [3] has used RSM for modelling and optimizing of surface roughness and flatness with end milling of Aluminum alloy by employing cutting speed, feed rate and depth of cut as machining parameters. The significant factor affecting surface roughness was feed rate and depth of cut for flatness. A full factorial design of experiment has been presented [-7] to study the effects of machining parameters on the surface roughness with milling of aluminum and aluminum alloys. The deviation error between measured and predicted values was satisfactory. The RSM and Radian Basis Function Network (RBFN) were used to optimize surface roughness in end milling of aluminum alloys. The authors have found that feed rate is the most significant factor [, 8]. Kadirgama et al. [9] integrated RSM and ant colony optimization (ACO) for modeling and optimization surface roughness with end milling of AA606- T6. Again, the most significant factor affecting surface roughness quality was the feed rate. The multiple regression analysis and artificial neural network (ANN) study has been implemented by Rashid and Abdul Lani [] to predict surface roughness for a CNC milling process. ANN provided prediction accuracy better than multiple regression. The ANN model has been presented by Zain et al. [] to create surface roughness model during end milling operation. As a result, to obtain very good surface finish, a high cutting speed with a low feed rate and radial rake angle are recommended. Razfar et al. [] coupled ANN with harmony search algorithm (HS) to optimize the cutting parameters, leading to obtain minimum surface roughness in face milling of X0Cr3 stainless steel. Kechagias et al. [3] used Taguchi design of experiment (DOE) for modeling surface roughness with end milling of Al alloy 83. The results showed that the cutting speed, the peripheral nd relief angle, and the core diameter are the most significant factors. Patel [] optimized surface roughness with CNC end milling process of AL 63 T6. L orthogonal array (OA) in the Taguchi parameter design was selected. The optimum values of factors were tool feed 0 mm/min, tool speed 0 rpm, tool diameter mm, and depth of cut. mm. In the present study, a RSM model was developed based on experimental data to predict cutting force, surface roughness and material removal rate with end milling of 83-Aluminum alloy. The machining input parameters were spindle speed, feed rate, axial and radial depth of cut. Full quadratic mathematical models were developed using MINITAB-

2 I Vol-0, Issue-0, January 0 software. An optimization process was performed for roughing and finishing cut to obtain the optimum machining parameters. II. EXPERIMENTAL PROCEDURES A. Design of Experiments A Design of Experiment (DOE) is a structured, organized method for determining the relationship between factors affecting a process and the output of that process. Among several DOE techniques, a central composite response surface methodology (RSM) method was selected for modelling and optimizing of cutting forces, surface roughness and material removal rate in end milling operation. Machining parameters such as spindle speed (n), feed rate (f), axial (a p) and radial (a e) depth of cut were selected as input parameters for this study. The machining parameters levels are listed in Table. The experiments were performed on a three axis CNC milling machine (EXTRON M8) which has a maximum spindle speed of 8000 rpm and 7. kw spindle motor. TABLE I MACHINING PARAMETERS LEVELS TABLE III MECHANICAL PROPERTIES OF 83-ALUMINUM ALLOY Si Fe Cu Mn Mg Cr Zn Ti Al C. Cutting Force Measurement Balance A three-component cutting force dynamometer was used for measuring the feed (Fx), main cutting (Fy), and thrust (Fz) components of the resultant tool force. The CNC machine tool used along with the forces measuring setup is shown in Fig.. The feed, main cutting, and thrust forces were calculated as the average of the recorded values in the steady-state cutting region. Machining parameters Spindle speed (n), (rpm) Level 00 Level Level 3 Level Level 0 Feed rate (f), (mm/min) Axial depth of cut (a p), Radial depth of cut (a e), B. Work Material and Cutting Tools Aluminum alloy 83 workpieces of ( x x 0 mm) were used in this investigation. The chemical composition and mechanical