Compression Behaviors of Parallel Bamboo Strand Lumber Under Static Loading

In order to investigate the influence of length and compression directions upon behaviour of parallel bamboo strand lumber (PBSL) specimens, 240 axial compression tests have been performed. With three similar one different typical failure modes, the mechanical performance for PBSL specimens under compression parallel to grain and perpendicular to grain are different as a whole. From the point of the characteristic values, the compression strength parallel to grain is 2.1 times of the compression strength perpendicular to grain. The elastic modulus for compression parallel to grain is 3.64 times of the compression strength perpendicular to grain. While the compression ratios along two compression directions are equal to each other. The bigger Poisson ratios for one typical side surface is 3.93 times of that for another typical side surface for PBSL specimens under compression perpendicular to grain, and the bigger value is equal to that for PBSL specimens under compression parallel to grain. Length can influence the mechanical properties of the PBSL specimens. The size 50 mm × 50 mm × 100 mm should be good choice for the standard or code to measure the compression strength. PBSL materials have better ductility under compression parallel to grain than that under compression perpendicular to grain. Stress-strain relationship models were proposed for PBSL specimens under compression parallel to grain and perpendicular to grain, respectively. These proposed models gave a good agreement with the test results.

Considering some influencing factors, the mechanical properties for PBSL have been studied. Huang [18] investigated the basic material performance and the aging resistant performance of PBSL. Liang Cheng [19] has studied the manufacturing technology for PBSL considering the bamboo bundle preparation, glue immersion, forming and hot pressing process. Naresworo Nugroho et al. [20][21] have performed a study to determine the suitability of zephyr strand from moso bamboo (Pyllostachys pubescens Mazel) for structural composite board manufacture. Cui et al. [22] found that the presence of prestress in bamboo slivers reinforced parallel strand lumber (PSL) could enhance its flexural stiffness and toughness but not the flexural strength. Pannipa Malanit et al. [23] found that the resin type has a significant effect on oriented strand lumber (OSL) board properties. Zhou et al. [24] studied the mode-I fracture properties of PBSL. Shangguan et al. [25] proposed strength models of PBSL for compressive properties. Zhong et al. [26][27] investigated the effects of temperature on the compressive strength parallel to the grain of PBSL as well as its design value. Ahmad et al. [28] have analyzed the physical and mechanical properties of parallel strand lumber (PSL) made from Calcutta bamboo. Xu et al. [29] have studied compressive and tensile properties of PBSL at elevated temperatures. Yu et al. [30][31] have carried out the experiments about the preparation, physical, mechanical, and interfacial morphological properties of PBSL. Considering some influencing factors, the mechanical performance for PBSL columns [16][17] and beams [32][33][34] were also investigated by the researchers.
As discussed above, even though the mechanical properties for PBSL have been studied considering many influencing factors, a good constitutive model for PBSL is still lack. None specific test standard for PBSL in structural use existed in the world. A good size for measuring the compression strength is very important to write the code. To achieve these aims, the study investigates in detail how length and compression directions influence upon behaviour of parallel bamboo strand lumber (PBSL) specimens.

Materials and Test Methods
The source Moso bamboo (Phyllostachys pubescens, from Feng-xin county in the Jiang-xi province) was harvested at the age of 3-4 years. Bamboo strips from the upper growth heights, of a 2000 mm tall culm were selected. The culms cut from assigned growth portions were then split into 20 mm wide strips, and the outer skin (epidermal) and inner cavity layer (pith peripheral) were removed using a planer. All culm strips were split into bamboo filament bundles by passing through a roller press crusher. Then these bamboo filament bundles were dried and charred under the temperature of 165 degree centigrade and air pressure of 0.3 MPa. Finally the bundles (Fig. 1) were made into parallel bamboo strand lumber lumbers by Fei-yu Co. Ltd. Phenol glue was used to manufacture the parallel bamboo strand lumber specimens. All bundles were put into many molds ( Fig. 1(b)), and these were then pressed together to form the blocks. A transverse compression of 10.3 MPa was applied for the blocks under the hot pressing temperature with the value of 160 o C. The final moisture content was 8.22% and the density was 1250 kg/m 3 for the laminate sourced from the upper portion.
Considering the length and compression directions, six groups of specimens were constructed and detailed information could be seen in Tab. 1. As for the naming rule for the groups, "CS + length" is for the group under compression parallel to grain and "CH + length" is for the group under compression perpendicular to grain. In addition, the "length" is the compression direction. Each group consisted of 40 identical specimens.  The total loading duration is controlled between 8-10 minutes. The load was applied initially through load control in the elastic stage, and then was changed to displacement control before the proportional limit. The load was ramped up linearly to 100 kN at a rate of 1.0 kN/sec, after which the testing process was changed to displacement control. The test continued at a displacement rate of 0.02 mm/sec until the specimen had sustained significant damage, at which time testing was halted.

