- Email: [email protected]
Temperature assessment when milling AISI D2 cold work die steel using tool-chip thermocouple, implanted thermocouple and finite element simulation
Accepted Manuscript TEMPERATURE ASSESSMENT WHEN MILLING AISI D2 COLD WORK DIE STEEL USING TOOL-CHIP THERMOCOUPLE, IMPLANTED THERMOCOUPLE AND FINITE ELEMENT SIMULATION Hugo V. Lima, Augusto F.V. Campidelli, Antônio A.T. Maia, Alexandre M. Abrão PII: DOI: Reference:
S1359-4311(18)32107-0 https://doi.org/10.1016/j.applthermaleng.2018.07.107 ATE 12465
To appear in:
Applied Thermal Engineering
Received Date: Revised Date: Accepted Date:
4 April 2018 10 July 2018 20 July 2018
Please cite this article as: H.V. Lima, A.F.V. Campidelli, A.A.T. Maia, A.M. Abrão, TEMPERATURE ASSESSMENT WHEN MILLING AISI D2 COLD WORK DIE STEEL USING TOOL-CHIP THERMOCOUPLE, IMPLANTED THERMOCOUPLE AND FINITE ELEMENT SIMULATION, Applied Thermal Engineering (2018), doi: https://doi.org/10.1016/j.applthermaleng.2018.07.107
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
TEMPERATURE ASSESSMENT WHEN MILLING AISI D2 COLD WORK DIE STEEL USING TOOL-CHIP THERMOCOUPLE, IMPLANTED THERMOCOUPLE AND FINITE ELEMENT SIMULATION
Hugo V. Lima, Augusto F.V. Campidelli, Antônio A.T. Maia, Alexandre M. Abrão*
Department of Mechanical Engineering, Universidade Federal de Minas Gerais, Av. Antônio Carlos, 6627, Pampulha, Belo Horizonte, MG, CEP 31270-901, Brazil. Phone: 55 31 3409 5138. E-mail: [email protected]
* Corresponding author
Abstract: Due to the cyclic mechanical and thermal loads imposed to the cutting tool during milling, the study of the process temperature is of utmost importance for the better understanding of various associated phenomena, such as tool life and wear mechanisms, cutting forces behaviour and workpiece subsurface metallurgical alterations. Nevertheless, temperature measurement during milling operations imposes a number of restraints to experimental methods, mostly related to the cutter rotational speed, variable chip thickness and intermittent action of the cutting edges. The principal goal of this work is to perform a comparative study of the cutting temperature during in end milling using implanted and tool-chip thermocouple methods under distinct operating parameters. Additionally, finite element simulation is employed to correlate the results provided by the experimental techniques. Tool-chip thermocouple and implanted thermocouple experimental methods were used in addition to three dimensional finite element simulation. The findings indicated that the developed system is capable to cope with the drawbacks associated with intermittent machining operations and to provide reliable temperature values for both experimental methods. Milling temperature increased with cutting speed, feed per tooth and both axial and radial depths of cut, however, the relevance of each factor varied in accordance with the measurement method. The average cutting temperature was not statistically affected by cutting direction and the experimental determination of the friction coefficient between tool and
workpiece was critical to the accurate temperature determination using three dimensional numerical simulation.
Keywords: milling; temperature; tool-chip thermocouple; implanted thermocouple; finite element simulation.
1. INTRODUCTION
The damage caused by the energy converted into heat during metal cutting operations is a critical factor restricting production rate owing to the fact that it drastically affects both the cutting tool performance and workpiece surface integrity. Therefore, cutting temperature measurement and control deserve constant attention, especially in the case of metals and alloys considered difficult to machine [1]. Nevertheless, investigations on cutting temperature face a series of restrictions caused by the nature of the contact between tool and workpiece, chip formation mechanism and occasional presence of cutting fluid [2]. According to Cui et al. [3], the assessment on the intensity and form as energy is converted into heat is more complex in milling due to its intermittent nature with a number of cutting edges being sequentially subjected to cyclic heating and cooling at high rates. Consequently, thermal stresses are imposed in the cutting tool, thus leading to the emergence of cracks that will impair the performance of the cutting tool [4]. Furthermore, Courbon et al. [5] state that while perfect thermal contact can be assumed to take place at the tool-chip interface where conditions of seizure are dominant, conditions of sliding at the tool-chip interface lead to a thermal contact resistance that drastically affects heat flux and temperature distribution in the cutting tool. Silva and Wallbank [6] state that the practical limitations associated with experimental methods used to determine temperature distribution in the tool-workpiece interface result in obtaining the average tool-workpiece temperature, nonetheless, such information is crucial to the development of advanced cutting tool materials. Despite the difficulties associated with long setup times and need of specific equipment, experimental techniques are considered quite reliable for the determination of cutting temperature and are often used to assess the performance of analytical and numerical
methods [7], which experience a number of difficulties to model either oblique or intermittent cutting. According to Maurel-Pantel et al. [8], numerical models are capable of describing only the essential aspects of machining operations based on orthogonal cutting, while complex operations, such as milling, are not sufficiently explored and required greater advances in numerical modelling. Jiang et al. [9] employed numerical and experimental methods to evaluate the influence of cutting parameters on temperature in orthogonal cutting of AISI D2 tool steel using polycystalline cubic boron nitride (PcBN) tools and noticed that cutting speed and the depth of cut are the most relevant factors, with the maximum temperature recorded near the tool tip ranging from 711º C to 792º C. Barrios et al. [10] investigated milling temperature of AISI H13 hot work die steel using numerical modelling and type K thermocouples implanted 3 mm beneath the surface. The results indicated that after 30 seconds the maximum workpiece temperature was 40º C for a cutting speed of 100 m/min and 35º C for 50 m/min the cutting temperature. According to Samy and Kumaram [11], the main drawback related to the implanted thermocouple method relies on the fact that the maximum temperature in the cutting region cannot always be determined. The fact that the tool rotates at high speeds during milling has prevented, for a long time, cutting tool temperature measurement using implanted thermocouples, however, a number of alternatives has been presented in recent years. Among them, Kerrigan et al. [12] developed a wireless temperature measurement system integrated into a cutting tool holder with a thermocouple embedded in the cutting tool. The results showed that the temperature stabilization at the thermocouple location requires a certain amount of time, which depends on the characteristics of the tool coating and substrate. A similar approach was employed by Guha et al. [13], who used a micro thin-film sensor embedded in the cutting tool. A Bluetooth module was used to acquire data and establish communication with the computer. The signal conditioning circuit was designed and developed to increase the signal-to-noise ratio of the embedded micro thin film sensors. Moreover, an averaging filter algorithm was implemented in the data acquisition software interface. The findings showed that the developed system possess a resolution of 1° C and was effective in minimizing noise when measuring low voltage signals.
The implanted thermocouple method was used by Le Coz et al. [14] to measure temperature when milling AA7075 aluminium alloy and drilling Ti6Al4V titanium alloy and the results indicated that the device is sensitive to changes in the cutting parameters (cutting speed, feed rate and tool geometry and coating) and work material (aluminium alloys, steel and cast iron). However, this technique presents disadvantages associated with the weakening of the cutting wedge and heat flow alteration caused by the hole produced. Furthermore, Sutter et al. [15] points out the importance of meticulous calibration of the pair and proper insulation of the cutter and workpiece when the tool-chip thermocouple method is used. Sun et al. [16] employed the tool-chip thermocouple technique to measure temperature when milling Ti6Al4V titanium alloy. The results indicated that milling temperature increased with the elevation of both cutting speed and feed rate, however, the former parameter was found to be the most relevant. As far as the influence of depth of cut is concerned, Kikuchi [17] reported that its elevation leads to higher milling temperatures, irrespectively of the work material being tested. Brandão et al. [18] compared the temperature values provided by experimental (implanted thermocouple) and numerical (finite element) approaches when milling hardened AISI H13 and AISI D2 tool steels using coated and uncoated polycrystalline cubic boron nitride tipped end mills. They found that although the finite element model was capable to accurately predict the energy transferred from the cutting zone to the workpiece, it was ineffective in estimating the temperature decrease when distinct cooling environments (compressed or cold air) were used in comparison with dry cutting. Liu et al. [19] employed finite element modelling (FEM) to predict the normal and tangential components of the grinding force when cutting Ti6Al4V reinforced with TiC particles using monolayer cubic boron nitride wheels and reported errors between predicted and experimental forces ranging from 6.1% to 9.4% for the normal force and from 7.4% to 9.8% for the tangential force. Furthermore, FEM results indicated that the maximum stress applied during the removal of the reinforcing particle varies considerably due to either crack propagation or particle breakage. Work by Dai et al. [20] comparing temperatures obtained experimentally and numerically when grinding Inconel 718 indicated that temperature increased with wheel
speed and the difference in the temperature values obtained by the two approaches was negligible. Considering the importance of milling operations for the metalworking industry and the fact the cutting temperature is a critical parameter affecting both productivity and the surface integrity of the machining component, the principal goal of this work is to perform a comparative study of the cutting temperature when end milling using implanted and tool-chip thermocouple methods under distinct operating parameters. Additionally, finite element simulation is employed to correlate the results provided by the experimental techniques. The relevance and novelty of the present work reside in the fact that two experimental and complementary techniques are employed for the determination of temperature in both the tool-chip interface and cutting tool, thus corroborating the accuracy of the numerical model to predict milling temperature. Consequently, the experimental approach, which is time consuming, can be replaced to a larger extent by numerical simulation.