properties of material used are given in Tables and 3. The cutting tool used in this investigation is high speed steel (HSS) flat end mill with four teeth, diameter 0 mm, helix angle º, overall length mm and flute length 38. mm. To eliminate tool wear effect, a new end mill was used for machining 3 workpieces only. TABLE II CHEMICAL COMPOSITION OF 83-ALUMINUM ALLOY Tesile strength 37 MPa Yield strength 8 MPa Shear strength 90 MPa Elasticity modulus 7 GPa Elongation % Fig. Forces measuring setup D. Surface Roughness Measurement A surface roughness tester (TR) with a pickup (TS0) was used to measure the arithmetical mean surface roughness (Ra) of each experiment. Surface roughness measurements were made in a direction parallel to the direction of the relative toolworkpiece motion (along feed) at four regions of the machined surface. The average value of Ra was recorded. The instrument resolution was 0.00 µm, and the cutoff length was 0.8 mm. III. RESULTS AND DISCUSSION After computing the average values of the three cutting force components, the resultant force () was calculated using the following equation: F r = F x + F y + F z () 3

3 I Vol-0, Issue-0, January 0 Also, the material removal rate () was calculated using the following equation: = f a p a e () The experimental and the calculated results are shown in Table. TABLE IV. EXPERIMENTAL AND CALCULATED RESULTS Run order n (rpm) 00 0 f (mm/min) 0 0 ap.6 ae 8 Ra (µm) (mm³/min) Fx Fy Fz A. Effects of Machining Parameters on Process Parameters Figure a-c show the main effects of machining parameters (spindle speed, feed rate, axial and radial depth of cut) on the resultant cutting force (), surface roughness (Ra) and material removal rate (), respectively. It has been observed from Fig. a that the cutting force decreases with an increase in spindle speed. This results is expected because as the spindle speed increases, the temperature at the shear zone rise very quickly and there is no chance for the generated heat to dissipate. The generated heat cause thermal softening of the machined material and decrease in shear strength which leads to decrease in cutting force []. It has also been observed that as the feed rate, axial and radial depth of cut increase, the cutting forces increases due to an increase in removed material from the workpiece and consequently increase in the energy required for this machining process. Figure b reveals that as the spindle speed increases the surface roughness decreases. When spindle speed increases the cutting edge removes more and more chips from the workpiece in one minute lead to significant improvement in the surface finish. The increasing of the feed rate leads to a deterioration in surface. The chip thickness increase with an increase in feed rate and the cutting edge cannot remove small chips and this cause more irregularities in the machined surface. The axial and radial depths of cut have little effect on surface roughness because the workpiece cutting resistance compared to the rigidity of the cutting tool used is very small. It can be seen from Fig. c that spindle speed has no effect on. Furthermore, the increases with increasing of feed rate, axial and radial depth of cut due to an increase in the volume of removed material from the workpiece.

4 I Vol-0, Issue-0, January 0 Cutting force 0 0 (a) Resultant force () Spindle speed (n), rpm Feed rate (f), mm/min Axial depth of cut (ap), mm Radial depth of cut (ae), mm appropriate for predicting and investigating machining parameters effect. (a) Percent Normal Probability Plot (response is resultant force, ) (b) Surface roughness (Ra) Residual Spindle speed (n), rpm Feed rate (f), mm/min (b) 99 Normal Probability Plot (response is surface roughness, Ra) Surface roughness Axial depth of cut (ap), mm 0 0 Radial depth of cut (ae), mm Percent Residual (c) Material removal rate () Spindle speed (n), rpm Feed rate (f), mm/min (c) 99 Normal Probability Plot (response is material removal rate, ) 0 9 Material removal rate Axial depth of cut (ap), mm 0 0 Radial depth of cut (ae), mm Percent Residual Fig. 3 Normal probability plots for process parameters. Fig. Main effects plot (data means) for process parameters Figures 3a-c show the normal probability plots for the resultant cutting force, surface roughness and material removal rate. The Figs. reveal that the residuals locate reasonably close to a straight line, and departure points do not exist. Therefore, the mathematical models are B. Mathematical Models Mathematical modelling involves examining data related to the problem with a view of formulating or constructing a mathematical relationship between the variables and the response in the problem using the available data. Minitab- software was used to carry out the regression analysis to find

