Cutting force,Vibration,and Temperature in Drilling on a Thermoplastic Material of PEEK

Abstract

Polyetheretherketone (PEEK) is one of the semi-crystalline thermoplastic polymers with excellent machinability and chemical stability applied to precise structural plates and electronic components. This study installed multiple sensors to analyze the machining characteristics in the PEEK drilling. According to the time domain signals, the effects of spindle speed and feed rate on the machining characteristics of cutting force and vibration were investigated. In addition, an infrared thermography was installed to record the temperature variation within the drilling area. The experimental hole was 2-mm diameter with a 4.5-mm depth. Experimental results showed that the effect of the feed rate on thrust force is greater than the spindle speed; drilling by a low-level spindle speed with a low-level feed rate can obtain the smallest cutting force and acceleration amplitude in the spindle axis; the temperature within the drilling area is inverse to the feed rate and a high-level feed rate is helpful for forming regular curl chips. When adequate airflow was applied during the drilling operation, the hole’s shrinkage ratio and roundness can be decreased. The data presented in this paper provide valuable references for realizing the drilling of the thermoplastic—PEEK.

Introduction

Engineering plastics have wide applications in industries as substitutions for traditional metallic materials due to their light weight, moderate strength, easy shaping, and environmental resistance.^1,2 In many cases, engineering plastics are used in making various machine parts because their superior specific strength compared with carbon steels. ³ High-performance engineering plastics commonly used include the crystal ones of polyphenylene sulfide (PPS), liquid crystal polymer (LCP), polyimide (PI), polytetrafluoroethylene (PTFE), polyetheretherketone (PEEK), polyacetal (POM), polyarylate (PAR) and the non-crystal of polyethersulfone (PSF), polyethersulfone (PES), polyetherimide (PEI), polyamide-imide (PAI). In addition to the better mechanical and/or thermal properties, their prices are competitive, very suitable for structural parts in different fields such as automobile, medical appliances, electronics, aerospace, construction machinery, and biomechanical applications.^1,4

Machinability is usually defined by the surface finish and integrity of the machined part, tool life, force and power required, and chip control. ⁵ However, the machinability of plastics depends on its mechanical, thermal, and rheological properties. Especially, the viscoelastic behavior of the plastics makes it difficult to realize the correlations between the cutting conditions and the materials.^1,6 Thermoset polymers generally have brittle behavior and are very sensitive to water, while thermoplastic polymers are generally ductile. ⁷ The heat cannot be quickly conducted to the cutting tool due to their low thermal conductivity, which results in localized heating at the machined surface. ⁸ Selections of tools and cutting parameters have a great effect on the machinability of the polymer materials. ⁹ In the drilling of thermoplastic matrix materials, undersized holes are produced unlike the metal due to the elastic recovery after machining. ⁷ In addition to the parameters of cutting speed and feed rate, the temperature variation also affects the thrust force, surface texture, and hole’s dimensions and quality.

The main problem in twist drilling is the premature tool breakage caused by chips clogging, tool deflection, and tool wear. Several signals have been successfully used in mechanical drilling, such as force,^2,7,10–15 vibration,^10,16–18 temperature,^8,10,19–23 motor power,^15,24 acoustic emission, ²⁵ etc. The selection of sensors depends on the signal or characteristic to be monitored. According to the signals detected from the sensing elements, the machining characteristics can be derived directly or indirectly.

In general, the cutting force is the most sensitive indicator of machining performance including the tool condition and chip formation. ¹⁰ Özden et al. ⁴ indicated that machinability of PEEK and its composites can be improved based on the cutting force induced. An appropriate cutting force during machining results in less tool wear and better dimensional accuracy. The damages created in the drilling of polymeric composites depends on the thrust force and torque induced during the operation, which varies by the cutting speed, feed rate, and tool geometry. ¹¹ From the drilling experiments of graphite-PEEK composite, ⁷ both thrust force and torque increased with increasing feed. While the cutting energy of the cutting tip increases as the feed increases; as a result, the materials may get softened at the elevated cutting temperature. Anand and Patra ¹² indicated the minimum cutting forces and hole quality error were obtained when the feed value is equal to the tool edge radius based on the experiment of carbon fiber reinforced polymer (CFRP). Luo et al. ¹³ believed that excessive cutting forces lead to high temperature and excessive tool wear, also deteriorate the hole quality such as burr, delamination, and interlayer chip in the drilling of CFRP/Ti stacks. Wang et al. ¹⁴ analyzed the effect of pilot holes in the drilling of CFRP composites. The thrust force can be reduced by approximately 55% compared with none of pilot holes in the drilling. Pervaiz and Deiab ¹⁵ reported the effect of peck drilling on aluminum alloy Al-6061. The average thrust force of the peck drilling was 22% lower than that of the conventional drilling. Time domain signals are raw signals collected from the sensors. Harun et al. ¹⁰ adopted the fast Fourier transform to conduct the signals frequency domain analyses of tri-axial forces in the tool steel drilling.

