HUNGARIAN JOURNAL OF INDUSTRY AND CHEMISTRY Vol. 49(2) pp. 47–52 (2021) hjic.mk.uni-pannon.hu DOI: 10.33927/hjic-2021-21 COMPARATIVE ANALYSIS OF BURRS IN UD-CFRP COMPOSITES USING ADVANCED HOLE MACHINING TECHNOLOGIES ANNA NIKOLETTA HELLE1 , CSONGOR PERESZLAI1 , ANDRÁS GÖDRI1 , AND NORBERT GEIER*1 1Department of Manufacturing Science and Engineering, Budapest University of Technology and Economics, Műegyetem rkp. 3, Budapest, 1111, HUNGARY Machining of carbon fibre reinforced polymer (CFRP) composites is challenging due to their inhomogeneous and anisotropic structure as well as the strong effect of the carbon fibres on wear. Burrs are critical machining-induced macro-geometrical defects in the case of the machining of CFRP composites, which may lead to assembly difficulties. Nowadays, although novel hole-machining technologies reduce the likelihood of burrs occurring, these technologies are often more costly and require longer machining times. The current experimental study focuses on the analysis of burrs in- duced by advanced hole machining technologies (helical milling, tilted helical milling and wobble milling) and comparison with a conventional one (conventional drilling). A total of 32 experiments were carried out in a VHTC 5-axis machining centre using uncoated solid carbide end mills. Furthermore, these technologies are compared and discussed based on the burrs experienced and average material removal rate (AMRR). Experimental results show that conventional drilling caused the lowest amount of burrs, followed by wobble milling, tilted helical milling and helical milling. Even though wobble milling is one of the most advantageous technologies in terms of burrs, the AMRR of conventional drilling is twenty times larger than that of wobble milling, therefore, the further development of wobble milling is recommended. Keywords: carbon fibre reinforced polymer, exit burr, drilling, digital image processing 1. Introduction Nowadays, the demand for fibre reinforced polymer com- posites is increasing and their key role in industry is unde- niable. Carbon fibre reinforced polymer (CFRP) compos- ites are no exception. According to a market report made in 2018 by M. Sauer and M. Kühnel [1], their demand by 2017 has more than doubled (114k tons) compared to that of 2010 (51k tons). The reason for this increas- ing trend lies behind the outstanding material and spe- cific mechanical properties of these composites. CFRPs possess larger strength-to-weight ratios compared to met- als, making them exceptionally suitable for parts built into assemblies connected to the aviation, aerospace and automotive industries, reducing fuel consumption [2, 3]. They also exhibit good levels of corrosion resistance and are very strong as well as stiff. However, to be a part of an assembly, these composites must be machinable, that is, can be drilled or milled. CFRPs are difficult to ma- chine, since the carbon fibres significantly contribute to tool wear as well as make the material anisotropic and inhomogeneous. These aspects lead to several difficulties such as delamination, burrs, matrix degradation, micro- cracks or fibre pullouts [3, 4]. Since the present of burrs renders post-machining almost inevitable, in order to re- duce the resources necessary for further manufacturing *Correspondence: geier.norbert@gpk.bme.hu and achieve the quality that the aviation and automotive industries require, novel technologies have been intro- duced. In the current study, the most frequently used and promising novel UD-CFRP machining technologies are compared, namely conventional drilling, helical milling, tilted helical milling and wobble milling. The kinematics of these technologies are illustrated in Fig. 1. The kinematics of conventional drilling (Fig. 1a) is the simplest of the four technologies investigated: the axis of the tool is coincident to the axis of the hole along which the tool is moving downwards while rotat- ing around this axis. Therefore, the diameter of the hole is determined by the diameter of the tool [2, 5]. This sim- plicity renders machining the least time-consuming and results in a large material removal rate (MRR), however, in the case of drilling, it is highly likely that the surface will be damaged when the tool enters and exits the com- posite, leading to separation of the laminated layers. This phenomenon is referred to as delamination, which can render parts unsuitable for further assemblies. To prevent such outcomes, a novel technology, namely helical milling (Fig. 1b) has been introduced. This technology is also known as orbital drilling [6], since the tool is moving on a helical path while rotating around its own axis, which is shifted from the axis of the hole. The kinematics can be carried out by circular inter- https://doi.org/10.33927/hjic-2021-21 mailto:geier.norbert@gpk.bme.hu 48 HELLE, PERESZLAI, GÖDRI, GEIER Figure 1: The examined technologies: a) conventional drilling, b) helical milling, c) wobble milling, d) tilted helical milling [3] polation even with a tool with a smaller diameter to the hole being machined. According to Denkena et al. [5], al- though smaller forces and tool wear