Machining is one of the most practical processes for finishing operations of composite components, allowing high-quality surface and controlled tolerances. The high-precision surface milling of carbon fiber-reinforced plastics (CFRP) is particularly applicable in the assembly of complex components requiring accurate mating surfaces as well as for surface repair or mold finishing. CFRP surface milling is a challenging operation because of the heterogeneity and anisotropy of these materials, which are the source of several types of damage, such as delamination, fiber pullout, and fiber fragmentation. To minimize the machining problems of CFRP milling and improve the surface quality, this research focuses on the effect of multiaxis machining parameters, such as the feed rate, cutting speed, and lead angle, on cutting forces and surface roughness. The results show that the surface roughness and cutting forces increase with the feed rate, whereas their variations are not uniform when changing the cutting speed. Generally, a lower surface roughness was achieved by using a lower cutting feed rate (0.063 mm/rev) and higher cutting speeds (250–500 m/min). It was also found that the cutting forces and surface roughness vary significantly and nonlinearly with the lead angle of the cutting tool with respect to the surface.
In recent years, the use of carbon fiber-reinforced plastics (CFRP) has increased considerably, especially in aerospace industries. Nowadays, many aircraft parts are made of this composite material. For example, approximately 50% of the weight of the Boeing 787 aircraft is made of composite materials, such as carbon/epoxy and graphite/titanium . CFRP composites are widely used for different parts of aircrafts, such as wing boxes, fuselages, ailerons, wings, spoilers, vertical stabilizers, traps, and struts . CFRP materials present many advantages compared to other materials, including higher strength and stiffness, longer fatigue life, low density, and better corrosion and wear resistance. Because of a negative coefficient of thermal expansion along the axis of carbon fibers, carbon-reinforced composites can be patterned to minimize the thermal expansion over a wide range of temperatures. This is very important for aerospace structures .
CFRP components are usually produced to near net-shape, but machining is often required to remove excess material and produce high-quality surfaces with controlled tolerances. In particular, drilling and trimming are extensively used to remove excessive material, to produce cutouts, or holes that are required for the product function, or to assemble components. The high-precision surface milling of CFRP is particularly useful for the assembly of complex components requiring accurate mating surfaces as well as for surface repair and mold finishing. CFRP surface milling is a challenging operation because of the heterogeneous and anisotropic nature of these composites that can cause some damages such as delamination, fiber pullout, fiber fragmentation, burring, fuzzing, or thermally affected matrix, which in turn may affect the surface finish and properties of the material [4, 5]. In addition, these composites are extremely abrasive; consequently, tool wear is one of the major problems encountered in CFRP machining. Poor cutting conditions produce increased specific cutting energies and higher tool temperatures, resulting in higher tool wear rates . Choosing the appropriate conditions, such as feed rate, cutting speed, and lead angle, in the case of multiaxis machining is thus very important.
In recent years, many studies have been carried out to provide a better understanding regarding the effects of cutting conditions in CFRP machining on the quality of machined surfaces.
Davim and Reis  investigated the effects of milling parameters on surface roughness and machining damage; they concluded that surface roughness (Ra) increases with the feed rate and decreases with the cutting speed. It was also found that the feed rate presents the highest statistical and physical influence on surface roughness and delamination factor, respectively. In another study, El-Hofy et al.  investigated the effects of different slotting parameters, such as tool materials (WC & PCD) and the cutting environment (chilled air and dry) on the surface roughness and integrity, using 3D roughness parameters (arithmetical mean height Sa and maximum peak to valley height St). According to the results of their research, the combination of low cutting speeds and high feed rates was recommended in view of improving surface roughness, with the feed rate being a significant factor. The effect of the feed rate on the surface roughness was also found to be significant from a study that was carried out by Chatelain et al. . Sheikh-Ahmad et al.  carried out an experimental study aimed to determine the effects of cutting conditions on machining quality during the edge trimming of CFRP; they demonstrated that the surface roughness and average delamination depth increase with an increase in the feed rate and decrease with an increase in the spindle speed.