Vibration is another proper parameter to monitor tool conditions. In the drilling operation, chips are produced in a closed and blind area. Uncontrollable vibrations created during the machining process generate a poor drilled surface, shorten tool life, reduce material removal rate, cause early wear and breaking, damage machine elements, and bring much noise.^16,17 Esim and Yildirim ¹⁶ believed that the undesired effect at the beginning of drilling is a vibration problem. They installed an accelerometer on the working table to analyze the vibration in drilling of the steel and aluminum workpiece. Kaplan et al. ¹⁷ indicated that the increases in cutting speed, number of holes drilled, and length of tool overhang make the acceleration amplitude increased in the drilling of tool steels (AISI D2 and D3). Arunkumar et al. ¹⁸ installed an accelerometer on the workpiece for monitoring the vibrations produced during the gun drill operations to decide the regrinding point. Online tool monitoring is a defensive and practical method which aims at improving the hole quality and avoiding drastic change in tool wear and breakage. Harun et al. ¹⁰ showed that vibrations are more proper than the force to be used as standalone for tool condition monitoring.

In the machining of polymer materials, the cutting heat cannot be conducted to the cutting tool quickly. Localized heating happened at the tool tip may cause excessive heating at the machined surface such as burning for thermoset polymers and gumming for thermoplastic polymers. ⁸ The drilling process can be more efficient as the drilling zone is properly cooled and lubricated. This is because friction generates nearly 33% of the heat produced during the process. ¹⁸ Temperature progression during the drilling affects the thrust force, torque, tool wear, surface texture, and hole quality.

Infrared thermography (IR) technique has been applied in the studies on drilling process.^8,19–22 Ramirez et al. ¹⁹ investigated the influences of tool wear on cutting forces and cutting temperatures in the drilling of CFRP. Results showed that the temperature gradually increases with the abrasive wear of the cutting edges and the abrasion is the main wear mechanism. Park et al. ²⁰ indicated if the chips are in poor removal, they may be accumulated and compacted in the hole; therefore, they cause additional heat and hamper heat dispersion. Erturk et al. ²¹ showed that increasing spindle speed or feed rate did not induce a significant temperature rise of the drill bit. While the temperature of the workpiece was slightly increased with increasing spindle speed in the drilling of glass fiber reinforced polymer (GFRP) composites. Uysal ⁸ pointed out cutting temperature increased with the increase in cutting speed in the drilling of carbon black-reinforced polymer composite. However, cutting temperature decreased with increasing feed due to the more heat was transmitted from the drill to the polymer material. Kumaran et al. ²² demonstrated that the increasing spindle speed reduced the thrust force and surface roughness in the drilling of aluminum-based composite; but increased the tool temperature. The highest temperature was measured during the period that the drill and workpiece are in full-contact. Xu et al. ²³ indicated when the high-temperature drill edges attack the PEEK polymers, the glass transition inevitably take place. Then the PEEK polymer becomes thermally softened and gets rubbery. The glass transition temperature (T_g) of PEEK is 143°C. They measured the machining temperature in the carbon/PEEK composite drilling by thermocouples embedded beneath the drill flanks of the inner coolant holes.

Characteristics of hole quality mostly discussed in research include the hole size, circularity or roundness error, burr formation, and surface roughness. ²⁶ The error in hole size is defined as the difference between the drill size and the hole size. Unlike the metal drilling, there are undersized holes produced in the drilling of thermoplastics. It is because that a high springback, compressive, or downward deformation of the workpiece always occurs during the thermoplastic cutting. ⁷ Aamir et al. ²⁶ indicated when the drill reaches the exit side of a hole, some of the material is pushed out by the thrust force without being cut; thus, forming burrs. Pervaiz and Deiab ¹⁵ showed thick chips are difficult to break by the action in the drilling process whereas continuous chips tend to enlarge in the holes and affect the surface finish. Peck drilling is easy to eliminate chips, reduces chip clogging, and has a high degree of controlling the chip removal process.

Thermoplastic material is generally fabricated by a near-net-shape production process, so only some final machining processes are needed such as drilling and edge trimming for the component assembly in the mechanical structures. ⁷ Rubio et al. ² indicated selections of the drilling parameters are very important due to the high ductility and low melting point of engineering polymers. In the drilling of a small hole, a large load cannot be put on the drills due to its low strength and rigidity, and the removal of drilled chips is obstructed owing to the small drill flute area. ³ The thermal softening effects arising from the high machining temperatures particularly when higher cutting speeds are employed in the drilling of PEEK polymer. ²³ As Thiruchitrambalam et al. said in the review literature, ²⁷ PEEK is useful for manufacturing several niche applications such as satellite components, bio implants, and wear resistant components due to its many interesting physical, mechanical, electrical properties. However, few studies on the machining of PEEK thermoplastic and its composites are reported recently.^4,7,23

This paper carried out a series of drilling experiments of the PEEK—MDS 100. Two parameters of the spindle speed (three levels) and the feed rate (six levels) were analyzed here. Tri-axial cutting forces, acceleration amplitude, and temperature were investigated for minimizing them and derived the optimal drilling conditions. Finally, air-cooling experiments were implemented to observe the effect of air-cooling on the hole quality and chips in the thermoplastic drilling.

