Experimental study on wear laws and mechanisms of end cutting edge in end-milling of carbon fiber reinforced polymer

End cutting edges of tool determine machining quality of the bottom surface of the slot in end-milling of CFRP. However, due to anisotropy and heterogeneous of CFRP, and also semi-closed characteristic of blind slots, the end cutting edges are more vulnerable to wear than the peripheral edges under the strong abrasive fibers, leading to the poor machining quality of the bottom surface of the slots, and thereby decreases the assembly performance of the components. This paper aims to reveal the wear laws and mechanisms of the end cutting edge which serves under the poor machining conditions in end-milling of CFRP. In order to obtain major wear forms of the end cutting edge, the tool’s structure and the removal characteristics in end-milling of CFRP are analyzed. For acquiring the wear laws accurately, slot-milling experiments are conducted, in which the quantitative characterization of each wear form is proposed. According to the wear laws obtained from the experiments, combining with the associated relationships of each wear form, the wear mechanisms of end cutting edge are revealed, and also, the influences of the tool wear on the surface’s machining quality are got. The results show that there are three wear forms in all including the corner wear, the cutting-edge wear, and the flank wear. Concretely, the corner wear is rapid; the cutting-edge radius is fluctuating and tends to decrease, while the flank face is wearing constantly. The conclusions of this study can provide foundations for further studies on suppression of the tool wear.


Introduction
Carbon fiber reinforced polymer (CFRP) has the advantages of high specific strength and outstanding designability, which promotes its applications in aerospace structures greatly. These CFRP components are complex entities made up of many features such as streamlined and special-shaped surfaces, which can be designed and manufactured directly by the technology combining the structure with function, but in order to meet the accuracy requirements and the assembly needs, milling processes including edge milling processes and slot milling processes are commonly used [1][2][3]. However, due to the anisotropy and the poor machinability of CFRP, the tools are vulnerable to wear severely in these milling processes, causing the declined machining quality of the surface, which not only decreases the assembly performance of the components, but also prolongs the auxiliary time and thereby increases the production costs. Hence, finding the effective way for suppressing the tool wear in milling of CFRP has been a hotspot in recent years.
In the milling process of CFRP, the main origin for the rapid tool wear is the severe friction imposed by the strong abrasive fibers and high adhesive resin. The blunt tool after wear will impose a smaller contact stress on the fibers, which leads to the ineffective cutting of the fibers, causing the machining damage such as burrs and fibers pull-out [4,5]. Particularly in the milling process of the blind slots, because the fibers and resin are cut in semi-closed space, the heat generated is more difficult to give out and thereby the resin is more likely to soften, leading to the more severe tool wear and the worse surface quality. Since the qualities of the bottom surface of these blind slots are crucial to the assembly performance of the CFRP components, it is difficult to ensure the reliable service of the components without the deeply reveal on the wear laws and mechanisms of the tool [6][7][8][9].
Concerning this issue, some researches have been conducted in recent years. Voss et al. [10] explored the effects of tool geometry on tool wear in edge trimming of CFRP. Based on the characterization method of cutting edge rounding and flank wear, a tool wear model considering the tool structure, cutting distance, and process parameters was established, and the accuracy of the model was verified by experimental methods. Nor et al. [11] revealed the tool wear mechanism in dry milling of multidirectional CFRP laminate. It was shown that the wear of the tool's corner is indicated by the polished area under the abrasion of powdery chips and fibers. Han et al. [12,13] investigated tool wear patterns and mechanisms about coated multi-tooth tool and double-helix tool based on milling experiments on carbon fiber composites. In their work, they found that the main failure modes of coated cutters included the breakthrough of coating, the coating flaking, and the abrasive wear of substrate. For the reason that the double-helix tool is easier to gather and store chips, the wear of double-helix tool is more serious comparing with the multi-tooth tool. Hamedanianpour and Chatelain [5] investigated the variation of tool wear rate under different machining parameters. Based on microscopic observations, the correspondence between surface defects and tool wear was established. It was shown that surface defect rate is increasing with the tool wear. These studies are mainly in terms of the wear laws and mechanisms of the peripheral edges of the tool. Since the bottom surface quality of the slot is mainly determined by the end cutting edges of the tool, in order to ensure the assembly performance of the CFRP components, a study on the wear laws and mechanisms of end cutting edge which serves in the worse machining conditions in milling of CFRP should be conducted.
The objective of this paper is to study the wear laws and mechanisms of the end cutting edge in end-milling of CFRP. The multi-tooth tool, which can improve the milling quality of CFRP and thereby widely applied in the actual production, is selected as research subject in this study. According to its structure, combining with the removal characteristics in end-milling of CFRP, the major wear forms of the end cutting edge are analyzed. Through the slot-milling experiments in which the quantitative characterization of each wear form is proposed, the changing laws of the wear extent of each form with the increase of the cutting distance are summarized. By combining the results with the associated relationships of each wear form, the wear mechanisms of end cutting edge in end-milling of CFRP are revealed. In addition, by comparing the wear extent of the end-edge and the roughness of the machined surface, the influences of the tool wear on the machining quality are got. The conclusions drawn from this study can provide foundations for optimal designing of the tool.
The rest of this paper is organized as follows. "Section 2" analyzes the major wear forms of the end edge based on the structure of the multi-tooth tool and the removal characteristics in end-milling of CFRP. "Section 3" reveals the wear laws and mechanisms of the end edge through the endmilling experiments. "Section 4" analyzes the influences of the tool wear of the end cutting edge on the surface machining quality. The conclusions are summarized in "Section 5."

