Quality and Bending Properties of Scots Pine ( Pinus Sylvestris L.) Sawn Timber

: The paper presents the partial research results of a mechanical properties study conducted on Scots pine from the Silesian Forestry Area in Poland. The scope of research included the visual strength grading of the timber according to the PN-D-94021:2013 standard, mechanical testing (according to EN 408:2012) and an attempt to assign visuals to the C strength classes. The EN 1912: 2012 standard assigns the visual sorting classes of individual wood species (according to national sorting standards) to the C strength classes introduced by the EN 338: 2018 standard. At the moment, this standard does not assign Polish visual sorting classes (KW, KS, and KG) to C strength classes. The obtained MOE (modulus of elasticity) and MOR (static bending strength) values were corrected according to the EN 384:2018 standard, and their characteristic values were later calculated. On this basis, we proposed a classiﬁcation of the grading classes determined on the basis of PN-D-94021 into C classes described in EN 338:2018, which is necessary in the process of the transposition of Polish strength grades (KW, KS, KG, and Reject) and the qualiﬁcation of them in line with EN 1912:2012. The calculated characteristic values of density, MOR and E 0 (the modulus of elasticity parallel to the grain) allowed us to assign Polish visual grades KW, KS, and KG to C35, C30, and C20, respectively. The pine timber under research had high physical and mechanical parameters, which translated into high C classes to which the KW, KS, and KG timber categories were assigned.


Introduction
Wood, as a natural material, has an anisotropic structure. For this reason, and due to possible wood defects, wood can have very diverse mechanical and physical properties, even within the same species. Therefore, only timber with guaranteed resistance-that is, strength graded timber-can be used in civil engineering for construction applications, according to the legal requirements binding in Poland and in other member states of the European Union. There are two methods of strength grading: by visual inspection and with the use of a grading machine. There are two types of grading machines. One of them works on the basis of the determination of the dynamic modulus of elasticity (using resonance frequency or ultrasounds), and the second one on the basis of the static modulus of elasticity within bending. In this situation, timber is machine-graded by passing it between rollers and measuring either the load required to give a fixed deflection or the deflection within fixed load.
Strength grading based on visual inspection consists of examining every piece of timber and classifying it within the appropriate strength class on the basis of visible defects of wood structure,

Material
The research material consisted of 210 pieces of Scots pine timber from the Silesian Forestry Region in Poland. The timber was acquired from raw wood aged about 120 years. The timber originated from logs of trees that grew in a fresh mixed forest under the Regional National Forest Directorate of Katowice (Olesno Forest District, Sternalice Forest Division, unit 14 d, geographical coordinates: 50 • 53 55.1" N and 18 • 25 26.1" E). The sawn timber was obtained from logs 3.5 m-long. Logs were cut on a sawline consisting of a canter, 4 vertical bandsaws, and an edger for side boards. The line was operated in a "merry go round" system. In the first pass, 150 mm-wide central lumber and side boards were obtained. In the second pass, planks were cut into mine boards with a depth of 50 mm and side boards. Only the main boards was used for the tests. Construction elements (according to PN-94021:2013 with the cross section at least 2000 mm 2 ) with nominal cross-section dimensions after drying and planning-40 × 138 × 3500 mm-were used in the tests. This cross-section is often used for the production of roof trusses in popular MiTek system (gang nail connections).

Methods
During the study, the batch of Scots pine timber from the Silesian Forestry Region was sorted into KW, KS, KG, and "Reject" classes according to the visual strength grading method described in PN-D-94021:2013. In the process of strength grading based on visual inspection according to PN-D-94021:2013, defects of wood structure, shape, and processing are taken into account. The following structural wood defects are analyzed: knots, twisted fibers, cracks, rot, insect tunnels, compression wood, blue stain, and graininess. The defects of shape and processing taken into account during visual strength grading are the following: wane, bows, cups, springs, warping, and other defects due to wrong processing.