Experimental details

Material and workpiece

The experimental material was a thermoplastic polymer of PEEK, MDS 100 produced by Quadrant Systems Ltd, UK, which has advantages of a low moisture absorption, good strength and stiffness, easily machined to precise dimensions, and available in thin cross-sections. It is suitable for making the fixtures for electronics testing, and positioning platforms for miniature motion devices, and sockets for the semiconductor industry. Table 1 presents the important properties of MDS 100. Figure 1 shows an experimental plate with the size of 80 × 30 × 4.5 mm³ finished 30 drilling trials. The holes marked with red circle were position holes (2-mm diameter) machined in pairs for setting the drilling area.

Table 1.

Material properties of the PEEK—MDS 100.

Property	Typical average value
Specific Gravity @73oF (22.8oC)	1.51 g/cm3
Tensile Strength	14,700 psi (101.4 MPa)
Elongation, at break	1,500,000 psi (10.3 GPa)
Flexural Strength	1.5%
Flexural Modulus of Elasticity	20,500 psi (141.3 MPa)
Hardness, Rockwell	R121
Coefficient of Linear Thermal Expansion (−40°F to 300°F; −40oC to 148.9oC)	1.1 × 10−5 in/in/ oF (1.98 × 10−5 mm/mm/ oC)

Source: www.mcam.com

Figure 1.

An experimental plate finished 30 drilling trials.

Experimental installation

The experimental machine was a vertical machining center, NXV-560A, made by YCM Ltd, Taiwan. This machine has a positioning precision of 0.01 mm and a repeatability of 0.007 mm. A tungsten carbide twist drill of 2-mm diameter was used, with the point angle of 118^o, made by Sphinx Tools Ltd, Swiss. Figure 2 shows the schematic diagram of the signals detection system. A working base made of bakelite was designed on the machine platform for mounting the force sensor under the spindle axis exactly, followed by an acrylic fixture screwed upon the base of the force sensor for installing the workpiece plate. This study used a force sensor (PCB 261A01) and an accelerometer (PCB 356A45), made by PCB Ltd, USA. The sensitivity of the force sensor of the Z axis is 0.56 mV/N, the broadband resolution is 0.027 N-rms, and its low frequency response is 0.01 Hz. As to the accelerometer, the sensitivity is 10.2 mV/(m/s²), the measuring range is within ±490 m/s² pk, the resonance frequency is greater than 30.000 Hz, and its wideband resolution is 0.005 m/s² rms. Both sensors work by a piezoelectric transducer, which can detect the tri-axial signals in real-time. Their installation is shown in Figure 3. One air-pipe was opened in the experiments. The cooling air blew from the side to the drilling area with an incident angle of 60 degrees. The amount of air was controlled by an adjustable valve and measured by a flow meter. When the valve is fully open, the flow rate is 90 L/min, and that of the half-open is 45 L/min.

Figure 2.

Schematic diagram of the signals detection system.

Figure 3.

Installation of the experimental sensors.

In the experimental operation, the workpiece was clipped upon the acrylic fixture by four clamps and positioned by two pins (inserted on the fixture) coupled with the two position holes on the workpiece as shown in Figure 4. It can ensure that each drilling trial to be operating over the center of the force sensor exactly, refer to Figure 2. The crosshair mark in the figure is the drilling location. The accelerometer was installed on the workpiece side close to the center of the acrylic fixture for monitoring the vibrations during the operation.

Figure 4.

Fixation of the workpiece plate and signal lines.

The force and acceleration signals were collected by two data acquisition cards of NI 9234 and a compact data acquisition (DAQ) of cDAQ-9171. The sampling frequency was 5120 Hz. After the signals transformed from analog to digital, import them into the computer, and then conduct the signal processing procedure. This procedure was executed by a dynamic signal acquisition software of m+p analyzer v. 5.2.1 (made by m+p international, Germany) to acquire the force and acceleration data for making the time domain diagrams and obtaining the information required in analyses.