Cutting forces are among the important factors in machining, which influence the process stability, part quality, cutting temperature, and tool wearing condition . Colligan and Ramulu  studied the edge trimming of graphite/epoxy with diamond abrasive cutters and demonstrated that cutting forces increase with the material removal rate=V×f×d, where V is the cutting speed, f is the feed rate, and d is the depth of cut.
The experiments of Sreejith et al.  examining the face turning of fiber-reinforced plastics showed that the variations of cutting forces/specific cutting pressure is not uniform over the cutting speed and the moderate cutting speeds (200–300 m/min) are more suited for the machining of CFRP. Zhang  investigated the machining of long fiber-reinforced polymer matrix composites and found that cutting forces became greater when the depth of cut increases. Rusinek  studied the milling process of CFRP and concluded that the cutting force rises with an increase in the feed rate. Wang et al.  studied CFRP milling using a PCD tool and showed that good surface quality and low delamination could be achieved in high-speed milling of CFRP by using PCD tool. These authors found that cutting forces are an important factor for controlling surface roughness; they also observed that the surface roughness tends to increase with the cutting forces up to 250 N followed by a decrease when the cutting forces continue to rise from 250 to 400 N.
The lead angle is the rotation of the tool axis about the cross-feed axis  (Figure 1). This angle has a significant effect on process mechanics and dynamics, which have not been studied in CFRP milling until now. The study of the effect of the lead angle on metal milling has shown that the cutting geometry, mechanics, and dynamics vary drastically and nonlinearly with the lead angle .
Despite all researches that have been carried out to provide a better understanding of the machining of fiber-reinforced polymers, there are still many challenges with CFRP machining. This work presents some experiments that have been carried out on CFRP to study the optimum condition for the multiaxis milling of these materials and investigates the effects of different parameters such as the cutting speed, feed rate, and lead angle on the resulting cutting forces, surface quality, and machining damages.
A set of experiments was carried out to provide a better understanding of the effects of machining parameters on surface quality and cutting forces. A high-performance carbon fiber epoxy prepreg having a 64% fiber volume content was used to produce stacks of 24 plies that were autoclave-cured to obtain composite plates with a final average thickness of approximately 3.5 mm. Quasi-isotopic laminates are an important class of composites and those that are most familiar to aerospace industries. With such laminates, the elastic properties are independent of orientation, and stiffness, compliance, and all engineering constants are almost identical in all directions [1, 18]. The symmetric stacking sequence [90/-45/45/0/(±45)2/0/-45/45/90]s of the plies was such as to provide a laminate with in-plane quasi-isotropic properties (Figure 2). This layup is balanced and symmetric, and as a result, extension/bending coupling (Bij) and shear coupling stiffnesses (Ais) are zero; because of the fine ply distribution, the torsion coupling (Dis) is relatively low (i, j=x, y, s; subscript s denotes shear stress in the x-y plane and subscripts x and y denote normal strains in the x- and y-directions, respectively). Because of these characteristics, warpage and unexpected distortion are avoided and interlaminar stresses reduce .
The experiments were carried out using a Huron K2X8 five-axis CNC machine with a maximum spindle speed of 24,000 rpm under different cutting speeds, feed rates, and lead angles under dry cutting condition while keeping the axial depth of cut and radial depth of cut (or width of cut: distance between milling passes) constant and equal to 1.4 and 0.71 mm, respectively.
The cutting mode was up-milling with a 3/8″ diameter ball (LMT. ONSRUD, Waukegan, USA) end mill having two flutes with polycrystalline diamond (PCD) brazed inserts (Figure 3). Table 1 details the tool geometry.
