Abstract
The aim of the conducted experiments was to determine the effect of selected machining parameters on power consumption and surface quality obtained during the milling of beech wood using a computerized numerical control woodworking machine. Surface roughness was tested using the contact roughness measurement method, while roughness parameters Ra and Rz were recorded and cutting energy was determined. Tests were conducted for two variants of cutting speed (7.5 and 15 m·s-1) as well as three variants of chip thickness (0.10, 0.06, and 0.02 mm); additionally, the tests examined different cross-sections of wood. It was found that greater chip thickness and feed speed caused an increase in surface roughness and cutting power. In turn, cutting speed had no effect on surface roughness, whereas its increase resulted in increased cutting power. Surface roughness at the radial and tangential cross-sections was comparable, while it was greater at the transverse cross-section. It was also found that cutting power was lowest at the radial cross-section, while it was greater at the tangential and the greatest at the transverse cross-section.
Download PDF
Full Article
Effect of Chip Thickness, Wood Cross-sections, and Cutting Speed on Surface Roughness and Cutting Power during Up-Milling of Beech Wood
Magdalena Piernik,* Grzegorz Pinkowski, and Andrzej Krauss
The aim of the conducted experiments was to determine the effect of selected machining parameters on power consumption and surface quality obtained during the milling of beech wood using a computerized numerical control woodworking machine. Surface roughness was tested using the contact roughness measurement method, while roughness parameters Ra and Rz were recorded and cutting energy was determined. Tests were conducted for two variants of cutting speed (7.5 and 15 m·s-1) as well as three variants of chip thickness (0.10, 0.06, and 0.02 mm); additionally, the tests examined different cross-sections of wood. It was found that greater chip thickness and feed speed caused an increase in surface roughness and cutting power. In turn, cutting speed had no effect on surface roughness, whereas its increase resulted in increased cutting power. Surface roughness at the radial and tangential cross-sections was comparable, while it was greater at the transverse cross-section. It was also found that cutting power was lowest at the radial cross-section, while it was greater at the tangential and the greatest at the transverse cross-section.
DOI: 10.15376/biores.18.4.6784-6801
Keywords: Milling; Fagus sylvatica L.; Surface roughness; Chip thickness; Feed speed; Cutting speed; Wood cross-section; Cutting power
Contact information: Poznan University of Life Sciences, Department of Woodworking Machines and Fundamentals of Machine Design, Wojska Polskiego Str. 28, 60 -637 Poznan, Poland;
* Corresponding author: magdalena.piernik@up.poznan.pl
INTRODUCTION
Woodworking is executed using a variety of cutting methods, e.g., milling, sawing, drilling, turning, or sanding. Milling is one of the machining techniques most commonly applied for wood and wood-based materials. At present wood milling processes are frequently performed on computerized numerical control (CNC) woodworking machines. This provides high accuracy of machining parameter settings and ensures their maintenance at the assumed level. High rotational and feed speeds available in CNC machines facilitate advantageous efficiency of the process; however, machining quality also needs to be considered.
Surface roughness is one of the most commonly applied machining quality indicators, and as such it has been investigated in many studies concerning wood and engineered wood materials. Surface roughness is influenced by many variables, such as the condition of the cutting edge, feed speed, the manner of chip formation, or the type of machining and the resulting surface deformation. It is important to control surface roughness during production processes, e.g., those involving bonding of wood elements, or appraising quality of final products (Kilic et al. 2006). This control in most cases is a technical requirement for machined products (Guo et al. 2015). Surface quality is a complex problem, dependent on many factors, including the heterogeneous wood structure and machining conditions (Malkoçoğlu 2007). Tests have been conducted on surface roughness of wood and wood-based materials, considering the effects of various factors such as cutting force (Aguilera and Martin 2001; Wang et al. 2015), thermal wood modification (Budakçı et al. 2013; Kvietková et al. 2015; Pinkowski et al. 2016), type of material (Davim et al. 2009; Candan et al. 2012; Bekhta et al. 2014; Guo et al. 2015; Stefanowski et al. 2020), machining parameters (Kilic et al. 2006; Malkoçoğlu 2007; Barcík et al. 2009; Pinkowski et al. 2018), tool wear (Gilewicz et al. 2010; Aguilera et al. 2016), species-specific characteristics (Thoma et al. 2015), analysis of acoustic signals (Aguilera and Barros 2012), among others.
