## Wednesday, 5 October 2016

### Compressive strength parallel to the fiber of spruce with high moisture content

Published Date
Volume 74, Issue 4, pp 527–542

Title

# Compressive strength parallel to the fiber of spruce with high moisture content

• Gordian Stapf

3 Review of published tests

Materials, specimens, and test program.

The conducted test program was designed to establish a consistent verification of the factor ${k}_{moistgreen}$ for small clear and full-scale round wood piles, at the mean and 5 % quantile level. In order to reveal the effect of moisture on strength between the 12 % moisture content and the green state, matched specimens (cut in a specific manner from individual logs, see below) were used at the small and the full-scale pile level

### 4.1 Log material

The test material comprised 2  21 log/stem-segments from individual trees which were purchased from a selected forest stand in Southern Germany, close to Stuttgart. The grade of the log-segments conformed to NEN 5491 (1999) with maximum knot diameters of 19 to 42 mm and knot area ratios of 0.11 to 0.31, hereby also conforming to Güteklasse II according to DIN 4074-2 (1958) and to Güteklasse B according to RVR (2015). The log segments, hereby known as “logs” only had a length of 3.5 m each and the average bottom and top diameters were ${d}_{bottommean}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}230\phantom{\rule{thickmathspace}{0ex}}\text{mm}$ and ${d}_{topmean}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}185\phantom{\rule{thickmathspace}{0ex}}\text{mm}$, respectively. The felling of the trees had been undertaken one month prior to purchase; between felling and delivery to the Institute, the trees were stored unsheltered at the sawmill. The logs were received in the green state. After grading, 4 pairs were excluded due to excessive knots/bow, so 217 pairs remained for full sized compression testing. One of the rejected logs was used for detailed moisture measurements.
From each log, two structural sized pile specimens with equal lengths of 1.2 m were cut almost successively, leaving only a narrow slice of about 150 mm in between. One of each matched, round wood specimens from each log was tested in wet, water-saturated state (see below) and the other in dry state. From the intermediate log slices, matched small clear specimens for wet and dry testing were taken. Figures 2 and 3 reveal the cutting scheme of the delivered logs into structural sized pile specimens and the cutting scheme of the matched small scale specimens.

### 4.2 Pile specimens

The aspect ratio of the $217$ full-scale pile specimens tested in compression parallel to the fiber was chosen according to the respective provisions in the European test standard for structural sized lumber, EN 408 (2010). The length of the pile specimens was nominally 6 times the smallest diameter of the employed conical log sections. The diameters of the dry and wet pile specimens varied in the range of 170 to 255 mm at the bottom end and 162 to 241 mm at the top end. The average mid-length diameters of the dry and wet tested specimens were 213 and 197 mm, respectively (see Table 4). The specimens tested in the wet state were taken throughout from the thinner, tree top-oriented part of the stem section. This selection procedure, slightly violating the otherwise rigorously pursued matching procedures, is owed to the limited diameter of the pressure-vacuum-vessel that was used to increase the moisture of the wet logs (see below) before they were tested.

### 4.3 Small clear specimens

Two different geometries and slightly different sizes were investigated with regard to effect of moisture on strength. One set of small clear specimens conformed to the aspect ratio provisions given in EN 408 (2010) where the cross-sectional dimensions were 20 mm  20 mm perpendicular to the fiber direction and the length was 120 mm. The other set of specimens conformed to the previous German standard DIN 52185 (1976) for small clear specimens where the much smaller aspect ratio of 2 resulted from a cross-section of 30 mm  30 mm and a length of 60 mm.
The specimens were cut as shown schematically in Figs. 2 and 3 from each intermediate log slice in such a manner that matching was best possible between dry and wet specimens of each specific specimen type (DIN 52185 or EN 408). Further, the spatial arrangement of the specimens should allow for the assessment of the effect of the position within the stem cross-section on the moisture strength relationship. In order to enable this, one group of the specimens was cut from the heart wood area at a distance of 20 mm from the pith and the other group was cut from the sapwood region with a distance of 20 mm from the periphery. The matching of the dry and wet specimens according to EN 408 and DIN 52185 was performed as shown schematically in Figs. 2 and 3. Due to the shorter length of the DIN specimens (60 mm), the matching with regard to effect of moisture was performed by cutting both specimens from one longer stick ($l\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}125\phantom{\rule{thickmathspace}{0ex}}\text{mm}$), enabling the highest possible similarity with regard to annual ring orientation within the cross-section. The matching of the longer sized EN 408 specimens with regard to the moisture effect was performed by cutting the specimens at the same radial distances ${r}_{i}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}20\phantom{\rule{thickmathspace}{0ex}}\text{mm}$ and ${r}_{a}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}D220\phantom{\rule{thickmathspace}{0ex}}\text{mm}$at a circumferential angle distance of $\pi 2$. The 21 logs together with the applied cutting scheme delivered for each geometry configuration (DIN 52185 and EN 408) 42 dry and 42 wet small clear specimens, whereby 50 % were from the heartwood and sapwood, respectively.