occur during machin- ing, the drawback of this technology is that the effective cutting speed can reduce to zero [7]. To avoid an effective cutting speed of zero, further novel technologies have been introduced, which require a 5-axis machining centre or an industrial robot to be ex- ecuted. One of them is tilted helical milling (Fig. 1d), dur- ing which the tool moves along a helical path while being tilted to the axis of the hole with an angle of ∆ [7]. The surface of the hole becomes threaded, which leads to a high incremental increase of surface roughness compared to the other technologies. The novel technology, wobble milling (Fig. 1c), is also based on the principle of tilted axes. The tool cham- fers the top and bottom sides of the composite while the axes of both the hole and the tool enclose an angle of ∆. This tilted motion compresses the layers of the compos- ite, before the material is eliminated between the cham- fers. The advantage of this process is that the surface is less likely to delaminate. In the case of technologies that can only be exe- cuted on the 5-axis machining centres or using industrial robots, since the time required for machining is higher than needed for drilling, a novel characteristic factor shall be introduced for their comparison, namely the average material removal rate (AMRR), which gives the average amount of material that is removed over a unit of time [3]. 2. Experimental and theoretical methods This chapter is comprised of three subchapters. The first summarizes the information about the workpieces, tools and machines required for the experiments. In the second, details regarding the methods applied during the experi- ment and analysis are provided such as the experimental design, measurements of burrs, AMRR and analysis of variance (ANOVA). 2.1 Experimental setup The workpieces of the present study were composed of UD-CFRP composite plates containing a vinyl ester ma- trix and long carbon fibres to provide reinforcement. The plates were cut by a water-jet cutting machine into cylin- drical parts with a diameter of d = 40 mm and thickness of h = 5 mm. The experiments were executed on a VHTC-130M- 5HT 5-axis CNC machining centre with a tilting head. Two cutting tools were used for the machining: both tools were uncoated solid carbide end mills with only one cut- ting edge and a helix angle of λ = 25◦. For conventional drilling and milling technologies, a tool with a diameter of D1 = 6 mm (TIVOLY 82329710600) and of D2 = 4 mm (TIVOLY 82329710400) were used, respectively. A Nilfisk GB733 industrial vacuum cleaner was used to eliminate chips from the working environment. No coolant was applied during the experiment. 2.2 Methods Full factorial experimental design By focusing on the impact of the technologies (category factor) and feed rate (continuous factor) on the burrs and AMRR, an experimental design was needed that can be applied in the case of category factors as well. In this study, a full factorial experimental design [8] was applied, Hungarian Journal of Industry and Chemistry COMPARATIVE ANALYSIS OF BURRS IN UD-CFRP COMPOSITES 49 Table 1: Parameters and their levels Feed rate (mm/min) Technologies (-) Levels 50 100 150 T 1 T 2 T 3 T 4 Figure 2: Digital photo capturing moreover, the impact of the feed rate on the burrs and AMRR was examined in the case of the following tech- nologies: conventional drilling (T 1), helical milling (T 2), tilted helical milling (T 3) and wobble milling (T4). Dur- ing the experiment, a constant cutting speed of vc = 120 m/min was set, thus the spindle speeds of the tools with diameters of D1 = 6 mm and D2 = 4 mm were n1 = 6,366 RPM and n2 = 9,549 RPM, respectively. The feed rate was set at the following three levels: vf,1 = 50 mm/min, vf,2 = 100 mm/mi and vf,3 = 150 mm/min. At least one of the settings was repeated five times for each technology to ensure the experiment was reproducible. In order to eliminate the hidden errors, the 32 experimental settings were randomized. The parameters and their lev- els are summarized in Table 1. Evaluating burrs A Dino-Lite AM0413MT digital microscope was used with a 50x magnification and DinoCapture 2.0 software installed to examine the burrs. The photos were taken from the top and bottom sides of the composites which, being denoted by their individual numbers written on top, were identified with ease. A light source was placed be- neath the composites to illuminate them, thereby enhanc- ing the view of the uncut fibres as presented in Fig. 2. In order to create as clear photos as possible, the dis- tance between the composites and the microscope varied at each setting. As a result, since the holes in the pho- tos were of different sizes, a universal evaluation method was needed to eliminate any errors when compared. This was achieved by creating an image of an ideal hole (not containing any black pixels where the white circle and black square meet) before segmenting it and calculating the number of black and white pixels found. The photos of the machined holes, after being seg- mented, were cropped given the absence of any black pix- els at each side of the square and circular