Drilling parameters and experimental plan

For the drilling process, most studies show that cutting speed and feed rate are the most significant parameters affecting the machining characteristics and hole quality.^7,11 In the drilling of a specific diameter, the cutting speed is determined by the spindle speed. Therefore, the spindle speed and feed rate were designated as the factors analyzed in our experiments. In the drilling of engineering plastic, Endo and Marui ³ indicated that the heating up of the workpiece due to the build-up of swarf on the drill flutes is an obstacle to the drilling of engineering plastics, the feed rate must be small. In the drilling of PEEK, the material may get softened at the elevated cutting temperature produced at a high speed. ⁷ In this study, both drilling parameters of spindle speed and feed rate were investigated for realizing the machining characteristics in the drilling of PEEK.

According to the parameters suggested by the material vendor (www.mcam.com), when the tool diameter is greater than 1/32 in (0.8 mm), the spindle speed should be below 2500 rpm and the feed rate is 0.004 in/rev (0.102 mm/rev). Since a high feed rate results in a larger cross-sectional area of the undeformed chip, the thrust force and torque are significantly increased when the feed is increased. ⁶ Moreover, a smaller feed rate increases the drilling time, and it may increase the friction heat between the tool and workpiece in the blind hole. Therefore, the factor of feed rate was discussed more particularly here. An upscaling increment of the feed rate was schemed based the suggested value of 0.004 in/rev. After several preliminary tests, the spindle speed was designed as a three-level factor, and the feed rate was expanded to a six-level factor with drilling a 2-mm-diametr hole as presented in Table 2.

Table 2.

Factorial levels of the drilling experiments.

Factor	Spindle speedunit: rpm			Feed rateunit: mm/rev (inch/rev)
Level	A1	A2	A3	B1	B2	B3	B4	B5	B6
Value	1750	2000	2250	0.102 (0.004)	0.127 (0.005)	0.152 (0.006)	0.178 (0.007)	0.203 (0.008)	0.229 (0.009)

PS: The depth of drilling cycle was 1 mm.

To observe the machining characteristics objectively, all experiments were operated by dry drilling. In the drilling operation, the drill is moving forward to the specific depth and then retreating to the start point above the workpiece surface for removing the chips and reducing the impact of cutting heat on the hole’s characteristics. This specific depth is set as the depth of the drilling cycle; and the value was 1 mm in our experiments, half of the tool diameter. Eighteen drilling trials were performed by a full-factorial scheme, and each trial implemented three replicas for replicate tests. As a result, a total of 54 drilling trials (3 × 6 × 3) were obtained. During the drilling operation, the force, acceleration, and temperature signals were detected in real-time for the following investigations.

Experimental results and discussion

Analysis of the cutting force

Thrust force (F_z) is the perpendicular force to the workpiece surface which is required to keep the drill in the workpiece during its translational motion. Cutting force components generated in the radial directions of X and Y are relatively small in comparison with the F_z in the drilling process. ²² For this study, the experimental hole was machined by a 2-mm-diameter twist drill and the depth of drilling was 1 mm. Since the thickness of the workpiece plate was 4.5 mm, the number of drilling cycles to finish a hole was 6 due to the tool geometry. Figure 5(a) shows the time domain diagram of the tri-axial cutting forces of Trial-1-1. Trial-1-1 indicates the first replica of the factorial level set of Trial-1. From the thrust force signals of F_z (draw by the blue in the figure), several points are found as follows:

Figure 5.

Time domain diagram of the cutting forces in Trial-1-1: (a) tri-axial components, F_z shows by the blue, (b) component of F_x, and (c) component of F_y.

• During the first drilling cycle, while the drill touched the workpiece surface, the thrust force F_z was rising to the peak immediately. As the increasing of the drilling depth, the contact area between the tool and workpiece was increased gradually. It caused the thrust force to decrease continuously in a period of time. When the tool reached the specified depth, the tool was returning, and the thrust force relieved at once. The maximum thrust force peak was 12.48 N in this cycle.

• During the second to the fourth cycles, when the tool re-touched with the workpiece surface, the drill flanks were almost in full-contact. The thrust force ascended to the peak in a very short time and descended continuously, while they were more stable than the first cycle. The peaks of the three cycles were 13.42, 12.63, and 12.30 N, respectively.

• When the process proceeded to the fifth and the sixth cycle, the drill pierced through the bottom surface. The contacts between the drill flanks and work piece become a ring-type, and the thrust force peaks were obviously lower than the formers. The thrust force peaks dropped to 11.60 and 8.47 N, respectively. According to the drilling conditions in Trial-1, the time for finishing this job was 4.4 seconds.

Figures 5(b) and 5(c) show the time domain diagrams of the radial components of F_x and F_y, respectively. From these force signals, we find:

During the first drilling cycle, both radial force components were slightly unstable. When the process entered full-contact regions (the 2nd to 4th cycles), the radial components became stable. The variation of the component Fx was ranged from −1.133 to 1.198 N and that of Fy was from −1.188 to 1.152 N.

During the last two cycles, the tool had pierced through the workpiece bottom, the contacts between the drill flanks and workpiece were decreased continuously, but the force variations of radial components were increased contrarily. The range of Fx was from −1.037 to 1.431 N and that of Fy was from −1.157 to 1.434 N.