Knots are assessed by their total knot area ratio (TKAR-the ratio of the sum of the total projected cross-sectional areas of all knots intersected by any cross-section to the total cross-sectional area of the piece) and their margin knot area ratio (MKAR-the ratio of the sum of the projected cross-sectional areas of all knots or portions of knots in a margin intersected at any cross-section to the cross-sectional area of margin). In each piece of timber with width h, there are two marginal zones (with width h/4) on both sides of the timber. Table 1 shows the values of the knot area ratio and the average width of annual rings according to PN-D-94021:2013. Wane is a typical defect occurring at the face of sawn timber-in the tested batch of sawn timber, there were no pieces of the sawn timber, in which there could have been wane in the central part that was subject to visual grading.
Another important structural feature that determined the sorting class of timber was the twist of the fibers. The PN/D-94021:2013 standard allows for the maximum slope of grain in the KW class up to 7% (1:14), up to 10% (1:10) in the KS class, and up to 16% (1:6) in the KG class. Completely unacceptable, regardless of the class, are rot and insect galleries. Blue stain, on the other hand, is admissible without restrictions. The admissibility of the remaining structural features taken into account for sorting is subject to the rule that in the highest class (KW), the admissibility of these defects is the lowest compared to the other classes, and it is the highest in the lowest class (KG).
The grading was carried out for the middle parts of the timber pieces located between the load points of the strength testing machine during bending tests, and the wood quality at the ends of timber pieces was not taken into account. This was due to the fact that the quality at the ends of timber pieces, located outside of the strength testing machine measurement area, does not influence the measured values of the modulus of elasticity and bending strength determined by the machine.
The next step consisted of a four-point static bending strength test according to EN 408:2010 + A1:2012, with the use of a TIRA Test 2300 universal strength testing machine (TIRA GmbH, Schalkau, Germany). In four-point bending tests, the test piece normally has a minimum length of 19 times the depth of the cross section. In the conducted research span, the depth ratio was 18:1. Additionally, to prevent buckling, lateral restraints were used. The span-to-depth ratio is an important parameter because it can affect structural behavior [30]. The weakest section of each piece was determined by visual inspection. When possible the weakest section was located within the load points of the strength testing machine. Bending strength and the modulus of elasticity were determined. Finally, according to the guidelines of EN 384:2018, moisture content adjustment was carried out for density and the modulus of elasticity parallel to the grain (E 0 ), and size adjustment was carried out for bending strength. Later, a free-from-failure-defects sample was taken from the rupture area to test its density and moisture with the gravimetric method (drier and scale).
The next stage consisted of calculating the characteristic values: the density, MOR, and E 0 of the timber under research. On the basis of the characteristic values, we proposed a manner of assigning the KW, KS, and KG grades to the C strength classes according to EN 338:2018, which is indispensable in order to introduce Polish strength grades to the EN 1912:2012 standard.

Statistical Analysis
The statistical analysis of test results was carried out in the Statistica v.13.3 software (StatSoft, Inc., Tulsa, USA). The statistical analysis involved an ANOVA to determine whether there were differences within specific grades (KW, KS, KG, or Reject) concerning the values of physico-mechanical parameters (wood density, modulus of elasticity parallel to the grain, and bending strength). Bearing this in mind, we tested whether the distribution of the variable (physico-mechanical parameter) under analysis was normal. For all the analyzed cases, in each strength grade, the normal distribution of results was verified with the Kolmogorov-Smirnov test, with the minimum level of significance p = 0.05. Subsequently, the homogeneous variance of the tested parameter was verified with Levene's test. If the test confirmed the fulfilment of the zero hypothesis, we carried out a post-hoc test with the Tukey method. If the Levene's test revealed a lack of variance homogeneity for a given parameter, we used a corrective coefficient and an adequate method of post-hoc test: the Dunnett's test.

Results and Discussion
As a result of visual strength grading, while considering the middle part (within the test span of the strength testing machine), the batch of 210 timber pieces was divided into strength grades in the following manner: 83 pieces were classified as KW (39.5% of the whole batch), 31 pieces were classified as KS (14.8% of the whole batch), 60 pieces were classified as KG grade (28.6% of the whole batch), and 36 pieces (17.1%) were classified as Reject ( Table 1). Features that had more influence in the downgrading process were knots (share of over 92%), followed by twisted fibers (share around 5%) and then rot (share around 3%).
In reference to the research conducted in previous years concerning timber from five different forestry regions [20], timber originating from the Silesian Forestry Region was characterized by a very high share of timber pieces classified as KW (24.6% higher than the average value of timber from all five regions classified as KW on the basis of the visual strength grading of the middle part of the timber piece, located within the span of loading points of the machine head during the bending strength test of each timber piece).
