1 Introduction

Fibre-reinforced polymer (FRP) composites are popular in the high-tech sectors, mainly due to their excellent specific mechanical properties and corrosion resistance [1,2,3]. Despite the high-strength carbon fibres being the most applied reinforcing structures, the existing difficulties in manufacturing, reusing, or recycling carbon fibres encourage the application of more sustainable reinforcing materials; therefore, researchers investigate recently novel solutions like basalt fibres [4,5,6]. Although the material scientists are working on the material characterisation of basalt fibre-reinforced polymer (BFRP) composites and the results are promising [5,6,7], their application is still not widespread, and the existing experiences in their machining behaviour are severely deficient.

Amuthakkannan et al. [8] carried out drilling experiments to analyse the drilling-induced delamination of BFRP composites. They analysed the effect of cutting speed, feed rate, and point angle, each on three levels. Their results show that the cutting speed has the most significant influence on the delamination factor; nevertheless, the effect of the cutting speed was not considerable. Navarro et al. [9] performed edge trimming experiments with an uncoated carbide cutting inserts tool in BFRP composites to investigate the tool wear. They found that the cutting speed is the most crucial variable during edge trimming, which seems to be a contradictory finding relative to Amuthakkannan’s, as the tool wear resulted in a larger cutting edge radius has a significant impact on the cutting force and thus on the delamination. Navarro-Mas et al. [10] investigated the edge trimming-induced geometrical defect formation (delamination type I, II, and III) of BFRP composites. Their results show that the larger the cutting speed and feed rate, the larger the value of all three types of delamination in BFRP composites is. Navarro-Mas et al. [11] compared different parameters to evaluate delamination in BFRP composites. They found that–despite the delamination has a quasi-random nature–the delamination formation probability is lower at lower feed rates. The main reason is that lower feed rates result in smaller chip cross-sections and thus lower cutting force that is responsible for layer deconstruction of composites. The number of available studies on BFRP composite machining is limited, and their information on the machinability of BFRP is deficient. Therefore, the design and optimisation of machining processes of BFRP and thus the spread of these sustainable composites is not supported properly.

Even though the machinability of BFRP composites is not studied widely yet, their machining behaviour may be expected based on the extensive expertise in the machining of carbon fibre-reinforced polymer (CFRP) composites [12]. With the similarity in the material properties of the applied matrix materials (that are often epoxy resins) and the thin-fibrous high-strength reinforcing structures, the dominant chip removal mechanisms of BFRP composites are expected to be similar to the CFRP. Li et al. [13] orthogonally machined CFRP and observed that the chip removal mechanisms and chip morphology depend significantly on the fibre orientation and the depth of cut. They showed that the machined surface quality is significantly worst at larger fibre cutting angles. Wang et al. [14] highlighted that the cutting edge radius is also a significant parameter affecting the cutting mechanisms and surface generation. Hintze et al. [15] conducted drilling experiments in CFRP and confirmed that fibre orientation has a significant impact on the cutting mechanism. Furthermore, they concluded that the cutting force is a good response value to monitor, as it correlates to the sizes and amounts of machining-induced geometrical defects (e.g. delamination and burrs) around the hole. The fibre-orientation-based failure mechanisms of CFRP were investigated by Calzada et al. [16]. They highlighted that the fibre fracture is crushing-dominated between fibre cutting angles of 45° and 90° and is bending-dominated between 0° and 135°. Turki et al. [17] proved that each area around the drilled hole corresponds to the observations gained through orthogonal cutting experiments. In summary, the bending-dominated fibre fracture at unfavourable fibre cutting conditions (θ = 135° ± δ°, where δ denotes error) around the machined hole significantly increases the risk of machining-induced geometrical error formation like delamination, burrs, fibre-pull-outs, and tearings [18].