Different cutting conditions were studied including the cutting speed (100–500 m/min), the feed rate (0.063–0.254 mm/rev), and the lead angle (-10° to +10°), as can be seen in Table 2. In this table, the cutting speed levels are calculated from the tool shank diameter. Each experimental run was repeated three times, with the same conditions, to evaluate the repeatability of the experiments. A Kistler 9255B (#3) three-axis dynamometer table (Kistler Group, Winterthur, Switzerland), connected to charge amplifiers, type Kistler 5010, was used for measuring the cutting forces during machining. The experimental setup is shown in Figure 1.
|Tool material||Number of flouts||Shank diameter||Flute length||Overall length||Helix angle||Rake angle||Overall length|
|PCD brazed inserts||2″||3/8″||1/2″||4″||0||24°||4″|
|Cutting speed (m/min)||Spindle speed (rpm)||Feed rate (mm/rev)||Lead angle (°)|
Commercially available dynamometers typically specify a bandwidth below the first natural frequency of the dynamometer structure . The Kistler 9255B dynamometer table has a nominal natural frequency (fn) equal to 2 kHz in x- and y-directions and 3.3 kHz in the z- direction . Any machining operations reaching this range may lead to cutting force signals that are distorted because of the influence of the dynamic behavior of the dynamometer. Thus, the determination of the passing bandwidth is a very important step for an accurate force measurement during milling with high cutting speed. Zaghbani et al.  studied the dynamometer behavior of the Kistler 9255B (3#) dynamometer table with the same setup and calibration method as the ones used in the present study; they showed that the cutting force measurement setup has a passing bandwidth <1 kHz in the z-direction and 2 kHz in x-and y-directions. In the case of the z-direction (lower passing bandwidth), the highest tooth passing frequency should therefore not be higher than 1 kHz, which corresponds to a spindle speed of 450 Hz for a two-tooth cutter (27,000 rpm). In this study, all spindle speeds were lower than 27,000 rpm (800 m/min for a tool with 3/8″ diameter), according to Table 2.
The roughness of the machined surfaces was measured using a Mitutoyo SJ400 contact profilometer (Mitutoyo Corporation, Tokyo, Japan) (Figure 4). Three readings were taken for each surface over an evaluation length of 12.5 mm, at regular intervals in a transverse direction to the cutting (feed direction), and their average was calculated. The measured values of Ra (arithmetic average height) and Rt (total height of the roughness profile) in different cutting conditions were compared to investigate the effect of cutting conditions on the surface quality. Table 3 indicates the average of measured resultant cutting forces and surface roughness for three times repetition of each condition.
The surfaces were also examined using a Keyence VHC-500F-type digital microscope (Keyence Corporation, Osaka, Japan) as well as Hitachi S-3600N electronic microscope (Hitachi Science Systems Ltd, Tokyo, Japan) [scanning electron microscopy (SEM)].
|Test no.||Cutting speed (m/min)||Feed rate (mm/rev)||Lead angle (°)||Fc (N)||Ra (μm)|
Surface morphology and integrity depend on the machining process and workpiece characteristics such as the cutting speed, the feed rate, the fiber type and volume content, the fiber orientation, and the matrix type .
Figures 5 and 6 show the effects of the feed rate and cutting speed on the average surface roughness (Ra) and total roughness (Rt), respectively. When comparing both figures, it is obvious that the variations of Rt and Ra with the cutting speed follow the same trends. All roughness results will therefore be discussed for Ra values alone. As can be seen, Ra increases with an increase in the feed rate. The dependence of the surface roughness on the cutting speed is more complex. However, it could generally be concluded that, for lower cutting speeds (100 and 175 m/min), the surface roughness decreases by increasing the cutting speed.
Increasing the cutting speed to more than 250 m/min does not have a significant effect on the surface roughness for lower feed rates (0.063 and 0.158 mm/rev). The minimum surface roughness values were achieved with a low feed rate (0.063 mm/rev) and higher cutting speed (250–500 m/min). For a higher feed rate (0.254 mm/rev), the roughness diagram has a minimum point at 175 m/min and a maximum point (point 2 in Figure 7) at 375 m/min cutting speed. Increasing the feed rate and cutting speed increases the cutting temperature , which can lead to the softening and burning of the matrix material . Therefore, decreasing the surface roughness for higher feed rates (0.158 and 0.254 mm/rev) at a 500 m/min cutting speed might be explained by the adhering of the uncut fibers to the softened matrix under high cutting temperatures.