The timber under research (Table 2) had high average density (541 kg/m 3 ), modulus of elasticity parallel to the grain (12,680 MPa), and static bending strength (47 MPa) values compared to the values obtained previously in the study of timber from five different Polish forestry regions [20]. A higher average density value was related to the higher density values of timber, compared to the density values of timber from the aforementioned five different Polish forestry regions, qualified to each grading class (KW, KS, KG, and Reject), respectively: by 30, 48, 38, and 58 kg/m 3 . Similarly, the values of the modulus of elasticity parallel to the grain determined for Silesian timber for the KW, KS, KG, and Reject classes corresponded to, respectively, 118%, 119%, 124%, and 136% of the average values of the modulus of elasticity parallel to the grain obtained for timber from the five Polish forestry regions. Regarding the static bending strength (MOR), the values determined for timber from the Silesian Forestry Region for each class (KW, KS, KG, and Reject) respectively corresponded to 123%, 105%, 114%, and 130% of the average MOR determined for timber from five different Polish forestry regions [20]. Silesian timber has very good physico-mechanical parameters, which has also been confirmed by studies conducted on Polish-grown Scots pine timber by other researchers. For example, a study of 219 Polish pine timber [12] resulted in an average density value that was by 4% lower than density obtained within reported study, analogously modulus of elasticity parallel to the grain, and static bending strength that was 12% lower than the values determined within reported study for Silesian timber. At the same time, the coefficients of variation of the physico-mechanical properties (density, modulus of elasticity parallel to the grain, and static bending strength) were similar in both studies. Analogous relations were confirmed after comparing the physico-mechanical parameters determined for Silesian timber with the results obtained within the framework of the Gradewood project carried out by the VTT Technical Research Centre of Finland. The average density of Scots pine timber from the Silesian Forestry Region in Poland was approximately 4% higher than the average density of Polish-grown timber tested within the framework of the Gradewood project and 12% higher than Swedish pine timber [21]. Regarding the modulus of elasticity parallel to the grain, Polish-grown Scots pine timber from the Silesian Forestry Region had a similar average value of the modulus of elasticity parallel to the grain as Polish timber in general, and it had a value 9% higher than Swedish timber. Regarding static bending strength, the average value obtained in the research was 11% higher than the average static bending strength value for Polish-grown timber in general and 2% lower than the value for Swedish pine timber [21].
The statistical analysis (one-way analysis of variance) revealed that there were no statistically significant differences between the average density values of wood classified in the KW and KS grades. In this respect, the strength grade did not determine the density. A similar situation concerning wood density was also observed between the two lower grades of KG and Reject. What is more, the analysis of variance revealed that the KW and KS strength grades were directly correlated with the values of the modulus of elasticity parallel to the grain and static bending strength (MOR). On the other hand, KG and Reject grades did not present statistically significant differences in the values of MOR and E 0 . Figure 1 presents the correlations between the modulus of elasticity parallel to the grain and wood density (the values were corrected according to the guidelines from EN 384). The coefficient of determination for the entire timber batch (210 pieces) amounted to 0.31. This value was significantly lower than the coefficient of determination between the modulus of elasticity and density determined in a previous study of 764 pieces of Scots pine timber from various Polish forestry regions, which amounted to 0.43 [20]. In a case of Finnish pine timber, the coefficient of determination between the modulus of elasticity parallel to the grain and wood density amounted to 0.54 [21]. In turn, studies of Scots pine timber pieces with different cross-section dimensions have shown that the coefficient of determination between the modulus of elasticity parallel to the grain and static bending strength