DETERMINATION OF PROCESSING CHARACTERISTICS OF WOOD MATERIALS DENSIFIED BY COMPRESSING

The main objective of this study is to determine optimum cutting parameters in order to specify the effect of densification by compressing on the processing properties of solid wood material and to achieve the best surface quality in materials densified at different rates. In line with this goal, the widely grown and low-density black poplar ( Populus nigra ) tree species were selected as the experimental material. Samples, which were compressed and densified by Thermo-Mechanical method at 0 %, 20 % and 40 % ratios, were processed at 1000 mm/min, 1500 mm/min and 2000 mm/min feed speeds and in 12000 rpm, 15000 rpm, 18000 rpm rotation speed on a computer numerical control machine by using two different cutters. Surface roughness values ( Ra and Rz ) were measured in order to evaluate surfaces obtained. Smoother surfaces were obtained in computer numerical control machining of densified samples. The lowest surface roughness values occurred in 40 % densified samples, which were the densest. The lowest surface roughness was obtained when 40 % densified samples were processed with cutter no.1 (two-flutes straight end mill), at 1000 mm/min feed speed and at 18000 rpm. Ra and Rz roughness values increased with the increase of the feed rate.


INTRODUCTION
From past to present, different "Wood Modification Methods" have been developed as a result of all scientific studies and research to eliminate some of the negativities of massive wood material. Modification of wood is applied to change or improve negative properties of wood material (Senol 2018, Senol andBudakcı 2016).
Wood material is often considered too soft or too weak for use in structures requiring high strength, hardness and durability. However, wood material with increased density can be used as an alternative to other materials (Blomberg and Persson 2004, Homan et al. 2000, Kutnar and Šernek 2007. Density of wood material significantly affects its mechanical (Blomberg and Persson 2004, Kamke 2006, Kutnar and Šernek 2007, Rautkari 2012) and machining properties (Blomberg and Persson 2004, Kamke 2006, Kutnar and Šernek 2007, Rautkari 2012. A lot of research has been done to improve densification process since increasing density of wood material increases its mechanical properties and hardness (Blomberg and Persson 2004). Density of low-density wood materials can be increased by densification process and their commercial value can be increased. Tree species with high density can be increased in density even more, and their properties can be improved by making them more resistant (Blomberg et al. 2005, Kutnar and Sernek 2007).
Many types of densified wood materials have been produced around the world until today. Also, in recent years, there have been restrictions in the use of protective impregnate materials that is harmful to humans and the environment with the increase of environmental priority perspective, this has led the development of new environmentally compatible methods that protect wood material against biological degradation and increase its dimensional stability (Korkut andKocaefe 2009, Senol andBudakci 2016). Deformations occur in cell wall of compressed wood material under normal atmospheric conditions. In densified wood materials; There are studies in the literature on density analysis and various mechanical properties (Ábrahám et al. 2010, Ábrahám and Németh 2012, Arruda and Del Menezzi 2013, Fang et al. 2012, Gong et al. 2010, Hajihassani et al. 2018, Kamke 2006, Kariz et al. 2017, Laine et al. 2013, Laskowska 2017, Pelit 2014, Pelit and Sonmez 2015, Rautkari et al. 2009, Senol et al. 2017, Skyba et al. 2009, Ulker et al. 2012, Cruz et al. 2018, Gao et al. 2019, Laskowska 2020, Pertuzzatti et al. 2018, determination of changes in surface hardness (Laine et al. 2013, Laskowska 2017, Pelit et al. 2015b, Rautkari et al. 2009, Rautkari et al. 2013, Senol and Budakci 2019, Skyba et al. 2009, Ulker et al. 2012, microscopic analysis (Blomberg and Persson 2004, Budakci et al. 2016, Rautkari et al. 2010), wettability (Arruda and Del Menezzi 2013, Bekhta and Krystofiak 2016, determination of spring-back amounts (Kariz et al. 2017, Laine et al. 2013, Pelit and Sonmez 2015, determination of surface properties by applying upper surface processes to surfaces obtained (Pelit 2014, Pelit et al. 2015a), resistance to fungi and termites (Esteves et al. 2017, Wehsener et al. 2018, color or gloss changes (Ábrahám et al. 2010, Cruz et al. 2018, Laskowska 2020, Pelit et al. 2015a, Tenorio et al. 2021. In addition, analysis on densification was made using finite element method (Fleischhauer et al. 2019). Many of these studies are related to densified Populus tree species (Ábrahám et al. 2010, Ábrahám and Németh 2012, Ahmed et al. 