Although the drilling-induced burrs do not weaken the resultant strength of the composite, burrs may become a potential starting point for further damage formations (e.g. delamination) and often result in assembly difficulties [19, 20]. Composite burrs are often measured and processed through digital image processing of optically captured images [21]. Xu et al. [22] observed that CFRP burrs in unidirectional fibrous laminates located symmetrically around the contour of the hole, depending significantly on the fibre cutting angle. Xu et al. [23] studied drilling-induced burrs in CFRP and found that the tool geometry has the most significant influence on the burr factor, followed by the cutting speed and feed rate, respectively. In contrast, some researchers found no relevant correlation between the feed and the burr characteristics [24,25,26]. Geier et al. [27] highlighted that drilling-induced burr formation in chopped CFRP is significantly influenced by fibre cutting angle and the mechanical supporting properties of the laminate.

Although the range of sustainable BFRP applications is getting more comprehensive, the experience and general understanding of the machining of basalt fibrous composites are lacking; therefore, the present study focuses on the experimental investigation of BFRP composites. The main objective of the present study is to analyse the machinability of BFRP composites through mechanical drilling experiments, focused on the analysis of drilling-induced burrs, surface roughness, and microstructure. The experimental results are compared to the observations of drilling CFRP composites. This study is structured as follows: first, the experimental setup and applied methods are presented in Sect. 2. Then, the results concerning burrs, surface roughness, and microstructure are presented in Sects. 3.1 and 3.2. Finally, optimal machining conditions are presented; and the results, limitations, and future aspects are discussed in Sect. 3.3.

2 Experimental setup and methods

2.1 Materials, tools, and machines

Two different composites were drilled in the experiments, provided by an industrial partner of the authors. One of them was a carbon fibre reinforcement polymer (CFRP) composite plate with a unidirectional (UD) reinforcement structure, vinyl-ester matrix, and nominal thickness of 5 mm. The second composite is a basalt fibre reinforcement polymer (BFRP) composite plate with a biaxial structure, epoxy resin, and nominal thickness of 10 mm. The main material characteristics–from the point of view chip removal mechanism–were measured five times, then averaged and listed in Table 1. The tensile strength, interlaminar shear, shore D hardness, and Charpy impact strength were measured by a Zwick Z250 universal tensile tester (according to the ISO 527–5:1997 standard), Zwick Z005 universal tensile tester (MSZ EN ISO 14130:1998), Zwick H04.3150 hardness tester (MSZ EN ISO 868:2003), and Ceast Resil Impact Junior impact tester with a 15 J hammer (MSZ EN ISO 179–1:2010), respectively. Considering that the two composites differ not only by the fibre material, but in the reinforcement structure and matrix material also, the results have to be critically handled; i.e. the findings and observations cannot be explained by the change of a single “fibre material” factor but a material category factor having two levels: CFRP and BFRP.

Table 1 Material properties of the applied CFRP and BFRP composites
Full size table

The drilling experiments were carried out on a Kondia B640 three-axis CNC machining centre. A high-performance Nilfisk GB733 industrial vacuum cleaner was used during the experiments to remove the chips from the cutting space. A Tivoly Polaris 150 Sim Dim 6537K (82415011000) Ø10 mm diameter twist drill with titanium aluminium-nitride coating was used, as shown in Fig. 1b. The tool geometry is based on the DIN 6537k standard. The tool has a tip angle of 145°, helix angle of 35°, and transverse cutting edge length of 1 mm. A special fixture was used to fix the CFRP and BFRP composite plates. Therefore, the holes were drilled under the same mechanical support conditions. The tool condition was monitored during experiments using a Dino-Lite AM413ZT digital microscope. We did not observe the increase of the cutting edge radius; thus, the tool wear status did not change during the experiments. Images of the holes were taken with a Mitutoyo Quick Image QI-A505 optical microscope to examine the burr around the holes. The surface roughness of the holes was measured using a Mitutoyo SJ-400 contact surface tester. The surface integrity of the holes was investigated through a Zeiss Evo MA 10 scanning electron microscope (SEM). The experimental machining setup can be seen in Fig. 1a.

