*3.4. Dimensional Analysis*

In the aeronautical sector, very tight tolerance ranges are applied to the hole diameters [25]. In order to characterize the new drill-milling process, the produced hole diameters were measured with a digital calliper. The inner size of each hole was measured at a half laminate thickness along four directions (0◦, ±45◦, and 90◦). The diameters were calculated as the arithmetic average of the measured values.

### *3.5. Surface Finish*

To characterize the hole surface finish, roughness measurements were carried out using a stylus profilometer with tri-axial movement (Taylor–Hobson Form Talysurf-50, Taylor-Hobson, Leicester, UK) by dragging it along the x-axis a probe with a diamond tip radius of 2 μm. The roughness evaluations were performed through software code Ultra ver. 4.6.8, which allows for viewing and processing the data.

To perform the measurements, the holes were sectioned; this was done every ten holes (number 1-10-20-30-40-50-60). The measurement length was 4 mm, with a probe travel speed of 0.5 mm/s, and the roughness tests were carried out according to standard UNI ISO 4288-2000 [26], a Gaussian filter with a 0.8 mm cut off was selected.

The roughness tests were performed three times for each hole, along three separate lines spaced by 0.05 mm.

The drilling of the CFRP laminates results in a series of micro-fibre fractures, fibre pull-outs, and matrix cracking; for this reason, the roughness parameter Ra (average) alone was not considered sufficient, and roughness parameters Rz (maximum peak to valley height) and RPc (peaks per mm count) were also evaluated [27].

#### **4. Results and Discussion**

#### *4.1. Cutting Forces*

The values and trends of the Fx and Fz cutting force signals were studied to analyse the cutting forces that develop during the drill-milling process. It was observed that Fz is predominant in the first phase (drilling process), while Fx is predominant in the second phase, where it is possible to observe the characteristic sinusoidal pattern of the peripheral milling process. To perform signal smoothing, the moving average method (with a window length of 200 samples, corresponding to 0.02 s) was applied to the Fx and Fz signals, as reported in Figures 5 and 6, respectively. The goal of smoothing is to facilitate the extraction of the maximum values of the force signals, without taking into consideration high-frequency oscillations.

**Figure 5.** Signal of radial force, Fx, in the drill-milling process for tool T2.

The Fx signals during the milling phase show a sinusoidal trajectory, where each sine wave represents a complete revolution of the tool inside the work-material. The size increase of the hole after each revolution is constant. The sine wave period increases, as the peripheral speed of the tool is constant, while the path to complete the whole revolution increases, as the radius becomes greater after each revolution. The volume of material removed also grows after, and during, each revolution as the radius of curvature increases (Figure 7). Because the work-material thickness is constant, the volume of material removed grows proportionally to the area increase, that is, with the square of the radius, according to a non-linear correlation. Moreover, the same considerations allow for stating that the amplitude of the sine wave also increases.

**Figure 6.** Signal of axial force, Fz, in the drill-milling process for tool T2.

**Figure 7.** Scheme of the tool path during the drill-milling process. The removed material area increases at each revolution.

From Figure 7, it can be observed that, after a complete revolution, at the same polar coordinate of the previous revolution, the volume of material that the tool faces at the front cutting-edge is greater than for the previous revolution. Table 2 shows the values of the area of material removed after each revolution, calculated by means of a CAD software (Autocad 2017 version) tool.


**Table 2.** Area of material removed after each revolution.

The growth of the area that the tool must cut leads to an increase of forces during the milling phase up to the eighth revolution, while the last revolution is characterized by a path that leads to the closure of the spiral and to a decrease of the cutting forces.

Then, there are three boring revolutions during which the sinusoidal waves display a decreasing amplitude, as shown in Figure 5, because the tool will follow the same path three times, removing less material at each revolution.

Using the Pearson's correlation coefficient, a strong correlation (>0.7) was found between the area of the material removed and the maximum value of the force in the *X* direction achieved after each spiral revolution. Specifically, a strong correlation was found for both tool T1 (0.86) and tool T2 (0.93).

Figures 8 and 9 show the maximum Fx and Fz cutting force values measured for all of the 60 holes made during the experimental testing campaign with tools T1 and T2. In all of the cases, an increase of both the Fx and Fz cutting forces with an increasing hole number, that is, with tool wear progression, is visible. Moreover, it can be observed that, during the tests with tool T2, the recorded forces in both the *X* and *Z* directions are higher than the ones recorded with tool T1. Using tool T1, the maximum

thrust force, Fzmax, during the drilling phase is 201.27 N and the maximum radial force, Fxmax, during the milling phase is 134.05 N (hole number 60). Using tool T2, the maximum thrust force, Fzmax, during drilling is 188.41 N and the maximum radial force, Fxmax, during milling is 179.57 N (hole number 60). It was verified that the values of Fzmax are quite similar to those found in the literature [8,12] for a traditional drilling process.

**Figure 8.** Maximum cutting force in the *X* direction, Fxmax, measured for tools T1 and T2.

**Figure 9.** Maximum cutting force in the *Z* direction, Fzmax, measured for tools T1 and T2.

#### *4.2. Metrological Analysis*

With regards to the hole size, generally, hole diameter decreases with an increasing number of holes as a result of tool wear. Furthermore, this gradual decrease can be attributed also to the increase of tool deflection due to the growth of radial cutting force [9].

