Finite Element Analysis of Drilling of Titanium Alloy

Transcription

1 Available online at Procedia Engineering 10 (2011) ICM11 Finite Element Analysis of Drilling of Titanium Alloy Ozden Isbilir a, Elaheh Ghassemieh a a Department of Mechanical Engineering, University of Sheffield, Sir Fredrick Mappin Building, Mappin Street, Sheffield S1 3JD, UK Abstract Drilling is a highly demanding machining process due to complex tool geometry and the progressive material failure on the work piece. In this study, a 3D model is developed using commercial finite element software ABAQUS/Explicit. The proposed model simulates the drilling process by taking into account of the damage initiation and evolution of the work piece material, a contact model at the interface between drill bit and work piece and the process parameters. The results of the simulations demonstrate the effects of machining parameters on drilling. The results also confirm the capability and advantage of FE simulation of the drilling process Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Selection and peer-review under responsibility of ICM11 Keywords: Finite element analysis; titanium; drilling; force; torque; burr 1. Introduction Drilling is a one of the most required machining process and has considerable economical importance in the industry. Drilling process accounts for 40 60% of the total material removal processes and it is an essential technique in aerospace industries [1]. Drilling of titanium and its alloys has been utilized in different industries nevertheless the number of published research is very limited in the literature. A series of experiments in drilling of Ti6Al4V have been conducted by Sakurai et al. [2 4]. Cantero et al. [5] focused on the dry drilling tool wear and work piece subsurface damage. Sun and Guo [6] investigated the machining of Ti6Al4V in adiabatic conditions. Guo and Dornfeld [7-8] are the pioneers in FE analysis of 3D drilling. The authors investigated the burr formation. It is desired to have a predictive capability in drilling to enable optimization of the process parameters and drill bit geometry while taking into account work piece quality and tool wear. The technical difficulty * Corresponding author. Tel.: ; fax: address: E.Ghassemieh@sheffield.ac.uk Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Selection and peer-review under responsibility of ICM11 doi: /j.proeng

2 1878 Ozden Isbilir and Elaheh Ghassemieh / Procedia Engineering 10 (2011) of 3D modeling is on the complicated geometry of the drill, material deformation and the chip formation. In this paper, drilling of Ti6Al4V has been investigated using a 3D Lagrangian finite element model. The interaction between drill and work piece is integrated by explicit dynamic analysis. Simulations were performed with a general FE software ABAQUS. For validation purpose experimental tests were carried out and compared to the FE results. Nomenclature f feed rate A the initial yield strength of the material B the hardening coefficient ε the equivalent plastic strain rate ε 0 the reference strain rate T the current temperature of the material T melt the melting temperature of the material T room the room temperature n coefficient of strain hardening m coefficient of thermal softening C coefficient of strain rate d1-d5 the failure parameters of Johnson-Cook damage model p the pressure stress q the Mises stress w D Johnson Cook Damage criterion ε D pl the equivalent plastic strain at the onset of damage 2. Finite Element Model A 3D model is developed using commercial finite element software ABAQUS. The model aims to simulate the drilling process, to calculate the damage initiation and evolution in the work piece material; to predict induced cutting forces, torque, stress distribution in the work piece throughout the drilling process; and to predict the burr height at the entrance and at the exit sides while taking into account of the cutting parameters. The FE model is based on Lagrangian formulation with explicit integration method. Each drilling experiment was carried out with the use of coolant. It is assumed that the drilling induced heat is removed by coolant, thus thermal issues are not accounted in the model. While mass and inertia effects are included in the model. The overall dynamics are not taken into account into consideration in the analysis. However the tool is assumed fully elastic material, it is modeled as a rigid body. The contact and the friction parameters between the tool and work piece are influenced by a number of factor such as cutting speed, feed rate, geometry and the surface properties. In this study, the chip is not modeled due to computational cost therefore the friction between the chip and drill is ignored. The Coulomb friction model is used and a constant friction coefficient of 0.5 is used in the analysis. Interaction between the work piece and the tool is modeled by using surface-surface kinematic contact available in ABAQUS/Explicit. The overall FE model is shown in Figure 1.

