An Assessment of Environmental Impact on Glued Wood Building Elements

• The research investigated the impact of environmental factors (temperature and humidity) on pine glulam, oak glulam, and laminated veneer lumber (LVL) elements, all of which can be used in building structures. Elements underwent freezing, heating, drying, and wetting processes in different modes, thereby simulating different environmental conditions that could be encountered during the service period of the materials. Their mechanical properties (modulus of elasticity - MOE and coef ﬁ cient of damping) were recorded at each stage. It was determined that, in the case of dry construction elements (where a moisture content was between 7.0 and 14.0 %), the MOE increases by a few percentage points with decreasing temperature and humidity levels, and decreases with increasing temperature and humidity levels. The coef ﬁ cient of damping varied by 20 % - in most cases, when the modulus of elasticity increased, this decreased, and vice versa. Under extreme environmental changes (with the elements being soaked, frozen at -25 °C, and dried at 40 °C), the MOE of the glued timber decreased by 16 % when this parameter of LVL decreased by about 10 %. Alterations in viscous properties produced similar results (the coef ﬁ cient of damping increased by 50 % for the glued timber and by 66 % for the LVL). This is explained by the partial destruction of the element structure, the occurrence of cracks, and the decreased anisotropy of the LVL structure.


UVOD
Timber and timber elements are wide ly used in building structures (Rindler et  . The production of various timber articles (laminated timber) has a purpose both in economic and ecological terms. In order to reduce emissions it is appropriate to use timber materials instead of concrete and steel structures. The demand for these materials is projected to continuously increase (Risse et al., 2019, Hildebrandt et al., 2017. The construction of buildings with more than two fl oors is one of the latest trends in which more timber products could be used (Markström et al., 2018).
Constructions that are made using glued timber materials, such as laminated veneer lumber (LVL) and cross laminated timber (CLT), can be used to reinforce concrete and steel structures and ensure better protection levels against earthquakes (Sanscartier Pilon et al., 2019).
More lately, identifying, improving, and forecasting the properties of timber structures has become a topical subject. It is already known that, in most cases, timber becomes wet or dry when in service under the changing ambient temperature and humidity levels. Alterations in timber moisture content may result in alterations in terms of dimensions and shape of timber materials, or the occurrence of various defects and changes in mechanical properties (Wood Handbook, 2010). It has already been established that an increase in temperature from -30 °C to +30 °C will decrease the MOE in timber materials (Ayrilmis et al., 2010).
The impact of different factors on alterations in the moisture content of timber materials under ambient conditions was also something that has previously been investigated. The absorption properties of timber materials can be altered by applying thermal modifi cations, along with wax, oil, and biocide treatments (Žlahtič-Zupanc et al., 2018). It has been determined that the moisture content of glued timber materialswhich changes under the infl uence of the ambient temperature -also alters its mechanical properties. In addition, the change in these properties also depends on the glue that has been used (Rindler et al., 2018). The mechanical properties of one of the most commonly used construction materials -LVL -can be improved by using other materials, such as polymeric options (Subhani et al., 2017). Thanks to the digital model developed, it is possible to predict the LVL mechanical properties in various directions and to produce beams of the desired mechanical properties (Gilbert et al., 2017). Building structures often require holes to be made in the materials. It has been determined that both the diameter of those holes and, especially, their location affect the strength properties and durability of LVL beams (Musselman et al., 2018).
The aim of the study is to evaluate the impact of environment factors on the mechanical properties of glulam beams in both oak and pine, as well as LVL beams.

MATERIJALI I METODE
The research study used glued oak and pine glulam beams (each glued segment consisting of fi ve planks) and LVL beams (the thickness of the layer was about 3 mm) with the dimensions of 1200 mm × 75 mm ×100 mm, a humidity level was between 10.5-14.0 %, and the density was accordingly within the limits of 740-780 kg/m 3 , 520-550 kg/m 3 , and 590-670 kg/m 3 , respectively. The glulam beams were glued by using water-resistant polyurethane glues, and the LVL specimens were glued by using formaldehyde-based glues. There were 10 beams of each type. A schematic for specimen cross-sections is provided in Figure 1.