Conditioning and moisture content of specimens
The conditioning of the specimens with regard to the target moisture conditions was performed as follows. The small specimens used for tests in dry condition (nominal MC = 12 %), were stored for 12 weeks at a climate of 20 ${}^{}$C and 65 % rh, resulting in an average moisture content of 13.9 ± 0.3 % at test time. The dry pile specimens were first stacked in a climate chamber at 20 ${}^{}$C and 65 % rh for 25 weeks and subsequently subjected to a dryer (35 % rh) climate in the testing hall for 4 weeks. The pile specimens dried to an average MC of 12.5 ± 0.8 %, determined after testing from slab 5 at mid-length (see below), whereby a moisture gradient of about 4 % along the cross-sectional diameter remained.
The small and the pile specimens tested in wet condition, though having been cut from green stems, were vacuum pressure treated with cold water (10 ${}^{}$C) to obtain a higher degree of moisture uniformity over the cross-section. The water treatment comprised 20 successive cycles whereby each cycle consisted of a vacuum (absolute pressure of 20 kPa) held for 20 minutes, followed by 35 min of pressure treatment (absolute pressure: 500 kPa). This procedure resulted in very uniform moisture contents for the small specimens, with an average value of 192 % with variations of  ±20 % over cross-sectional dimensions.
In the case of the pile specimens however, considerable moisture gradients in radial and stem axis direction remained. The gradients in both directions were determined from one spare wet pile specimen (No. 21) used for moisture measurement purposes only. Six cross-sectional slabs, each with a thickness of 30 mm, were cut at different positions along the log axis and then each slab was subdivided into cubic specimens A to E arranged symmetrically to the pith in the slab (Fig. 4). Further, from each of the 17 pile specimens tested in wet state, a moisture control slab—then cut into cubic specimens A to E—was taken at mid-length (position of slab 5, Fig. 4) immediately after testing. The moisture content of the slab specimens was determined by the oven dry method.

Figure 5 shows the cross-sectional moisture distributions of slabs 1 to 6 along the axis of wet pile No. 21. Every slab reveals a very pronounced moisture gradient along the cross-sectional diameter. The outer sapwood area specimens A and E show consistently for all slabs moisture values in the range of 190 to 220 %. Towards the inner heartwood area, the moisture content decreases significantly, being rather radially symmetric apart from slabs 1 and 6 at the pile ends. In the lengthwise inner part of the log (slabs 3 to 5), the moisture content of the centroid (pith) specimen C is roughly 35 % reaching up to 150 % for the end-located slab 1.

Figure 6 presents the cross-sectional moisture distribution of all wet piles as determined from the individual mid-length slabs 5. Aside from pile No. 1, which showed a very high moisture content ($u150\phantom{\rule{thickmathspace}{0ex}}\mathrm{%}$) throughout the cross-section, a pronounced cross-sectional moisture gradient was seen in all piles similar to that of pile No. 21, depicted in Fig. 5. In the center of the stems, moisture specimens B, C, D, the moisture content was between 29 and 151 % with a median at 34 % (pile No. 1 excluded). More detailed, the medians for both specimens B and D were 38 % whereas at mid-width (specimen C), the median moisture content was 32.5 %. For about half of the wet pile specimens, moisture values of at least two adjacent specimens B and C or C and D were in the range of 29 to 34 %. The moisture content at the stem periphery, moisture specimens A and E, was denoted by median and mean values of 144 % and 130 ± 55 %. The three lowest moisture values (in ascending order) were: 36, 41 and 51 %.