2013, Budakci et al. 2016, Diouf et al. 2011, Fang et al. 2012, Gaff and Gašparík 2013, Gao et al. 2019, Gong et al. 2010, Hajihassani et al. 2018, Bami and Mohebby 2011, Lamason and Gong 2007, Lykidis et al. 2020, Mania et al. 2020, Ozdemir 2020, Pelit et al. 2018, Senol 2018, Senol et al. 2017, Wehsener et al. 2018. In general, physical and mechanical properties improve, surface roughness and wettability decrease, hardness increases, and spring-back may occur as a negative situation in wood species that are compressed and densified depending on density increase. However, there have been no studies on the change in the processing properties of massive wood materials whose structure and density are changed by increasing density with compressing and the optimum processing properties of obtained materials. Before densified materials are turned into final product, they must be processed with machines used in machining of classical wood and wood-based materials, as well as with modern computer numerical control (CNC). In this context, with this work parameters will be determined to obtain lowest surface quality, productivity will increase and next processes such as sanding etc. will be reduced or unnecessary.
Black poplar (Populus nigra) that are produced and used frequently in the world by using CNC, with various machining parameters that will affect surface quality in different values. Therefore, the scopes of this study were organized as the following: To determine the effect of densification on machining properties and to investigate optimum processing parameters to obtain the smoothest surface for black poplar, for the machined surface, stylus tracing surface roughness profilometer was used to measure surface roughness (Ra and Rz) values, and statistical software was used to analyze the results.

MATERIALS AND METHODS
Black poplar (Populus nigra L.), which is one of the broad-leaved tree species with a wide usage area, low density and widely grown, was chosen as the experimental material in the study. Samples were all randomly selected from Afyonkarahisar, Turkey. The test specimens were cut from the parts of air-dried (approximately 15 % MC) sapwood. They were conditioned at temperatures of 20 °C ± 2 °C and 65 °C ± 5 °C, with a relative humidity to moisture content (MC) of about 12 %. The density of poplar tree species at 12 % humidity was determined as 850 kg/m 3 according to ISO 3130 (1975) and ISO 13061 (2014) standards. Experimental process of the study is given in Figure 1. Samples in the dimensions given in Table 1 were densified by compressing with thermo-mechanical (TM) method in open system (Total time = Heating time + 15 min). With a specially designed hydraulic press (Gazi University, Ankara/Turkey) with table dimensions of 60 cm x 60 cm, which can control temperature and pressure (100 t-250 atm), compression process in radial direction was carried out with automatic control at 60 mm/min loading speed. Densified test samples were kept under the press for 15 minutes and at the end of this period, samples were taken from the press, and they were allowed to cool down to room temperature under 5 kg/cm 2 pressure in order to minimize spring-back effect. After that; experiments were carried out on a Reksis Rekspeed 2137 3 axis CNC milling machine (Cözüm ahsap¸ Afyonkarahisar, Turkey) with 9 kW spindle power, a maximum spindle speed of 24000 rpm and a maximum feed rate of 60 m/min. Experiments were carried out with two router cutters (Netmak, Double-edged straight end mill and Two-flutes helisel end mill that was 8 mm in diameter) ( Figure 2). New and sharp cutters were used in each cutting test. Four parameters were used in the experiment and one of them had 2 levels, while others had three levels ( Table 2).   A total of 54 pieces with dimensions of 55 x 55 mm 2 were grooved on wood materials by a CNC router ( Figure 3). Surface roughness measurements were performed on a radial surface parallel to grain at 3 separate points on each specimen. Measuring parameters (Ra and Rz) are described in ISO 468 (1982). Measurement of surface roughness was conducted according to the protocols in ISO 468 (1982), ISO 3274 (2017), and ISO 4287 (2015. Surface Roughness Tester Time TR200 (Time Group Inc., China) type surface roughness measurement equipment was used for the determination of surface roughness values via a contact stylus trace method. Sampling length was taken as 0,8 mm (Ra: 0,32 µm -2,50 µm). Surface roughness values were measured with an accuracy of ±0,01 µm. Stylus probe speed was chosen as 10 mm/min, diameter of measurement needle was 4 µm, and needle tip was 90°. Care was taken to have a measurement environment around 18 ºC -22 °C, and without vibrations. Tool was calibrated prior to measurement, and calibration was checked at established intervals.