In order to simulate changes in environmental parameters, the climatic conditions were modelled in the climatic chamber (the appropriate air temperature and relative humidity are determined in the chamber, all samples are stacked in chamber and the change in their mechanical properties is recorded every 7 days). In the chamber, the temperature was maintained within an accuracy level of 1 °C, and the relative humidity level was maintained within an accuracy level of 2 %. The length of the specimens was measured using a tape measure within an accuracy of 1 mm, while the width and thickness were measured using sliding callipers within an accuracy of 0.05 mm, and the mass was measured using a balance within an accuracy of 0.1 g. The moisture content for the specimens was measured with an electric resistance moisture meter. 4 cycles were simulated: "cold" -samples frozen in air at -25 °C, "dry" -samples kept at 40 °C at 40 % relative humidity, "wet" -samples irrigated at 20 °C at 85 % humidity and "extreme" -the samples were soaked in water at 20 °C, frozen in air at -25 °C and dried at 40 °C and 40 % relative humidity.
Before carrying out the research study, all specimens were conditioned in a climatic chamber for a total of fourteen days at 20 °C and with a relative air a b humidity of 60 %. After completing the conditioning process, the moisture content of the specimens was measured at between 11.0-12.5 %.
The special test stand ( Figure 2) was used to determine the MOE and the coeffi cient of damping on the basis of the non-destructive testing (transverse resonant vibrations) method, which also allowed assessing the mechanical properties of the specimens (Albrektas and Vobolis, 2003; Albrektas and Vobolis, 2004;Timoshenko et al., 1985). The studies were performed at a frequency of 20-2000 Hz.
The MOE was calculated based on the following Eq. 1 (Timoshenko et al., 1985): Where: E-modulus of elasticity, f rez -frequency of transverse vibrations, ρ -density of wood, s -crosssectional area, l -beam length, I -cross-sectional moment of inertia, A -method of fastening represented by a coeffi cient The viscous properties of studied specimens were evaluated based on the following Eq. 2 (2) Where: f rez -frequency of transverse vibrations, Δf -frequency bandwidth when amplitude of vibrations decreases by 0.7 times.
The MOE and the coeffi cient of damping for each specimen were determined in two directions: with the glue joints on the specimens orientated as shown in Figure 1 and the specimen turned at an angle of 90°.
For example, the oak glulam beam that falls within code 'O1', should be coded 'O1.1' when placed on a test stand as shown in Figure 1, and 'O1.2' when turned at an angle of 90°. The average values for all ten oak timber beams are marked as 'OX.1' and 'OX.2', respectively. Meanwhile, the groups into which the pine glulam beams have been organised are marked as 'PX.1' and 'PX.2', respectively, and the LVL specimens are marked as 'LX.1' and 'LX.2', respectively.

REZULTATI I RASPRAVA
The measurements for the MOE and the coefficient of damping for the specimens was carried out after completing the conditioning process. These values are displayed in Table 1.
It was estimated that the MOE of oak glulam beams altered within the limits of 11500-14200 MPa, and for pine glulam beams it was within the limits of It should be noted that the MOE of the LVL specimens is slightly higher, although it corresponds to the MOE of natural wood. In addition, having carried out a statistical analysis of the MOE and coeffi cient of damping values for each group, it was found that the coefficient of variation for the MOE values in the LVL specimens was approximately 2 %, and the coeffi cient of variation of the coeffi cient of damping was approximately 5 %. These coeffi cients of variation for oak and pine specimens altered within the limits of 6-8 %.
In order to simulate the various operating conditions of their structures, the specimens underwent a freezing phase at -25 °C and were observed for changes in their mechanical properties. Variations in the MOE are displayed in Figure 3.
The results show that, during a freezing phase, the MOE increased for all specimens. The most significant increase was recorded after the fi rst seven days. For the oak specimens the increase amounted to 1.5 %, for the pine specimens it was 4.9 %, and for the LVL specimens it was 5.5 %. After a period of 21 days, the MOE for the oak specimens increased by 3.1 %, for the pine specimens by 5.5 %, and for the LVL by 6.6 % in comparison to the original value. In all cases the increase in the MOE for group '1' is higher than that of   ., 2010). The viscous properties of the specimens also altered during the freezing phase. The variations in the coeffi cient of damping are shown in Figure 4.