meeting point, as is shown in Fig. 3. Then the number of pixels was de- termined using the program Wolfram Mathematica, be- fore being listed in Microsoft Excel and used for further calculations in Minitab. The burrs resulting from different technologies were compared based on a burr ratio factor (Fbr) and the length of the longest burr of the holes (LB). The burr ratio factor Fbr = 100  1 − Wi Ti Wid Tid   (1) is based on the quotient of the ratio Wi (the number of white pixels of the given hole) to Ti (total number of pix- els of the given hole) and the ratio of Wid (the number of white pixels of the ideal hole) to Tid (total number of pixels of the ideal hole). The larger Fbr is, the more burrs found in the hole. By following this method, since the difference in the size of the photos does not affect the results, the comparison of holes does not depend on the distance set between the microscope and the composites. The length of the longest burr in the case of each setting examined was determined using AutoCAD 2022 by measuring their lengths based on the photos created by DinoCapture 2.0 (Fig. 3a). On the one hand, mea- surements were made by connecting the endpoint of the longest uncut fibre to the edge of the hole with a straight line and determining the distance between them three times before calculating their mean according to LB = 1 3 3∑ j=1 lj (2) where lj (µm) denotes the measured length of the longest burr and LB (µm) represents the average length of a given burr used for comparison. On the other hand, a corrected burr length (LB,corr) was calculated by measuring the length of the burr via the creation of a polyline: the five break points of these poly- lines were nearly equidistant from each other and each hole was measured three times. The lengths of the poly- lines were calculated by LB,corr = 1 3 1 5 3∑ j=1 5∑ i=1 lij (3) where lij (µm) denotes a segment of the polyline of the longest burr and LB,corr (µm) represents the length of the burr used for comparison. Average material removal rate The AMRR (mm3/min), which gives the average amount of material removed, was calculated using AMRR = hd2π ∆t (4) 49(2) pp. 47–52 (2021) 50 HELLE, PERESZLAI, GÖDRI, GEIER Figure 3: Digital image processing: a) photo captured by the digital microscope, b) segmented image, c) cropping the image, d) image used for the analysis Figure 4: Burr ratio factor – feed rate diagram by measuring the required time for machining. In Eq. 4, h = 5 mm is the height of the composite, d = 6 mm is the diameter of the holes, and ∆t (min) denotes the time required for machining. Furthermore, assuming the holes have the same geometry at each setting, the average material removal rate could be determined. Since the tool follows a longer path without machin- ing in the case of 5D technologies, material is removed over a longer period of time. Therefore, the average ma- terial removal rate is needed to precisely compare these technologies. Analysis of variance In this study, the impact of technologies on burrs was in- vestigated by ANOVA. The null hypothesis states that the technologies have no significant effect on the burrs at a significance level of α = 0.05. 3. Results and discussion 3.1 Analysis of burrs The results show that the holes machined by conventional drilling and wobble milling are the most burr-free com- pared to the ideal hole, whereas the holes machined by Figure 5: Main effect of the technologies on the burr ratio factor at vf = 100 mm/min. tilted helical milling and helical milling contain a larger number of burrs as is shown in Fig. 4. Helical milling resulted in the greatest amount of burrs, contradictory to the statements made by Denkena et al. [5] as well as Eguti and Trabasso [7] that helical milling can yield less burrs compared to conventional drilling. Given the main effect of technologies shown in Fig. 5, the null hypothesis was rejected, that is, these technologies have a significant im- pact on the burr ratio factor. There is a significant difference between the length of burrs achieved by different technologies as can be seen in Fig. 6. The length of burrs was less than 1 mm for holes machined by conventional drilling and wobble milling, while the length of burrs in holes machined by tilted he- lical milling exceeded 2 mm at all settings. The longest burrs were produced by tilted helical milling and conven- tional helical milling, the latter resulted in burrs that were more than four times as long compared to conventional drilling and wobble milling at the greatest examined feed rate. Wobble milling, causing short burrs and a low burr ratio factor, yielded outstanding results compared to tech- nologies based on helical milling. Analysis of variance showed that with a significance level of α = 0.05 the null hypothesis was rejected, thus Hungarian Journal of Industry and Chemistry COMPARATIVE ANALYSIS OF BURRS IN UD-CFRP COMPOSITES 51 Figure 6: Length of burrs – feed rate the technologies have a significant effect on the length of burrs. 