Another important factor (also economically) is connected with power consumption (electrical energy demand) of machining. Cutting power is commonly used as the primary indicator for energy consumption of machining processes. This aspect has been investigated by numerous researchers (Aguilera and Martin 2001; Marthy and Cismaru 2009; Barcík et al. 2010; Kvietková 2015; Krauss et al. 2016; Ispas et al. 2016; Koleda et al. 2019; Chuchala et al. 2020).
The primary parameters of milling affecting both machining quality and energy consumption include machining diameter, cutting speed, feed speed, the number of cutters, feed per tooth, width ae and cutting depth ap. Changing the milling parameters causes different effects in terms of machining quality, efficiency, energy consumption, etc. Milling may be performed with the same mean chip thickness at different feed speeds and rotational speeds of the cutter tool obtaining different efficiencies, which is primarily determined by feed speed.
Many studies have already been conducted to determine dependencies between individual factors and their effect on machining quality and cutting power. The influence of feed speed on roughness (at a constant cutting speed) has been investigated by many researchers, who at an increasing feed speed found an increase in surface roughness of solid wood (Škaljić et al. 2009; Vanco et al. 2016) and wood-based materials (Davim et al. 2009; Guo et al. 2015; Wang et al. 2015), while some studies also showed a negligible effect of feed speed on roughness (Malkoçoğlu 2007).
The effect of cutting speed on surface roughness has also been investigated; however, in this case the dependencies were not consistent. Some authors indicated a positive effect of an increase in cutting speed on surface roughness of machined wood (Han et al. 2004; Kvietková et al. 2015; Ispas et al. 2016; Vanco et al. 2016) – although it was not always and with a slight increase only in some cases – as well as wood-based materials (Guo et al. 2015). Some analyses were conducted showing no effect of cutting speed on machining quality (Aguilera and Martin 2001; Aguilera et al. 2016; Rajko et al. 2021).
The influence of cross-section type was also partly investigated, with some authors reporting that “a tangential direction in the planing process produces a smoother surface compared to a radial direction” (Sogutlu 2010). An increase in the angle between feed direction and the orientation of wood fibers leads to increased roughness (Cyra and Tanaka 2000). However, these studies were conducted excluding the effect of cutting speed and feed.
Increasing the cutting speed leads to an increase in cutting power (Aguilera and Martin 2001; Barcík et al. 2010; Ispas et al. 2016; Koleda et al. 2019; Rajko et al. 2021). The same effect on cutting power also has been observed for feed speed (Marthy and Cismaru 2009; Barcík et al. 2010; Kvietková 2015). Thus, the dependencies are evident; however, the cited studies were conducted without taking into consideration wood cross-sections, which is crucial, because during milling of curvilinear solid wood pieces on a CNC woodworking machine the cross-section is altered.
The research gap indicates the lack of data on the unambiguous shape of the relationship between individual machining parameters and their impact on surface roughness and cutting power. The aim of this study was to determine machining quality and cutting power at a constant chip thickness applying various variants of feed speed and cutting speed. Moreover, these dependencies were established in relation to individual wood cross-section variants.
EXPERIMENTAL
Materials
Tests were conducted on beech wood (Fagus sylvatica L.). It is a species producing hard wood, diffuse-porous surface, lacking sapwood, and characterized by no differentiation of annual rings into early and late wood. Beech is a species commonly used in furniture making. It is characterized by a relatively uniform, homogeneous anatomical structure, which is one of the reasons it is often used in research.