It is clear that the coeffi cient of damping for all specimens tended to decrease during the freezing phase, i.e. the specimens lost plasticity. The most signifi cant decrease was recorded after the fi rst seven days. The most signifi cant decrease in the coeffi cient of damping was observed in pine specimens (with an average decrease of 17.4 %), while the least signifi cant decrease was observed in the LVL specimens (at an average of 7.1 %). During further freezing, the coeffi cient of damping decreased less. For pine specimens, it decreased by an average of 20.0 %, for oak specimens by an average of 16.5 %, and for the LVL specimens by an average of 10.6 %. Unlike the MOE, the coeffi cient of damping in oak and pine specimens was particularly unaffected by the orientation of the glue joint. For the LVL specimens alone, the average coeffi cient of damping of group '1' specimens decreased by 10.6 %, and by 5.1 % for the specimens in group '2'.
Afterwards, the specimens underwent a drying phase at 40 °C, at 40 % relative air humidity, and with about 1 m/s forced circulation. During the drying process, the mass of the specimens decreased by between  Figure 5. It was established that, at the beginning of the drying phase (after the fi rst seven days), the MOE for the specimens decreased (by 5.9 % for oak specimens, 6.6 % for pine specimens, and 12.7 % for LVL specimens). This could be explained by the fact that the moisture content decreased by an insignifi cant level during the fi rst seven days (the specimen mass decreased on average by about 52 g and the average moisture content decreased by about 1.0-1.2 %); however, the timber warmed up and became more fl exible. Later, during the drying process, the mass of the specimen decreased, and the MOE increased at a similar rate. When compared to the MOE of the frozen specimens before drying, after a period of 21 days the average MOE for the oak specimens decreased by 2.1 %, for the pine specimens by 3.6 %, while the LVL specimens decreased by 9.3 %. In this case the orientation of the glue joint did not have any signifi cant impact (the difference did not exceed 1 %).
The variation in the coeffi cient of damping during the drying process is shown in Figure 6.
The results show that, after the fi rst seven days of drying, the coeffi cient of damping for all specimens increased (by 13 % for the pine and the LVL specimens, and by 9 % for the oak specimens); this was due to the  Coeff. of damping, r. u. koeficijent prigušenja, r.u.

Figure 6
Variation in coeffi cient of damping for specimens during the drying process Slika 6. Varijacije koefi cijenta prigušenja na uzorcima tijekom faze sušenja fact that the timber warmed up and became more fl exible. During further drying, the coeffi cient of damping decreased by an insignifi cant amount. When compared to the primary coeffi cient of damping (after freezing), the coeffi cient of damping for all specimens, which were dried for 21 days, managed to increase by between 2.0-6.9 %. The specimens were dampened in a climate chamber at 20 °C and at a relative air humidity of 85 %. The variation in the MOE during the wetting process is shown in Figure 7.
It was found that the MOE for all specimens decreased by an insignifi cant amount during the wetting process. The greatest decrease in the MOE was observed in the LX.1 specimen group (by an average of 3.8 % during the fi rst seven days and by an average of 5.5 % during a period of 21 days). Variations for other groups of specimens were found to be 3.0 % and between 4.0-4.7 %, respectively. The differences are very small (the mass of the specimens pretty much increased in an identical manner); therefore it can be stated that the humid air had an equal effect on the elastic properties of all of the specimens. The variations in viscous properties (the coeffi cient of damping) are shown in Figure 8.
It was established that the coeffi cient of damping for the specimens increased during the wetting process. The highest growth levels were observed during the fi rst seven days (for all groups of specimens the coeffi cient of damping increased by between 10-19 %). The lowest growth in coeffi cient of damping after 21 days was observed for the LVL specimens (14.7 % and 15.7 %, respectively, for individual groups). The average growth of the coeffi cient of damping in the oak specimens was at 21.4 % and 24.0 %, respectively, whereas the coeffi cient of damping for the pine specimens was at 18.6 % and 24.2 %. It is likely that the slightly different behaviour of the LVL specimens was prompted by the relatively large number of glue joints, which ensured a more uniform structure in the entire specimen.
