Study on hole making process of thick-section Carbon Fiber Reinforced Plastics

： In this paper, the mechanism of helical milling was analyzed at first. Then, the orthogonal experiments on the 24mm thick carbon fiber reinforced composite material (CFRP) that affect the quality of hole making, such as spindle speed, feed per tooth, and number of tool’s blades were conducted. The influence of process parameters on the quality of hole making was analyzed. Finally, the range analysis method was used to obtain the best hole-making parameters: spindle speed is 4500rpm, feed per tooth is 0.02mm/z, the number of tool’s blades is 2. Based on the obtained optimal parameters, the step-by-step hole making of thick-section CFRP with variable parameters was studied, and two sets of parameters were designed. Through the measurement of the hole morphology, roughness, axial force and aperture deviation after the experiment, it is concluded that hole making process with variable parameters can meet the technical requirements of thick-section CFRP.


Introduction
CFRP has been widely used in the main bearing structure of large aircraft due to the advantages of high specific strength, high specific modulus, good rigidity, small specific gravity, good fatigue resistance, high temperature resistance, corrosion resistance, and wear resistance [1][2][3]. In order to meet the assembly requirements between the CFRP main loadbearing components in large aircraft, a large number of riveting and bolt connections are required. Therefore, the hole making process is the most commonly used processing technology in the manufacture of CFRP structural parts [4]. There has been a long period of research on CFRP around the world. Koplev et al. proposed for the first time that the fiber and matrix fracture caused the formation of chips through a series of unidirectional carbon fiber composite cutting tests. This is also the most recognized theory in the industry at this stage [5]. Zhang L et al. used the drill bit of different structure to study the hole making of CFRP/Al laminated material, and the law of the influence of the tool on the hole making quality has been obtained [6]. Qiu X et al. studied the influence of processing parameters and drill bit geometry on cutting force and hole wall damage [7]. Wang XD et al. processed large thick-section CFRP with variable parameters, which effectively reduced the cutting heat and axial force, and improved the quality of hole making [8].
Compared with the traditional drilling process, helical milling adopts eccentric processing, which eliminates the effect of the traditional twist drill chisel edge, reduces the axial force, and greatly improves the processing quality. There are more problems and defects in hole making on the thick-section CFRP than on the conventional thickness CFRP. In this paper, the mechanism of helical milling was studied, orthogonal experiments were designed, the results were measured to analyze the influence of different processing parameters on the quality of hole making. The best hole making parameters was found, and on this basis, some researches on helical milling of thick-section CFRP with variable parameters were carried out.

Research on Milling Mechanism
Helical milling is a brand-new hole making method based on the milling machine processing method. On the CNC milling machine, the helical feed motion is composed of two independent motions, including the rotary motion of the tool around the axis of the machined hole is called revolution motion, and the linear feed motion of the tool along the axis is called axial feed motion. In addition, the tool itself rotates at a high speed, that is the main cutting motion, also known as the rotation motion [13]. The plan schematic diagram of helical milling is shown in Figure 3.
The main parameters shown in Figure 2 and Figure 3 are: spindle speed -N, helical speedω, axial feed -f, eccentricity -e, aperture -D, tool's diameter -d, pitch -a and the number of blades(teeth) -Zn. It can be seen from the figures that the pitch is: = / (1) And can be seen from the geometric relationship that: According to formula (1) and formula (2): In helical milling, the axial feed -a is small. When the helix angle β is very small, tanβ≈β, then: The hypotenuse of the right-angled triangle in Figure 3 is the length of the milling line. During the cutting process, the amount of material removal -Ψ can be expressed as: When the helix angle β is very small, sinβ ≈ β. Substituting formula (4) into formula (5): It can be seen from formula (6) that the material cutting amount -Ψ is inversely proportional to the spindle speed -N. When the spindle speed increases, the material cutting amount decreases, and the cutting force also decreases accordingly. According to theoretical analysis, the selection of technological parameters for experimental research is: spindle speed, feed per tooth and number of blades.