As the statuses of cutting force above mentioned, the last two cycles that the tool pierced through the workpiece bottom should be excluded in the force analysis. Table 3 presents the data of the cutting force components calculated by the 1st–4th drilling cycles. The factorial level set of Trial-1 presented the smallest thrust force (F_z). On the contrary, Trial-18 was the largest one.

Table 3.

Experimental data of the tri-axial cutting force components.

			Avg. of the max. thrust force peaks—Fz (N)				Avg. of the max. radial force peaks—Fx (N)				Avg. of the max. radial force peaks—Fy (N)
Trial	A	B	Fz-1	Fz-2	Fz-3	Mean	Fx-1	Fx-2	Fx-3	Mean	Fy-1	Fy-2	Fy-3	Mean
Trial-1	1	1	12.71	13.35	13.76	13.273	1.18	1.27	1.42	1.290	1.17	1.20	1.26	1.212
Trial-2	1	2	15.94	15.63	15.78	15.781	1.31	1.33	1.20	1.282	1.26	1.27	1.20	1.246
Trial-3	1	3	19.78	19.62	18.28	19.225	1.31	1.30	1.33	1.313	1.25	1.26	1.28	1.265
Trial-4	1	4	22.94	23.93	23.69	23.521	1.19	1.42	1.18	1.261	1.31	1.40	1.21	1.305
Trial-5	1	5	29.09	28.80	28.73	28.874	1.22	1.35	1.32	1.295	1.12	1.22	1.15	1.159
Trial-6	1	6	33.60	33.84	34.07	33.836	1.14	1.12	1.19	1.150	1.05	1.19	1.18	1.137
Trial-7	2	1	13.27	13.31	13.43	13.334	1.31	1.21	1.16	1.230	1.26	1.21	1.16	1.211
Trial-8	2	2	17.01	17.06	16.77	16.945	1.29	1.32	1.27	1.297	1.20	1.26	1.22	1.227
Trial-9	2	3	21.15	20.91	20.44	20.833	1.43	1.40	1.29	1.377	1.21	1.33	1.26	1.266
Trial-10	2	4	24.75	24.95	24.99	24.895	1.44	1.41	1.49	1.445	1.35	1.43	1.23	1.337
Trial-11	2	5	29.49	29.57	29.50	29.520	1.34	1.34	1.30	1.325	1.12	1.24	1.20	1.186
Trial-12	2	6	33.59	33.39	33.59	33.524	1.28	1.46	1.44	1.395	1.19	1.34	1.10	1.212
Trial-13	3	1	13.98	13.75	13.53	13.753	1.28	1.17	1.30	1.248	1.25	1.21	1.25	1.238
Trial-14	3	2	17.03	17.39	17.01	17.146	1.35	1.41	1.18	1.313	1.22	1.21	1.20	1.212
Trial-15	3	3	21.36	21.16	21.35	21.292	1.36	1.35	1.45	1.389	1.29	1.27	1.23	1.264
Trial-16	3	4	25.33	25.16	25.18	25.226	1.36	1.37	1.36	1.362	1.28	1.17	1.25	1.235
Trial-17	3	5	29.28	29.01	28.99	29.094	1.28	1.28	1.40	1.319	1.19	1.31	1.22	1.241
Trial-18	3	6	33.89	33.78	33.69	33.786	1.20	1.18	1.08	1.156	1.21	1.14	1.14	1.162

Figure 6 shows the time domain diagram of the tri-axial cutting forces of Trial-18-1. To compare them with the trial of Trial-1-1 (see Figure 5), both cases present a similar trend in the thrust force (F_z), but the mean of the averages of maximum thrust force peaks was increased from 13.273 N to 33.786 N. The whole cutting time was shortened to 2.2 seconds only. In terms of the radial forces, both were slightly decreased as shown in Figures 6(b) and 6(c). The mean of the component F_x was 1.156 N and that of F_y was 1.162 N.

Figure 6.

Time domain diagrams of the cutting forces in Trial-18-1: (a) tri-axial components, F_z shows by the blue, (b) component of F_x, and (c) component of F_y.

For a drilling process, the principal cutting force is the thrust force (F_z), that is, the component along the spindle axis. The factorial level effects of the thrust force are shown in Figure 7(a), that the factorial effect of the feed rate (Factor B) was obviously greater than the spindle speed (Factor A). Nevertheless, for the radial force components (F_x and F_y), their effects of the feed rate were far smaller than the thrust force (F_z), and both presented an approximate scale, refer to Figures 7(b) and 7(c).

Figure 7.

Factorial level effects of the cutting force components: (a) F_z, (b) F_x, and (c) F_y.