In summary, it can be stated that the minimum moisture content of the pile specimens was in majority along pile length and cross-section beyond the chosen fiber saturation limit of 30 %. However, since the saturation level might be higher for some specimens and clear wood tests indicate that compressive strength drops after reaching the upper realistic saturation level of 35 %, the obtained strengths of the wet log specimens should be rather higher than lower compared to long-time water submerged piles.
Compression tests parallel to the fiber
• ### 6.1 Small clear specimens

Prior to testing, the weights and dimensions were measured to determine the density ${\rho }_{u}$ at test time. The axial compression tests parallel to the fiber with the small clear specimens following the provisions in EN 408 (2010) and DIN 52185 (1976), were performed displacement-controlled with a constant rate of cross-head motion of 4 mm/min in a computer controlled electro-mechanical test machine (type: Zwick/Roell Smart.pro 100 kN). The modulus of elasticity was not measured. The tests were conducted at about 20 ${}^{}$C; relative humidity (rh) was in the range of 40 to 55 %. Immediately after testing, all specimens were oven dried for determination of moisture content and densities ${\rho }_{0}$ and ${\rho }_{12}$.

### 6.2 Pile specimens

Prior to testing, all dry and wet pile specimens were weighed and measured (circumferences at top, bottom and mid length). The tests were performed with a computer controlled servo-hydraulic test machine (Ernst HPM 1600) with a load capacity of 1600 kN in a non climatized environment at a temperature of 18 to 21 ${}^{}$C and a relative humidity of about 40 %. Figure 7 shows the test scheme and Fig. 8 gives views of the realized test set-up. The butt-oriented end of the pile specimens was supported unhinged (flat end condition) by a rigid horizontal steel platform. The small/top end was loaded by the piston of the test machine via an intermediate steel plate connected with a spherically supported compression hinge whose rotation could be blocked by screws (Figs. 78b). According to EN 408, the specimens were loaded with an unblocked upper hinge to a preload level of 0.5 % of ${F}_{c0maxestimated}$, after which point the hinge was restrained from further rotation.

For all wet and dry pile specimens, the axial compression displacement was measured over a constant length of 660 mm in the center section of the specimen length by four LVDTs mounted at circumferential distances of $\pi 2$ (see Figs. 78a). Depending on the slightly varying log diameters, the displacement measurement length was in the range of about 2.5 to 3 times of the specimen diameter at mid-length.
The tests were performed displacement-controlled in accordance with EN 408 (2010) at a constant piston speed of 20 mm/min. Failure was obtained within 300 ± 120 s.
Test results for small clear specimens
• ### 7.1 Data and observations

The failure of the rather stout, small clear DIN specimens occurred for both the dry and wet specimens throughout via the well-known inclined kink band formation damage (Roš  1936; Hoffmeyer 1990). In the case of the wet specimens, the kink bands were wider. The load compression deformation curves were linear until approx. 80 % of peak load. After peak load, the stress dropped about 20 % before reaching a pronounced post-peak plateau in many cases. The green specimens could be tested to a much higher strain (often 30 % and more as determined from global displacement) than the dry specimens before total failure. In the case of the more slender EN 408 compression specimens, the failure occurred partly as pure kink band formation, however for about 40 % of the specimens the kink band formation was overlaid by lateral bending associated with large (several mm) deformations perpendicular to the loading axis. Figure 9 shows the lognormal distributions of the compressive strengths of the dry and wet DIN and EN specimens, respectively. Table 2contains the statistical evaluation of all test series regarding compressive strengths, densities and moisture contents. Further, ratios $$f_{c, 0, EN}/f_{c, 0, DIN}$$ for the dry and green moisture state are given to reveal the influence of the different specimen aspect ratios.