Tests were conducted on beam-shaped samples of 19 mm × 70 mm × 250 mm. Three sets of samples were produced, which were oriented anatomically in relation to wood fibers. Sample thickness and at the same time cutting width ap in all the three cases was 19 mm. The diagrams of radial and transverse cross-sections are given in Fig. 1. Samples were produced from the same location in a log to ensure the smallest possible variability in terms of anatomical structure and density. Samples were seasoned. Their moisture content during the tests was 8 ± 1%, and density was on average 685 kg/m3.
Fig. 1. Milling diagram with cross-section variants marked; milling operations at: (a) radial; (b) transverse cross-section; vf – feed speed, n – rotational speed, cross-section: Rad – Radial, Tan – Tangential, Tra – Transverse
Methods
Wood milling
Machining operations were performed using a computerized numerical control FLA 16 CNC router (PIAP-OBRUSN, Toruń, Poland). An end radial cutter was the cutting tool used. Angle parameters for the HW cemented carbide cutter were: rake angle 20° and wedge angle 55°.
Average chip thickness (mm) was calculated according to the Eq. 1,
(1)
where vf is feed speed (mm/min), n is rotational speed (min-1), z is number of teeth on tool, ae is cutting depth (mm), and D is cutting diameter (mm).
It follows from this formula that the same mean chip thickness may be obtained using various combinations of parameters given in the formula, i.e., feed speed, rotational speed, the number of cutter teeth, cutting depth, and tool diameter. To determine the effect of variable chip thickness on the measured parameters of surface roughness and cutting power during milling, appropriately selected values of feed speed and rotational speed were applied.
In these tests, two variants of up-milling were used for each of the three analyzed mean chip thickness levels. Values of the parameters for individual variants are given in Table 1. The radial, tangential, and transverse surfaces were milled. For each milling option one rotational speed of the tool and one cutting speed were adopted. The other constant parameters for all the variants were as follows: the number of cutters – 1, cutting diameter D – 16 mm, and cutting depth ae – 2.06 mm.
Table 1. Milling Parameters
Determination of surface roughness parameters
Surface roughness parameters were determined using the profile method according to the ISO 4287 (1997) standard. It is the most common and objective method to determine surface roughness. It was applied to measure the most typical roughness parameters: Ra – mean roughness and Rz – mean peak-to-valley height or total roughness. Various authors often provide either parameter Ra (Škaljić et al. 2009; Budakçı et al. 2013; Sofuoǧlu and Kurtoğlu 2015; Vanco et al. 2016; Koleda et al. 2019; Rajko et al. 2021) or Rz (Malkoçoğlu 2007; Aguilera et al. 2016), or both of them jointly as indicators of surface quality (Pinkowski et al. 2018). In some publications, to ensure a more extensive analysis of tested surface parameters, Ra and Rz are supplemented with parameter Rt (Davim et al. 2009; Sogutlu 2010; Pinkowski et al. 2016) as well as other additional parameters (Kilic et al. 2006; Ispas et al. 2016; Gurau et al. 2017; Hacibektasoglu et al. 2017). For each analyzed case a total of 5 roughness measurements were taken parallel to the direction of feed during machining. These measurements were obtained using a Mitutoyo Surftest SJ-210 surface roughness tester (Mitutoyo, Japan, Kawasaki) equipped with a diamond tip stylus. The following parameters were applied in the course of measurements: measuring force – 0.75 mN, feed speed of the stylus – 500 μm·s-1, radius of stylus curvature – 2 μm, apex angle of the stylus – 60°, cut-off – 2.5 mm, length of measured section – 12.5 mm.
Cutting power
Cutting power was measured for each milling case in five replications. Cutting power was measured using a Rohde & Schwarz HMC8015 gauge (Rohde & Schwarz, Munich, Germany). Milling width ap was 19 mm (sample thickness) and milling depth ae was 2.06 mm. The analyses included only cutting power (Pc), which was the difference between cutting power generated during sample cutting (power while cutting) and power at idling (idling power), as presented in Fig. 2. The determined cutting power (Pc) was the average value from all collected measurement points during cutting for a particular process.