Analysis on wear forms of end cutting edge
Since the small cutting width and small cutting depth can effectively decrease the machining damage and improve the machining quality of CFRP [3], in the practical production the multi-tooth tools, which can divide the whole cutting width and cutting depth into several parts by the multiple cutting edges, are often adopted in the finish milling of CFRP. In particular, in the milling process of the blind slot, since the requirement of the surface quality of the bottom surface is rather high, the multi-tooth tools with end cutting edges are recommended, whose structure is shown in Fig. 1.
Considering the space in the end-milling process of the blind slot is semi-closed, in order to give out the cutting heat more effectively, the end cutting edge of the tool is often designed as the structure with a minor cutting edge angle and chip pockets. The minor cutting edge angle leads to the incline of the cutting edge of the end cutting edge relative to the bottom surface, which cause that only a small part of the end cutting edge near the tool's corner involve in cutting process before the wear of the tool. With the process of cutting, the inclined end cutting edge gradually abrades into a flat one, so the length of the end cutting edge involved in cutting process is increasing continuously, as the schematic diagram shows in Figs First, due to the minor cutting edge angle, the tool's corner is formed at the junction of the side cutting edge and the end cutting edge. This structure causes that only a small part of the end cutting edge near the tool's corner involve in cutting process before the wear of the tool. Since the contact area of the tool and the surface is rather small at this time, the corner wears rapidly and thereby turns into arc shape, as shown in Figs. 3b and 4b. This kind of wear of end cutting edge is categorized as "corner wear." Second, as the tool's corner wearing, the end cutting edge near the tool's corner is leveling, so the length of the end cutting edge involved in cutting process is increasing continuously. Due to the strong abrasive nature of the carbon fibers, the cutting edge of the end cutting edge will suffer severe friction from the fibers, causing the cutting edge radius increasing. Meanwhile, the particles of the cemented carbide on the cutting edge are vulnerable to fall off under the impact effect of the hard inclusion such as chip, which leads to the cutting edge tipping. This kind of wear of end cutting edge is categorized as "cutting edge wear." Third, at the beginning of the cutting edge wear, the cutting edge radius is increasing, leading to the declined cutting effects on the fibers and thereby the increase number of ineffective-cut fibers. These ineffective-cut fibers will spring back after the tool moving away, which will rub the flank of the end cutting edge and cause the flank wear, as shown in Figs. 5 and 6. But since the flank is wear, the flank angle increases, which in turn decreases the cutting edge radius. In other words, the wear on the flank is beneficial to sharpen the cutting edge to some extent. On the other hand, since the cutting chip is mainly evacuated from the chip pockets, as the high-speed rotating of the tool, the chip is likely to be brought into the narrow space between the flank and the machined surface, as shown in Fig. 6. This phenomenon intensifies the flank wear. In addition, due to the semi-closed space in the end-milling process of the blind slot, the accumulated cutting heat is more likely to soften the resin, which increases the adhesiveness of it and causes the adhesion of  According to the analysis above, in the end-milling process of the CFRP blind slot, the end cutting edge mainly suffers three forms of wear, including the corner wear, the cutting edge wear, and the flank wear. In the next section, the specific wear law of each form of wear will be analyzed and upon which, the mechanisms of the end cutting edge will be revealed. Since the worn tool may influence the machining quality of the bottom surface of the blind slot, which determines the assembly performance of the CFRP components, the concrete influences of the tool wear on the machining quality of the surface will be discussed in details as well.

Wear laws and mechanisms of end cutting edge
In order to obtain the specific wear laws of the end cutting edge, in this section the end-milling experiments of CFRP are conducted. By micro-observing the end cutting edge each time after it cutting for a certain distance, the wear extent of each wear form is quantitative characterized and summarized, so that the wear laws and mechanisms can be obtained.