Table 2
Compilation of results for matched small clear spruce specimens tested in compression parallel to the fiber in dry and green state according to two different test standards (DIN 52185 and DIN EN 408)
Physical, mechanical property or ratio
Statistical quantity
Small scale clear wood specimens
Dry (MC = 12 %)
Green (MC $$>$$ 30 %)
Specimen type
Specimen type
DIN 52185
EN 408
DIN 52185
EN 408
No. of specimens tested
42
42
42
42
No. of specimens used for evaluation
35
38
35
38
Moisture content in %
$$x_{mean}$$(COV in %)
13.9 (1.9)
13.9 (2.5)
200 (16)
181 (18)
Compressive strength $$f_{c, 0}$$ in MPa
$$x_{mean}$$(std)
44.3 (9.9)
41.6 (8.5)
21.9 (5.4)
20.1 (4.4)
COV in %
22.4
20.3
24.7
22.0
$$x_{min}$$
30.2
29.0
14.6
12.5
$$x_{max}$$
72.4
60.7
36.4
31.2
$$x_{05}$$
29.3
28.4
13.9
13.0
Density $$\rho _{12}$$ in kg/m$$^3$$
$$x_{mean}$$(std)
442 (75)
442 (58)
439 (64)
442 (60)
COV in %
16.9
13.0
14.7
13.6
$$x_{min}$$
349
334
353
348
$$x_{max}$$
653
580
659
604
$$x_{05}$$
326
347
337
344
Compressive strength ratio
$$f_{c, 0, EN}/f_{c, 0, DIN}$$
at $$x_{mean}$$level
0.94
0.95
at $$x_{05}$$level
0.97
0.94
The density results reveal that all test series had a closely conforming density distribution characterized by an average density $$\rho _{12}$$ of about 440 kg/m$$^3$$ and 5 %-quantile values in the range of about 330 to 350 kg/m$$^3$$. The compressive strength ratios $$f_{c, 0, EN}/f_{c, 0, DIN}$$ reveal on average (dry and wet, $$x_{mean}$$ and $$x_{05}$$-level) roughly 5 % lower strengths for the EN specimen shape. This reflects the superimposed bending and buckling effects of the more slender EN specimens. For both specimen types (DIN, EN) the mean compressive strength of the specimens that were taken from the inner part of the cross-sections was 83 % of the compressive strength of the outer specimens.

### 7.2 Evaluation of moisture impact

Figure 10 and the related regression equation
\begin{aligned} f_{c, 0, u=12\%} = 4.72 + 1.83 \cdot f_{c, 0, green} \end{aligned}
(9)
reveal a highly correlated ($$R^2\;=\;0.97$$) relationship of the compressive strength parallel to the fiber of the very closely matched water saturated (green) and dry DIN specimens. The strength data for the dry state ($$u\;=\;12\;\%$$) presented in Fig. 10, Eq. (9) and Table 2 were transformed from the experimentally obtained strength values with consideration of the moisture content of every single specimen ($$u_{mean, dry}\;=\;14 \; \%$$) by Eq. (8b). The relationship for the originally obtained test data $$f_{c, 0, u=14\%} = 4.45 + 1.64 \cdot f_{c, 0, green}$$ delivers about 10 % lower values than Eq. (9). This result coincides well with those obtained from the relationship $$f_{c, 0, 12}/f_{c, 0, u}\;=\;20/(32-u[\%])$$ specified in Kollmann (1955) for compressive strength adjustment of clear spruce/fir in the moisture range of 8 to 18 %.