Experimental conditions
The material selected for this experiment is CFRP, the thickness is 24mm, the diameter of the milling cutter used is 6mm, and the diameter of the hole is 8mm.

Experimental design
Orthogonal experiment is a method to deal with multi-factor and multi-level experiments [14]. This paper uses orthogonal experiment method to design the experiment. Through the experiment and consulting related materials, it is found that the main factors affecting the quality of hole making are the speed of the tool, the amount of feed, and the tool parameters. This article selects the spindle speed -N, feed per tooth fz, and the number of blades -Zn as the experimental factors, which have a direct effect on the axial force. Among them, the spindle speed and the feed per tooth select 4 levels, and the number of blades choose 3 levels. This experimental program adopts the orthogonal table design without interaction, and adopts the similar concept of the quasi-level method [15]. The threefactor four-level orthogonal table was listed first, and then eliminate all groups with 4 in the third factor (number of blades). The factor level table is shown in Table 1. The orthogonal experiment table was designed, and the data of the axial force during the helical milling process was recorded, as shown in Table 2. The resultant force of the workpiece in the X-axis and Y-axis directions is the radial force.
Through the comparison and calculation of the force change curve collected by the three-axis force sensor, it can be obtained that the radial force generated during the thick-section CFRP hole making process is very small, and the largest group is 35.35N, which accounts for about 20% of the axial force of Z-axis, so the next hole making experiment only needs to consider the influence of the axial force of Zaxis.

Data analysis
The purpose of processing the obtained experimental data is to determine the optimal level of each experimental factor and the optimal combination within the experimental range, and to determine the primary and secondary relationship between the experimental factors. The first is the direct comparison method. This method simply compares the experimental results, which is in line with people's conventional cognition. Although the direct comparison method provides a certain description of the test results, this description is qualitative and cannot certainly tell us whether the best combination is included in the above experiment. So we use the second method which called the range analysis method. This method can analyze the test results intuitively and determine the priority and optimal combination of each experimental factor.
In the range analysis method, Rj is the range of j factor, and its calculation formula can be defined as follows: = [ 1 ̅̅̅̅, 2 ̅̅̅̅, … ] − n[ 1 ̅̅̅̅, 2 ̅̅̅̅, … ] (7) Among them, yjk is the sum of experimental indicators corresponding to j factor and k level, ̅̅̅̅ is the average of yjk. The size of ̅̅̅̅ can judge the optimal level of j factor and its optimal combination. Rj reflects the fluctuation range of the experimental index when the factor level of the j-th column fluctuates. The larger the Rj, the greater the influence of the experimental factor on the index. Therefore, according to the value of Rj, the primary and secondary order of the influence of factors on experimental indicators can be judged.
Due to the inconsistent levels of the factors selected in the orthogonal table, the range value needs to be calculated by the conversion coefficient method, and the calculation is shown in formula (8): The value of the conversion coefficient di in the formula is shown in Table 3. The range R of the sum of the experimental levels can be calculated by the range ̅ of its average. Since = ̅ , it was put into formula (8), ′ = √ × × ̅ (9) In the formula: R'-conversion range; di-conversion coefficient; i-the number of levels; ri-the number of repetitions at each level.
The range value of the range analysis method in this paper needs to be corrected by formula (9). The range analysis method was used to analyze the axial force, as shown in Table 4. ̅̅̅̅̅, it means that when the number of blades is 2, the axial force is the smallest. The analysis of the range R'j shows that the spindle speed, the feed per tooth and the number of blades all have effects on the axial force. Among them, the spindle speed has the greatest influence and the number of blades has the least influence.
To sum up, when the spindle speed is 4500rpm, the feed per tooth is 0.02mm/z, and the number of blades is 2, the axial force is the smallest. That is, A4B1C1 is the best parameter combination for machining thick-section CFRP with helical milling.