Experimental setup for end-milling
In the study, T800 multi-directional CFRP workpieces with the stacking-sequence of (− 45/0/45/90)s are used. The K44 cement carbide tool is used whose specific structural parameters are listed in Table 1.
The experiments are conducted on the Mikron HSM500 machining center, as shown in Fig. 7, and the machining parameters are set as listed in Table 2. Each time after the tool    cutting for a certain distance, the end cutting edge is microobserved in order to measure the wear extent, and at the same time the machined surface quality is measured for analyzing the influence of the tool wear which will be discussed in "Section 4." The certain distance is defined as the cutting interval length in this study. In order to accurately acquire the initial wear laws of the tool's corner, the first nine cutting interval distance is set as 25 mm, while the one after that is set as 50 mm.

Quantitative characterization and law analysis on end cutting edge wear
According to the analysis in "Section 2," the minor cutting edge angle leads to the rapid wear of the tool's corner, as the micro-observing image shown in Fig. 8. Since the corner wear directly leads to the increase on the length of the end cutting edge involved in cutting process, which facilitates the wear on the cutting edge and the flank and thereby influences the machining quality of the bottom surface of the blind slots, the effects of the corner wear can be considered as an indirect effect on the surface quality [14]. Hence, in this study, the wear laws of the cutting edge and the flank are emphatically analyzed. Ultra depth-of-field microscope from KEYENCE was used to take pictures of corner wear of end cutting edge and wear extent flank face (for example, Fig. 8), and Alicona was used to measure wear extent of cutting edge (for example, Fig. 9), as shown in Fig. 10.

Wear laws of cutting edge
It can be found from Fig. 11 that the changing trend of the cutting edge radius is not monotone, but presents the changing laws that keeping stable at first (cutting distance from 0 to 50 mm) and then fluctuating (cutting distance further than 50 mm). This changing trend relates to the three wear forms of the end cutting edge analyzed in "Section 2." Namely, at the early stage of tool wear, the main wear form is the corner wear. Since the length of the end cutting edge involved in cutting process is relatively small, the worn part of the cutting edge has not reached the position marked in Fig. 10a, so the change of the cutting edge radius is not obvious [15,16]. On the other hand, the small contact length of the cutting edge on the unmachined surface cause the very rapid wear of the part of edge involved in cutting, which turns the tool's corner into the arc shape, as Fig. 10b shows.
With the wear of the tool's corner, the length of the end cutting edge involved in cutting process is increasing. At this time, the part of end cutting edge involved in cutting will suffer the wear both on the cutting edge and the flank. First, the cutting edge radius will increase rapidly due to the severe friction of the strong abrasive carbon fibers on the cutting edge, causing the blunting of the cutting edge and thereby decline the cutting effect on the fibers. These ineffective-cut fibers will spring back after the tool moving away, which will rub the flank of  the end cutting edge and cause the flank wear. But since the flank is wear, the flank angle increases, which in turn decreases the cutting edge radius. Taken together, in the end-milling process of the CFRP, the cutting edge of the end cutting edge will be under two opposite effects. One is the abrasive effects derived from the fibers, the other one is the sharpening effects derived from the flank wear. Hence, when the cutting edge wears to some extent, namely after the cutting distance reaching 50 mm as shown in Fig. 11, the cutting edge radius will fluctuate with the increase of cutting distance. In addition, according to the analysis in "Section 2," except for the rub from the spring back of the ineffectivecut fibers, the chip which is brought into the narrow space between the flank and the machined surface will also lead to the flank wear. Since the chip is continuously formed in the end-milling process of CFRP, the flank wear induced by the rub of chip is ongoing, so the cutting edge is sharpening all the time during the end milling process [15,17]. Therefore, the cutting edge radius is decreasing overall, as shown in Fig. 11.
On the other hand, with the decrease of the cutting edge radius and the continuous wear on the flank, the stiffness of the cutting edge will decline, which leads to the tipping of cutting edge, as shown in Fig. 12.