2. EXPERIMENTAL PROCEDURE
Annealed AISI D2 cold work die steel (average hardness of 180 HV) was selected as work material. Indexable tungsten carbide inserts ISO grade P15-P30 (Sandvik Coromant code R390-17 04 08M-PM 1025) coated with TiN+TiCN were mounted on an end milling cutter coded R390-025A25-17L from the same manufacturer (Ø25 mm and two teeth). Dry down end milling tests were performed on a machining centre with 9 kW power and 7,500 rpm maximum rotational speed. Table 1 presents the cutting parameters tested. A full factorial experimental design with one replicate for each trial was employed, thus resulting in 72 tests performed at random.
Table 1. Conditions employed in the cutting temperature measurement tests Cutting speed vc (m/min)
60 – 90 - 180
Feed per tooth fz (mm/rev)
0.05 – 0.1 – 0.15
Axial depth of cut ap (mm)
1–2
Radial depth of cut ae (mm)
12.5 – 25
The data acquisition system specifically developed for this research is schematically represented in Figure 1. This system was employed to measure temperature using both experimental methods (tool-chip thermocouple and implanted thermocouple). It can be noticed that the data acquisition system is constituted of three parts: acquisition circuit, Arduino Nano 3.0 microcontroller (responsible for serial communication with the A/D precision converter present in the data acquisition system and for receiving and sending digital data to the computer) and Bluetooth HC-06 device. The principal components of the acquisition circuit are the following integrated circuits manufactured by Linear Technology: LTC1091 A/D precision converter, LT1019A voltage reference, LTC1052 operational amplifier and LT1025 thermocouple cold junction compensator. A K-type thermocouple is connected to the A/D converter. The data acquisition system was set to operate at an acquisition rate of 500 Hz.
Figure 1. Schematic diagram of the temperature acquisition system
2.1 Tool-chip thermocouple
In order to form a thermocouple between the workpiece and cutting insert, both were electrically insulated from the machining centre. The milling cutter was covered with polyurethane paint used in electric motors and the workpiece was insulated from the machine tool table using a nylon plate, see Figure 2. Milling cutter and workpiece were connected to the data acquisition system shown in Figure 1 and a graphite brush was
used to make electric contact between the rotating milling cutter and the cable connected to the acquisition system (Figure 2).
Figure 2: Tool-chip thermocouple experimental setup
Calibration of the tool-workpiece thermocouple was conducted in a furnace. Silver brazing was used to connect a cutting insert to a sample of the work material. Copper wires were silver brazed at the opposite ends of both materials and connected to a voltmeter. In spite of its undesirable influence on the calibration results, silver alloy was selected due to its high melting temperature (approximately 700º C). The calibration curve is shown in Figure 3, where it can be noted that it can be satisfactorily represented by a straight line with a correlation coefficient of 0.984.
Figure 3. Calibration curve for AISI D2 steel and tungsten carbide grade 1025
2.2 Implanted thermocouple
In this case, holes were produced in the inserts using die sinking electrical discharge machining (EDM) and Ø1 mm copper electrodes. Holes were produced at the opposite side of the rake face and halted 1mm from the rake face. A fixture was designed and built to make sure holes were produced exactly in the same position for each insert. Figure 4(a) shows a detail of the EDM processes and Figure 4(b) presents an insert after EDM.
(a) Figure 4. (a) EDM hole drilling and (b) drilled insert
(b)
In order to measure the temperature of the rotating insert using a K-type implanted thermocouple, a wireless connection with the computer was necessary because the developed data acquisition system (Figure 1) was attached to the tool holder. Figure 5(a) shows the case (made by fused deposition modelling) attached to the tool holder to accommodate the data acquisition system shown in Figure 5(b).
(a)
(b)
Figure 5. (a) Case mounted on the tool holder and (b) data acquisition system
2.3 Three dimensional finite element simulation
Finite element simulation of end milling was performed using Deform Machining 3D software aiming at determining heat distribution in the cutting zone and confronting numerical and experimental temperature values. Deform uses implicit method with updated Lagrangian. Tetrahedral elements with four points of integration were used. The elements in the cutting tool were considered rigid, while the elements in the workpiece were considered rigid-plastic. Rigid-plastic elements allow automatic
remeshing. Owing to the considerable computation time required by numerical modelling (12 hours for each milling condition), it was not possible to simulate all milling conditions depicted in Table 1 and the parameters used in the numerical simulation procedure can be seen in Table 2. The model for transient temperature considers the effects of conduction and convection. Initial temperature of both the cutting tool and workpiece was 20° C. The value of the heat transfer coefficient was set to 10000 N/s/mm/°C. Heat transfer in the workpiece was considered on the machined surface only, while in the cutting tool it was considered on the entire volume. Inferior and lateral workpiece surfaces were fixed in nine degrees of freedom. Owing to the high strain rates involved, flow stress was simulated using Johnson-Cook model [3], which takes into account strain hardening, strain rate hardening and thermal softening effects during metal cutting. Furthermore, feeding the software with accurate values of the friction coefficient between tool and work materials is critical to obtain reliable results. Therefore, pin on disk tests were performed in a Microtest MT tribometer using the same materials employed in the machining trials (AISI D2 steel and titanium nitride coated carbide), see Figure . A load of 5 N was applied during six minutes together with two peripheral speeds (60 m/min and 90 m/min) used in the milling trials (Table 1) in order to provide the numerical model with friction coefficient values as close as possible to those observed in actual cutting conditions. This load was chosen based on the work by Jiang et al. [21], who investigated the behavior of AISI D2 steel against TiAlN coated carbide employing experimental and numerical techniques.
Table 2.Cutting parameters used in numerical simulation of milling temperature Cutting speed vc (m/min)
60
90
90
90
90
Feed per tooth fz (mm/rev)
0.05
0.05
0.1
0.15
0.05
Axial depth of cut ap (mm)
1.0
1.0
1.0
1.0
2.0
Radial depth of cut ae (mm)
12.5
12.5
12.5
12.5
12.5
(a)
(b)
Figure 6. (a) AISI D2 disc and (b) TiN coated tungsten carbide pins
The insert manufacturer did not provide the complete dimensions and geometry of the cutting tool, therefore, a Nanovea PS50 optical profilemeter was employed to obtain full details on the insert geometry. Figures 7(a) and 7(b) show insert and workpiece meshes with 5500 and 27000 elements (947 and 5167 nodes), respectively, and Figure 7(c) presents a snapshot of the three dimensional numerical simulation. The initial number of elements in the workpiece was 35126 and its final number of elements resulted from automatic remeshing aimed at describing stress, strain, strain rate and temperature at the tool-workpiece interface during simulation. (a)
(b)
(c)
Figure 7. (a) Insert mesh, (b) workpiece mesh and (c) simulation snapshot
3. RESULTS AND DISCUSSION
Firstly, the experimental results are presented with the corresponding statistical analysis, followed by the findings concerned with the friction coefficient and numerical simulations.
3.1 Experimental results
Figures 8 to 11 show the influence of the cutting parameters on milling temperature measured employing the tool-chip thermocouple (blue bars associated with the left axis of ordinates) and implanted thermocouple (red bars related to the right axis of ordinate) methods and the error bars indicate the highest and lowest temperature values obtained under each condition. The difference in temperature values recorded by the two techniques derives from the fact that the tool-chip thermocouple measures the interface temperature, which is expected to be higher since this is the region where energy is converted into heat by plastic deformation and friction. In contrast, the implanted thermocouple measures the temperature of the cutting tool 1 mm beneath the interface, which is affected by energy partition and thermal conductivity of the tool material. Temperature values represented in the graphs are related to a milling time of 7 s in order to avoid the influence of tool wear on temperature and owing to the fact that after such period the temperature elevation recorded by the implanted thermocouple method became less steep (23.2° C/s during the first 5 s and 8.5° C between 6 and 8 s of cutting). The effect of cutting speed and feed rate on temperature when down milling using an axial depth of cut ap=1 mm and a radial depth of cut ae=12.5 mm can be seen in Figure 8. Irrespectively of the method employed, temperature increased with cutting speed and feed rate, since the elevation of the former results in higher power and energy conveyed to the cutting zone and the increase of the latter means that higher milling forces and energy will be required to shear a larger area of material. Highest temperatures were recorded under the most severe cutting condition (vc=180 m/min and fz=0.15 mm/rev): 457° C for the tool-chip thermocouple and 139° C for the implanted thermocouple method.