Fig. 1
figure 1

The experimental machining setup: a working environment and b the applied coated solid carbide twist drill

Full size image

The experiments were designed with the central composite face-centred (CCF) design of experiment methods. The experimental design has three different factors: the material type as a categorical factor with two levels (BFRP and CFRP), and the cutting speed (vc) and feed (f) as continuous factors with three levels each, as shown in Table 2. The factor spaces and levels were determined based on suggestions of cutting tool producers and on previous studies [27, 28]. The drilling experiments were conducted in a dry condition, randomised order. The experimental setup in level zero (vc = 100 m/min, f = 0.10 mm/rev) was repeated five times to calculate reproducibility and analysis of variances (ANOVA).

Table 2 Factors and their levels
Full size table

2.2 Methods

After the drilling experiments, images of the holes were taken with the Mitutoyo microscope and were evaluated using the following main steps (Fig. 2): First, the original image of the hole is uploaded to a Wolfram Mathematica program, as shown in Fig. 2a, then the contour of the hole with the burr is detected using edge detection (Fig. 2b), and its length is calculated. Then, the area of the burr is detected by comparing the previously detected contour and the nominal diameter of the hole, as can be seen in Fig. 2c. Finally, the burr area is calculated. The burr factor (Fb) and the contour burr factor (Fbc) were calculated according to Eqs. (1) and (2) [19].

$${F}_{b}=\frac{{A}_{nom}-{A}_{free}}{{A}_{nom}}\cdot 100 \%=\frac{{A}_{b}}{{A}_{nom}}\cdot 100 \%$$
(1)
$${F}_{bc}=\frac{{C}_{b}-{C}_{nom}}{{C}_{nom}}\cdot 100 \%$$
(2)

where Ab (mm2) is the burr area, Anom (mm2) represents the nominal area of the hole, and Afree denotes the burr-free area of the hole; Cb (mm) is the length of the contour of the machined geometry, and Cnom (mm) is the length of the ideal geometry.

Fig. 2
figure 2

Workflow of digital image processing of drilling-induced burrs in fibrous composites: a original image, b detected contour of the burr (Cb), and c burr area (Ab)

Full size image

The surface roughness was measured according to the ISO 4287:1997 standard, but the evaluation length was decreased to Le = 4 mm because the length of the hole (5 mm in CFRP) was shorter than the standard. Each hole was measured five times in similar positions on the length of 2.5 mm. The speed of the detector was 1 mm/s, and the cut-off wavelength was set to 2.5 mm.

The micro-geometric properties of a machined surface carry several relevant information, such as the formation of fracture surfaces, the destruction mechanisms of fibres, or fibre matrix debonding [29]. A Zeiss Evo MA 10 scanning electron microscope was used for the investigation of these. Since the matrix materials of the workpieces are electrically insulating polymers, it was necessary to cover the surface with a conductive (golden) film with a BAL-TEC SCD 005 sputter coater machine (30-s cycles, with a current of 45 mA). The SEM images were taken with a 15 kV electron high tension (EHT) voltage. Several pictures were taken with different magnifications from the entry side (from which the tool penetrated the material), the middle section, and the exit side of the machined surface. Based on the images with good depth of field, the nature of the formed surfaces, the cutting and failure mechanisms, and the production quality of the workpieces were also established.

The quantitative results were evaluated using response graphs, which demonstrate the influences of factors on the corresponding responses. Response surface methodology (RSM) was used to develop these response surfaces. According to previous investigations [30], the effects of the process parameters on the analysed machining parameters are expected to be non-linear; therefore, a second-degree polynomial model was used to model, as shown in Eq. (3).

$$Y\left({x}_{1},{x}_{2}\dots {x}_{n}\right)={b}_{0}+\sum_{i=1}^{n}{b}_{i}{x}_{i}+\sum_{i=1}^{n}{b}_{ii}{x}_{ii}^{2}+\sum_{i=1}^{n-1}\sum_{j=i+1}^{n}{b}_{ij}{x}_{i}{x}_{j}+\delta$$
(3)

where Y is the corresponding response value; xi are the factors, b0, bi, bij, and bii are the regression coefficients of the parameters; and δ is a random experimental error. The analysis of variance (ANOVA) technique was used to analyse the significance level of the factor effects on the measured and calculated response values. The effect of a factor was considered significant if the ANOVA result has shown at least a significance level of α = 0.05. The prediction accuracy of the RSM models was characterised by the percentage error (PE) parameter, as expressed by Eq. (4).