Analysis of variance (ANOVA) at a 95 % confidence level was applied using Minitab 19 software on the data obtained from the study. Data were evaluated by obtaining normality, main effect, and interaction graphs.

RESULTS AND DISCUSSION
Experiments were carried out by machining of surfaces with CNC in order to determine the effect of cutter type, compression ratio, feed rate and speed on roughness parameters (Ra and Rz) in the study. Roughness parameter values measured on machined surfaces are given in Table 3.  Table 2 is examined, the lowest roughness values (Ra = 2,32 μm and Rz = 11,73 μm) were obtained in the cutter type no. 1 at 40 % compression ratio with 1000 mm/min feed speed and 18000 rpm. For Ra, the highest roughness value (Ra = 6,02 μm) was measured at 0 % compression ratio, 2000 mm/min feed rate and 18000 rpm speed for cutter type 2, and for Rz (Rz = 29,06 μm), it was obtained at 0 % compression ratio for cutter type 1 with 2000 mm/min feed rate and 18000 rpm.
Statistical analyses were performed by using MINITAB R19 software with a confidence level of 95 % (e.g., significance level of 0,05). The obtained data were subjected to normality test. As seen in Figure 4, average Ra and Rz values obtained in average roughness measurements show normal distribution at 95 % confidence level, since the P value is higher than 0,05 (P = 0,150 for Ra; P = 0,229 for Rz). According to variance analysis results, for Ra at a 95 % confidence level; it was seen that cutter type (0,05 < P = 0,911) and speed (0,05 < P = 0,738) did not make a statistically significant difference, and compression ratio (0,05 > P = 0,000) and feed speed (0,05 > P = 0,004) was found to be a statistically significant difference (Table 4).  In Figure 5, the interaction of cutter type, compression ratio, feed rate and speed in terms of Ra is shown in the main effect plot. In terms of cutter type, it is seen that values close to each other in terms of Ra were obtained in machining with the cutter number 1 and 2. Although there is no significant difference between the groups, it is possible to say that generally lower Ra values were obtained with the cutter number 1 and better results can be obtained. According to the results of the variance analysis, the factor with the highest effect in terms of Ra value (P = 0,00) is the compression ratio. It is seen that as the compression ratio increases, Ra value decreases linearly and it creates a statistically significant difference. It is also stated in the literature that smoother surfaces were obtained in massive wood and wood-based materials with high density (Hiziroglu 1996, Kilic et al. 2006, Lin et al. 2006, Malkocoglu 2007, Malkocoglu and Ozdemir 2006, Pinkowski et al. 2018, Zhong et al. 2013). Acoording to Kminiak and Gaff (2015), hardwoods induce better surface quality than softwoods (Kminiak and Gaff 2015) and Budakci et al. (2011), diffuse porous wood species within hardwoods are typically associated with lower surface roughness than ring porous wood species (Nasir and Cool 2020). The reason for the decrease in Ra value can be interpreted as the decrease in gaps in the anatomical structure of the wood material with the increase in density and accordingly, smoother surfaces can be obtained when machined. It is seen that Ra value increases as the feed rate increases and it creates a statistically significant difference (P = 0,004). In the literature, it was stated that roughness values increase as a general tendency as the feed rate increases in the processing of wood and wood-based materials (Aykac and Sofuoglu 2021, Bal 2018, Bal and Akçakaya 2018, Davim et al. 