Fig. 2. An example graph of cutting power
Statistical analysis
Statistical analyses included a factorial analysis of variance (ANOVA), followed by Tukey’s honest significant difference (HSD) test at α = 0.05. All the statistical analyses were performed using the TIBCO Software Inc., Statistica version 13.3 (Palo Alto, CA, USA). The analysis of variance makes it possible to determine the existence of differences between the mean values, i.e. it allows determination of whether there are statistically significant differences between the obtained results.
RESULTS AND DISCUSSION
Table 2 presents averaged results of analyzed roughness parameters Ra and Rz, as well as cutting power Pc for all the milling cases along with their coefficient of variation. When investigating direct values, the lowest values Ra (2.1 µm) and Rz (13.86 µm) were recorded for the lowest mean chip thickness (0.02 mm) and variant 1 of cutting speed (7.5 m·s-1) at the tangential cross-section. In turn, the highest values of Ra (10.25 µm) and Rz (80.72 µm) were obtained for the greatest chip thickness (0.1 mm) and variant 1 of cutting speed at the transverse cross-section. Differences observed between the minimum and maximum values need to be considered large, because they amounted to 8.15 µm and 66.86 µm, respectively, for Ra and Rz. The lowest value of cutting power Pc (36.56 W) was recorded for the smallest chip thickness (0.02 mm) and variant 1 of cutting speed for the radial cross-section, while it was greatest for the largest chip thickness (0.01 mm) and variant 2 of cutting speed (15 m·s-1) for the transverse cross-section.
Table 2. Means of Roughness Parameters Ra and Rz and Cutting Power (Pc)
Note: Results in parentheses are the coefficients of variation
The analysis of the coefficient values indicates that variation in the analyzed roughness parameters in a vast majority of cases for Ra and Rz is small (up to 25%), or in some cases average (from 25% to 36.9%), while power cutting Pc exhibits minimal variation (from 0.2 to 3.3%).
The ANOVA (Table 3) was conducted for roughness parameters and for cutting power, with the effect of individual factors as well as interactions (double and triple) being verified.
Table 3. Summary of ANOVA in a Factorial Arrangement of Roughness Parameters and Cutting Power
Note: ns: non-significant (95% confidence level)
Based on this analysis, in most cases differences were found in terms of mean values. Exceptions in this respect included cutting speed for Ra and Rz, as well as the double interaction between chip thickness and cutting speed also for Ra and Rz. For cutting power significant differences were recorded for all the analyzed factors as well as interactions between them.
Table 4 presents results for Ra and Rz as well as cutting power Pc in the function of main factors, i.e., chip thickness, cutting speed, and wood cross-section. Analysis of chip thickness as a main factor shows that all the analyzed chip thickness variants had a significant effect both on surface roughness (Ra and Rz) and cutting power Pc. This is indicated by various homogeneous groups identified by Tukey’s test. The dependence between chip thickness and roughness and cutting power is presented in Fig. 3.
Table 4. Results for Ra, Rz, and Pc as a Function of Main Factors
Note: Means followed by the same letter in the column do not differ statistically in Tukey’s test (P > 0.05)
Fig. 3. Surface roughness and power cutting as a function of chip thickness. Vertical bars denote 95% confidence intervals for the mean
Along with an increase in chip thickness, there was also an increase in surface roughness as shown by Ra and Rz. This is consistent with testing results provided by the authors at machining of wood flour/polyvinyl chloride composite (Guo et al. 2015). An increase in chip thickness is consistent with an increase in feed speed. The greater the feed, the greater the resulting surface roughness. In tests on beech wood this was also confirmed by other authors (Han et al. 2004). The demand for energy consumed by milling grows also at an increased chip thickness (feed speed) (Ispas et al. 2016).