Failure Phenomena Analysis
Similar to other kinds of engineered bamboo materials [3][4][5][6][7][8], all PBSL specimens experienced elastic stage firstly and then elastic-plastic stage. None obvious failure phenomena on the surfaces could be seen both for elastic stage and elastic-plastic stage. But a clear voice could be heard when the loading value increased about 70% of the ultimate load. Some minor cracks might be appeared in the inner part of the specimens. With the increase of loading, cracks appeared on the surface of the specimens. However, the final failure phenomena varied greatly. Detailed analysis about the final phenomena will be discussed in the following. Four side surfaces of the specimens were named as A, B, C, D (Fig. 3) so that the description of the failure phenomena for each specimens could be done easily [35].

Compression Parallel to Grain
According to the failure process for all specimens, two main failure phenomena series could be classified which are axial compression strength failure and buckling failure. Only strength failure happed for group CS100 and CS150, but both failures occurred for group CS200. Four typical failure modes could be divided into for axial compression strength failures, which were shown from Fig. 4. And detailed failure modes information for each group could be seen from Tab. 2.
As for mode I shown in Fig. 4(a), the shape for the main cracks looks like "Y". Cracks appeared firstly from the two top corners. With the increase of loading, more and more cracks appeared along the two corner angular surfaces. Two cracks surfaces met at the central in the middle axis line along the length of the specimens. Then the cracks surface developed from the cross point line to the bottom along the vertical axis line. There are 7 specimens damaged as this mode for group CS100, 11 for CS150 and 4 for group CS200. The total number for mode I is 22 which is 18.33% of 120 specimens.
(a) Mode Ⅰ (b) Mode Ⅱ (c) Mode Ⅲ (d) Mode Ⅳ Figure 4: Typical failure modes for axial compression strength failure Typical failure mode II could be seen from Fig. 4(b). Cracks appeared firstly just from one top corners or one side of the specimens and then extended along one corner angular surfaces with the increase of loading. Only one main cracks line could be seen from the surface of the specimen in the final failure photo and the line crossed through D surface. Group CS100 has 16 specimens damaged with mode II. Group CS150 and CS200 have 19 and 6 specimens respectively. With the total number of 41, specimens of mode II account for 34.17% of 120 specimens. Fig. 4(c) shows typical failure for mode III. Similar to mode II, cracks appeared firstly just from one top corners and extended along one corner angular surfaces with the increase of loading. However, this cracks line stoped along the original direction when it achieved close middle point and then developed along the central line or parallel to the central line from the middle to the bottom. The numbers for mode III are 15, 6, 5 for group CS100, CS150 and CS200 respectively. And the total number is 26 which is 21.67% of 120 specimens damaged with this failure mode.
Mode IV could be called split failure. All cracks developed along the vertical length direction of the specimen with the increase of loading, as shown in Fig. 4(d). With the number of 2 for group CS100, the number of 4 for CS150 and the number of 13 for group CS200, the total number for mode IV is 19. That is to say 15.83% of 120 specimens damaged with this failure mode. As can be seen from Tab. 2, 108 specimens damaged with axial compression strength failure which take the percentage of 90% in 120 specimens. And the left 12 specimens failed as the buckling failure. As for this failure mode, lateral deformation occurred for all 12 specimens firstly and then some cracks appeared in the specimens. All these specimens belong to group CS200. That is to say 50 by 50 by 200 mm is not good size for the specimen to measure the compression strength parallel to grain in a standard or code.