2009, Hazir et al. 2018, Ilter et al. 2002, Isleyen and Karamanoglu 2019, Karagoz 2010, Koc et al. 2017, Pinkowski et al. 2018, Sutcu and Karagoz 2012, Sofuoglu et al. 2022. It seems that the data obtained in the study are compatible with the literature. With the increase in the speed, it is seen that Ra value decreases although it is not considered as significant at 95 % confidence level (P = 0,738). While there was a slight decrease in the Ra value with the increase of speed from 12000 rpm to 15000 rpm, the decrease was higher with the increase of the speed from 15000 rpm to 18000 rpm. With the increase in the speed in rotary cutting cutters, roughness values decrease and smoother surfaces can be obtained (Aykac and Sofuoglu 2021, Davim et al. 2009, Hazir et al. 2018, Isleyen and Karamanoglu 2019, Karagoz 2010, Koc et al. 2017, Sofuoglu 2015, Sutcu and Karagoz 2012, Sutcu and Karagoz 2013. It is seen that the data obtained in the study show similar trends with the literature. The higher the number of cutter marks per unit distance on the material surface of the cutters, the smoother surfaces can be obtained (Malkocoglu and Ozdemir 2006, Sofuoglu 2008, Sofuoglu and Kurtoglu 2014. Accordingly, the impact of spindle speed directly depends on the dynamic behavior of the cutter when cutting at different spindle speeds. Higher or lower spindle speed results in getting close to a blade's natural frequency and increasing the tool's vibration, then it will result in higher roughness or waviness (Nasir andCool 2019, Nasir et al. 2020). However, if the speed exceeds a certain value, burning may occur on the material surface. In Figure 6, the interactions of cutter type, compression ratio, feed rate and speed in terms of Ra are given graphically. According to the interaction graphics in terms of Ra; It is seen that the cutter number 2 results in lower roughness values by providing a linear relationship with the increase in density. Although lower Ra values were obtained on machined surfaces by cutter number 1 with the increase in density, this was not as linear and distinct as much as cutter number 2. There was an increase in the Ra value linearly in machined with cutter no. 1, with the increase in the feed rate. This was slightly different in cutter no. 2, with the increase in the feed rate from 1000 mm/min to 1500 mm/min, the Ra value slightly increased, but with the increase in the feed rate to 2000 mm/min, the Ra value was slightly decreased. Again, when looking at cutters in terms of the speed, different reactions occurred in terms of cutters even if difference was very small. While there was a tendency to decrease in Ra values with the increase in the speed in machined surfaces with cutter number 1, there was a tendency to increase for cutter number 2. Approximately close Ra values were obtained for both cutters at 15000 rpm.
As the density increased in terms of compression ratio and feed rate, lower Ra values were obtained. As the feed rate increases, a linear increase occurs in the Ra value for each density value. While Ra values was close to each other in non-densified and 20 % densified samples at at 12000 rpm, much lower Ra values were obtained at 40 % concentration. Considering compression and speed relationship, the most obvious difference occurred at 18000 rpm. Nearly Ra values were obtained at each speed at a feed rate of 2000 mm/min. While very close Ra values were obtained at feed speeds of 1500 mm/min and 2000 mm/min at 12000 rpm, similar Ra values were obtained at 1000 mm/min and 1500 mm/min at 15000 rpm. The differences in feed rates were more apparent at 18000 rpm.