5 I Vol-0, Issue-0, January 0 the partial regression coefficients. Among several models tried (linear, linear + interaction, linear + squares, full quadratic), the full quadratic model is found to be the bestfit model. The mathematical models in uncoded levels are as follow: Mathematical model for resultant force F r = n 0.08 f a p + 87 a e + n 6.9 f +. a p a e.07 n f 0.0 n a p n a e f a p f a e +.0 a p a e The data correlated excellently with a correlation coefficient of 99.8 % Mathematical model for surface roughness R a =.96 7.E0 n f 0.86 a p a e n f 0.09 a p a e n f n a p n a e f a p f a e a p a e The data correlated excellently with a correlation coefficient of 97. % Mathematical model for material removal rate = 39.f a p 80 a e + f a p + f a e + a p a e The data correlated excellently with a correlation coefficient of %. C. ANOVA for process parameters ANOVA analysis was carried out for a level of significance (α) of %, i.e., for a level of confidence of 9%. The significant machining parameters can be determined by comparing the P-value from the ANOVA table with the level of significance (α= 0.0). The significant parameter has a P- value less than (α) and the insignificant parameter has a P- value larger than (α). the other significant factors are spindle speed (3.76 %), interaction of spindle speed and feed rate (.9 %), axial depth of cut (3. %), radial depth of cut (.7 %), and second order effect of spindle speed (.68 %),. The analysis of the material removal rate shows that the most significant factor is the feed rate (60.70 %). The factors which have less significance on the response are axial depth of cut (.09 %) and radial depth of cut (. %). IV. PROCESS OPTIMIZATION After developing models to predict the process parameters (, Ra, and ) the next logical step is optimizing them with respect to machining parameters. Selection of optimum machining parameters has always been a challenge to the manufacturing world. Optimization process of the cutting forces and surface roughness was carried out by Minitab software using (RSM). In the present investigation, two proposed process optimization were carried out for both roughing and finishing cut. Roughing cut case In the optimization of roughing cut, maximum material removal rate must be achieved more than minimum surface roughness. The proposed optimization problem is Objective: max.. Constrains: target Ra = µm, target =0 N. Finishing cut case The main goal of finishing cut process is obtaining a very smooth machined surface. The proposed optimization problem is Objective: min. Ra. Constrains: target = 000 mm³/min, target = N. Figures and show the optimum machining parameters for roughing and finishing cases, respectively. The Figs. show that the confidence coefficient is 8 % for roughing cut case and 97. % for finishing cut case. The optimum machining parameters and validation experiments results for these two cases are shown in Table. om the analysis of the resultant cutting force (), it can be seen that the most significant factors of the () model are the feed rate (8.9 %) and the spindle speed (8 %). The other factors which have significant effect on the response () are axial depth of cut (9. %) and radial depth of cut (3. %). The ANOVA analysis for surface roughness (Ra) indicates that the feed rate is the most significant factor (. %) and