Figure 8. Effect of cutting speed and feed per tooth on milling temperature measured with tool-chip thermocouple and implanted thermocouple (ap = 1 mm e ae = 12.5 mm)
A similar trend is observed when the axial depth of is elevated to ae=25 mm, see Figure 9, although higher temperature values are usually expected due to the fact that a higher radial depth of cut represents a higher volume of material to be removed in addition to a longer tooth path, which promotes more friction between the insert and the workpiece. The highest temperature recorded were 493° C using the tool-chip thermocouple and 181° C using the implanted thermocouple.
Figure 9. Effect of cutting speed and feed per tooth on milling temperature measured with tool-chip thermocouple and implanted thermocouple (ap = 1 mm e ae = 25 mm)
Figures 10 and 11 are related to a radial depth of cut ap=2 mm and axial depths of cut of, respectively, ae=12.5 mm and ae= 25 mm. Interestingly, the difference between the temperature values recorded using both techniques became smaller, probably due to the fact that an increase in the axial depth of cut means that the remote thermocouple location is closer to the cutting zone. Similarly to Figures 8 and 9, temperature increased with cutting speed and feed per tooth for the same reasons given above, however, comparing Figures 8 to 10 it can be noticed that the implanted thermocouple is more sensitive to changes in both the axial and radial depths of cut, i.e., when the axial depth of cut is elevated from ap=1 mm to ap=2 mm, the temperature measured by the tool-chip thermocouple increases 12° C and 15° C, respectively, for ae=12.5 mm and ae= 25 mm, while the values recorded using the implanted thermocouple increase 43° C and 62° C. A similar trend is noticed when the axial depth of cut increases from ae=12.5 mm to ae=25 mm: while the temperature measured by the tool-chip thermocouple increases 36° C and 39° C for ap=1 mm and ap=2 mm, respectively, elevations of 43° C and 61° C were recorded using the implanted thermocouple.
Figure 10. Effect of cutting speed and feed per tooth on milling temperature measured with tool-chip thermocouple and implanted thermocouple (ap = 2 mm e ae = 12.5 mm)
Figure 11. Effect of cutting speed and feed per tooth on milling temperature measured with tool-chip thermocouple and implanted thermocouple (ap = 2 mm e ae = 25 mm)
Comparing Figures 9 to 11 it can be noted that the difference in temperature decreases as the axial and radial depths of cut (ap and ae, respectively) are increased. This behavior can be explained by the fact that an increase in a p elevates the contact area between tool and chip and approximated the heat source to the thermocouple location. In turn, increasing ae results in longer active periods (when cutting takes place), thus allowing more conduction of heat from the source to the thermocouple location. In order to identify the relevance of each milling parameter on the temperature measured using each methods, Pareto charts for a significance level α=0.05 are presented in Figures 12(a) and 12(b). Figure 12(a) shows that cutting speed is by far the most influential parameter affecting milling temperature, followed, in this order albeit with less influence, by radial depth of cut, feed per tooth, axial depth of cut and by the interactions between feed per tooth and axial depth of cut the fourth order interaction among all investigated factors. This behaviour can be explained by the fact that the toolchip thermocouple measures the average temperature in the hot junction, i.e., in the tool-chip interface, which increases with cutting speed. Moreover, the elevation of cutting speed causes a reduction in chip thickness, thus concentrating heat in a smaller area and promoting high temperatures. The remaining factors possess similar effect: on the one hand, they contribute to the elevation of the cutting area, while on the other they represent larger shear area and lower average temperature in the tool-chip interface.
(a) Standardized effects for milling temperature using the tool-chip thermocouple method
(b) Standardized effects for milling temperature measured using the implanted thermocouple method Figure 12. Pareto char for milling temperature measured using: (a) tool-chip thermocouple and (b) implanted thermocouple (significance level α=0.05)
The Pareto chart with standardized effects for milling temperature measured with the implanted thermocouple is given in Figure 12(b). In this case, axial depth of cut is the most prominent factor, followed by radial depth of cut, cutting speed and feed per tooth. Increasing axial depth of cut makes the location of the implanted thermocouple closer to the heat source and the larger the radial depth of cut, the longer the active cutting time during which heat is conducted through the insert, thus leading to higher temperatures. Increasing cutting speed results in higher cutting temperature (higher power and energy required for shearing), however, the shorter active time and thinner chip thickness counterbalance this effect. As far as the feed per tooth is concerned, its elevation results in higher cutting temperature due to the larger shear area, however, in contrast to the evident influence of axial depth of cut, it seems that the elevation of feed per tooth from 0.05 mm/rev to 0.15 mm/rev is not sufficient to make the heat source come closer to the thermocouple location and cause a substantial temperature elevation. Figures 13 and 14 show the main effects plots for milling temperature measured with the tool-chip and remote thermocouples, respectively. Mean temperature increases with the elevation of any factor irrespectively of the measuring technique, nevertheless it responds differently for each cutting parameter. As indicated in the Pareto chart presented in Figure 12(a), cutting speed is by far the most significant factor (Figure 13a), followed by radial depth of cut (Figure 13d), feed per tooth (Figure 13b) and axial depth of cut (Figure 13c), due to the reasons previously discussed.
Figure 13. Main effects plots milling temperature measured using tool-chip thermocouple: (a) cutting speed, (b) feed rate, (c) axial depth of cut and (d) radial depth of cut
A different scenario is found when the main plots for mean temperature recorded with the remote thermocouple are observed, see Figure 14. As previously explained, in this case axial depth of cut is the most relevant factor (Figure 14c), followed by radial depth of cut (Figure 14d), cutting speed (Figure 14a) and feed per tooth (Figure 14b).
Figure 14. Main effects plots milling temperature measured using remote thermocouple: (a) cutting speed, (b) feed rate, (c) axial depth of cut and (d) radial depth of cut
The influence of cutting direction on milling temperature measured with the tool-chip thermocouple is shown in Figure 15. Irrespectively of the cutting condition employed, similar temperature values are recorded and in order to determine whether cutting direction statistically affects milling temperature, a paired t-test was performed with a significance level α=0.05. The average temperature value was 370.3° C for up milling and 360.3° C for down milling, however, the p-value was 0.051 (higher than the selected significance level), which means that there is no difference between the temperatures when up or down milling.
Figure 15. Effect of cutting direction on milling temperature measured with tool-chip thermocouple and implanted thermocouple under selected cutting conditions
3.2 Friction coefficient and three dimensional numerical simulation results
The friction coefficient between AISI D2 steel discs and coated tungsten carbide pins of the same grade used as cutting insert was determined under sliding speeds identical to cutting speed values used in the temperature measurement tests (60 m/min and 90 m/min) and are shown in Figures 16(a) and 16(b), respectively. Due to the scatter in the results, the friction coefficient used to feed the numerical simulations was the average value during the first 600 rotations, i.e., µ=0.54 for 60 m/min and µ=0.46 for 90 m/min. Figure 17 shows the importance of the proper determination of the friction coefficient for model accuracy. When simulating milling at a cutting speed of 90 m/min, feed per tooth of 0.05 mm/rev, axial depth of cut of 1 mm and radial depth of cut of 12 mm, the friction coefficient obtained experimentally (µ=0.46) resulted in an interface temperature of 311° C, which increased to 332° C for µ=0.56 and to 402° C when µ=0.66.
Fric tio n coe ffici ent
(a) Friction coefficient curve obtained for a sliding speed of 60 m/min Fric tio n coe ffici ent
(b) Friction coefficient curve obtained for a sliding speed of 90 m/min Figure 16. Results of the pin on disc tests performed at distinct sliding speeds: (a) 60 m/min and (b) 90 m/min
Figure 17. Effect of friction coefficient on the temperature obtained by FEM
The numerical simulation results for cutting speeds of v c=60 m/min and vc=90 m/min are presented, respectively, in Figures 18 and 19 and the remaining parameters are given in Table 2 (average simulation time was 12 hours for each cutting condition). The highest temperature value obtained numerically in the tool-chip interface was 313° C
for vc=60 m/min, fz=0.05 mm/rev, ap=1 mm and ae=12.5 mm, see Figure 18(a), which is slightly higher than that recorded using the tool-chip thermocouple method (304° C) under the same cutting condition, please refer to Figure 8. Similarly, the temperature value calculated in the implanted thermocouple location, represented by a black straight line in Figure 18(b), ranged from 50.2° C to 80.3° C, encompassing the experimental temperature of 76° C recorded by the implanted thermocouple (Figure 8).