Compression Perpendicular to Grain
Similar to the failure for compression parallel to grain, both axial compression strength failure and buckling failure occurred for the specimens under compression perpendicular to grain. However, only several specimens from group CH200 failed as the buckling failure and all other specimens have strength failure modes. Fig. 5 shows four typical failure modes for axial compression strength failures and detailed failure modes information could be seen from Tab. 3.
As shown in Fig. 5, the first three failure modes are similar to those for the specimens under compression parallel to grain. However, as for mode I, cracks appeared firstly from the middle parts of the specimens but not from the two top corners. There are 6 specimens failed as mode I for group CH100, 13 for group CH150 and 4 for group CH200. The total number for mode I is 24 which takes the percentage of 20% in 120 specimens. With the number of 22 for group CH100, 19 for group CH150 and 13 for group CH200, the total number for mode II is 54 which accounts for 45% of 120 specimens. The numbers for mode III are 2, 3, 4 for group CH100, CH150 and CH200 respectively. And the total number for mode III is 9 which is 7.5% of 120 specimens.
Mode IV for the specimens under compression perpendicular to grain is different from the failure mode IV for the specimens under compression parallel to grain. As the shape for the crack line looks like arc, mode IV could be called arc shape failure. Each specimen with this failure mode just has one arc shape crack line and the position for the line is random in one side. Group CH100 has 10 specimens damaged with mode IV. Group CH150 and CH200 have 5 and 14 specimens respectively. With the total number of 29, specimens of mode IV account for 24.17% of 120 specimens.
(d) Mode Ⅳ Figure 5: Typical failure modes for axial compression strength failure Seen from Tab. 3, 116 specimens belong to axial compression strength failure which account for 96.67% in 120 specimens. The other 4 specimens failed as the buckling failure and all these specimens belong to group CH200. Similar to the compression under compression parallel to grain, 50 by 50 by 200 mm is not good size for the specimen to measure the compression strength perpendicular to grain in a standard or code.

Calculation and Analysis of Test Results
According to the mechanics of materials and GB/T 50329-2012 [36], the final calculation results for compression strength fc, Ec, and Poisson's ratio v for group specimens can be seen in Tab. 4. As the specimens under compression perpendicular to grain have two kinds of surfaces which have different mechanical properties, vAC and vBD are chosen for the Poisson's ratios of two series of the surfaces. L is the length of the specimens and s is the axial displacement of the specimens. The compression ratio can be calculated by s/L. COV means coefficient of variation; SDV means the standard deviation; CHV means characteristic value, calculated on the basis that 95% of samples exceed the characteristic value (mean ultimate value -1.645 x standard deviation). "CS" stands for all specimens under compression parallel to grain and "CH" stands for all specimens under compression perpendicular to grain.   Fig. 7 plots the comparison of elastic modulus for two directions considering the length effect. Both the trend lines in Fig. 6

Comparison of Elastic Modulus
(a) Compression parallel to grain (b) Compression perpendicular to grain Figure 8: Comparison of Poisson's ratio Fig. 9 plots how the length values influence on the compression ratio under two compression directions. As can be seen from the figures, the values for compression ratio decrease with the increase of the length no matter which direction it is. The slope of the trend lines for the compression ratio is bigger than that for all parameters discussed above. The CHV value for group CS100 is 0.038 which is very close to that value 0.036 for group CH100. It is also interesting to find that both the CHV values for group CS200 and CH200 are 0.024. Biggest compression ratio value means that the specimen with the length of 100 mm undertook fully compression which is close to the ideal axial compression. The size 50 mm × 50 mm × 100 mm should be good choice for the standard or code to measure the compression strength compared with other sizes.