can fall in the range between 0.49 and 0.66 [17]. Taking the division of timber into strength grades (KW, KS, KG, and Reject) into account, the highest value of coefficient of determination (0.66) was obtained for timber from the KS grade. The coefficient was lower for both KW and KG grades (0.54), while the Reject grade had a coefficient of determination of 0.60. These values were clearly higher than the values obtained in a previous study of Polish-grown timber from five different forestry regions. The coefficients of determination obtained in that study were 0.59 for KW, 0.50 for KS, 0.52 for KG, and 0.48 for Reject [20]. The statistical analysis (one-way analysis of variance) revealed that there were no statistically significant differences between the average density values of wood classified in the KW and KS grades. In this respect, the strength grade did not determine the density. A similar situation concerning wood density was also observed between the two lower grades of KG and Reject. What is more, the analysis of variance revealed that the KW and KS strength grades were directly correlated with the values of the modulus of elasticity parallel to the grain and static bending strength (MOR). On the other hand, KG and Reject grades did not present statistically significant differences in the values of MOR and E0. Figure 1 presents the correlations between the modulus of elasticity parallel to the grain and wood density (the values were corrected according to the guidelines from EN 384). The coefficient of determination for the entire timber batch (210 pieces) amounted to 0.31. This value was significantly lower than the coefficient of determination between the modulus of elasticity and density determined in a previous study of 764 pieces of Scots pine timber from various Polish forestry regions, which amounted to 0.43 [20]. In a case of Finnish pine timber, the coefficient of determination between the modulus of elasticity parallel to the grain and wood density amounted to 0.54 [21]. In turn, studies of Scots pine timber pieces with different cross-section dimensions have shown that the coefficient of determination between the modulus of elasticity parallel to the grain and static bending strength can fall in the range between 0.49 and 0.66 [17]. Taking the division of timber into strength grades (KW, KS, KG, and Reject) into account, the highest value of coefficient of determination (0.66) was obtained for timber from the KS grade. The coefficient was lower for both KW and KG grades (0.54), while the Reject grade had a coefficient of determination of 0.60. These values were clearly higher than the values obtained in a previous study of Polish-grown timber from five different forestry regions. The coefficients of determination obtained in that study were 0.59 for KW, 0.50 for KS, 0.52 for KG, and 0.48 for Reject [20]. The coefficient of determination between the static bending strength and wood density corrected for the moisture content of 12% (Figure 2) for the entire batch of timber (210 pieces) amounted to 0.27. This value was significantly lower than the coefficient of determination between the bending strength and density corrected for a 12% moisture content determined in a previous study of 764 pieces of Scots pine timber from various Polish forestry regions, which amounted to 0.42 [20]. In a case of Finnish pine timber, the coefficient of determination between the static bending strength and wood density amounted to 0.21 [21]. In turn, studies of Scots pine timber pieces with different cross-section dimensions have shown that the coefficient of determination between the modulus of elasticity The coefficient of determination between the static bending strength and wood density corrected for the moisture content of 12% (Figure 2) for the entire batch of timber (210 pieces) amounted to 0.27. This value was significantly lower than the coefficient of determination between the bending strength and density corrected for a 12% moisture content determined in a previous study of 764 pieces of Scots pine timber from various Polish forestry regions, which amounted to 0.42 [20]. In a case of Finnish pine timber, the coefficient of determination between the static bending strength and wood density amounted to 0.21 [21]. In turn, studies of Scots pine timber pieces with different cross-section dimensions have shown that the coefficient of determination between the modulus of elasticity parallel to the grain and density can fall in the range between 0.43 and 0.56 [17].