Drilling by a low-level feed rate brings the tool sustained a smaller thrust force. For the case of Trial-1, drilled by the feed rate of 0.102 mm/rev (B1), the ratio of the thrust force (F_z) to the radial force (F_x) was about 10:1 (13.273:1.290). On the other hand, increasing the feed rate rises the material removal rate and reduces the cycle time, but the cutting force will be increased obviously. For the case of Trial-6, drilled by the feed rate of 0.229 mm/rev (B6) with the same spindle speed as Trial-1 (A1, 1750 rpm), the ratio of F_z to F_x was increased to 29:1 (33.836: 1.150), referred Table 3.

Analysis of the vibration

Factors induced cutting vibrations are plenty such as the tool lifting, platform movement, machinery vibration, tool rounding, etc. Hence, the noise and disturbance signals are easy to be mistaken as acceleration signals. To observe the signals distinctly and realize the machining status, a signal preprocessing was executed by the Butterworth low-pass filter before the signals analysis. Considering the spindle speed and the number of tool edges, the filtering frequency was set as 90 Hz here. After the filtering process, the time domain diagrams of the tri-axial accelerations were obtained.

Before the actual drilling experiments, a dummy drilling operation was conducted (no drill on the tool chuck) by using the drilling parameters of Trial-1 to monitor the acceleration variation along the three orthogonal axes. The vibration induced from the drilling operation can be estimated by the amplitudes of acceleration between the actual and dummy drilling. Figure 8 shows the time domain diagram of the axial acceleration component (A_z) in the dummy operation. The amplitude was varied by about ± 40 × 10⁻³ m/s².

Figure 8.

Time domain diagram of the axial acceleration (A_z) in the dummy drilling operation.

Figure 9 shows the time domain diagram of the acceleration of Trial-1-1 (the trial with the smallest F_z in experiments). To discriminate the vibration signals between the tool feeding and retreating, the thrust force signals were imported to the diagram for overlay analysis. During the period of tool retreating, the amplitudes of acceleration detected were higher than the tool feeding. For instance, during the first drilling cycle, the amplitude of acceleration component A_z during the tool feeding was within ± 20 × 10⁻³ m/s²; however, the maximum amplitude during the tool retreating was high to ± 80 × 10⁻³ m/s² as shown in Figure 9(a). This indicates that the cutting action during the feeding period inhibits the acceleration signals produced by the operating system. As to the radial components of A_x and A_y, they had the same trend as the A_z as shown in Figures 9(b) and 9(c). The acceleration amplitude of the long-axis of the machine platform (A_x) was obviously greater than that of the A_y and A_z. Five high-response regions were presented based on the acceleration signals during the six drilling cycles.

Figure 9.

Overlay analyses on the acceleration and thrust force signals in Trial-1-1: (a) A_z to F_z, (b) A_x to F_z, and (c) A_y to F_z.

In terms of the factorial level effects, the maximum acceleration peak of each drilling cycle during the tool feeding was first examined. Both the first initial cycle and the last piercing cycle were excluded. The average of the maximum peaks of each drilling trial was calculated, then the mean of the tri-axial accelerations was obtained as presented in Table 4. Figure 10 shows the factorial level effects of the tri-axial accelerations of the operating system. The level effect of the feed rate was greater than the spindle speed, especially in the direction of the long platform (X axis) as shown in Figure 10(b). The optimal factorial level set of the three acceleration components was identical to Trial-1, that is, the use of a low-level spindle speed (A1) with a low-level feed rate (B1) induced the smallest acceleration amplitude during the operating system.

Table 4.

Experimental data of the tri-axial acceleration components.