Comparison of Compression Ratio
Combined all test results together, the mean compression ratio values for compression parallel to grain and perpendicular to grain are 0.049 and 0.043, with the standard deviation of 0.013, 0.009 and coefficient of variation of 0.265, 0.222 respectively. However, it is interesting to find that the characteristic value of compression ratio for compression parallel to grain is equal to that for compression perpendicular to grain.
(a) Compression parallel to grain (b) Compression perpendicular to grain Figure 9: Comparison of compression ratio Fig. 10 plots the load-displacement curves for all six groups. The failure process for the specimens under compression parallel to grain could be divided into four main stages which are elastic stage, elastic-plastic stage, plastic stage and decrease stage. However, there are only three stages for the failure process of the specimens under compression perpendicular to grain which are elastic stage, elastic-plastic stage and decrease stage. As for the specimens under compression perpendicular to grain, the load-displacement curves decreased once achieving the ultimate load. While for the specimens under compression parallel to grain, the ultimate load value will be kept for a certain time and it looks like plateau in the curves. PBSL materials have better ductility under compression parallel to grain than that under compression perpendicular to grain. The length value does not have too much influence on the trend of the load-displacement curves. The load values decrease very quickly in the decrease stage for most of specimens no matter how long and which compression direction it is.  Fig. 11 plots the stress-strain curves for the test specimens from each group. The strain values were obtained from the strain gauges. As it is easy for the strain gauges to be broken or stop work when cracks appeared, none decrease stages could be seen from Fig. 11. However, the overall trend for the curves in the first three stages give a good agreement with that for load displacement curves. It could be seen clearly that longer specimens have bigger discreteness for the curves.

Stress-Strain Relationship Model 4.1 Stress-Strain Relationship Model Under Compressive Parallel to Grain
As discussed above, the whole failure process could be divided into four stages for the specimens under compression parallel to grain. However, the period for the decrease stage is short and it does not have too much significance from the view of design point. When dealing with the test data for proposing the stress-strain relationship model, the compression failure process can be simplified into three stages which are elastic stage, elastic-plastic stage and plastic stage as shown in Fig. 12.
where σ is the stress value of the PBSL under compression parallel to grain; c E is the modulus of elasticity for PBSL; p E is the secant modulus for peak point ( c0 ε , c0 f ) ; ε is the strain value of the PBSL; cy ε is the strain for the yield point; c0 ε is the compression peak strain value; c0 f is the compression peak stress value; cu ε is the ultimate maximum compression strain value.
The comparison between the proposed model and the test results for PBSL under compression parallel to grain can be seen from Fig. 13. As can be see clearly that the proposed model gives a good agreement with the rest results.

Stress-Strain Relationship Model Under Compressive Perpendicular to Grain
Similar reason to that for the specimens under compression parallel to grain, when dealing with the test data for proposing the stress-strain relationship model for PBSL under compression perpendicular to grain, the failure process can be simplified into two stages which are elastic stage, elastic-plastic stage as shown in Fig. 14.

Conclusions
In order to investigate the compression performance of parallel bamboo strand lumber manufactured from a new way, 240 axial compression tests have been carried out. Based on the analysis of the test data, the following conclusions can be drawn.
(1) Four typical failure modes could be classified both for the PBSL specimens under compression parallel to grain and perpendicular to grain. The former three modes are similar for the two compression directions. As the features for the fourth failure modes, all cracks developed along the vertical length direction of the specimen under compression parallel to grain while the final crack line shape for the specimen under compression perpendicular to grain looks like arc.
(2) As for the characteristic values, the compression strength parallel to grain is 2.1 times of the compression strength perpendicular to grain. The elastic modulus for compression parallel to grain is 3.64 times of the compression strength perpendicular to grain. While the compression ratios along two compression directions are equal to each other. The bigger Poisson ratios for one typical side surface is 3.93 times of that for another typical side surface for PBSL specimens under compression perpendicular to grain, and the bigger value is equal to that for PBSL specimens under compression parallel to grain.
(3) Length can influence the mechanical properties of the PBSL specimens. The size 50 mm × 50 mm × 100 mm should be good choice for the standard or code to measure the compression strength.
(4) PBSL materials have better ductility under compression parallel to grain than that under compression perpendicular to grain.
(5) Stress-strain relationship models were proposed for PBSL specimens under compression parallel to grain and perpendicular to grain, respectively. These proposed models gave a good agreement with the test results.