$${PE}_{i}=\left|\frac{{Y}_{i}-{Y}_{m}}{{Y}_{m}}\right|\cdot 100\%$$
(4)

3 Results and discussion

In this section, the experimental results concerning burrs, surface roughness, and microgeometry are presented, optimised, and discussed. The measured, calculated, and predicted response parameters (burr factor, contour burr factor, arithmetical mean height, and maximum height of profile) are summarised in the appendix (Table 5).

3.1 Analysis of drilling-induced burrs

The calculated burr factor (Fb) and contour burr factor (Fbc) values are summarised in the Appendix (Table 5) and illustrated in Fig. 3. The processed images suggest that the drilling-induced burrs were more severe in the BFRP than in the CFRP composite. This is in good agreement with the observations of Fu et al. [31], as they pointed out that the drilling-induced damage formation in MD composites is more severe than in UD composites. In the case of drilling MD composite, one cutting edge gets contact four times with the technologically disadvantageous fibre cutting angles (θ = 135° ± δ°), while this contact number is only two in the case of unidirectional composites. Therefore, the more the fibre reinforcement direction, the larger the probability of machining-induced defect formation is. Nevertheless, a non-expected difference in the influence of process parameters on the burr parameters was found, as shown in Fig. 4 and discussed below.

Fig. 3
figure 3

Processed images of drilled hole exit in BFRP and CFRP composites

Full size image
Fig. 4
figure 4

The effect of process parameters on the burr factor and the contour burr factor after drilling a, c BFRP and b, d CFRP composites

Full size image

The RSM-based mathematical models predicting the burr factor and the contour burr factor in the case of BFRP and CFRP are expressed by Eqs. (5)–(8). The visualisations of these mathematical models in 3D response diagrams are shown in Fig. 4. These response diagrams illustrate the effect of the process parameters on the burr factor and the contour burr factor. The average percentage errors for the RSM models predicting Fb and Fbc in BFRP are 36.42%, 20.80%, and in CFRP 28.95% and 16.89%, respectively. These drilling-induced burr prediction accuracies are not as prominent as they are influenced primarily by the tool condition, as pointed out by Hrechuk et al. [21].

$${{F}_{b}}^{BFRP}=12.2-30f+0.003{v}_{c}+536f\cdot f+0.000034 {v}_{c}\cdot {v}_{c}-0.224 f\cdot {v}_{c}$$
(5)
$${{F}_{b}}^{CFRP}=23.72-376f+0.060{v}_{c}+1233f\cdot f-0.000456 {v}_{c}\cdot {v}_{c}+0.516 f\cdot {v}_{c}$$
(6)
$${{F}_{bc}}^{BFRP}=-19 + 501 f + 1.35 {v}_{c} + 765 f\cdot f - 0.00511 {v}_{c}\cdot {v}_{c} - 3.58 f\cdot {v}_{c}$$
(7)
$${{F}_{bc}}^{CFRP}=61.4 - 886 f + 1.13 {v}_{c}+ 6487 f\cdot f - 0.00237 {v}_{c}\cdot {v}_{c} - 7.24 f\cdot {v}_{c}$$
(8)

ANOVA results (Table 3) and the main and cross-effect plots (Fig. 5) show that the type of the composite (whether it is BFRP or CFRP) significantly influences the Fb and Fbc parameters (F-value = 20.54, P-value = 0.000 for Fb, and F-value = 7.08, P-value = 0.016 for Fbc). Despite the P-values of the feed (0.566 and 0.903) in the ANOVA table suggesting that the effect of the feed is not significant, its interaction with the composite type is significant. This means that the effect of feed is significantly influenced by whether the CFRP or the BFRP is drilled. The larger the feed, the larger the probability of burr formation in the BFRP is. However, the opposite was observed in the CFRP. This reason may be that the larger the feed, the larger the chip thickness is; therefore, larger energy is needed to bend the last laminated CFRP layers. In the literature, there is still disagreement whether the larger feed results in larger burrs in fibrous composites or not, as the mechanical supporting circumstances and the tool condition play a more critical role in burrs formation [19].