According to the literature, cutting forces generally increase with an increase in the feed rate, but the dependence of cutting forces on the cutting speed is not uniform for different types of fiber-reinforced plastics . Figure 7 illustrates the effect of the feed rate and cutting speed on the resultant cutting force in our experiments. It can be seen that the cutting force increases with an increase in the feed rate, and there is a greater influence on the cutting force for higher cutting speeds. The variation of cutting forces is not uniform over the cutting speed and can be studied in three cutting speed ranges, including (I) low cutting speeds (100–175 m/min), (II) moderate cutting speeds (175–375 m/min), and (III) high cutting speeds (375–500 m/min). In range I, the effect of cutting speed on resultant cutting force is not significant; in range II, the cutting force rises with the cutting speed; and in range III, the cutting force diminishes when the cutting speed increases. The nonuniform variation of the cutting force to cutting speed is consistent with other studies [3, 14, 21]. The rate of variation of the cutting forces with the cutting speed is related to cutting temperatures. At low cutting speeds, the cutting temperatures are not high enough to soften the polymer matrix, and dry friction predominates. The softening/degrading of the matrix in the cutting zone occurs at a critical speed and causes a reduction in cutting forces . Figure 7 shows that this critical speed is probably reached in range III, where the cutting forces become almost independent of the cutting speed, and the minimum cutting force is achieved in this range.
Among the conditions resulting in lower roughness (feed rate 0.063 mm/rev and cutting speeds 250–500 m/min), point 1 in Figures 5 and 7 has the lowest cutting force, which produces greater process stability and part quality. Therefore, this condition is recommended for the surface machining of CFRP with this cutting tool. A comparison of the surface quality in point 1 with that in point 2 (the point with the highest surface roughness and cutting force) in Figure 8 shows that much damage occurs using a high feed rate.
The study of the effect of the lead angle on the surface roughness showed that it varies nonlinearly with the lead angle and that variation depends on the cutting speed. Figure 9 shows the effect of the lead angle on the surface roughness for different cutting speeds and a feed rate of 0.0635 mm/rev. The minimum Ra is achieved for a lead angle of 5° for low cutting speeds (100 and 175 m/min) and 0° for higher cutting speeds (250–500 m/min). The diagram in Figure 9 illustrates that the variability in roughness curves is higher for negative lead angles. It is shown that the roughness curves for lower cutting speeds (100, 175, and 250 m/min) have high amplitudes compared to those for higher cutting speeds (375 and 500 m/min). In other words, the sensitivity of roughness to the lead angle is higher for the low cutting speeds.
Figure 10 shows the effect of the lead angle on the resultant cutting forces for different cutting speeds and a feed rate of 0.063 mm/rev. As can be seen, the variation of the cutting force with the lead angle is not uniform for all cutting speeds. However, the minimum cutting forces were achieved at the lead angle 0° for low cutting speeds (100 and 175 m/min) and -10° for higher cutting speeds (250, 375, and 500 m/min). Figure 11 shows the SEM images of a machined surface with different lead angles. As can be seen, the best quality surface was achieved with a lead angle equal to 0° and -10°, where the roughness and cutting force are at minimum values, respectively. More damage, such as fiber breakage, fiber de-cohesion, and matrix damage, is observed in the case of 5° lead angle, whereas the roughness and cutting force have maximum values.
In this paper, surface milling experiments were carried out on carbon fiber-reinforced laminates to study the effects of cutting parameters on the cutting force and surface quality and to find the optimum conditions for this operation type using a PCD two-flute ball nose end mill. Based on the presented results, the following conclusions are drawn:
The surface roughness increases with an increase in the feed rate.
At lower cutting speeds (100 and 175 m/min), the surface roughness decreases with an increase in the cutting speed, whereas increasing the cutting speed to more than 250 m/min has no significant effect on the surface roughness for lower feed rates (0.063 and 0.158 mm/rev).
The cutting force increases with the feed rate, but the variation of cutting forces showed no consistent trend over the cutting speed range evaluated. However, the effect of the cutting speed on cutting force is more significant for moderate cutting speed values (175–375 m/min) while improving the cutting force.