Experimental design
Based on the conclusion of orthogonal experiment, this chapter takes the improvement of the quality of the entrance and exit of the hole as the experimental goal, using a 2-blades tool, and planning a hole making scheme with variable parameters, as shown in Figure 4. When the tool is close to the surface of the workpiece to the depth of 8mm, parameter 1 is used for processing; when the tool is to make a hole in the workpiece at a depth of 8-16mm, parameter 2 is used, and from 16mm to the completion of the hole, parameter 3 is used. Two sets of experimental studies were carried out under the cutting process parameters shown in Table 5, with 5 holes in each group, and the influence of hole making with various parameters scheme on the quality of the holes was analyzed.  Step 1 Step 2 Step 3  Step 1 Step 2 Step 3

Analysis of surface quality and axial force
The increase of the spindle speed can reduce the axial force, so the quality of the processed hole is improved, but the high speed will cause the hole diameter deviation too large which cannot meet the accuracy requirements.
Considering the problem of aperture deviation, the machining hole too large or too small also unable to meet the requirements of assembly accuracy. For this reason, the experimental method with various parameters was set up. The microscopic morphology of the holes processed were observed, and the morphology of the entrance and outlet of the obtained hole is shown in Figure 5.
The entrance of the first set The entrance of the second set The outlet of the first set The outlet of the second set It can be seen from Figure 5 that the quality at the exit of the hole is slightly better than the quality at the entrance, and there is no burr delamination. This is because the exit speed is higher than the entrance, so the quality of the exit will be higher than the quality of the hole processed at the same speed. The quality of the holes at the entrance and the exit of two sets of experiments is not much different. This is because the two sets of experiments have the same speed at the entrance and the same speed at the exit. The step-by-step hole making experiment can not only meet the requirement of not too high rotation speed at the entrance, but also meet the requirement of reducing burrs, tears and other damages at the exit. Figure 6 is the curve of the hole wall morphology and surface roughness corresponding to the typical holes (Ra is the largest of 5 holes ) of the two sets of experiments, and Figure 7 shows the measured axial force curve after the experiment. It can be seen from Figure 6 that after magnification of 100 times, the hole wall of the first set has gullies and scratches, and the surface roughness is 4.466, which does not meet the hole wall roughness requirements, as shown in Figure  6(a). Compared with the first hole, the wall of the second set is smoother, and the surface roughness is 2.899, which meets the requirements of hole wall roughness. This is caused by the axial force change during the hole machining process. As shown in Figure 7, when the hole is processed to the step 2, for the first step, the axial force in the middle of the hole is basically maintained at about 145~155N, while the axial force in the middle of the second set is basically stable at 120N. The fluctuation range of first set is relatively large, so the surface roughness value of the prepared hole is relatively high, and the quality of the hole is relatively poor. Since the selection of parameters of step-by-step experiment is irregular, the roughness of the hole is too large. These parameters can be used as a reference to process the hole which meet the roughness requirements on thick-section CFRP.

Analysis of aperture deviation
The use of this scheme can effectively avoid the problem of aperture deviation caused by excessively high speed. Under the two sets of parameters hole making scheme, 5 holes were processed each, the diameters of the holes at the entrance (3mm away from the entrance), middle and exit (3mm away from the exit) were measured and compared with the theoretical value, the aperture deviation is shown in Figure 8. It can be seen that the hole-making with various parameters effectively reduces the aperture deviation, and the deviation at the entrance, exit and the middle part is significantly reduced from before. The two sets of holes' apertures have little difference at the entrance. This is because the rotation speed and feed rate at the entrance is the same as that at the exit. However, there are some changes in the middle because of a sudden change in the rotation speed, but they all meet the accuracy requirements, and the deviation can be controlled between -0.05~0.1. It can be seen that the step-by-step scheme is of great help to reduce the aperture deviation.