The effect of cutting speed as the main factor for Ra and Rz was not significant for surface roughness, as no difference was observed in mean values of Ra and Rz for cutting speeds of 7.5 and 15 m·s-1. In literature on the subject, this lack of any dependence has been confirmed in some studies. Aguilera et al. (2016) investigated two cutting speeds of 44 m/s and 56 m/s at machining of pine wood, and they found a difference in roughness as specified by parameter Rz. Another study by Aguilera and Martin (2001) tested beech wood while applying three cutting speeds and observed comparable roughness for all the conducted tests. Some reports have also been published where researchers stated an effect of cutting speed on surface roughness (Kvietková et al. 2015), whereas other authors confirmed this dependence only in some cases (Ispas et al. 2016; Vanco et al. 2016) or reported only a slight effect (Han et al. 2004). In turn, cutting speed significantly affects cutting power. This dependence is directly proportional, i.e., cutting power grows with an increase in cutting speed. This dependence has been confirmed in other studies on beech wood (Barcík et al. 2010). In turn, while machining birch wood (Kvietková 2015) an increase in cutting power was recorded with an increase in cutting speed; however, no significant differences were found between all the applied values of cutting speed.
When analyzing the effect of wood cross-section on surface roughness parameters Ra and Rz, no differences were observed between the radial and tangential cross-sections, while a difference was recorded between the transverse cross-section and the other two cross-sections. For the effect of this factor on cutting power, a difference was found between all the three cross-sections. These dependencies are presented in Fig. 4.
Parameter Ra for the radial (3.83 µm) and tangential cross-sections (4.24 µm) was comparable, whereas the value obtained for the transverse cross-section (6.40 µm) was higher than Ra recorded for the radial and tangential cross-sections by 67% and 51%, respectively. An analogous situation was observed for the parameter Rz, with an increase for the transverse cross-section compared to the radial and tangential cross-sections, amounting to 97% and 90%, respectively. A lack of variability in roughness was also reported in studies on beech wood during planing (Kilic et al. 2006); however, this referred to the radial and tangential cross-sections only, as the transverse cross-section was not included in those investigations.
The effect of wood cross-section on cutting power Pc was significant for all the analyzed cross-sections. The lowest value of cutting power was obtained for the radial cross-section (76.7 W), while it was higher for the tangential cross-section (117.7 W) and the highest for the transverse cross-section (231.6 W). In the course of cutting operations along the grain, a chip was separated as a result of the disruption of adjacent fibers (which may be described as comparable to overcoming tensile strength across the grain). When performing transverse cutting operations it is necessary to overcome forces required to cut perpendicular fibers. Thus, machining along the grain is connected with the consumption of lesser power to separate the chip from the sample than is the case in transverse cutting.
Fig. 4. Surface roughness and power cutting as a function of wood cross-section. Vertical bars denote 95% confidence intervals for the mean.
The obtained ANOVA results also indicate differences in interactions. For parameters Ra and Rz, interactions involving two factors differences were found for the pairs of chip thickness and wood cross-section, as well as cutting speed and wood cross-section. For cutting power, differences were recorded for all the pairs of factors. A list of obtained means in terms of homogeneous groups is given in Tables 5, 6, and 7.
Table 5. Results for Ra, Rz, and Pc as a Function of the Double Interaction between Chip Thickness and Cutting Speed
Note: Means followed by the same letter in the column do not differ statistically in Tukey’s test (P > 0.05)
Table 5 and data presented in Fig. 5 illustrate that parameters Ra and Rz showed no differences within the same chip thickness using different variants of cutting speed and feed. This is confirmed by tests conducted on beech wood (Aguilera and Martin 2001), although then it was recorded at different values of chip thickness and cutting speed. In contrast, there were differences between surface roughness recorded for different chip thicknesses. The lowest values were obtained for the smallest chip thickness (0.02 mm), while they were highest for the greatest chip thickness (0.1 mm). The dependence between chip thickness and feed speed was proportional, because when feed speed was increased, chip thickness also increased. Some authors have changed chip thickness by applying different feed speeds at a constant rotational speed. Thus, with an increase in feed speed surface roughness will increase. This dependence has been confirmed by many authors both for solid wood (Kvietková et al. 2015; Vanco et al. 2016) and for wood-based materials (Davim et al. 2009; Guo et al. 2015).