6 I Vol-0, Issue-0, January 0 Optimal D High Cur 0.88 Low Composite Desirability 0.88 Targ: 0.0 y =.69 d = Ra along Targ:.0 y =.9909 d = Maximum y = 3E+0 d = Optimal D High Cur Low Composite Desirability Targ:.0 y = d = Ra along Minimum y = 0.79 d = Targ: y =.E+0 d = n f ap ae 0.0 [88.89] 0.0 [0.0].0000 [3.908] 8.0 [.] Fig. Optimum machining parameters for roughing cut case. n f ap ae 0.0 [0.0] 0.0 [9.960].0000 [.60] 8.0 [8.0] Fig. Optimum machining parameters for finishing cut case. TABLE V OBTAINED OPTIMIZATION VALUES AND VALIDATION EXPERIMENTS RESULTS FOR ROUGHING AND FINISHING CASES Optimization process Roughing Finishing cut case cut case n (rpm) 89 0 f (mm/min) 0 60 ap ae. 8 Ra (µm) (mm³/min) Predicted value Exp. value Error (%) Predicted value Exp. value Error (%) Predicted value Exp. value 8 38 Error (%) V. CONCLUSION The purpose of this research is to determine the optimum machining parameters for the rough and finish cutting process. RSM method was used to model the cutting force and surface roughness during end milling of 83- Aluminum alloy. The ANOVA analysis was conducted to indicate the effect of the four machining parameters, namely; spindle speed, feed rate, axial and radial depth of cut on the cutting force and surface roughness. In light of these results, the following conclusions may be summarized: om the analysis of regression, the data correlated excellently with a correlation coefficient of 99.8, 97. and % for the cutting force, surface roughness and, respectively. om the ANOVA analysis, the most significant factors of the F r model are the feed rate (8.9 %) and the spindle speed (8 %). The feed rate is the most significant factor of surface roughness and models with., %, respectively. The RSM optimization procedure is efficient and sufficient to obtain the optimum machining parameters during rough and finish cutting process. REFERENCES [] Routara, B., A. Bandyopadhyay, and P. Sahoo, Roughness modeling and optimization in CNC end milling using response surface method: effect of workpiece material variation. The International Journal of Advanced Manufacturing Technology, 9. 0(-): p [] Del Prete, A., A. De Vitis, and A. Spagnolo, Experimental development of RSM techniques for surface quality prediction in metal cutting applications. International Journal of Material Forming, 0. 3(): p [3] Tammineni, L. and H.P.R. Yedula, Investigation of influence of milling parameters on surface roughness and flatness. International Journal of Advances in Engineering & Technology, 0. 6(6): p. -6. [] Hayajneh, M.T., M.S. Tahat, and J. Bluhm, A study of the effects of machining parameters on the surface roughness in the end-milling process. Jordan Journal of Mechanical and Industrial Engineering, 7. (): p. -. [] Rawangwong, S., et al., An Investigation of Optimum Cutting Conditions in Face Milling Aluminum Semi Solid 0 Using Carbide Tool. Energy Procedia, 03. 3: p [6] Rawangwong, S., et al., An Investigation of Optimum Cutting Conditions in Face Milling Semi-Solid AA 707 Using Carbide Tool. International Journal of Innovation, Management and Technology, 0. 3(6): p [7] Rawangwong, S., et al., An investigation of optimum cutting conditions in face milling aluminum 707-t6 using design of experiment. Lecture Notes in Management Science, 0. : p. -3. [8] Kadirgama, K., et al., Optimization of surface roughness in end milling on mould aluminium alloys (AA606-T6) using response surface method and radian basis function network. 7

7 I Vol-0, Issue-0, January 0 Jourdan Journal of Mechanical and Industrial Engineering, 8. (): p [9] Kadirgama, K., M. Noor, and A.N.A. Alla, Response ant colony optimization of end milling surface roughness. Sensors, 0. (3): p [] MFF, A.R., M. Rizal, and A. Lani, Surface roughness prediction for CNC milling process using artificial neural network. Proceedings of the World Congress on Engineering, 0. [] Zain, A.M., H. Haron, and S. Sharif, Prediction of surface roughness in the end milling machining using Artificial Neural Network. Expert Systems with Applications, 0. 37(): p [] Razfar, M.R., R.F. Zinati, and M. Haghshenas, Optimum surface roughness prediction in face milling by using neural network and harmony search algorithm. The International Journal of Advanced Manufacturing Technology, 0. : p [3] Kechagias, J. et al. Parameter optimization during finish end milling of Al alloy 83 using robust design. Proceedings of the World Congress on Engineering. 0. [] Patel, K., Experimental Analysis on Surface Roughness of CNC End Milling Process Using Taguchi Design Method. International Journal of Engineering Science & Technology, 0. (): pp. -. [] Bäker, M., Finite element simulation of high-speed cutting forces. Journal of Materials Processing Technology, 6. 76(): p