Trial	A	B	Avg. of the max. acceleration peaksAz (m/s2)				Avg. of the max. acceleration peaksAx (m/s2)				Avg. of the max. acceleration peaksAy (m/s2)
			Az-1	Az-2	Az-3	Mean	Ax-1	Ax-2	Ax-3	Mean	Ay-1	Ay-2	Ay-3	Mean
Trial-1	1	1	0.041	0.045	0.047	0.045	0.096	0.086	0.110	0.097	0.073	0.072	0.081	0.075
Trial-2	1	2	0.054	0.058	0.062	0.058	0.102	0.104	0.099	0.102	0.080	0.074	0.072	0.075
Trial-3	1	3	0.053	0.059	0.060	0.057	0.122	0.105	0.127	0.118	0.072	0.070	0.084	0.075
Trial-4	1	4	0.063	0.064	0.062	0.063	0.128	0.132	0.129	0.130	0.085	0.087	0.076	0.083
Trial-5	1	5	0.057	0.071	0.061	0.063	0.128	0.141	0.132	0.134	0.081	0.088	0.081	0.083
Trial-6	1	6	0.070	0.075	0.069	0.071	0.160	0.160	0.132	0.150	0.090	0.083	0.078	0.084
Trial-7	2	1	0.040	0.055	0.036	0.044	0.089	0.103	0.107	0.100	0.073	0.081	0.080	0.078
Trial-8	2	2	0.052	0.055	0.061	0.056	0.105	0.103	0.126	0.111	0.082	0.075	0.078	0.078
Trial-9	2	3	0.063	0.058	0.062	0.061	0.125	0.129	0.138	0.131	0.081	0.083	0.082	0.082
Trial-10	2	4	0.083	0.074	0.064	0.074	0.147	0.152	0.150	0.149	0.094	0.088	0.083	0.088
Trial-11	2	5	0.072	0.065	0.065	0.067	0.176	0.127	0.140	0.148	0.091	0.089	0.093	0.091
Trial-12	2	6	0.080	0.076	0.069	0.075	0.184	0.176	0.129	0.163	0.095	0.092	0.083	0.090
Trial-13	3	1	0.047	0.055	0.053	0.052	0.104	0.101	0.106	0.104	0.073	0.075	0.074	0.074
Trial-14	3	2	0.055	0.057	0.058	0.057	0.131	0.120	0.123	0.125	0.074	0.070	0.082	0.075
Trial-15	3	3	0.074	0.064	0.071	0.069	0.139	0.139	0.147	0.142	0.083	0.082	0.076	0.080
Trial-16	3	4	0.076	0.070	0.062	0.069	0.165	0.167	0.139	0.157	0.089	0.084	0.079	0.084
Trial-17	3	5	0.077	0.086	0.072	0.078	0.183	0.183	0.189	0.185	0.090	0.085	0.088	0.088
Trial-18	3	6	0.072	0.076	0.072	0.074	0.182	0.182	0.175	0.180	0.092	0.087	0.089	0.089

Figure 10.

Factorial level effects of the acceleration components: (a) A_z, (b) A_x, and (c) A_y.

Temperature within the drilling area

Cutting heat produced with varied drilling conditions is very different, which affects the quality characteristics of the hole drilled, especially for the plastic material. For the drilling of polymers, ²⁰ when a higher feed for a given cutting speed was applied, more cutting heat was transmitted from the drill bit to the material, then the cutting temperature was decreased. An IR of InfReC R300SR made by Avoi, Japan, was used in this study to monitor the temperature variation within the drilling area during the process. The sensitivity of this device is 0.03^oC at 30^oC and the accuracy is ±1^oC. Figure 11 shows a temperature image captured. At this moment, the highest temperature was 76.0^oC occurred at the workpiece surface near the drill. Averages of the highest temperature measured from the 54 trials of 18 different drilling conditions are presented in Figure 12. Figure 13 shows the factorial level effects of the drilling temperature. The effect of feed rate (Factor B) was slightly greater than the spindle speed (Factor A), and the temperature almost decreased linearly as the increasing of the feed rate.

Figure 11.

Temperature image captured by the IR camera.

Figure 12.

Averages of the highest temperature within the drilling area of the 18 drilling conditions.

Figure 13.

Factorial level effects of the drilling temperature.

Air-cooling drilling experiment

Thermoplastics generally have low thermal conductivity and elastic modulus, and they are thermally softening. External cooling of the cutting zone may be necessary to keep the chips from becoming gummy and sticking to the tools. ⁹ In order to realize the effect of air-cooling in the PEEK drilling, this study implemented air-cooling experiments by the drilling condition of Trial-6 (A1: 1750 rpm and B6: 0.229 mm/rev) which presented the lowest temperature in the former experiments (refer to Figure 12). Three air conditions were set in the experiment including the full-air of 90 L/min, half-air of 45 L/min, and none-air, respectively. Each one also performed three plicas for replicate tests. The temperature progression was recorded by the IR with a sampling rate of 60 Hz to obtain the highest temperature during the drilling operation.

Figure 14 presents the variations of the highest temperature within the drilling area in the air-cooling experiments. The dashed vertical lines indicate the start time of each drilling cycle, and their corresponding locations of the drill are presented on the lateral axis. In the two cases of full-air and half-air, the temperature variations responded to the sequence of the drilling cycles clearly, that the temperature peaks were ascending with the increasing of drilling depth during the 1st to 4th cycles. The highest temperatures were detected in the fourth cycle. Since the drill had pierced through the bottom surface in the fifth and sixth cycles, the temperatures rose in a short period and dropped presently, and then the work was finished. As to the case of none-air, the temperature was obviously higher than the two air-cooling cases, and the highest temperature occurred during the fifth cycle before the drill throughout the workpiece bottom.

Figure 14.

Temperature variations within the drilling area in the air-cooling experiments.

On the whole, the drilling temperature of the full-air case was the lowest as the red curve shown in Figure 14. Comparing the cases of the full-air to the none-air, their highest temperatures were 57.5^oC and 72^oC, respectively. The air-cooling effect in this experiment achieved a temperature reduction of 14.5^oC.