Table 6. Results for Ra, Rz, and Pc as a Function of the Double Interaction between Wood Cross-section and Cutting Speed
Note: Means followed by the same letter in the column do not differ statistically in Tukey’s test (P > 0.05)
Table 7. Results for Ra, Rz, and Pc as a Function of the Double Interaction between Chip Thickness and Wood Cross-section
Note: Means followed by the same letter in the column do not differ statistically in Tukey’s test (P > 0.05)
For cutting power, an increase in this parameter was observed with an increase in chip thickness and cutting speed. The lowest values (63.9 W) were recorded for chip thickness of 0.02 mm and variant 1 of cutting speed, while they were highest (231.7 W) for chip thickness of 0.1 mm and variant 2 of cutting speed. These differences were significant, as confirmed by Tukey’s test. Similar dependencies were shown by other authors, who investigated cutting power depending on cutting speed (Aguilera and Martin 2001; Barcík et al. 2010; Ispas et al. 2016) and feed speed (Barcík et al. 2010; Ispas et al. 2016).
Fig. 5. Surface roughness and power cutting as a function of chip thickness and cutting speed. Vertical bars denote 95% confidence intervals for the mean
It follows from data presented in Table 6 that for the radial and tangential cross-sections no differences were found for surface roughness, whereas for the transverse cross-section Ra and Rz were markedly greater than analogous values recorded for the other two cross-sections. For the transverse cross-section a lower cutting speed resulted in greater values of Ra and Rz. It has been reported previously in literature that a lack of differences in surface roughness was confirmed for different cutting speeds (Aguilera and Martin 2001); however, those studies made no distinction between various anatomical profile cross-sections of wood. No differences were also observed when testing surface roughness at the radial and tangential cross-sections of beech wood at constant cutting conditions (Kilic et al. 2006). Other tests confirmed an increase in surface roughness along with an increase in the angle between feed direction and wood fibers (Cyra and Tanaka 2000). Those tests were conducted at a constant cutting speed (disregarding the effect of cutting speed). Some studies also confirmed a positive effect of increased cutting speed on surface roughness; however, this dependence was not consistent (it was not found in all the analyzed cases) (Vanco et al. 2016).
Cutting power Pc for each of the cross-sections showed values greater for variant 2, i.e. for greater cutting speed and feed values. For beech wood this was confirmed by earlier studies, although it was with no division into anatomical profile cross-sections (Aguilera and Martin 2001; Ispas et al. 2016).
Table 7 shows that for chip thickness of 0.02 mm the wood cross-section used had no effect on surface roughness, whereas it influenced cutting power, which was lowest for the radial cross-section (56.8 W) and highest for the transverse cross-section (113.2 W). For chip thickness of 0.06 mm cutting power values were analogous, whereas there was no difference in parameters Ra and Rz for the radial and tangential cross-sections. Beech wood is a diffuse porous wood with a uniform and homogeneous structure, with no division into early and late wood; thus, characteristics of surface at the radial and tangential cross-sections in terms of roughness may be similar. Roughness at the transverse cross-section had greater values in relation to the other two cross-sections. Variability in roughness between the radial cross-section and the other cross-sections may be caused by the anatomical structure of wood, and it results, among other things, both from the deeper penetration of the stylus tip into the lumen of wood cells at the transverse cross-section, as well as the deformation of wood fibers generated as a result of transverse cutting.
For the greatest chip thickness (0.1 mm), cutting power was changing analogously as it was at smaller chip thicknesses, whereas Ra values differed for all the cross-sections, while those of Rz differed only between the transverse cross-section and the other two cross-sections. Additionally, the greatest values of roughness Ra (9.68 µm) and Rz (78.47 µm) were observed for the transverse cross-section. The double interaction, i.e., that between all the three analyzed factors, is presented in Table 8 and Figs. 6 and 7.
Table 8. Results for Ra, Rz, and Pc as a Function of the Triple Interaction between Chip Thickness, Cutting Speed, and Wood Cross-section
Note: Means followed by the same letter in the column do not differ statistically in Tukey’s test (P > 0.05).
Fig. 6. Surface roughness Ra and Rz as a function of chip thickness, wood cross-section, and cutting speed. Vertical bars denote 95% confidence intervals for the mean.