The revealed high correlation of $$f_{c, 0, green}$$ and $$f_{c, 0, u=12 \; \%}$$ of the almost defect free specimens suggests that the strength decreasing influence of the high moisture content is rather unaffected by the density of the timber. This is substantiated by a regression of the strength ratio $$f_{c, 0, green}/f_{c, 0, 12}$$ vs. density $$\rho _{12}$$ which gives no ($$R^2\;=\;0.2$$) correlation (Note: the EN 408 specimens revealed a slightly different relationship of $$f_{c, 0, u=12\%} = 9.32 + 1.14 f_{c, 0, green}$$ with a considerably worse coefficient of correlation of $$R^2=0.67$$, which is presumably due to the less closely matched specimens).
Table 3 specifies the ratio $$f_{c, 0, green}/f_{c, 0, 12}$$ at the mean, 5 %-quantile and minimum strength level. The results for the DIN and EN specimens conformed closely. On average, $$k_{moist, green}$$-values of about 0.48 were obtained at the $$x_{mean}$$- and $$x_{05}$$-level. The obtained mean moisture modification factors are about 10 % lower as compared to the literature evaluation (Eq. 8b), though agree well, when considering higher moisture contents of $$u =$$ 50 % or higher.
Table 3
Ratios of compressive strength and modulus of elasticity parallel to the fiber at water saturated green state ($$u > 30\;\%$$) vs. dry state ($$u\;=\;12\;\%$$) of small defect free and structural sized pile/round wood specimens
Specimen size
Specimen type
MOR and MOE ratios
$$k_{moist, green}\;=\;\frac{f_{c, 0, green}}{f_{c, 0, dry}}$$
$$\dfrac{E_{c, 0, green}}{E_{c, 0, dry}}$$
$$x_{mean}$$
$$x_{05}$$
$$x_{min}$$
$$x_{mean}$$
Small defect free
DIN 52185
0.49
0.47
0.48
EN 408
0.48
0.46
0.43
Structural sized
Round wood
0.57
0.66
0.60
0.89
Test results for the pile specimens
• ### 8.1 Data and observations

The load displacement behavior of the dry pile specimens was strictly linear up to about 85 to 90 % of the maximum compression load ${F}_{c0max}$, whereas for the wet specimens, the linear range ended somewhat lower at about 75 to 85 % of ${F}_{c0max}$. Beyond the linear stiffness region, progressive nonlinearity occurred throughout until peak load. For both dry and wet specimens, stable strain softening with gradually decreasing load occurred after peak load. In the case of the wet specimens, the quasi plastic peak load plateau was more expressed and the slope of the compressive stress-strain curves on the descending branch was not as steep. Figure 11 a and b show typical stress-strain curves encountered for the dry and wet specimens, respectively.

The failure of the pile specimens showed in principle similar features for the case of the dry and wet logs, however the ductility of the local and global failure mechanisms was far more expressed in the case of the wet specimens. In general, three different failure mechanisms occurred in a successive manner, first in the nonlinear pre-peak range and then in the softening post-peak domain, being (1) splitting parallel to the fiber axis starting often but not generally at a knot, (2) kink band formation as in the case of the small clear specimens and (3) localized buckling of a larger stem part at the periphery. Figures 12 and 13 show examples of the observed failures.

Figure 14 gives the compressive strength results of the individual dry and wet pile specimens as cumulative frequencies together with the fitted lognormal distribution functions. The strength values were derived from peak load and pile diameters at failure plane. Table 4presents the statistical evaluation of the experimental data regarding compressive strength, modulus of elasticity and density. Similarly, as in the case of the small clear specimens, the densities of the dry and wet pile specimens due to the performed matching procedure conform closely. The average densities ${\rho }_{12}$ of the dry and wet sample are 453 and 447 kg/m${}^{3}$, respectively, and fit very well with the density results of the clear wood specimens, especially when the correction factor of (1/1.05) according to EN 384 (2010) accounting for the density contribution of the knots is applied to the stem specimens.