Thermocouple hole
a)
b) Figure 18. Milling temperature simulation (vc=60 m/min, fz=0.05 mm/rev, ap=1 mm and ae=12.5 mm)
Temperature simulation results for vc=90 m/min, fz=0.05 mm/rev, ap=1 mm and ae=12.5 mm) are presented in Figure 19, where it can be noted that the maximum temperature at the tool-chip interface reached 332° C (Figure 19a) and the temperature at the implanted thermocouple location stretched from 52.3° C to 84.7° C (Figure 19b). These values can be considered quite reliable since under the same cutting condition, the temperature of the tool-chip interface obtained experimentally was 328° C and the temperature recorded by the implanted thermocouple was 78° C, please refer to Figure 8.
Thermocouple hole
a)
b) Figure 19. Milling temperature simulation (vc=90 m/min, fz=0.05 mm/rev, ap=1 mm and ae=12.5 mm) Finally, Figure 20 summarizes the results obtained from the experimental work (using both techniques) and those achieved through numerical simulation for the cutting conditions depicted in Table 2. Experimental and numerical results concerned with the tool-chip thermocouple method are represented in blue, while the results associated with the implanted thermocouple (remote temperature) method are represented in red. A quite satisfactory agreement between experimental and numerical results was obtained, especially when the tool-chip thermocouple was employed. In the case of the implanted thermocouple, however, the experimental temperature value tended to stray from the range stipulated by the isothermal curves where the thermocouple was confined. One possible reason for such behaviour may be changes in heat partition fractions (for chip, insert and workpiece) as the cutting parameters associated with the shear plane area (feed per tooth or axial depth of cut) are increased.
Figure 20. Comparison between experimental and numerical temperature results obtained at distinct cutting conditions
4. CONCLUSIONS
Cutting temperature was assessed during dry milling of annealed AISI D2 cold work tool steel using indexable tungsten carbide inserts under various cutting conditions. Additionally, finite element modelling was performed and the results compared with those obtained experimentally. The following conclusion can be drawn from this work:
A wireless data acquisition system constituted of a temperature acquisition
system, Arduino microcontroller and Bluetooth device was successfully developed to measure milling temperature using the tool-chip and implanted thermocouple methods.
Milling temperature increased with cutting speed, feed per tooth and both axial
and radial depths of cut, nevertheless, the relevance of each factor varied according to the measurement technique: for the tool-chip thermocouple method, cutting speed was by far the most relevant factor, followed by radial depth of cut, feed per tooth and axial depth of cut. However, when the implanted thermocouple method was employed, axial depth of cut was the most influential factor, being followed radial depth of cut, cutting speed and feed per tooth at more regular intervals.
Although the average cutting temperature when up milling had been slightly
higher in comparison with down milling (370.3° C against 360.3° C), the paired t-test indicated that cutting direction did not affect milling temperature measured by the toolchip thermocouple.
The experimental determination of the friction coefficient between tool and
work materials was critical to the successful three dimensional numerical simulation of milling temperature, which was capable of satisfactorily estimating temperature distribution in the cutting zone.
ACKNOWLEDGEMENTS
The authors are grateful to Prof. Elaine C.S. Corrêa (Centro Federal de Educação Tecnológica de Minas Gerais, CEFET-MG) for the support during the calibration procedure. Additional thanks go to the following research agencies in Brazil: CAPES, CNPq and FAPEMIG.
FUNDING
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
REFERENCES
[1] G.M. Pittalà, M. Monno, A new approach to the prediction of temperature of the workpiece of face milling operations of Ti-6Al-4V, Appl. Therm. Eng. 31 (2011) 173180. [2] N.A. Abukhshim, P.T. Mativenga, M.A. Sheikh, Investigation of Heat Partition in High Speed Turning of High Strength Alloy Steel, Int. J. Mach. Tools Manuf. 45 (2005) 1687–1695. [3] X. Cui, J. Zhao, Z. Pei, Analysis of transient average tool temperatures in face milling, Int. Comm. Heat Mass Transfer 39 (2012) 786-791.
[4] M.A. El Hakim, M.A. Shalaby, S.C. Veldhuis, G.K. Dosbaeva, Effect of secondary hardening on cutting forces, cutting temperature, and tool wear in hard turning of high alloy tool steel, Meas. 65 (2015) 233-238. [5] C. Courbon, R. Mabrouki, J. Rech, D. Mazuyer, E. D’Eramo. On the existence of a thermal contact resistance at the tool-chip interface in dry cutting of AISI 1045: formation mechanisms and influence on the cutting process, Appl. Therm. Eng. 50 (2013) 1311-1325. [6] M.B. Silva, J. Wallbank, Cutting temperature: prediction and measurement methods a review, J. Mater. Process. Technol. 88 (1999) 195-202. [7] F. Benabid, H. Benmoussa, M. Arrouf, A thermal modeling to predict and control the cutting temperature, The simulation of face-milling process. Procedia Eng. 74 (2014) 37-42. [8] A. Maurel-Pantel, M. Fontaine, S. Thibaud, J.C. Gelin, 3D FEM simulations of shoulder milling operations on a 304L stainless steel, Simul. Modell. Pract. Theory 2 (2012) 13–27. [9] F. Jiang, L. Yan, Y. Rong, Orthogonal cutting of hardened AISI D2 steel with TiAlN-coated inserts—simulations and experiments, Int. J. Adv. Manuf. Technol. 64 (2013) 1555-1563. [10] A.N.S. Barrios, J.B.C. Silva, A.R. Rodrigues et al., Modeling heat transfer in die milling, Appl. Therm. Eng. 64 (2014) 108-116. [11] G.S. Samy, T. Kumaram, Measurement and analysis of temperature, thrust force and surface roughness in drilling of AA (6351)-B4C composite, Meas. 103 (2017) 1-9. [12] K. Kerrigan, J. Thil, R. Hewison, G.E. O’Donnell, An integrated telemetric thermocouple sensor for process monitoring of CFRP milling operations, Procedia CIRP 1(2012) 449-454. [13] A. Guha, H. Li, Z. Sun, C. Ma, D. Werschmoeller, Xiaochun Li, Wireless acquisition of temperature data from embedded thin film sensors in cutting insert, J. Manuf. Process. 14 (2012) 360-365. [14] G. Le Coz, M. Marinescu, A. Devillez, D. Dudzinski, L. Velnom, Measuring temperature of rotating cutting tools: Application to MQL drilling and dry milling of aerospace alloys, Appl. Therm. Eng. 36 (2012) 434-441.
[15]G. Sutter, A. Faure, A. Molinari, N. Ranc, V. Pina, An experimental technique for the measurement of temperature fields for the orthogonal cutting in high speed machining, Int. J. Mach. Tools Manuf. 43 (2003) 671-678. [16] Y. Sun, J. Sun, J. Li, Q. Xiong, An experimental investigation of the influence of cutting parameters on cutting temperature in milling Ti6Al4V by applying semiartificial thermocouple. Int. J. Adv. Manuf. Technol. 70 (2014) 765-773. [17] M. Kikuchi, The use of cutting temperature to evaluate the machinability of titanium alloys, Acta Biomater. 5 (2009) 770-775. [18] L.C. Brandão, R.T. Coelho, A.R. Rodrigues, Experimental and theoretical study of workpiece temperature when end milling hardened steels using (TiAl)N-coated and PcBN-tipped tools, J. Mater. Process. Technol. 199 (2008) 234–244. [19] C. Liu, W. Ding, T. Yu, C. Yang, Materials removal mechanism in high-speed grinding of particle reinforced titanium matrix composites, Precis. Eng. 51(2018) 68-77. [20] C. Dai, W. Ding, Y. Zhu, J. Xu, H. Yu, Grinding temperature and power consumption in high speed grinding of Inconel 718 nickel-based superalloy with a vitrified CBN wheel, Precis. Eng. 52(2018) 192-200. [21] F. Jiang, L. Yan, Y. Rong, Orthogonal cutting of hardened AISI D2 steel with TiAlN-coated inserts—simulations and experiments, Int. J. Adv. Manuf. Technol. 64(2013) 1555–1563.