Wear laws of flank
In addition to the measurement of the wear extent of the cutting edge, since the initial flanks gully-like striations after the action of the grinding wheel, the worn flank does not have a gully-like streak and shows a bright band. The wear of the flank is also quantitative characterized through the  Results on the changing law of cutting edge radius with cutting distance micro-observation. As Fig. 13 shows, the flank wear width can be calculated through Eq. (1), just: In which, 20 is the real length of the ruler, while 0.0322 is the measuring length of the ruler in the figures. L is the measuring length of the flank wear width. The units of these values in Eq. (1) are all micrometer. By this way, the changing law of the flank wear width with the increase of the cutting length is summarized in Fig. 14.
According to the results in Fig. 14, it is obvious that the overall changing trend of the VB is increasing, which is consistent with the results in "Section 3.2.1" and the analysis in "Section 2." In addition, it can also be found that with the same increasement of the cutting distance, the increasement of the VB varies. Namely, at the early stage of the flank wear, the wear is very rapid, as segment AB shown in Fig. 14. But when the cutting distance exceeds 75 mm, the wear rate of the flank appears to be fast and slow alternant, as segment BC. At the late stage of the wear, the VB nearly keep constant.
The changing laws of VB above can be explained through combining the wear laws of the cutting edge and the two origins of the flank wear. At the early stage of the wear, due to the small contact length of the cutting edge on the unmachined surface, the wear is rapid. With the process of cutting, as the tool's corner becomes blunt, the length of the cutting edge involved in cutting increases, which facilitates the cutting edge wear and the flank wear, as analyzed above. At this time, the cutting edge radius fluctuates significantly which in turn influences the wear rate of the flank. That is, when the cutting edge radius is relatively small, the tool is sharp and the most of the fibers will be cut effectively. In this situation, the flank wear mainly derives from the rub of the chip, so the wear rate is small. However, with the cutting edge radius increasing due to the severe rub from the fibers, the cutting effects on the fibers are declining and thereby the origins of the flank wear are not only the abrasion of the   chip, but also the rub of the spring-back ineffective-cut fibers. In this situation, the wear rate of the flank is relatively high, leading to the sharpening trend of the cutting edge. Hence, the wear rate of the flank appears to be fast and slow alternant.
At the late stage of the tool wear, since the cutting edge radius is rather small according to Fig. 10, the main source of the flank wear is the rub of chip. Because the length of the cutting edge involved in cutting is very large, the contact area between the flank and the chip is large enough that the flank is rubbed evenly and mildly. But on the other hand, the large contact area with the chip facilitates the adhesion of the chip on the flank under the adhesive effect of the softening resin [7,16,18], as shown in Fig. 15.

The effect of tool wear on the surface roughness
Since the surface roughness is a direct index reflecting the machining quality of the surface [19][20][21][22], in this study the surface roughness of the machined surface is measured through the ZYGO interferometer each time after the tool cutting for the set length introduced in "Section 3.1." The sampling length and width are all set as 1.8 mm, and the result of each measurement is summarized in Fig. 16. From Fig. 16, it can be found that with the wear of the end edge, the surface roughness of the bottom surface of the blind slot decreases first, and then fluctuates, finally increases. When the cutting distance is short, the surface roughness is high. That is because at the early stage of the tool wear, due to the exist of the sharp corner of the end edge, the bottom surface of the blind slot is vulnerable to be scratched, forming into pits on the surface, as shown in Fig. 16a. With the increase of the cutting distance, the tool's corner is turning into the arc shape, causing the increase of the length of the cutting edge involved in cutting. That makes the cutting process become smooth, which is beneficial to the decrease of the surface roughness. After the corner wear, according to the analysis in "Section 3.2," due to the continuous wear on the flank, the cutting edge tends to keep sharp, so the surface roughness overall remains small. However, at the late stage of the tool wear, since the cutting edge radius is rather small and the flank is worn seriously, the stiffness of the cutting edge is so poor that it is vulnerable to be smashed like Fig. 12. In this case, the stiffness of the cutting edge will decline further which causes the cutting process unstable. That is the reason why the surface quality will decrease obviously when the cutting distance reaches 500 mm, as shown in Fig. 16d and e.
The surface quality of the workpiece after cutting was analyzed, and Alicona was used to photograph its surface morphology, as shown in Fig. 17. Figure 17a is the surface topography of the workpiece when the tool is in the stable wear section, and its surface roughness is Sa = 1.948. Figure 17b is the surface topography of the workpiece when the tool is in the severe wear section, and its surface roughness is Sa = 2.329. The results show that after tool wear, the surface quality of CFRP decreases due to the decrease of cutting ability. At the same time, the surface features of small damage appear.

Conclusions
In order to reveal the wear laws and mechanisms of the end edge in end-milling of CFRP, in this study the major wear forms of the end edge are analyzed based on the structure of the multi-tooth tool and the removal characteristics in endmilling of CFRP. Through quantitative characterization on each wear form in the actual end-milling experiments, the specific wear laws and mechanisms are revealed. In addition, by comparing the wear extent of the end cutting edge and the roughness of the machined surface, the influences of the tool wear on the machining quality are got. The study leads to the following conclusions: (1) The major wear forms of the end cutting edge in endmilling of CFRP include the corner wear, the cutting edge wear, and the flank wear. (2) The corner wear is rapid. After that, the machined surface quality will improve. (3) The changing trend of the cutting edge radius with the increase of cutting distance is not monotone, but presents the changing laws that keeping stable at first, and then fluctuating. The overall changing trend of the flank wear is increasing, but its wear rate is rapid at first, and then fluctuating, finally keeping stable. (4) With the wear of end cutting edge, the cutting edge radius tends to decrease finally, while the wear extent of the flank is continuously increasing. These phenomena cause the stiffness of the cutting edge decline significantly, which leads to the tipping of the cutting edge. The tipping cutting edge will make the cutting process unstable, causing the increase of the surface roughness.