Chips observation

All experiments in this study were operated by dry drilling. The chips of each drilling trial were collected separately for observing the influences of the tool feeds on the chips produced. Figure 15 shows the chips in the drilling trials from Trial-1 to Trial-6 which were drilled by an identical spindle speed of 1750 rpm with an upscaling feed rate of 0.0254 mm/rev. The chips of Trial-1 were thinner due to the lowest feed rate of 0.102 mm/rev (B1) and presented in an irregular shape as shown in Figure 15(a). Removals of these chips were uneven and might rub the hole surface, and the temperature was the highest among the six trials (87.30°C, see Figure 12). When the feed rate increased to 0.178 mm/rev (B4), the spiral chips started forming as shown in Figure 15(d). The temperature within the drilling area was dropped to 84.27°C (refer to Figure 12). In the case of Trial-6, the feed rate was increased to 0.229 mm/rev, the chips were in a regular curl shape as shown in Figure 15(f). These chips could be easily removed along with the drill flutes and the cutting heat would be relieved fluently. As a result, the highest temperature detected was 80.87 °C, comparing with Trial-1 the temperature dropped of 6.43°C. According to the temperatures presented in Figure 12, for the six trials drilled by the high-level spindle speed of 2250 rpm (A3), the temperature dropped was more enlarged to 15.23 °C. Thus, increasing the feed rate can help the generation of regular curl chips, and the removal of chips affects the drilling temperature obviously.

Figure 15.

Chips observations: (a) Trial-1, B1—0.102 mm/rev, (b) Trial-2, B2—0.127 mm/rev, (c) Trial-3, B3—0.152 mm/rev, (d) Trial-4, B4—0.178 mm/rev, (e) Trial-5, B5—0.203 mm/rev, and (f) Trial-6, B6—0.229 mm/rev.

Quality characteristics of the hole drilled

Because of the elastic recovery of the thermoplastic matrix, the compressed layer on the drilled surface might be expanded while the cutting edge passes over the material. Therefore, undersized holes always occur in the plastic drilling process. This study used the image measurement instrument (Vision Hawk 6, made by Baty, UK) to measure the characteristics of hole diameter and roundness. Measurement data of the air-cooling experiments are presented in Table 5.

Table 5.

Data of the hole’s characteristics of air-cooling experiments.

Characteristics	Full-air(90 L/min)	Half-air(45 L/min)	None-air	Effect ofair-cooling
Hole diameter measured (mm)	1.993	1.996	1.991	0.005
Shrinkage rate, SR	1.51%	1.00%	1.83%	0.83%
Roundness (µm)	13.00	7.00	16.33	9.33

PS: The data presented are the average value of the three replicate trials.

Air-cooling can reduce the temperature within the drilling area, but the high-pressure air might make the tool unstable. Especially for a slender tool, it is very likely to affect the quality characteristics of the hole drilled. Figure 16 shows the hole images captured from the measurement device under three air-cooling conditions. From these images, some burrs (shadows around the hole’s profile) were observed. The half-air presented the best performance (see Figure 16(b)) and the worst one was the none-air, refer to Figure 16(c).

Figure 16.

Hole images of the air-cooling experiments: (a) full-air, (b) half-air, and (c) none-air.

The shrinkage ratio (SR) of the hole drilled is given by

$SR=\left(1-\frac{A_{mea}}{A_{tar}}\right)\times 100\%,$

where A_mea is the area measured from the hole image, A_tar is the target area of the hole. The case of the half-air presented the best performance in both SR and roundness (see Table 5). Comparing them with the none-air, the cooling effects of the shrinkage rate was 0.83% and the roundness was 9.33 µm.

Conclusions

This study implemented drilling experiments of the engineering plastic PEEK to investigate the effects of spindle speed and feed rate on the cutting force, vibration, and temperature during the operation of a 2-mm-diamter hole. From the experimental results, several points are concluded as follows.

In terms of the cutting force, the effect of the feed rate was more significant than the spindle speed. When the drill entered the full-contact period, the thrust force was decreasing gradually as the increasing of drilling depth; nevertheless, the radial forces was increasing slightly. The ratio of the axial thrust force to the radial force component can be controlled to 10:1.

Because the acceleration amplitudes detected in the tool feeding were larger than that in the retreating, the cutting vibration was suppressed the vibration of the operating system contrarily. The long axis of the working platform (X axis) presented the largest variation, and the use of a low-level spindle speed with a low-level feed rate resulted in the smallest acceleration amplitude in the principal cutting axis (Z axis).

The increasing of feed rate can help for reducing the drilling temperature and bring the regular curl chips. In addition, supplying proper air can lower the temperature within the drilling area. The process becomes stable and obtains a better performance in the hole shrinkage and roundness.

Results of this research are useful in realizing the cutting characteristics in the drilling of thermoplastic PEEK. The heat dissipation for a small hole drilling is more severe, it will be the following topic to be proceeding.