Table 3 ANOVA table for the burr factor and contour burr factor
Full size table
Fig. 5
figure 5

Main effect plots of the type of the composite for a Fb and b Fbc and the interaction of feed vs composite on the c Fb and d Fbc

Full size image

Generally, more drilling-induced burrs were formed in the BFRP, possibly due to the more-difficult cutting nature and multidirectional reinforcement structure than the CFRP. The BFRP and CFRP composites seem to be similarly destroyed (the contour of the burrs have similar lengths), as the similar Fbc parameters and images in Fig. 3 prove. The ANOVA resulted that the cutting speed has no relevant influence on the drilling-induced burr formation neither in the BFRP nor the CFRP.

3.2 Analysis of surface integrity

Although the machined surface characteristics are often not the critical parameters to be optimised during the drilling of fibre-reinforced composite materials, it is a general characterisation of the machinability of a material [32,33,34,35]. In this section, we report the results obtained from investigating the integrity of the drilled surfaces by examining the effects associated with the measured surface roughness values and analysing the formed surfaces through scanning electron microscopy.

3.2.1 Surface roughness

The results of the surface roughness measurement, such as arithmetical mean height (Ra) and maximum height of profile (Rz), are listed in the Appendix (Table 5). The surface roughness parameters were calculated based on the ISO 4287:1997 standard. The experimental results demonstrate that the measured surface roughness is much less favourable for the BFRP composite than for the CFRP composite. The reason for this may be found in the biaxial reinforcement structure of the BFRP, as there are far more uncut or crushed fibres (as discussed later in the SEM investigations) in the BFRP composite holes, thus resulting in rougher surface quality. The developed RSM models predicting the Ra and Rz values are expressed by Eqs. (9)–(12). The visualisations of these models representing the effect of the process parameters are shown in Fig. 6 in 3D response graphs. The average percentage errors for the RSM models predicting Ra and Rz in BFRP are 15.80% and 10.67%, and in CFRP are 13.69% and 17.40%, respectively. These prediction accuracies are better than the prediction accuracy of drilling-induced burr, as the selected factors (feed and speed) have theoretically a more vital role in surface generation than burr formation.

Fig. 6
figure 6

The effect of process parameters on the arithmetical mean height of hole wall surface in a BFRP and b CFRP composites and on the maximum height of profile of hole wall in c BFRP and d CFRP composites

Full size image
$${{R}_{a}}^{BFRP}=14.76+15f-0.18{v}_{c}-473f\cdot f+0.000711{v}_{c}\cdot {v}_{c}+0.928 f\cdot {v}_{c}$$
(9)
$${{R}_{a}}^{CFRP}=-3.31+99.9f+0.0306{v}_{c}-353f\cdot f+0.000127{v}_{c}\cdot {v}_{c}-0.223 f\cdot {v}_{c}$$
(10)
$${{R}_{z}}^{BFRP}=70.0+258f-0.687{v}_{c}-3150f\cdot f+0.00288{v}_{c}\cdot {v}_{c}+3.80 f\cdot {v}_{c}$$
(11)
$${{R}_{z}}^{CFRP}=25.5-252f+0.177{v}_{c}+1083f\cdot f-0.00007{v}_{c}\cdot {v}_{c}+0.32 f\cdot {v}_{c}$$
(12)