Highlights
A wireless data acquisition system was developed to measure milling temperature Tool-chip and implanted thermocouple techniques were used together with FEM The influence of cutting parameters on temperature depends on the selected method Cutting direction does not statistically affect milling temperature Determination of friction coefficient is critical to the success of FE simulation
S1359-4311(18)32107-0 https://doi.org/10.1016/j.applthermaleng.2018.07.107 ATE 12465
To appear in:
Applied Thermal Engineering
Received Date: Revised Date: Accepted Date:
4 April 2018 10 July 2018 20 July 2018
Please cite this article as: H.V. Lima, A.F.V. Campidelli, A.A.T. Maia, A.M. Abrão, TEMPERATURE ASSESSMENT WHEN MILLING AISI D2 COLD WORK DIE STEEL USING TOOL-CHIP THERMOCOUPLE, IMPLANTED THERMOCOUPLE AND FINITE ELEMENT SIMULATION, Applied Thermal Engineering (2018), doi: https://doi.org/10.1016/j.applthermaleng.2018.07.107
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
TEMPERATURE ASSESSMENT WHEN MILLING AISI D2 COLD WORK DIE STEEL USING TOOL-CHIP THERMOCOUPLE, IMPLANTED THERMOCOUPLE AND FINITE ELEMENT SIMULATION
Hugo V. Lima, Augusto F.V. Campidelli, Antônio A.T. Maia, Alexandre M. Abrão*
Department of Mechanical Engineering, Universidade Federal de Minas Gerais, Av. Antônio Carlos, 6627, Pampulha, Belo Horizonte, MG, CEP 31270-901, Brazil. Phone: 55 31 3409 5138. E-mail: [email protected]
* Corresponding author
Abstract: Due to the cyclic mechanical and thermal loads imposed to the cutting tool during milling, the study of the process temperature is of utmost importance for the better understanding of various associated phenomena, such as tool life and wear mechanisms, cutting forces behaviour and workpiece subsurface metallurgical alterations. Nevertheless, temperature measurement during milling operations imposes a number of restraints to experimental methods, mostly related to the cutter rotational speed, variable chip thickness and intermittent action of the cutting edges. The principal goal of this work is to perform a comparative study of the cutting temperature during in end milling using implanted and tool-chip thermocouple methods under distinct operating parameters. Additionally, finite element simulation is employed to correlate the results provided by the experimental techniques. Tool-chip thermocouple and implanted thermocouple experimental methods were used in addition to three dimensional finite element simulation. The findings indicated that the developed system is capable to cope with the drawbacks associated with intermittent machining operations and to provide reliable temperature values for both experimental methods. Milling temperature increased with cutting speed, feed per tooth and both axial and radial depths of cut, however, the relevance of each factor varied in accordance with the measurement method. The average cutting temperature was not statistically affected by cutting direction and the experimental determination of the friction coefficient between tool and
workpiece was critical to the accurate temperature determination using three dimensional numerical simulation.
Keywords: milling; temperature; tool-chip thermocouple; implanted thermocouple; finite element simulation.
1. INTRODUCTION
The damage caused by the energy converted into heat during metal cutting operations is a critical factor restricting production rate owing to the fact that it drastically affects both the cutting tool performance and workpiece surface integrity. Therefore, cutting temperature measurement and control deserve constant attention, especially in the case of metals and alloys considered difficult to machine [1]. Nevertheless, investigations on cutting temperature face a series of restrictions caused by the nature of the contact between tool and workpiece, chip formation mechanism and occasional presence of cutting fluid [2]. According to Cui et al. [3], the assessment on the intensity and form as energy is converted into heat is more complex in milling due to its intermittent nature with a number of cutting edges being sequentially subjected to cyclic heating and cooling at high rates. Consequently, thermal stresses are imposed in the cutting tool, thus leading to the emergence of cracks that will impair the performance of the cutting tool [4]. Furthermore, Courbon et al. [5] state that while perfect thermal contact can be assumed to take place at the tool-chip interface where conditions of seizure are dominant, conditions of sliding at the tool-chip interface lead to a thermal contact resistance that drastically affects heat flux and temperature distribution in the cutting tool. Silva and Wallbank [6] state that the practical limitations associated with experimental methods used to determine temperature distribution in the tool-workpiece interface result in obtaining the average tool-workpiece temperature, nonetheless, such information is crucial to the development of advanced cutting tool materials. Despite the difficulties associated with long setup times and need of specific equipment, experimental techniques are considered quite reliable for the determination of cutting temperature and are often used to assess the performance of analytical and numerical
methods [7], which experience a number of difficulties to model either oblique or intermittent cutting. According to Maurel-Pantel et al. [8], numerical models are capable of describing only the essential aspects of machining operations based on orthogonal cutting, while complex operations, such as milling, are not sufficiently explored and required greater advances in numerical modelling. Jiang et al. [9] employed numerical and experimental methods to evaluate the influence of cutting parameters on temperature in orthogonal cutting of AISI D2 tool steel using polycystalline cubic boron nitride (PcBN) tools and noticed that cutting speed and the depth of cut are the most relevant factors, with the maximum temperature recorded near the tool tip ranging from 711º C to 792º C. Barrios et al. [10] investigated milling temperature of AISI H13 hot work die steel using numerical modelling and type K thermocouples implanted 3 mm beneath the surface. The results indicated that after 30 seconds the maximum workpiece temperature was 40º C for a cutting speed of 100 m/min and 35º C for 50 m/min the cutting temperature. According to Samy and Kumaram [11], the main drawback related to the implanted thermocouple method relies on the fact that the maximum temperature in the cutting region cannot always be determined. The fact that the tool rotates at high speeds during milling has prevented, for a long time, cutting tool temperature measurement using implanted thermocouples, however, a number of alternatives has been presented in recent years. Among them, Kerrigan et al. [12] developed a wireless temperature measurement system integrated into a cutting tool holder with a thermocouple embedded in the cutting tool. The results showed that the temperature stabilization at the thermocouple location requires a certain amount of time, which depends on the characteristics of the tool coating and substrate. A similar approach was employed by Guha et al. [13], who used a micro thin-film sensor embedded in the cutting tool. A Bluetooth module was used to acquire data and establish communication with the computer. The signal conditioning circuit was designed and developed to increase the signal-to-noise ratio of the embedded micro thin film sensors. Moreover, an averaging filter algorithm was implemented in the data acquisition software interface. The findings showed that the developed system possess a resolution of 1° C and was effective in minimizing noise when measuring low voltage signals.
The implanted thermocouple method was used by Le Coz et al. [14] to measure temperature when milling AA7075 aluminium alloy and drilling Ti6Al4V titanium alloy and the results indicated that the device is sensitive to changes in the cutting parameters (cutting speed, feed rate and tool geometry and coating) and work material (aluminium alloys, steel and cast iron). However, this technique presents disadvantages associated with the weakening of the cutting wedge and heat flow alteration caused by the hole produced. Furthermore, Sutter et al. [15] points out the importance of meticulous calibration of the pair and proper insulation of the cutter and workpiece when the tool-chip thermocouple method is used. Sun et al. [16] employed the tool-chip thermocouple technique to measure temperature when milling Ti6Al4V titanium alloy. The results indicated that milling temperature increased with the elevation of both cutting speed and feed rate, however, the former parameter was found to be the most relevant. As far as the influence of depth of cut is concerned, Kikuchi [17] reported that its elevation leads to higher milling temperatures, irrespectively of the work material being tested. Brandão et al. [18] compared the temperature values provided by experimental (implanted thermocouple) and numerical (finite element) approaches when milling hardened AISI H13 and AISI D2 tool steels using coated and uncoated polycrystalline cubic boron nitride tipped end mills. They found that although the finite element model was capable to accurately predict the energy transferred from the cutting zone to the workpiece, it was ineffective in estimating the temperature decrease when distinct cooling environments (compressed or cold air) were used in comparison with dry cutting. Liu et al. [19] employed finite element modelling (FEM) to predict the normal and tangential components of the grinding force when cutting Ti6Al4V reinforced with TiC particles using monolayer cubic boron nitride wheels and reported errors between predicted and experimental forces ranging from 6.1% to 9.4% for the normal force and from 7.4% to 9.8% for the tangential force. Furthermore, FEM results indicated that the maximum stress applied during the removal of the reinforcing particle varies considerably due to either crack propagation or particle breakage. Work by Dai et al. [20] comparing temperatures obtained experimentally and numerically when grinding Inconel 718 indicated that temperature increased with wheel
speed and the difference in the temperature values obtained by the two approaches was negligible. Considering the importance of milling operations for the metalworking industry and the fact the cutting temperature is a critical parameter affecting both productivity and the surface integrity of the machining component, the principal goal of this work is to perform a comparative study of the cutting temperature when end milling using implanted and tool-chip thermocouple methods under distinct operating parameters. Additionally, finite element simulation is employed to correlate the results provided by the experimental techniques. The relevance and novelty of the present work reside in the fact that two experimental and complementary techniques are employed for the determination of temperature in both the tool-chip interface and cutting tool, thus corroborating the accuracy of the numerical model to predict milling temperature. Consequently, the experimental approach, which is time consuming, can be replaced to a larger extent by numerical simulation.