When examining the variance analysis table for Rz in Table 5, although different P values were obtained according to Ra in terms of the significance of factors, it is seen that results are the same in terms of whether they are meaningful or not. According to the variance analysis results, for Rz at the 95 % confidence level; cutter type (0,05 < P = 0,542) and speed (0,05 < P = 0,674) did not make a statistically significant difference, while it was observed that compression ratio (0,05 > P = 0,000) and feed speed (0,05 > P = 0,012) made a statistically significant difference ( Table 5). The factor with the highest effect for Ra and Rz values was the compression ratio, the second effective factor was the feed rate.
In Figure 7, the interactions of cutter type, compression ratio, feed rate and speed in terms of Rz are shown in the main effect plot. When the graph is examined in terms of the cutter type, it is seen that values close to each other in terms of Rz were obtained in the cutter number 1 and 2. However, although there is no statistically significant difference, it is possible to say that, unlike the Ra value, lower Rz values were obtained with the cutter number 2 and better results can be obtained.
An evaluation similar to Ra can be made for compression ratio and speed in terms of Rz. However, while the Rz value decreased with increasing the speed from 12000 rpm to 15000 rpm, there was a slight increase in the Rz value with increasing the speed from 15000 rpm to 18000 rpm, unlike the Ra value. The results obtained for Rz were similar to Ra in terms of feed rate, speed and density as general tendency and gave similar values as tendency with the literature. In general, similar relationships occurred in the interaction graphs in Ra and Rz. However, when the interaction graphs are evaluated in terms of Rz, while the cutter number 1 and 2 have close Rz values in non-densified samples and 40 % densified samples in terms of the compression ratio with the cutter, this difference is slightly more in 20 % densified samples. In the relation of 20 % compression ratio and feed rate, the graph did not show a linear relationship as in Ra. While the roughness value of Rz increased when the feed rate was increased from 1000 mm/min to 1500 mm/min, a slight decrease in the Rz value was observed when the feed rate was increased to 2000 mm/min. However, this drop remained at a value between 1000 mm/min and 1500 mm/min (Figure 8).

CONCLUSIONS
In this study, black poplar (Populus nigra) tree species, which is an intensively and rapidly growing tree species in our country, was preferred in order to determine the effect of thermo-mechanical densification on machining properties of massive wood material. Its low density and insufficient mechanical properties have limited the use of this material. With the thermo-mechanical densification method, it was aimed to improve properties of the test specimens, which were densified in various ratios, Machining was performed CNC machine in different parameters and the surface quality of the obtained areas was evaluated by considering surface roughness. When the obtained roughness values are examined: The surface roughness, both R a and R z , increased with increasing feed rate, but decreased with increasing spindle speed. From the ANOVA, it can be seen that compression ratio was the most dominating factor for surface roughness.
Statistically, the difference of the cutters used in machining of massive wood materials did not affect the roughness of the machined surfaces.
With the increase in the compression ratio (from 0 % to 40 %), roughness values decreased. Denser materials give smoother surfaces in machining.
With the increase in the speed, there was a decrease in surface roughness values in general.
Keeping cutter feed rate low affects the surface quality positively. However, this situation causes an increase in machining time. In this case, the intersection should be ensured at the optimum point by evaluating intersection time and feed rate.
Increasing speed over a certain value was not found to be statistically significant in surface roughness values. In order for the cutter to be used efficiently and for a long time, the cutter speed should be applied taking into account the values specified on the cutter.
The lowest roughness value for Ra occurred in the cutter type number 1% at 40 % compression ratio, 1000 mm/min feed rate and 18000 rpm.
The highest roughness value for Ra occurred at 0 % compression ratio, 2000 mm/min feed rate, 18000 rpm speed in cutter number 2.
The lowest roughness value for Rz occurred in the cutter type number 1 % at 40 % compression ratio, 1000 mm/min feed rate and 18000 rpm.
The highest roughness value for Rz occurred at 0 % compression ratio, 2000 mm/min feed rate and 18000 rpm in the cutter type number 1.
Further investigations should be conducted to identify optimum machining parameters for different wood machining methods and densified wood species.