The specimens were then immersed in water at a temperature of 20 °C and were then stored for a period of fourteen days; then they were removed from the water and stored at -25 °C for fourteen days; subsequently they were dried at 40 °C for fourteen days with a relative air humidity of 40 % and forced air circulation. This was done in order to intentionally cause defects in the specimens.
Pictures of several specimens after the immersion/freezing/drying process are shown in Figure 9.
It is evident that the drastic environmental impact in most cases caused a serious level of damage to the glulam specimens tested. Clearly-visible cracks emerged when exposed to sources of tension. Visually signifi cant defects were not observed in the LVL specimens. In most cases there were shallow cracks (singlelayer thickness) that became visible on the face. Small peelings on the cross section were also observed. Table 2 shows the values for the mechanical properties of the specimens following the conditioning process (initial) and after the immersion/freezing/drying process (fi nal).
It was established that the extreme humidity and temperature fl uctuations had a signifi cant impact on the specimens. The results show that the cracks that appeared resulted in a decrease in the MOE of glulam beams by between 13-16 %, while for the LVL specimens this value was a little lower at around 10 %. However, the coeffi cient of damping for the specimens changed much more signifi cantly. The uniform struca b c d e f Figure 9 Examples of typical defects in specimens: a) face of oak specimen; b) cross section of oak specimen; c) face of pine specimen; d) cross section of pine specimen; e) face of LVL specimen; and f) cross section of LVL specimen Slika 9. Primjeri tipičnih grešaka na uzorcima: a) lice uzoraka hrastovine; b) poprečni presjek uzoraka hrastovine; c) lice uzoraka borovine; d) poprečni presjek uzoraka borovine; e) lice LVL uzoraka; f) poprečni presjek LVL uzoraka ture of the specimens was destroyed, causing the specimen that underwent vibration treatment to behave like a system involving several bodies rather than as a single body. The coeffi cient of damping for the natural timber specimens increased by about 50 % and the one for the LVL specimens increased even more.
All of the results were statistically processed. The coeffi cient of variation for the MOE and the coeffi cient of damping for different groups after each test cycle did not exceed 9.6 % (Pekarskas, 2007).

CONCLUSIONS 4. ZAKLJUČAK
It was established that, due to different mechanical properties of timber in different directions (perpendicular and parallel to the positioning directions of the joints), the same glulam beam had different mechanical properties. This difference for individual glulam beams was up to 4.8 %.
During freezing process, the dry glulam beams (at between 11.0-12.5 %) become more elastic. This is shown by the MOE which increased by 6.6 % and by the coeffi cient of damping which decreased by 20.0 %.
After starting the drying process in frozen specimens, during the fi rst seven days their moisture content decreased by an insignifi cant amount (by 1.2 %). However, the timber warmed up and became more fl exible. Therefore, during the period of seven days of drying, the MOE decreased by 12.7 %, whereas the coeffi cient of damping increased by 13.0 %.
As the moisture content of glulam beams decreases, its MOE increases and the coeffi cient of damping decreases (with a moisture content reduction of about 3 % the MOE increased by an average of between 2.8-3.2 %, whereas the coeffi cient of damping decreased by between 2.3-8.5 %).
During the process of wetting, glulam specimens in the air had a moisture content that increased by an average of between 3.0-4.0 %, while their MOE decreased by 5.5 %, and the coeffi cient of damping increased by 24.2 %. This is mainly due to the increase in surface humidity and plasticity.
Coeffi cient of damping for the LVL specimens increased the least only during the wetting process (on average by 15.7 %). It is likely that the slightly different behaviour of the LVL specimens was prompted by the relatively large number of glue joints, which ensured a more uniform structure in the entire specimen.
Freezing wet glulam specimens and their subsequent drying (which served to cause the defects) showed the lowest levels of deterioration in LVL elastic properties (by an average of 10.4 %); however, the viscous properties increased signifi cantly (the coefficient of damping increased by 66.1 %).
In general it can be stated that the glulam beams and the LVL beams reacted similarly to changing environmental conditions.