Table 4
Statistical evaluation of test results MOR, MOE and density of structural sized spruce pile specimens tested in axial compression at dry and wet (water saturated) state
Physical, mechanical property or ratio
Statistical quantity
Moisture state of specimen
Ratio
Dry
Wet
Wet/dry
No. of specimens tested
17
17
17
Moisture content in %
${x}_{mean}$
12.5
89.4
Pile diameter at failure plane ${d}_{f}$in mm
${x}_{mean}$
214
196
Bottom/top diameter ${d}_{bottomtop}$in mm
${x}_{mean}$
220/205
204/190
Compressive strength ${f}_{c0}$ in MPaa
${x}_{mean}$
30.8
17.6
0.57
std(x)
6.1
2.3
COV in %
19.8
13.1
0.66
${x}_{05}$
20.3
13.4
0.66
${x}_{min}$
21.9
13.3
0.60
${x}_{max}$
39.9
21.0
0.53
Density ${\rho }_{12}$ in kg/m${}^{3}$
${x}_{mean}$
453 (48)
447 (34)
0.99
COV in %
10.6
7.5
0.71
${x}_{min}$
398
389
0.98
${x}_{max}$
566
550
0.97
${x}_{05}$
370
387
1.05
MOE E c ,0 a in GPa
${x}_{mean}$
12.1
10.7
0.89
std(x)
1.9
1.6
COV in %
15.9
14.9
0.93
${x}_{05}$
8.6
7.9
0.92
${x}_{min}$
8.9
7.7
0.87
${x}_{max}$
15.0
14.3
0.95
aDry values adjusted from experimental MC to MC = 12 % by provisions given in EN 384 (2010), i.e. 3 % and 1 % change per 1 % MC change in case of MOR and MOE, respectively
The values obtained for the 5 % quantile of compressive strength, mean modulus of elasticity and characteristic density of the dry round wood specimens ${f}_{c005}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}21.5\phantom{\rule{thickmathspace}{0ex}}\text{MPa}$${E}_{c0mean}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}12.1\phantom{\rule{thickmathspace}{0ex}}\text{GPa}$ and ${\rho }_{1205}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}350\phantom{\rule{thickmathspace}{0ex}}{\text{kg/m}}^{3}$ conform very well with the numbers specified for strength class C24 of the European strength class standard EN 338 (2010) for structural timber, which designates values of ${f}_{c0k}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}21\phantom{\rule{thickmathspace}{0ex}}\text{MPa}$${E}_{0mean}$ = 11 GPa, ${\rho }_{mean}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}420\phantom{\rule{thickmathspace}{0ex}}{\text{kg/m}}^{3}$ and ${\rho }_{k}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}350$ kg/m3 (Note: The standardized values refer implicitly to sawn timber with rectangular cross-sections; for round wood no European strength class standard exists).

### 8.2 Evaluation of moisture impact

The ratios of compressive strength and MOE of wet ($u30\phantom{\rule{thickmathspace}{0ex}}\mathrm{%}$) vs. dry ($u\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}12\phantom{\rule{thickmathspace}{0ex}}\mathrm{%}$) state are given in Table 3 for relevant statistical levels (MOE: ${x}_{mean}$ only). The moisture modification factor at the mean strength level

$\begin{array}{r}{k}_{moistmeangreen}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}\frac{{f}_{c0meangreen}}{{f}_{c0drymean}}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}0.57\end{array}$
(10)
is roughly 15 % higher as compared to the clear wood samples. The higher ${k}_{moistmeangreen}$possibly results from the lower moisture content of the wet specimens that just reaches 30 % in the inner cross-sections of most wet piles, see Sect. 4.1. Since strength reductions over 30 % moisture content are debatable and against the contemporary technical understanding, the results will not be adjusted in the following.
It should be kept in mind, however, that even higher reduction factors could result from future research.
In contrast to the small clear specimens, where the strength ratios of wet to dry specimens were rather similar at different statistical distribution levels, the ratio at the lower tail of the distributions is now considerably higher, i.e.

$\begin{array}{r}{k}_{moist05green}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}\frac{{f}_{c005green}}{{f}_{c005dry}}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}0.66.\end{array}$
(11)
The mathematical explanation for the increased value for ${k}_{moist05green}$ is related to a considerably lower COV of 13.1 % of the wet pile specimens as compared to a COV = 19.8 % of the dry pile specimens. This is in contrast to the small clear specimens where the wet specimens gave on average about 12 % higher COVs. One possible reason for the reduced scatter of structural sized round wood/pile specimens in wet state as compared to dry conditions and the hereof resulting increase of ratio ${f}_{c0green}{f}_{c0dry}$ at the ${x}_{05}$-level might be that the tree branches allow for a better distribution of the stresses, especially in the fully water saturated material state, leading to a more homogeneously behaving branch-reinforced bulk composite. Although the higher strength ratios at the ${x}_{05}$ and ${x}_{min}$ level seem plausible, there remains the uncertainty that the reason is primarily statistical due to the limited number of specimens.
With regard to modulus of elasticity, the moisture bound decrease obtained at the mean level (see Table 3) was