In the case of both composites, the response graphs show a similar characteristic. This is in a good correlation with results of ANOVA (Table 4), which shows that the cutting speed (F-value = 21.93, P-value = 0.000 for Ra, and F-value = 22.78, P-value = 0.000 for Rz) and the type of the composite (F-value = 45.72, P-value = 0.000 for Ra, and F-value = 68.77, P-value = 0.000 for Rz) have significant effect on both surface roughness metrics. However, ANOVA results show that the feed does not significantly influence the surface roughness parameters. It seems contradictory to the machining theory, as the larger the feed, the larger the distance between roughness valleys; thus, the worse the surface quality is [36,37,38]. The possible reason for the neutral effect of the feed on surface roughness may be that the variation interval of the feed factor was small. A wider range of feeds would be beneficial to analyse in the future. ANOVA results prove that none of the interaction terms is significant on the surface roughness parameters.

Table 4 ANOVA table for arithmetical mean height and maximum height of profile
Full size table

The main effect plots of factors that significantly influence the responses are shown in Fig. 7. The larger the cutting speed, the worse the Ra and Rz parameters are in both composites. The higher speeds often result in higher cutting power and, therefore, in higher cutting temperatures; thus, there is a possibility that the matrix is more likely to plastically deform and smear instead of being cut [39]. Furthermore, the surface was rougher in the BFRP than in the CFRP. The machined surfaces are analysed through scanning electron microscopy and discussed in the next section to find the proper reason for this.

Fig. 7
figure 7

Main effect plots of significant factors (vc and composite): on the arithmetical mean height of a BFRP and b CFRP, and on the maximum height of profile of c BFRP and d CFRP composites

Full size image

The ratio of Rz to Ra is a representative quality parameter showing whether the amount of uncut materials (fibre, matrix, non-reinforcing fibres, etc.) is remarkable compared to the average surface quality [32]. The scatterplot of Rz versus Ra for the BFRP and CFRP composites is shown in Fig. 8. The slopes of the fitted linear models to the measurements show that the Rz/Ra for the BFRP composite (5.86) is not significantly different from the CFRP’s (5.10). This suggests that the specific amount of uncut fibres (bent instead of being cut at the nominal depth) on the machined surfaces of BFRP and CFRP composites significantly does not differ.

Fig. 8
figure 8

Scatterplot of maximum height of profile versus arithmetical mean height in BFRP and CFRP composites

Full size image

3.2.2 SEM investigations

The SEM images show all the essential properties (feed direction, value of the fibre cutting angle –θ, etc.) and characteristic elements (debris, cavities, fracture surfaces, etc.) separately. Figure 9 shows the entry side of the drilled composites. The machined surface of the CFRP material is less rough, but the surface properties change enormously with the change of the fibre cutting angle, while in the case of the BFRP composite, the surface is almost uniform but rougher. This means that the significant difference in the previously presented surface roughness between the BFRP and CFRP composites is not caused primarily by the different reinforcing structures (and thus the fibre cutting angle) but by the composite material characteristics (i.e. matrix rigidity, reinforcing fibre strength). The presence of debris can be observed on both composites. In the case of the BFRP composite, due to the structure of the fabric, the fibres at the edge of the workpiece point towards the inside of the hole resulting in uncut fibres, while in the case of the CFRP material, this phenomenon does not occur at the same fibre cutting angles. This partly explains why the BFRP material was more prone to higher burr formation. In the case of CFRP, although the interior structure of the surface is highly dependent on the direction of the fibres, there is no significant difference at the edge of the material.

Fig. 9
figure 9

SEM images on the entry side of drilled composite holes in a, b BFRP and c, d CFRP composites