2. EXPERIMENTAL PROCEDURE
Annealed AISI D2 cold work die steel (average hardness of 180 HV) was selected as work material. Indexable tungsten carbide inserts ISO grade P15-P30 (Sandvik Coromant code R390-17 04 08M-PM 1025) coated with TiN+TiCN were mounted on an end milling cutter coded R390-025A25-17L from the same manufacturer (Ø25 mm and two teeth). Dry down end milling tests were performed on a machining centre with 9 kW power and 7,500 rpm maximum rotational speed. Table 1 presents the cutting parameters tested. A full factorial experimental design with one replicate for each trial was employed, thus resulting in 72 tests performed at random.
Table 1. Conditions employed in the cutting temperature measurement tests Cutting speed vc (m/min)
60 – 90 - 180
Feed per tooth fz (mm/rev)
0.05 – 0.1 – 0.15
Axial depth of cut ap (mm)
1–2
Radial depth of cut ae (mm)
12.5 – 25
The data acquisition system specifically developed for this research is schematically represented in Figure 1. This system was employed to measure temperature using both experimental methods (tool-chip thermocouple and implanted thermocouple). It can be noticed that the data acquisition system is constituted of three parts: acquisition circuit, Arduino Nano 3.0 microcontroller (responsible for serial communication with the A/D precision converter present in the data acquisition system and for receiving and sending digital data to the computer) and Bluetooth HC-06 device. The principal components of the acquisition circuit are the following integrated circuits manufactured by Linear Technology: LTC1091 A/D precision converter, LT1019A voltage reference, LTC1052 operational amplifier and LT1025 thermocouple cold junction compensator. A K-type thermocouple is connected to the A/D converter. The data acquisition system was set to operate at an acquisition rate of 500 Hz.
Figure 1. Schematic diagram of the temperature acquisition system
2.1 Tool-chip thermocouple
In order to form a thermocouple between the workpiece and cutting insert, both were electrically insulated from the machining centre. The milling cutter was covered with polyurethane paint used in electric motors and the workpiece was insulated from the machine tool table using a nylon plate, see Figure 2. Milling cutter and workpiece were connected to the data acquisition system shown in Figure 1 and a graphite brush was
used to make electric contact between the rotating milling cutter and the cable connected to the acquisition system (Figure 2).
Figure 2: Tool-chip thermocouple experimental setup
Calibration of the tool-workpiece thermocouple was conducted in a furnace. Silver brazing was used to connect a cutting insert to a sample of the work material. Copper wires were silver brazed at the opposite ends of both materials and connected to a voltmeter. In spite of its undesirable influence on the calibration results, silver alloy was selected due to its high melting temperature (approximately 700º C). The calibration curve is shown in Figure 3, where it can be noted that it can be satisfactorily represented by a straight line with a correlation coefficient of 0.984.
Figure 3. Calibration curve for AISI D2 steel and tungsten carbide grade 1025
2.2 Implanted thermocouple
In this case, holes were produced in the inserts using die sinking electrical discharge machining (EDM) and Ø1 mm copper electrodes. Holes were produced at the opposite side of the rake face and halted 1mm from the rake face. A fixture was designed and built to make sure holes were produced exactly in the same position for each insert. Figure 4(a) shows a detail of the EDM processes and Figure 4(b) presents an insert after EDM.
(a) Figure 4. (a) EDM hole drilling and (b) drilled insert
(b)
In order to measure the temperature of the rotating insert using a K-type implanted thermocouple, a wireless connection with the computer was necessary because the developed data acquisition system (Figure 1) was attached to the tool holder. Figure 5(a) shows the case (made by fused deposition modelling) attached to the tool holder to accommodate the data acquisition system shown in Figure 5(b).
(a)
(b)
Figure 5. (a) Case mounted on the tool holder and (b) data acquisition system
2.3 Three dimensional finite element simulation
Finite element simulation of end milling was performed using Deform Machining 3D software aiming at determining heat distribution in the cutting zone and confronting numerical and experimental temperature values. Deform uses implicit method with updated Lagrangian. Tetrahedral elements with four points of integration were used. The elements in the cutting tool were considered rigid, while the elements in the workpiece were considered rigid-plastic. Rigid-plastic elements allow automatic
remeshing. Owing to the considerable computation time required by numerical modelling (12 hours for each milling condition), it was not possible to simulate all milling conditions depicted in Table 1 and the parameters used in the numerical simulation procedure can be seen in Table 2. The model for transient temperature considers the effects of conduction and convection. Initial temperature of both the cutting tool and workpiece was 20° C. The value of the heat transfer coefficient was set to 10000 N/s/mm/°C. Heat transfer in the workpiece was considered on the machined surface only, while in the cutting tool it was considered on the entire volume. Inferior and lateral workpiece surfaces were fixed in nine degrees of freedom. Owing to the high strain rates involved, flow stress was simulated using Johnson-Cook model [3], which takes into account strain hardening, strain rate hardening and thermal softening effects during metal cutting. Furthermore, feeding the software with accurate values of the friction coefficient between tool and work materials is critical to obtain reliable results. Therefore, pin on disk tests were performed in a Microtest MT tribometer using the same materials employed in the machining trials (AISI D2 steel and titanium nitride coated carbide), see Figure . A load of 5 N was applied during six minutes together with two peripheral speeds (60 m/min and 90 m/min) used in the milling trials (Table 1) in order to provide the numerical model with friction coefficient values as close as possible to those observed in actual cutting conditions. This load was chosen based on the work by Jiang et al. [21], who investigated the behavior of AISI D2 steel against TiAlN coated carbide employing experimental and numerical techniques.
Table 2.Cutting parameters used in numerical simulation of milling temperature Cutting speed vc (m/min)
60
90
90
90
90
Feed per tooth fz (mm/rev)
0.05
0.05
0.1
0.15
0.05
Axial depth of cut ap (mm)
1.0
1.0
1.0
1.0
2.0
Radial depth of cut ae (mm)
12.5
12.5
12.5
12.5
12.5
(a)
(b)
Figure 6. (a) AISI D2 disc and (b) TiN coated tungsten carbide pins
The insert manufacturer did not provide the complete dimensions and geometry of the cutting tool, therefore, a Nanovea PS50 optical profilemeter was employed to obtain full details on the insert geometry. Figures 7(a) and 7(b) show insert and workpiece meshes with 5500 and 27000 elements (947 and 5167 nodes), respectively, and Figure 7(c) presents a snapshot of the three dimensional numerical simulation. The initial number of elements in the workpiece was 35126 and its final number of elements resulted from automatic remeshing aimed at describing stress, strain, strain rate and temperature at the tool-workpiece interface during simulation. (a)
(b)
(c)
Figure 7. (a) Insert mesh, (b) workpiece mesh and (c) simulation snapshot
3. RESULTS AND DISCUSSION
Firstly, the experimental results are presented with the corresponding statistical analysis, followed by the findings concerned with the friction coefficient and numerical simulations.
3.1 Experimental results
Figures 8 to 11 show the influence of the cutting parameters on milling temperature measured employing the tool-chip thermocouple (blue bars associated with the left axis of ordinates) and implanted thermocouple (red bars related to the right axis of ordinate) methods and the error bars indicate the highest and lowest temperature values obtained under each condition. The difference in temperature values recorded by the two techniques derives from the fact that the tool-chip thermocouple measures the interface temperature, which is expected to be higher since this is the region where energy is converted into heat by plastic deformation and friction. In contrast, the implanted thermocouple measures the temperature of the cutting tool 1 mm beneath the interface, which is affected by energy partition and thermal conductivity of the tool material. Temperature values represented in the graphs are related to a milling time of 7 s in order to avoid the influence of tool wear on temperature and owing to the fact that after such period the temperature elevation recorded by the implanted thermocouple method became less steep (23.2° C/s during the first 5 s and 8.5° C between 6 and 8 s of cutting). The effect of cutting speed and feed rate on temperature when down milling using an axial depth of cut ap=1 mm and a radial depth of cut ae=12.5 mm can be seen in Figure 8. Irrespectively of the method employed, temperature increased with cutting speed and feed rate, since the elevation of the former results in higher power and energy conveyed to the cutting zone and the increase of the latter means that higher milling forces and energy will be required to shear a larger area of material. Highest temperatures were recorded under the most severe cutting condition (vc=180 m/min and fz=0.15 mm/rev): 457° C for the tool-chip thermocouple and 139° C for the implanted thermocouple method.