$\begin{array}{r}\frac{{E}_{c0meangreen}}{{E}_{c0drymean}}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}0.89\end{array}$
and on the ${x}_{05}$ and ${x}_{min}$ level very similar ratios of 0.90 and 0.87, respectively, were determined. The results conform closely to data by Kühne et al. (1955) and Fischer (1960) on the moisture-MOE relationship of spruce and fir in compression parallel to fiber resulting in a similar ratio of 0.89 at the mean level. Finally, it can be stated that the experimentally obtained ratios of compressive strength parallel to the fiber and MOE for water saturated vs. dry structural sized coniferous (European spruce) pile specimens comply well with respective adjustment factors specified in the North American standards and design provisions for round wood piles.
• Conclusion
• The conducted work confirmed that a moisture modification factor of 0.8 as specified implicitly in the current Eurocode 5 for service class 3 conditions is too little of a reduction for the compression strength parallel to the fiber of spruce and fir in the very wet, fiber saturated state. The test results for clear wood—together with a comprehensive literature evaluation—gave a moisture reduction factor in the range of 0.5 to 0.55 at the mean strength level. The investigations with well matched wet and dry structural sized pile specimens revealed a somewhat less pronounced strength decrease of about 35 % vs. the dry state (MC = 12 %) on the 5 % quantile level. Hence, for design of wet, water saturated piles and sawn softwood lumber, a pure moisture modification factor of ${k}_{moistgreen}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}0.65$ seems appropriate.
The superimposed impact of moisture and loading time on strength is considered in several design codes product-wise. Based on the time effect for permanent duration of load, which can be derived from Eurocode 5 as ${k}_{time}0.6$, at present irrespective of moisture level, a modification factor ${k}_{mod}\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}0.650.6\phantom{\rule{thickmathspace}{0ex}}\phantom{\rule{thickmathspace}{0ex}}0.4$ would result for lumber and piles subjected to permanent duration of load in the water saturated state. A value of this order of magnitude is proposed as a safe adjustment factor for a revised version of Eurocode 5. A more moderate value of ${k}_{mod}$ in the range of 0.5 might be justifiable if proven by adequate experimental and theoretical approaches i.a. including FORM/SORM (First/Second Order Reliability Method) calibration.
Although the above is derivation-wise rigorously confined to European spruce and fir, it may be assumed to hold for other softwood species like larch and Douglas fir as well. In the case of modulus of elasticity of fiber saturated clear and structural sized softwood timbers loaded in compression parallel to fiber, a reduction factor of 0.85 is recommended.
• References
1. ANSI/AWC NDS-2015 (2015) National design specification (NDS) for wood construction 2015 edition. American National Standards Institute (ANSI), Washington, D.C., USA and American Wood Council (AWC), Leesburg, VA, USA
2. ASTM D2555—06 (Reapproved 2011) (2011) Standard practice for establishing clear wood strength values. ASTM International, West Conshohocken, PA, USA
3. ASTM D2899—12 (2012) Standard practice for establishing allowable stresses for round timber piles. ASTM International, West Conshohocken, PA, USA
4. ASTM D3200—74 (Reapproved 2012) (2012) Standard specification and test method for establishing recommended design stresses for round timber construction poles. ASTM International, West Conshohocken, PA, USA
5. Austrian Institute of Construction Engineering (OiB) (2013a) VIGAM-Glued laminated timber of oak. ETA-13/0642 European Technical Approval, valid until 27.06.2018
6. Austrian Institute of Construction Engineering (OiB) (2013b) Sierolam - Glued laminated timber of chestnut. ETA-13/0646 European Technical approval, valid until 27.06.2018
7. Buiten H, Rijsdijk JF (1982) Compressive strength of larch, douglas fir and spruce piles (in dutch). Rapport HI 82.1140, Houtinstituut TNO, Postbus 151, 2600 AD Delft, order Nr.: 30.01.1.0002
9. DIN 1052-1 (1988) Timber structures; design and construction. German Institute for Standardization (DIN), Berlin, Germany