Full size image

Figure 10 shows the exit side of the holes. More significant material defects tend to occur on the exit side of the hole after drilling fibre-reinforced composites [40], as was observed in this study also. An independent manufacturing defect in the form of an air bubble can be observed in Fig. 10a. The possible reason for this error is that during the manual lamination, it was not possible to remove all the air from the material, so during the curing of the matrix, it remained there, causing a material shortage. With mass production technology (like in the case of the CFRP material), such defects are less typical than in the case of manual lamination. However, Fig. 10b, d show typical exit side defects, delamination in the case of the BFRP material, and a brittle fracture surface of the matrix material in the case of the CFRP material. Delamination is a characteristic and severe defect of fibre-reinforced polymer composites [41]. The reason for delamination formation is that the connection between the layers of reinforcing material cannot provide sufficient rigidity for the tool to cut (i.e. the thrust force reaches the critical thrust force [42]), but the force in the feed direction pushes the material in front of it, the layers separate, and the separated layers remain in this shape on the resulting machined surface after cutting ceases. This phenomenon may also be related to the formation of burrs, and here, too, the fibres point towards the inside of the hole. In the case of CFRP, the brittle fracture surface results from the brittleness of the vinyl ester matrix material. There may have fewer reinforcing fibres and more matrix material around the fracture surface, so the feed force broke the remaining material instead of being able to cut it. The surface of the inner part of the hole depends on the fibre direction for CFRP, but less for BFRP. However, the properties of the immediate environment of the exit do not depend on the fibre direction. The inner part of the exit side is similar to the inner part of the entry side. The surface of the BFRP material is rougher, more debris and protruding fibres are observed than in the case of CFRP, and the surface quality of CFRP is highly dependent on the fibre cutting angle.

Fig. 10
figure 10

SEM images on the exit side of drilled composite holes in a, b BFRP and c, d CFRP composites

Full size image

Figure 11 shows the middle section of the hole in both composites. In Fig. 11a1, a2, the layers of the biaxial reinforcement can be observed in the BFRP material. Although there is no significant effect of the direction of the fibres on the fabric structure, it can be observed that during cutting, the tool (i) cuts the reinforcing fibres, (ii) cuts off the matrix material from the fibres and makes the fibres bundles visible, and (iii) pulls out the fibres from the matrix and form voids on the surface in some part of the material. These parts do not cause a very different surface structure, the overall surface quality is similar, and this phenomenon does not cause any significant effect during the practical use of the material. In Fig. 11a3, a4, the adhesion between the elementary fibres and the matrix and the generally sheared surface on the elementary fibres can also be observed in the BFRP material. There is also an interesting phenomenon in which the tool did not sheared the fibres but penetrated them and did not cut the entire cross-section, but only twisted part of the fibres. This phenomenon can be explained by the relationship between cutting and fibre direction. However, it can still be said that in the case of BFRP with a biaxial reinforcement structure, this fibre direction has no significant effect on either the machinability or the practical applicability.

Fig. 11
figure 11

SEM images on the middle section of drilled composite holes in a1–a4 BFRP and b1–b4 CFRP composites

Full size image

Figure 11b1–b4 shows the middle section of the CFRP material. As on the BFRP material, the adhesion between the elementary fibres and the matrix material can be observed here. In this case, the cutting angle significantly affects the surface formed after cutting. Going from left to right in the figures, there is a significant change in the quality of the machined surface due to the change in the fibre cutting angle. At a cutting angle of 90°, the reinforcing fibres are sheared and have no particular effect on the surface quality (Fig. 11b1). At a cutting angle of 110°, only the matrix material is removed, and the elementary fibres are shown. Typically, sheared surfaces can be observed here. Since the diameter of the carbon fibres is smaller than that of the basalt fibres, no twisted fibres were found during the surface examination. In Fig. 11b3–b4, at a cutting angle of 135°, the tool cuts the reinforcing fibres, then pushes them in front of them and pulls them out of the matrix material, creating a highly rough, uneven surface. When the tool starts to push the fibres in front of it, much greater forces occur than in other cases. Thus, with a strong exaggeration, it can also be considered a kind of interrupted cutting. Furthermore, this area also explains why the surface roughness is so different in different parts of the hole. For CFRP, the best surface quality is obtained for fibre cutting angles less than 90°. A very smooth, uniform surface can be observed in these areas (except debris) due to the smearing of the matrix material on the surface.

Comparing Fig. 11a2, b2, the degree of adhesion between the matrix material and the reinforcing fibres is very similar for the two composites. The difference between fibre production technologies was not a relevant factor in this respect.