Forests 2020, 11, x FOR PEER REVIEW 7 of 13 the KS and KG grades, the coefficients of determination amounted to 0.12 and 0.13, respectively. In the case of timber classified as "Reject," the coefficient of determination amounted to 0.18. These values were clearly lower than the values obtained in a previous study of Polish-grown timber from five different forestry regions. The coefficients of determination obtained in that study were: 0.55 for KW, 0.37 for KS, and 0.25 for KG and Reject [20]. Those lower values of the coefficient of determination obtained in the cited research could have been due to lower numbers of timber pieces in the entire batch and pieces sorted into strength grades. The coefficient of determination between the modulus of elasticity parallel to the grain and static bending strength (values corrected in line with EN 384) (Figure 3) for the entire batch of timber (210 pieces) amounted to 0.38. This value was lower than the coefficient of determination between the modulus of elasticity and the static bending strength determined in a previous study of 764 pieces of Scots pine timber from various Polish forestry regions, which amounted to 0.67 [20]. In the case of Finnish pine timber, the coefficient of determination between the static bending strength and modulus of elasticity parallel to the grain amounted to 0.53 [21]. In turn, studies of Scots pine timber pieces with different cross-section dimensions have shown that the coefficient of determination between the modulus of elasticity parallel to the grain and static bending strength can fall in the range between 0.72 and 0.79 [17]. Taking the division of timber into strength grades into account, the highest value of coefficient of determination (0.68) was obtained for timber of the "Reject" grade. The coefficient of determination amounted to 0.40 for the KW grade, 0.31 for the KS grade, and 0.38 for the KG grade. The obtained values of the coefficient of determination were lower than the values obtained during tests of Scots pine timber from five different forestry regions classified into different grades (KW-0.56; KS-0.50; KG-0.52; and Reject-0.48). According to [26], the coefficient of determination between the corrected values of the modulus of elasticity and static bending strength for timber from Tuchola (Kujawsko-Pomorskie voivodeship, Poland) amounted to: KW-0.87; KS-0.44; and KG-0.41. Taking the division of timber into strength grades (KW, KS, KG, and Reject) into account, the highest value of coefficient of determination (0.27) was obtained for timber from the KW grade. For the KS and KG grades, the coefficients of determination amounted to 0.12 and 0.13, respectively. In the case of timber classified as "Reject," the coefficient of determination amounted to 0.18. These values were clearly lower than the values obtained in a previous study of Polish-grown timber from five different forestry regions. The coefficients of determination obtained in that study were: 0.55 for KW, 0.37 for KS, and 0.25 for KG and Reject [20]. Those lower values of the coefficient of determination obtained in the cited research could have been due to lower numbers of timber pieces in the entire batch and pieces sorted into strength grades.
The coefficient of determination between the modulus of elasticity parallel to the grain and static bending strength (values corrected in line with EN 384) (Figure 3) for the entire batch of timber (210 pieces) amounted to 0.38. This value was lower than the coefficient of determination between the modulus of elasticity and the static bending strength determined in a previous study of 764 pieces of Scots pine timber from various Polish forestry regions, which amounted to 0.67 [20]. In the case of Finnish pine timber, the coefficient of determination between the static bending strength and modulus of elasticity parallel to the grain amounted to 0.53 [21]. In turn, studies of Scots pine timber pieces with different cross-section dimensions have shown that the coefficient of determination between the modulus of elasticity parallel to the grain and static bending strength can fall in the range between 0.72 and 0.79 [17]. Taking the division of timber into strength grades into account, the highest value of coefficient of determination (0.68) was obtained for timber of the "Reject" grade. The coefficient of determination amounted to 0.40 for the KW grade, 0.31 for the KS grade, and 0.38 for the KG grade.
The obtained values of the coefficient of determination were lower than the values obtained during tests of Scots pine timber from five different forestry regions classified into different grades (KW-0.56; KS-0.50; KG-0.52; and Reject-0.48). According to [26], the coefficient of determination between the corrected values of the modulus of elasticity and static bending strength for timber from Tuchola     Figures 5 and 6 show a clear, typical, and high variance of the parameter within a given strength grade. This variance can be reduced by a more careful selection of research material, as well as by sub-dividing strength grades into smaller sub-grades [31,32]. Another possibility consists of improving the sorting methods, especially by preparing models that would provide higher correlations between the physicomechanical properties of timber and timber strength grades.