Figure 8. Effect of cutting speed and feed per tooth on milling temperature measured with tool-chip thermocouple and implanted thermocouple (ap = 1 mm e ae = 12.5 mm)
A similar trend is observed when the axial depth of is elevated to ae=25 mm, see Figure 9, although higher temperature values are usually expected due to the fact that a higher radial depth of cut represents a higher volume of material to be removed in addition to a longer tooth path, which promotes more friction between the insert and the workpiece. The highest temperature recorded were 493° C using the tool-chip thermocouple and 181° C using the implanted thermocouple.
Figure 9. Effect of cutting speed and feed per tooth on milling temperature measured with tool-chip thermocouple and implanted thermocouple (ap = 1 mm e ae = 25 mm)
Figures 10 and 11 are related to a radial depth of cut ap=2 mm and axial depths of cut of, respectively, ae=12.5 mm and ae= 25 mm. Interestingly, the difference between the temperature values recorded using both techniques became smaller, probably due to the fact that an increase in the axial depth of cut means that the remote thermocouple location is closer to the cutting zone. Similarly to Figures 8 and 9, temperature increased with cutting speed and feed per tooth for the same reasons given above, however, comparing Figures 8 to 10 it can be noticed that the implanted thermocouple is more sensitive to changes in both the axial and radial depths of cut, i.e., when the axial depth of cut is elevated from ap=1 mm to ap=2 mm, the temperature measured by the tool-chip thermocouple increases 12° C and 15° C, respectively, for ae=12.5 mm and ae= 25 mm, while the values recorded using the implanted thermocouple increase 43° C and 62° C. A similar trend is noticed when the axial depth of cut increases from ae=12.5 mm to ae=25 mm: while the temperature measured by the tool-chip thermocouple increases 36° C and 39° C for ap=1 mm and ap=2 mm, respectively, elevations of 43° C and 61° C were recorded using the implanted thermocouple.
Figure 10. Effect of cutting speed and feed per tooth on milling temperature measured with tool-chip thermocouple and implanted thermocouple (ap = 2 mm e ae = 12.5 mm)
Figure 11. Effect of cutting speed and feed per tooth on milling temperature measured with tool-chip thermocouple and implanted thermocouple (ap = 2 mm e ae = 25 mm)
Comparing Figures 9 to 11 it can be noted that the difference in temperature decreases as the axial and radial depths of cut (ap and ae, respectively) are increased. This behavior can be explained by the fact that an increase in a p elevates the contact area between tool and chip and approximated the heat source to the thermocouple location. In turn, increasing ae results in longer active periods (when cutting takes place), thus allowing more conduction of heat from the source to the thermocouple location. In order to identify the relevance of each milling parameter on the temperature measured using each methods, Pareto charts for a significance level α=0.05 are presented in Figures 12(a) and 12(b). Figure 12(a) shows that cutting speed is by far the most influential parameter affecting milling temperature, followed, in this order albeit with less influence, by radial depth of cut, feed per tooth, axial depth of cut and by the interactions between feed per tooth and axial depth of cut the fourth order interaction among all investigated factors. This behaviour can be explained by the fact that the toolchip thermocouple measures the average temperature in the hot junction, i.e., in the tool-chip interface, which increases with cutting speed. Moreover, the elevation of cutting speed causes a reduction in chip thickness, thus concentrating heat in a smaller area and promoting high temperatures. The remaining factors possess similar effect: on the one hand, they contribute to the elevation of the cutting area, while on the other they represent larger shear area and lower average temperature in the tool-chip interface.
(a) Standardized effects for milling temperature using the tool-chip thermocouple method
(b) Standardized effects for milling temperature measured using the implanted thermocouple method Figure 12. Pareto char for milling temperature measured using: (a) tool-chip thermocouple and (b) implanted thermocouple (significance level α=0.05)
The Pareto chart with standardized effects for milling temperature measured with the implanted thermocouple is given in Figure 12(b). In this case, axial depth of cut is the most prominent factor, followed by radial depth of cut, cutting speed and feed per tooth. Increasing axial depth of cut makes the location of the implanted thermocouple closer to the heat source and the larger the radial depth of cut, the longer the active cutting time during which heat is conducted through the insert, thus leading to higher temperatures. Increasing cutting speed results in higher cutting temperature (higher power and energy required for shearing), however, the shorter active time and thinner chip thickness counterbalance this effect. As far as the feed per tooth is concerned, its elevation results in higher cutting temperature due to the larger shear area, however, in contrast to the evident influence of axial depth of cut, it seems that the elevation of feed per tooth from 0.05 mm/rev to 0.15 mm/rev is not sufficient to make the heat source come closer to the thermocouple location and cause a substantial temperature elevation. Figures 13 and 14 show the main effects plots for milling temperature measured with the tool-chip and remote thermocouples, respectively. Mean temperature increases with the elevation of any factor irrespectively of the measuring technique, nevertheless it responds differently for each cutting parameter. As indicated in the Pareto chart presented in Figure 12(a), cutting speed is by far the most significant factor (Figure 13a), followed by radial depth of cut (Figure 13d), feed per tooth (Figure 13b) and axial depth of cut (Figure 13c), due to the reasons previously discussed.
Figure 13. Main effects plots milling temperature measured using tool-chip thermocouple: (a) cutting speed, (b) feed rate, (c) axial depth of cut and (d) radial depth of cut
A different scenario is found when the main plots for mean temperature recorded with the remote thermocouple are observed, see Figure 14. As previously explained, in this case axial depth of cut is the most relevant factor (Figure 14c), followed by radial depth of cut (Figure 14d), cutting speed (Figure 14a) and feed per tooth (Figure 14b).
Figure 14. Main effects plots milling temperature measured using remote thermocouple: (a) cutting speed, (b) feed rate, (c) axial depth of cut and (d) radial depth of cut
The influence of cutting direction on milling temperature measured with the tool-chip thermocouple is shown in Figure 15. Irrespectively of the cutting condition employed, similar temperature values are recorded and in order to determine whether cutting direction statistically affects milling temperature, a paired t-test was performed with a significance level α=0.05. The average temperature value was 370.3° C for up milling and 360.3° C for down milling, however, the p-value was 0.051 (higher than the selected significance level), which means that there is no difference between the temperatures when up or down milling.
Figure 15. Effect of cutting direction on milling temperature measured with tool-chip thermocouple and implanted thermocouple under selected cutting conditions
3.2 Friction coefficient and three dimensional numerical simulation results
The friction coefficient between AISI D2 steel discs and coated tungsten carbide pins of the same grade used as cutting insert was determined under sliding speeds identical to cutting speed values used in the temperature measurement tests (60 m/min and 90 m/min) and are shown in Figures 16(a) and 16(b), respectively. Due to the scatter in the results, the friction coefficient used to feed the numerical simulations was the average value during the first 600 rotations, i.e., µ=0.54 for 60 m/min and µ=0.46 for 90 m/min. Figure 17 shows the importance of the proper determination of the friction coefficient for model accuracy. When simulating milling at a cutting speed of 90 m/min, feed per tooth of 0.05 mm/rev, axial depth of cut of 1 mm and radial depth of cut of 12 mm, the friction coefficient obtained experimentally (µ=0.46) resulted in an interface temperature of 311° C, which increased to 332° C for µ=0.56 and to 402° C when µ=0.66.
Fric tio n coe ffici ent
(a) Friction coefficient curve obtained for a sliding speed of 60 m/min Fric tio n coe ffici ent
(b) Friction coefficient curve obtained for a sliding speed of 90 m/min Figure 16. Results of the pin on disc tests performed at distinct sliding speeds: (a) 60 m/min and (b) 90 m/min
Figure 17. Effect of friction coefficient on the temperature obtained by FEM
The numerical simulation results for cutting speeds of v c=60 m/min and vc=90 m/min are presented, respectively, in Figures 18 and 19 and the remaining parameters are given in Table 2 (average simulation time was 12 hours for each cutting condition). The highest temperature value obtained numerically in the tool-chip interface was 313° C
for vc=60 m/min, fz=0.05 mm/rev, ap=1 mm and ae=12.5 mm, see Figure 18(a), which is slightly higher than that recorded using the tool-chip thermocouple method (304° C) under the same cutting condition, please refer to Figure 8. Similarly, the temperature value calculated in the implanted thermocouple location, represented by a black straight line in Figure 18(b), ranged from 50.2° C to 80.3° C, encompassing the experimental temperature of 76° C recorded by the implanted thermocouple (Figure 8).