3.2 Analysis of AMRR AMRR was greatest in the case of conventional drilling. The AMRRs resulting from milling technologies were lower but similar. As is illustrated with the second axis of the diagram in Fig. 7, the AMRRs of the milling tech- nologies are far greater than that of the drilling technol- ogy (4−5% of the AMRR). The AMRR values are shown in Fig. 7 with a feed rate of vf = 50 mm/min. 4. Conclusions In this study, 32 holes were machined by conventional drilling, helical milling, tilted helical milling and wobble milling technologies using solid carbide end mills. The conclusions are summarized as follows. • The experimental results show that the hole machin- ing technology has a significant effect on the burr ratio factor defined at the exit of the holes. • It was observed that holes machined by conventional drilling and wobble milling are characterized by the smallest amount of burrs compared to the other ex- amined technologies. Since the burr ratio factors of holes machined by tilted helical milling and heli- cal milling were larger, these holes contain a larger amount of burrs. Holes machined by helical milling are the least ideal of the four technologies examined concerning burrs. • ANOVA results show that the technologies have a significant effect on the length of burrs in the holes. • The length of burrs was the longest in the case of holes machined by helical milling, followed by tilted helical milling. However, with a feed rate of vf = 50 mm/min, holes machined by tilted helical milling contained the longest burrs. Holes machined Figure 7: AMRR of the technologies compared to conven- tional drilling by conventional drilling and wobble milling con- tained shorter burrs compared to the other milling technologies. Burrs of holes machined by wobble milling were the shortest, independent of the feed rate. • The AMRR of conventional drilling is the largest due to the simplicity of its kinematics and the AMRR of the milling technologies is around 4−5% of that of drilling for the parameter set applied dur- ing the experiment. • Although wobble milling is a promising novel hole machining technology, further development is re- quired to increase its material removal rate. Acknowledgement The research reported in this paper and carried out at BME was partly supported by the National Research, De- velopment and Innovation Office (NKFIH) No. OTKA- PD20-134430, the NRDI Fund (TKP2020 NC, Grant No. BME-NC) based on the charter of bolster issued by the NRDI Office under the auspices of the Ministry for Innovation and Technology and the project “Cen- tre of Excellence in Production Informatics and Con- trol” (EPIC) No. EU H2020-WIDESPREAD-01-2016- 2017-TeamingPhase2-739592. The authors acknowledge the assistance of Dániel István Poór in their experimental work. REFERENCES [1] Composites Market Report 2018. https://www.avk-tv.de download date: 2021 [2] Boccarusso, L.; Fazio, D.D.; Durante, M.; Langella, A.; Capece Minutolo, F.M.: CFRPs drilling: com- parison among holes produced by different drilling strategies. Procedia CIRP, 2019, 79, 325–330 DOI: 10.1016/j.procir.2019.02.075 [3] Geier, N.; Davim, J. P.; Szalay, T.: Advanced Cut- ting Tools and Technologies for Drilling Carbon Fi- bre Reinforced Polymer (CFRP) Composites: A Re- view. Compos. A Appl. Sci. 2019, 125, 105552 DOI: 10.1016/j.compositesa.2019.105552 49(2) pp. 47–52 (2021) https://www.avk-tv.de/files/20181115_avk_ccev_market_report_2018_final.pdf https://doi.org/10.1016/j.procir.2019.02.075 https://doi.org/10.1016/j.procir.2019.02.075 https://doi.org/10.1016/j.compositesa.2019.105552 https://doi.org/10.1016/j.compositesa.2019.105552 52 HELLE, PERESZLAI, GÖDRI, GEIER [4] Poór, D. I.; Geier, N.; Pereszlai, C.; Xu, J.: A Critical Review of the Drilling of CFRP Compos- ites: Burr Formation, Characterisation and Chal- lenges. Compos. B Eng. 2021, 223, 109155 DOI: 10.1016/j.compositesb.2021.109155 [5] Denkena, B.; Boehnke, D.; Dege, J. H.: Helical milling of CFRP–titanium layer compounds. CIRP J. Manuf. Sci. Technol., 2008, 1(2), 64–69 DOI: 10.1016/j.cirpj.2008.09.009 [6] Pereira, R. B. D.; Brandăo, L. C.; de Paiva, A. P.; Ferreira, J. R.; Davim, J. P.: A review of helical milling process. Int. J. Mach. Tools Manuf. 2017, 120, 27–48 DOI: 10.1016/j.ijmachtools.2017.05.002 [7] Wang, Q.; Wu, Y.; Bitou, T.; Nomura, M.; Fujii, T.: Proposal of a tilted helical milling technique for high quality hole drilling of CFRP: kinetic analy- sis of hole formation and material removal. Int. J. Adv. Manuf. Technol. 2018, 94(9–12), 4221–4235 DOI: 10.1007/s00170-017-1106-3 [8] Full Factorial Design - an overview. ScienceDirect Topics. 2018 https://www.sciencedirect.com Hungarian Journal of Industry and Chemistry https://doi.org/10.1016/j.compositesb.2021.109155 https://doi.org/10.1016/j.compositesb.2021.109155 https://doi.org/10.1016/j.cirpj.2008.09.009 https://doi.org/10.1016/j.cirpj.2008.09.009 https://doi.org/10.1016/j.ijmachtools.2017.05.002 https://doi.org/10.1007/s00170-017-1106-3 https://www.sciencedirect.com/topics/engineering/full-factorial-design Introduction Experimental and theoretical methods Experimental setup Methods Full factorial experimental design Evaluating burrs Average material removal rate Analysis of variance Results and discussion Analysis of burrs Analysis of AMRR Conclusions