.    Figures 5 and 6 show a clear, typical, and high variance of the parameter within a given strength grade. This variance can be reduced by a more careful selection of research material, as well as by sub-dividing strength grades into smaller sub-grades [31,32]. Another possibility consists of improving the sorting methods, especially by preparing models that would provide higher correlations between the physico-mechanical properties of timber and timber strength grades.    Figures 5 and 6 show a clear, typical, and high variance of the parameter within a given strength grade. This variance can be reduced by a more careful selection of research material, as well as by sub-dividing strength grades into smaller sub-grades [31,32]. Another possibility consists of improving the sorting methods, especially by preparing models that would provide higher correlations between the physicomechanical properties of timber and timber strength grades.    Polish-grown Scots pine timber from the Silesian Forestry Region proved to have a quality comparable to Finnish spruce timber. In a study performed by Ranta-Maunus et al. [33], only five timber pieces (out of the 600 pieces tested) had a static bending strength lower than 20 MPa. An analysis of Figure 5 shows that a total of nine timber pieces had bending strength lower than 20 MPa: one of them classified as KG and eight classified as Reject. At the same time, we can notice that the highest range of distribution of static bending strength was achieved by the timber classified as KW (from 28 to 94 MPa) in comparison to the timber from other strength grades. On the basis of the cumulative distribution of the modulus of elasticity parallel to the grain (Figure 6), we can notice that two timber pieces (one in KW and one in Reject) had significantly different values of the modulus of elasticity parallel to the grain than other timber pieces from the given grade. Additionally in this case, the KW grade had the highest range of distribution for the modulus of elasticity parallel to the grain   Polish-grown Scots pine timber from the Silesian Forestry Region proved to have a quality comparable to Finnish spruce timber. In a study performed by Ranta-Maunus et al. [33], only five timber pieces (out of the 600 pieces tested) had a static bending strength lower than 20 MPa. An analysis of Figure 5 shows that a total of nine timber pieces had bending strength lower than 20 MPa: one of them classified as KG and eight classified as Reject. At the same time, we can notice that the highest range of distribution of static bending strength was achieved by the timber classified as KW (from 28 to 94 MPa) in comparison to the timber from other strength grades. On the basis of the cumulative distribution of the modulus of elasticity parallel to the grain (Figure 6), we can notice that two timber pieces (one in KW and one in Reject) had significantly different values of the modulus of elasticity parallel to the grain than other timber pieces from the given grade. Additionally in this case, the KW grade had the highest range of distribution for the modulus of elasticity parallel to the grain Polish-grown Scots pine timber from the Silesian Forestry Region proved to have a quality comparable to Finnish spruce timber. In a study performed by Ranta-Maunus et al. [33], only five timber pieces (out of the 600 pieces tested) had a static bending strength lower than 20 MPa. An analysis of Figure 5 shows that a total of nine timber pieces had bending strength lower than 20 MPa: one of them classified as KG and eight classified as Reject. At the same time, we can notice that the highest range of distribution of static bending strength was achieved by the timber classified as KW (from 28 to 94 MPa) in comparison to the timber from other strength grades. On the basis of the cumulative distribution of the modulus of elasticity parallel to the grain (Figure 6), we can notice that two timber pieces (one in KW and one in Reject) had significantly different values of the modulus of elasticity parallel to the grain than other timber pieces from the given grade. Additionally in this case, the KW grade had the highest range of distribution for the modulus of elasticity parallel to the grain (from 7100 to 20,750 MPa). In the case of single pieces of classified sawn timber, for example, in the KW class, very low E 0 and MOR values were obtained. This was due to crack initiation in the zone outside the area to be sorted by the visual method.
On the basis of the characteristic values of density, MOR, and MOE for each strength grade, the grades were assigned to C strength classes while considering the minimum requirements specified by the EN 338 standard. According to Table 3, Scots pine timber from the Silesian Forestry Region graded as KW corresponds to the C35 class, KS corresponds to C30, KG corresponds to C20, and Reject is also classified as Reject. On the basis of other studies conducted on Scots pine [10], the grades were transposed as: KW-C35 class; KS-C24 class; and KG-C20 class. On the basis of the obtained MOR and MOE values, the study of Polish-grown Scots pine timber from five different Polish Forestry Regions [20] (for the entire batch of timber under research, determined) that the KW grade corresponded to C30, KS corresponded to C24, and KG corresponded to C18, while transposition made on the basis of density assigned KW to C45, KS to C40, and KG to C30. Considering the obvious differences in the transposition of visual strength grading classes to the C strength classes, depending on timber origin and its physical and mechanical parameters, it would be convenient to perform resistance tests for a larger set of timber from different forestry regions. The final result of such research would consist of a transposition proposal between the Polish grades (KW, KS, and KG) and the universal system of C strength classes (in line with EN 1912).

1.
The calculated characteristic values of density, the modulus of elasticity, and bending strength allowed us to assign the visual grades KW, KS, and KG to C35, C30, and C20, respectively. 2.
The Scots pine timber under research had high physico-mechanical parameters that translated into high C classes to which the KW, KS, and KG timber grades were assigned.

3.
Despite the classification of timber into strength grades (KW, KS, KG, and Reject), there was still a high variance of the tested properties (wood density, MOR, and E 0 ). This proves that the visual strength grading method is not perfect.