3 Ozden Isbilir and Elaheh Ghassemieh / Procedia Engineering 10 (2011) Fig 1 Finite Element Model 2.1. Material and Constitutive Models In order to perform FE simulation of drilling process, an accurate and reliable flow stress model is highly necessary to represent work material constitutive behavior under large deformation due to cutting conditions. In this study, the Johnson-Cook constitutive material model [9] and related damage model are used for titanium alloy. The flow stress is calculated according to Eq. (1). ( [ ] n ε T T σ = A+ B ε )(1+ Cln[ ])(1 [ room ] m ) Tmelt Troom ε0 (1) When the equivalent plastic strain has reached to the criteria value, the damage initiated. The equivalent plastic strain at the onset of damage is calculated from the Eq. (2). pl pl p T T = room D d d d 1 d 4 ln ε ε 1 2exp 3 1 d 5 q T T melt room ε 0 (2) The Johnson-Cook constitutive material model and damage model parameters of Ti6Al4V are given in the Table 1. Table 1. Johnson-Cook constitutive material model and damage model parameters of Ti6Al4V [10-11] A (MPa) B(MPa) C n m d1 d2 d3 D4 d Experiments The drilling trials were conducted using TiAlN coated carbide twist drills with 8 mm diameter, 140 point angle geometry on a Mori Seiki SV-500 CNC milling machine. Tests were carried out at 1400 rpm spindle speed and 95, 119, 142, 171 mm/min feed rates with the use cutting fluid. The thrust force and

4 1880 Ozden Isbilir and Elaheh Ghassemieh / Procedia Engineering 10 (2011) torque were measured using a dynamometer ( Kistler 9255B ), burr width was measured by an optical microscope and burr height was measured with the use of surface profilometer ( Mitutoyo SV-602 ). 4. Results and Discussion The figure 2 shows the thrust forces collected from experimental study and finite element model which were carried out at 95 mm/min feed rate. As shown in the figure, the FE model provides a very good estimation of thrust force. The thrust force is underestimated by 1 %. As it can be from the figure, drilling can be divided into 3 stages: entrance, stable and exit. Fig. 2 Validation of the FE Model ( f=95 mm/min ) 4.1. Thrust Force and Torque The figure 3(a) shows that FE model provides a good estimation for thrust force. The discrepancy between experiments and the model is less than 10 %. In experiments, a large amount of coolant was used and the effect of heat was minimized. Due to this, the model is based on mechanical loading and does not enable thermal softening. Another possible reason for the difference is friction model. A constant coefficient of friction was used in the FE model. The lack of accurate friction modeling is a limitation in FE analysis. Inaccuracy of the material can also affect the results of the FE analysis. (a) (b) Fig. 3 (a) Comparison of FE Model and Experimental Thrust Force (b) Comparison of FE Model and Experimental Torque

5 Ozden Isbilir and Elaheh Ghassemieh / Procedia Engineering 10 (2011) The FE model provides a reasonable estimation for torque as shown in the figure 3 (b). The FE model overestimated the torque about 20%. The mechanistic model may be one of the reason since there is not softening in the material. As mentioned in discussion about thrust force, the friction model and inaccurate material can also affect the torque. The lack of chip formation may also have an influence on the torque Burr The figure 4 displays the burr height. The FE model provides a moderate estimation for burr height. The model underestimated the burr height 50% and 75% at the entrance and exit sides, respectively. The lack of thermal model is thought to be predominant effect Burr Height (um) Feed Rate (mm/min) Burr Height at Exit Surface (Experiment) Burr Height at Entrance Surface (Experiment) Burr Height at Exit Surface (FEA) Burr Height at Entrance Surface (FEA) Fig. 4 Comparison of FE Model and Experimental Burr Height 4.3. Workpiece Stress The distribution of Von Mises stress of the work piece is shown in the figure 5. The maximum Von Mises stress is estimated around 1.1 GPa. It shows that Mises stress increased gradually in the entrance stage, then the maximum Mises Stress was obtained in steady state, later it decreased gradually until the hole was drilled through. Fig. 5 Stress Distribution of FE Model ( f=95 mm/min )