This behaviour is confirmed by the measurements of the hole diameters, which always show a decreasing trend of the measured values for 60 consecutive holes made with the same parameters and tools, as shown in Figure 10. Diameter reduction is due to heavy tool wear during the drilling process, which is also caused by the decision to employ uncoated WC tools. This decision was made to verify the tool geometry and the cutting parameters independently of the tool material.

As shown in Figure 10, up until hole number 48, the tool T1 has a lower size reduction, whereas from hole number 48, both tools behave in a similar way.

**Figure 10.** Hole diameter as a function of the number of holes.

With regards to the roundness of the CFRP holes, Figure 11 shows that the holes made by tool T1 keep their roundness up to hole number 30, and lose their roundness in one direction at hole number 60, whereas the holes made by tool T2 do not display any loss of roundness. Figures 12 and 13 show holes number 1, number 30, and number 60 realized with tools T1 and T2, respectively. No presence of uncut fibers was verified in the holes, even in the case of high tool wear, contrarily to what often happens for traditional drilling [14,25]. This was true for both the T1 and T2 tools, as it can be observed in Figures 12c and 13c.

**Figure 11.** Hole diameter (mm) measured in four directions for holes number 1, number 30, and number 60, made by (**a**) tools T1 and (**b**) tool T2.

**Figure 12.** Pictures of (**a**) hole number 1, (**b**) hole number 30, and (**c**) hole number 60, made by tool T1.

**Figure 13.** Pictures of (**a**) hole number 1, (**b**) hole number 30, and (**c**) hole number 60, made by tool T2.

#### *4.3. Surface Roughness*

Figure 14 reports the Ra values for tools T1 and T2, showing in both cases an increase with the increasing hole number. The holes made with tool T2 have internal surface Ra values greater than those made with tool T1. For the holes made with tool T1, the difference of Ra values between hole number 60 and number 1 is 2.96 μm, whereas for the holes made with tool T2, the difference of Ra values between hole number 60 and number 1 is 4.62 μm. As identical cutting conditions were employed with both tools, the Ra difference can be related to the different wear level of the tool cutting-edge [28,29].

**Figure 14.** Ra measured for every ten holes made by tool T1 and tool T2.

Figure 15 reports the Rz values measured for every ten holes made by both tool T1 and tool T2. This roughness parameter demonstrates that, although the Ra values are slightly different, tool T1 responds with a better cut of the fibres. Taking into consideration hole number 1, the Rz values for the

two tools differ by 13.92 μm, whereas for hole number 60, they differ by 21.07 μm, showing that tool T2 generates a severer fibre pull-out phenomenon.

**Figure 15.** Rz measured for every ten holes made by tool T1 and tool T2.

Figure 16 reports the trend of the RPc parameter measured for every 10 holes. It can be observed that the number of peaks per mm is not high for both tools (only peaks exceeding ±0.5 μm were considered). This roughness parameter indicates that the surface finish of the hole is good, but there are gaps due to the fibre pull-out phenomenon, as observed in the literature on CFRP drilling [30,31].

**Figure 16.** RPc measured for every ten holes made by tool T1 and tool T2.

#### **5. Conclusions**

This study presented an innovative hole-making process for CFRP laminates, where the hole is generated by a combination of drilling and peripheral milling processes. To develop this technique, innovative drill-milling tools were designed and realized, and experimental machining tests on CFRP laminates were performed using two different tool configurations, T1 and T2. During the drill-milling process, the thrust cutting force and the radial cutting force were detected through a sensor system to analyse the cutting force behaviour. The holes made with the two cutting tool configurations

employing identical cutting conditions were characterized in terms of size, roundness, and surface finish. The following conclusions can be drawn from the obtained results:


Overall, this process represents an interesting opportunity to reduce the delamination damage produced by traditional drilling and, although to a minor extent, by orbital drilling. The advantages of this new hole-making process are represented by the ease of programming, the absence of coolants, with benefits in terms of green technology, and the good surface finish, as shown by the roughness measurements, with particular reference to the RPc parameter. As a future development, the use of coated WC tools will be tested to further improve the surface finish in terms of all of the measured roughness parameters. Moreover, different path strategies will be tested in order to verify the potential benefits in terms of the hole quality and tool wear improvement.

**Author Contributions:** Conceptualization and Methodology, L.N; Investigation, L.N. and I.I.; Data Curation, A.C. and L.N.; Writing-Original Draft Preparation, A.C., I.I. and L.N.; Writing-Review & Editing, A.C., I.I., L.N; Project Administration, L.N.; Funding Acquisition, L.N.

**Funding:** This research was funded by MISE, Ministry of Economic Development, Italy, PON03PE\_00129\_1 B56D12000520007.

**Acknowledgments:** The authors gratefully acknowledge LAER Srl, Acerra (Naples) facility, for providing access to the five-axis CNC machine tool used for the drill-milling tests.

**Conflicts of Interest:** The authors declare no conflict of interest.