10. DIN 4074-2 (1958) Building timber for wood building components; quality conditions for building logs (softwood). German Institute for Standardization (DIN), Berlin, Germany
11. DIN 52185 (1976) Testing of wood; compression test parallel to grain. German Institute for Standardization (DIN), Berlin, Germany
12. EN 14358 (2007) Timber structures—calculation of characteristic 5-percentile values and acceptance criteria for a sample. European Committee for Standardization (CEN), Brussels, Belgium
13. EN 1995-1-1 (2010) +AC (2006)+A1 (2008) Eurocode 5: design of timber structures—part 1-1: General—common rules and rules for buildings. European Committee for Standardization (CEN), Brussels, Belgium
14. EN 338 (2010) Structural timber—strength classes. European Committee for Standardization (CEN), Brussels, Belgium
15. EN 384 (2010) Structural timber—Determination of characteristic values of mechanical properties and density state. European Committee for Standardization (CEN), Brussels, Belgium
16. EN 408 (2010)+A1 (2012) Timber structures—structural timber and glued laminated timber—determination of some physical and mechanical properties. European Committee for Standardization (CEN), Brussels, Belgium
17. Fischer H, Kühne H (1960) Buckling strength of swiss spruce, fir, larch, red beech and oak (in German). Bericht 190, EMPA
18. Frese M, Enders-Comberg M, Blaß HJ, Glos P (2011) Strength of spruce glulam subjected to longitudinal compression. International Council for Research and Innovation in Building and Construction, Working Commission W18—Timber Structures, Alghero, Italy, paper 44–12–2
19. Glass SV, Zelinka SL (2010) Wood handbook—wood as an engineering material (2010 Edition), Forest Products Society, Madison, Wisconsin, USA, chap Moisture relations and physical properties of wood, pp 4–1–4–19
20. Glos P (1978) Determination of the strength behavior of glulam under compressive stress from material and action parameters (in German). PhD thesis, Technische Universität München, Fachbereich Bauingenieur-und Vermessungswesen
21. Hoffmeyer P (1990) Failure of wood as influenced by moisture and duration of load. PhD thesis, State University of New York, College of environmental Science and Forestry
22. ISO 20152-1 (2010) Timber structures—bond performance of adhesives—part 1:basic requirements. International Organization for Standardization (ISO), Geneva, Switzerland
23. Jorissen A (2007) Module 7: the strength of foundation piles (in dutch). In: SHR Course Timber Research during Foundation Inspections. SHR Timber Research Wageningen, The Netherlands
24. Kollmann F (1955) Technology of wood and wood-based materials. Springer, Berlin, Germany
25. Kühne H, Fischer H, Vodoz J, Wagner T (1955) Influence of water content, bulk density, fiber position and tree ring position on the strength and ductility of swiss spruce, fir, larch, red beech and oak (in German). Report 183, Swiss Federal Laboratories for Materials Science and Technology (EMPA), Zurich, Switzerland
26. NEN 5491 (1999) Quality requirements for timber—piles—coniferous timber. Netherlands Standardisation Institute (NEN), Delft, Netherlands
27. NEN 6760 (2008) TGB 1990—timber structures—basic requirements and determination methods. Netherlands Standardisation Institute (NEN), Delft, Netherlands
28. NEN-EN 1990+A1+A1/C2 (2011) Eurocode: basis of structural design. Netherlands Standardisation Institute (NEN), Delft, Netherlands
29. NEN-EN 1995-1-1 NB (2011) National annex to NEN-EN 1995-1-1. Eurocode 5: design of timber structures—part 1-1: General—common rules and rules for buildings (includes C1:2006 and A1:2008). Netherlands Standardisation Institute (NEN), Delft, Netherlands
30. NZS 3603 (1993) Timber structures standard. Amendments No 1 & 2 & 4 Appended. Standards New Zealand, Wellington, New Zealand
31. Roš M (1936) Wood as a building material (in German). 1. Schweiz. Kongress zur Förderung der Holzverwertung, Bern, Switzerland
32. RVR (2015) Framework agreement for lumber trade in Germany (RVR) (in German). German Forestry Council (DFWR) and German Wood Council (DHWR), Berlin, Germany
33. van de Kuilen JWG (1994) Determination of the characteristic compressive strength of timber piles. TNO-report 94-CON-R0271, TNO
• For further details log on website :