Thermocouple hole
a)
b) Figure 18. Milling temperature simulation (vc=60 m/min, fz=0.05 mm/rev, ap=1 mm and ae=12.5 mm)
Temperature simulation results for vc=90 m/min, fz=0.05 mm/rev, ap=1 mm and ae=12.5 mm) are presented in Figure 19, where it can be noted that the maximum temperature at the tool-chip interface reached 332° C (Figure 19a) and the temperature at the implanted thermocouple location stretched from 52.3° C to 84.7° C (Figure 19b). These values can be considered quite reliable since under the same cutting condition, the temperature of the tool-chip interface obtained experimentally was 328° C and the temperature recorded by the implanted thermocouple was 78° C, please refer to Figure 8.
Thermocouple hole
a)
b) Figure 19. Milling temperature simulation (vc=90 m/min, fz=0.05 mm/rev, ap=1 mm and ae=12.5 mm) Finally, Figure 20 summarizes the results obtained from the experimental work (using both techniques) and those achieved through numerical simulation for the cutting conditions depicted in Table 2. Experimental and numerical results concerned with the tool-chip thermocouple method are represented in blue, while the results associated with the implanted thermocouple (remote temperature) method are represented in red. A quite satisfactory agreement between experimental and numerical results was obtained, especially when the tool-chip thermocouple was employed. In the case of the implanted thermocouple, however, the experimental temperature value tended to stray from the range stipulated by the isothermal curves where the thermocouple was confined. One possible reason for such behaviour may be changes in heat partition fractions (for chip, insert and workpiece) as the cutting parameters associated with the shear plane area (feed per tooth or axial depth of cut) are increased.
Figure 20. Comparison between experimental and numerical temperature results obtained at distinct cutting conditions
4. CONCLUSIONS
Cutting temperature was assessed during dry milling of annealed AISI D2 cold work tool steel using indexable tungsten carbide inserts under various cutting conditions. Additionally, finite element modelling was performed and the results compared with those obtained experimentally. The following conclusion can be drawn from this work:
A wireless data acquisition system constituted of a temperature acquisition
system, Arduino microcontroller and Bluetooth device was successfully developed to measure milling temperature using the tool-chip and implanted thermocouple methods.
Milling temperature increased with cutting speed, feed per tooth and both axial
and radial depths of cut, nevertheless, the relevance of each factor varied according to the measurement technique: for the tool-chip thermocouple method, cutting speed was by far the most relevant factor, followed by radial depth of cut, feed per tooth and axial depth of cut. However, when the implanted thermocouple method was employed, axial depth of cut was the most influential factor, being followed radial depth of cut, cutting speed and feed per tooth at more regular intervals.
Although the average cutting temperature when up milling had been slightly
higher in comparison with down milling (370.3° C against 360.3° C), the paired t-test indicated that cutting direction did not affect milling temperature measured by the toolchip thermocouple.
The experimental determination of the friction coefficient between tool and
work materials was critical to the successful three dimensional numerical simulation of milling temperature, which was capable of satisfactorily estimating temperature distribution in the cutting zone.
ACKNOWLEDGEMENTS
The authors are grateful to Prof. Elaine C.S. Corrêa (Centro Federal de Educação Tecnológica de Minas Gerais, CEFET-MG) for the support during the calibration procedure. Additional thanks go to the following research agencies in Brazil: CAPES, CNPq and FAPEMIG.
FUNDING
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
REFERENCES
[1] G.M. Pittalà, M. Monno, A new approach to the prediction of temperature of the workpiece of face milling operations of Ti-6Al-4V, Appl. Therm. Eng. 31 (2011) 173180. [2] N.A. Abukhshim, P.T. Mativenga, M.A. Sheikh, Investigation of Heat Partition in High Speed Turning of High Strength Alloy Steel, Int. J. Mach. Tools Manuf. 45 (2005) 1687–1695. [3] X. Cui, J. Zhao, Z. Pei, Analysis of transient average tool temperatures in face milling, Int. Comm. Heat Mass Transfer 39 (2012) 786-791.
[4] M.A. El Hakim, M.A. Shalaby, S.C. Veldhuis, G.K. Dosbaeva, Effect of secondary hardening on cutting forces, cutting temperature, and tool wear in hard turning of high alloy tool steel, Meas. 65 (2015) 233-238. [5] C. Courbon, R. Mabrouki, J. Rech, D. Mazuyer, E. D’Eramo. On the existence of a thermal contact resistance at the tool-chip interface in dry cutting of AISI 1045: formation mechanisms and influence on the cutting process, Appl. Therm. Eng. 50 (2013) 1311-1325. [6] M.B. Silva, J. Wallbank, Cutting temperature: prediction and measurement methods a review, J. Mater. Process. Technol. 88 (1999) 195-202. [7] F. Benabid, H. Benmoussa, M. Arrouf, A thermal modeling to predict and control the cutting temperature, The simulation of face-milling process. Procedia Eng. 74 (2014) 37-42. [8] A. Maurel-Pantel, M. Fontaine, S. Thibaud, J.C. Gelin, 3D FEM simulations of shoulder milling operations on a 304L stainless steel, Simul. Modell. Pract. Theory 2 (2012) 13–27. [9] F. Jiang, L. Yan, Y. Rong, Orthogonal cutting of hardened AISI D2 steel with TiAlN-coated inserts—simulations and experiments, Int. J. Adv. Manuf. Technol. 64 (2013) 1555-1563. [10] A.N.S. Barrios, J.B.C. Silva, A.R. Rodrigues et al., Modeling heat transfer in die milling, Appl. Therm. Eng. 64 (2014) 108-116. [11] G.S. Samy, T. Kumaram, Measurement and analysis of temperature, thrust force and surface roughness in drilling of AA (6351)-B4C composite, Meas. 103 (2017) 1-9. [12] K. Kerrigan, J. Thil, R. Hewison, G.E. O’Donnell, An integrated telemetric thermocouple sensor for process monitoring of CFRP milling operations, Procedia CIRP 1(2012) 449-454. [13] A. Guha, H. Li, Z. Sun, C. Ma, D. Werschmoeller, Xiaochun Li, Wireless acquisition of temperature data from embedded thin film sensors in cutting insert, J. Manuf. Process. 14 (2012) 360-365. [14] G. Le Coz, M. Marinescu, A. Devillez, D. Dudzinski, L. Velnom, Measuring temperature of rotating cutting tools: Application to MQL drilling and dry milling of aerospace alloys, Appl. Therm. Eng. 36 (2012) 434-441.
[15]G. Sutter, A. Faure, A. Molinari, N. Ranc, V. Pina, An experimental technique for the measurement of temperature fields for the orthogonal cutting in high speed machining, Int. J. Mach. Tools Manuf. 43 (2003) 671-678. [16] Y. Sun, J. Sun, J. Li, Q. Xiong, An experimental investigation of the influence of cutting parameters on cutting temperature in milling Ti6Al4V by applying semiartificial thermocouple. Int. J. Adv. Manuf. Technol. 70 (2014) 765-773. [17] M. Kikuchi, The use of cutting temperature to evaluate the machinability of titanium alloys, Acta Biomater. 5 (2009) 770-775. [18] L.C. Brandão, R.T. Coelho, A.R. Rodrigues, Experimental and theoretical study of workpiece temperature when end milling hardened steels using (TiAl)N-coated and PcBN-tipped tools, J. Mater. Process. Technol. 199 (2008) 234–244. [19] C. Liu, W. Ding, T. Yu, C. Yang, Materials removal mechanism in high-speed grinding of particle reinforced titanium matrix composites, Precis. Eng. 51(2018) 68-77. [20] C. Dai, W. Ding, Y. Zhu, J. Xu, H. Yu, Grinding temperature and power consumption in high speed grinding of Inconel 718 nickel-based superalloy with a vitrified CBN wheel, Precis. Eng. 52(2018) 192-200. [21] F. Jiang, L. Yan, Y. Rong, Orthogonal cutting of hardened AISI D2 steel with TiAlN-coated inserts—simulations and experiments, Int. J. Adv. Manuf. Technol. 64(2013) 1555–1563.
Highlights
A wireless data acquisition system was developed to measure milling temperature Tool-chip and implanted thermocouple techniques were used together with FEM The influence of cutting parameters on temperature depends on the selected method Cutting direction does not statistically affect milling temperature Determination of friction coefficient is critical to the success of FE simulation