6 1882 Ozden Isbilir and Elaheh Ghassemieh / Procedia Engineering 10 (2011) Effect of Feed Rate Fig 3 (a), (b) and Fig 4 show the effect of feed rate on thrust force, torque and burr height, respectively. An increase in the feed rate leads an increase in the thrust force, torque and burr height. The peak thrust force and torque were obtained at 171 mm/min feed rate in both experiments and FE model. Since there is not a thermal softening effect, the strain hardening effect will be dominant in the stress analysis. The burr height found to increase linearly with the increasing feed rate in FE model however this was not obtained in experiments. 5. Conclusion A 3D finite element model which includes complex tool geometry, constitutive models appropriate for high strain rate, and process parameters is presented. Drilling tests are performed to present the efficiency of the FE model of drilling of Ti6Al4V. The thrust force, torque are collected, burr height is measured and compared to the FE results for validation. The model provides good estimation of thrust force whereas torque is overestimated by 20% and burr height is underestimated between 50 and 75%. The study shows that the FE model of drilling is able to predict changes in cutting force, torque and stresses with respect to drilling process parameters. Acknowledgement Authors greatly acknowledge the provision of materials by Airbus and the access to the drilling facilities of the Advanced Machining Research Centre (AMRC) at University of Sheffield. The funding of the PhD scholarship of Mr. Ozden Isbilir supplied by the Ministry of National Education of the Republic of Turkey is highly appreciated. Reference [1] Schroeder P.T. Widening interest in twist drill. Modern Mach. Shop 1998; 71 (4): p [2] Sakurai, K., Adachi, K., Ogawa, K., and Niba, R. Drilling of Ti-6Al-4V Alloy. J. Jpn. Inst. Light Met., 1992; 42 (7): p [3] Sakurai, K., Adachi, K., and Ogawa, K,. Low Frequency Vibratory Drilling of Ti-6Al-4V alloy, J. Jpn. Inst. Light Met. 1992; 42 (11): p [4] Sakurai, K., Adachi, K., Kamekawa, T., Ogawa, K., and Hanasaki, S.. Intermittently Decelerated Feed Drilling of Ti-6%Al-4%V Alloy. J. Jpn. Inst. Light Met ; 46 (3): p [5] Cantero, J., Tardío, M., Canteli, J., Marcos, M., and Miguélez, M.. Dry drilling of alloy Ti-6Al-4V. Int. J. Mach. Tools Manuf., 2005; 45 (11): p [6] Sun J. and Guo Y. B. A New Multi-View Approach to Characterize 3D Chip Morphology and Properties in End Milling Titanium Ti-6Al-4V. Int. J. Mach. Tools Manuf., 2008; 48: p [7] Guo, Y.B., Dornfeld, D.A. Finite Element analysis if drilling burr minimization with a back up material. Proceedings of the XXVI NAMRC Conference, 1998; p [8] Guo, Y.B., Dornfeld, D.A. Finite Element Modelling of Drilling Burr Formation Process. Journal Manufacturing Science and Enginnering. 2000; 122: [9] Johnson, R., Cook, W.K. A constitutive model and data for metals subjected to large strains high strain rates and high temperatures. 7th International Symposium on Balistics. 1983; pp [10] Lesuer, D.R. Experimental investigations of material models for Ti 6Al 4V Titanium and 2024-T3 Aluminium. U.S. Department of Transportation Federal Aviation Administration, 2000; DOT/FAA/AR-00/25 [11] Meyer, H.W., Kleponis, D.S. Modeling the high strain rate behavior of titanium undergoing ballistic impact and penetration. International Journal of Impact Engineering. 2001; 26: p