3.3 Optimisation, discussion, and outlook

The optimal drilling conditions to minimise each analysed quantitative response parameter (Fb, Fbc, Ra, Rz) are calculated through multi-objective optimisation in the Minitab software, and it is characterised by a feed of 0.05 mm/rev and cutting speed of 72.2 m/min in the BFRP, and feed of 0.11 mm/rev and cutting speed of 50 m/min in the CFRP. The multi objective-optimisation is illustrated in Fig. 12. In the case both materials can meet the requirements of the designers (having appropriate strength, dimensional stability, etc.), the application of the CFRP is recommended from the point of view of minimising the machining-induced geometrical defect formation. Furthermore, if sustainability is one of the main critical criteria of composite material selection, the BFRP is possibly the right choice. The analysis of cutting energetics (cutting force, torque, and vibrations) is one of the next steps in this research project.

Fig. 12
figure 12

Multi-objective optimisation of drilling in a BFRP and b CFRP composites to minimise Fb, Fbc, Ra, and Rz parameters

Full size image

It is essential to highlight that the applied BFRP and CFRP composites cannot be compared correctly enough, as their optimal layering sequence, reinforcement structures, applied matrix materials, etc. differ in industrial applications. Thus, if these composites are compared from a certain point of view (burr formation, surface generation, etc.), these composites are better to consider as category factors (like A and B material). Otherwise, if only the reinforcement material (carbon and basalt) is varied systematically in the design of experiments, the obtained results would fail to present relevant information for the industrial applications. Nevertheless, the analysis of the influences of reinforcing fibres on the cutting mechanism is essential in the future to deeper understand the machinability of BFRP composites.

Although the present study suggests that the BFRP composite machining is even more challenging than the machining of CFRP, the interpretation of the findings and observations into industrial scales needs further and more detailed investigations on the chip removal mechanisms, and delamination and burr formation mechanisms of BFRP. Furthermore, as the tool condition has the most significant influence on the machining mechanisms, the tool wear mechanisms and speed of tool wear in the machining of BFRP are recommended to be analysed in the future.

4 Conclusions

In the present experimental study, the machinability of basalt fibre reinforced polymer (BFRP) was investigated by analysing drilling-induced burrs, surface roughness, and microstructure. The results were then compared with the observations gained in carbon fibre-reinforced polymer (CFRP) composite drilling. According to the present study, the following conclusions can be drawn:

  • More severe drilling-induced burr formation was observed in the BFRP composite than in the CFRP. The composite-feed interaction influenced the burr formation in the composites. The larger the feed, the larger the drilling-induced burrs are in the BFRP. However, the opposite phenomenon was observed in the CFRP composite. The influence of the cutting speed was not considerable; thus, the feed is recommended to be optimised to minimise burrs.

  • ANOVA results show that the composite type has the most significant influence on the surface roughness, followed by the cutting speed. However, the effect of the feed was not considerable in each composite. Both the arithmetical mean height and maximum height of profile of BFRP was worse than those of CFRP. The interaction terms were not significant.

  • It was observed through SEM investigation that the main reasons for the differences between the BFRP and CFRP surfaces are the differences between the structure, the matrix material, and the diameter of the fibres. A less rough surface is obtained for CFRP, except for the region around the 135° cutting angle. In the case of the vinyl-ester matrix material, the smear of the material was characteristic, while in the case of the epoxy, this was not typical; much larger particles were observed. Respectively, in the case of CFRP, the shear surface was observed in the fibres, while in the case of BFRP, the twisted elementary fibres were also visible due to the thicker fibres.

  • The optimal conditions to drill suitable quality holes in BFRP are characterised by the feed of 0.05 mm/rev and cutting speed of 72.2 m/min, and feed of 0.11 mm/rev and cutting speed of 50 m/min in the CFRP. From the point of view of drilling-induced micro and macro geometrical damage formation, the CFRP is recommended to apply. However, the analysis of chip removal and tool wear mechanism in BFRPs is recommended in the future to obtain more significant regularities and a generous understanding of BFRP machining.