The variation of the cutting force and surface roughness with the lead angle is nonlinear, and the minimum values are found at the 250 m/min speed and 0.0635 mm/rev feed rate for lead angles equal to 0° and -10°, respectively. This latter value is unexpected because it is quite an unusual lead angle in multiaxis machining.
Instability in the roughness diagram increases when using a negative lead angle. On the contrary, using a positive lead angle produces higher cutting forces.
This research was funded by the Consortium for Research and Innovation in Aerospace in Quebec (CRIAQ) and its partners the Natural Sciences and Engineering Research Council of Canada (NSERC), MITACS, Bombardier Aerospace, Avior Integrated Products, Delastek, and AV&R Vision & Robotics.
 Daniel IM, Ishai O. Engineering Mechanics of Composite Materials, 2nd ed., Oxford University Press: New York, 2006. Search in Google Scholar
 Gay D, Hoa SV. Composite Materials, 2nd ed., CRC Press: Boca Raton, 2007. Search in Google Scholar
 Sheikh-Ahmad JY. Machining of Polymer Composites, Springer: New York, 2008. Search in Google Scholar
 Ferreira R, Coppini NL, Miranda GWA. J. Mater. Process. Technol. 1999, 92–93, 135–140. Search in Google Scholar
 Wang XM, Zhang LC. Int. J. Mach. Tools Manuf. 2003, 43, 1015–1022. Search in Google Scholar
 Boothroyd G, Knight, WA. Fundamentals of Machining and Machine Tools, 3rd ed., CRC Press: New Delhi, 2006. Search in Google Scholar
 Davim JP, Reis P. J. Mater. Process. Technol. 2005, 160, 160–167. Search in Google Scholar
 El-Hofy MH, Soo SL, Aspinwall DK, Sim WN, Pearson D, Harden P. Proc. Eng. 2011, 19, 94–99. Search in Google Scholar
 Chatelain JF, Zaghbani I, Monier J. World Acad. Sci. Eng. Technol. 2012, 68, 1204–1210. Search in Google Scholar
 Sheikh-Ahmad J, Urban N, Cheraghi H. J. Mater. Manuf. Processes 2012, 27, 802–808. Search in Google Scholar
 Zaghbani I, Chatelain JF, Songmene V, Berube S, Atarsia A. J. Compos. Mater. 2011, 46, 1955–1971. Search in Google Scholar
 Colligan K, Ramulu M. ASME Publ. PED 1999, 121, 648–655. Search in Google Scholar
 Sreejith PS, Krishnamurthy R, Malhotra SK, Narayanasamy K. J. Mater. Process. Technol. 2000, 104, 53–58. Search in Google Scholar
 Zhang LC. J. Mater. Process. Technol. 2009, 209, 4548–4552. Search in Google Scholar
 Rusinek R. Int. J. Non-Linear Mech. 2010, 45, 458–462. Search in Google Scholar
 Wang YG, Yan XP, Chen XG, Sun CY, Liu G. Adv. Mater. Res. 2011, 215, 14–18. Search in Google Scholar
 Ozturk E, Tunc LT, Budak E. Int. J. Mach. Tools Manuf. 2009, 49, 1053–1062. Search in Google Scholar
 Soden PD, Hinton MJ, Kaddour AS. Compos. Sci. Technol. 1998, 58, 1011–1022. Search in Google Scholar
 Burton D, Duncan S, Ziegert JC, Schmitz TL. 19th Annu. Meet. Am. Soc. Precision Eng. (ASPE 2004) 2004, pp. 221–224. Search in Google Scholar
 Kistler Group, Cutting Force Measuring Catalog, Data sheet: 9255B_000-148, 2009. Search in Google Scholar
 Hamedanianpour H, Chatelain JF. 2nd International Conference on Manufacturing Engineering and Process, Vancouver, 2013. Search in Google Scholar
©2016 Walter de Gruyter GmbH, Berlin/Boston
This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.