An experimental investigation on Parallel Bamboo Strand Lumber specimens under quasi static and impact loading

Construction and Building Materials 228 (2019) 116724

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

An experimental investigation on Parallel Bamboo Strand Lumber specimens under quasi static and impact loading Xin Li a,b, Mahmud Ashraf a,b,⇑, Haitao Li a,⇑, Xiaoyan Zheng a, Hongxu Wang c, Safat Al-Deen c, Paul J. Hazell c a

College of Civil Engineering, Nanjing Forestry University, Nanjing 210037, China Deakin University Geelong, School of Engineering, Waurn Ponds, VIC 3216, Australia c School of Engineering and Information Technology, The University of New South Wales Canberra, ACT 2612, Australia b

h i g h l i g h t s
PBSL showed significantly higher compression strength than other bio-composite materials.
Specimens with height/width ratio more than 2 produced consistent results under quasi-static testing.
Height of specimen showed approximately negligible effect on material responses under drop-weight.
Better dynamic deformability was observed for specimen impacted perpendicular to the fibre.
The compression response of PBSL was deemed to be strain-rate sensitive.

a r t i c l e

i n f o

Article history: Received 13 March 2019 Received in revised form 3 July 2019 Accepted 14 August 2019

Keywords: Parallel Bamboo Strand Lumber (PBSL) Failure mechanism Dynamic deformation mechanism Drop-weight impact

a b s t r a c t This paper investigates the compression behaviour of 42 Parallel Bamboo Strand Lumber (PBSL) specimens under quasi-static and drop-weight impact loading. 25 mm  25 mm square specimens with varying heights and fibre orientations were tested to investigate compression and impact resistance of PBSL. Quasi-static results indicated a distinct 5-stage failure pattern for the complete quasi-static loading history with a 45° failure plane in all specimens when the compression load was applied parallel to the fibres. Specimen height did not affect the ultimate load carrying capacities but showed considerable influence on the initial stiffness as well as the post-ultimate loading regime. Experimental results showed that the deformation ratio and the energy absorption ratio for longer specimens were not affected by fibre orientations. Low-velocity drop-weight impacts were conducted under 25.4 J and 101.5 J energy levels. Specimen height did not play any significant factor during impact testing. The initial impact level dominated the maximum deformation and dynamic energy absorption observed in specimens that failed showing a 4-stage deform procedure. Although the deformation patterns between fibre orientations were distinctly different, similar energy absorption capacities were obtained regardless of specimen height. Comparison of experimental evidences obtained from quasi-static and drop-weight impact tests clearly showed that the compression response of PBSL is strain-rate sensitive. Ó 2019 Elsevier Ltd. All rights reserved.

1. Introduction Bamboo is an environmentally friendly construction material that is used worldwide for various types of structural applications [1–4]. As a green material, bamboo is considered as an efficient and effective carbon sink [5]. Whilst traditional timber products typically need decades between planting and harvesting, bamboo

⇑ Corresponding authors at: College of Civil Engineering, Nanjing Forestry University, Nanjing 210037, China (M. Ashraf). E-mail addresses: [email protected] (M. Ashraf), lhaitao1982@126. com (H. Li). https://doi.org/10.1016/j.conbuildmat.2019.116724 0950-0618/Ó 2019 Elsevier Ltd. All rights reserved.

is very fast growing and can be harvested in 3–4 years from planting [6,7]. Its strength-to-weight ratio is reported to be higher than those for common wood, aluminium alloy, cast iron, structural steel and GFRP [8]. However, the natural tubular shape with limited control over diameter and low rigidity of some bamboobased materials limit its widespread use in construction [9]. Parallel Bamboo Strand Lumber (PBSL) is an innovative composite material manufactured from bamboo culm and adhesive. Bamboo culm is disassembled into loosened reticular fibre bundled with identical length and original fibre orientation. After dying to the specific moisture content, the fibre bundle is glued together using adhesive, and then assembled by hot-pressing technique to

2

X. Li et al. / Construction and Building Materials 228 (2019) 116724

form standard structural members. It offers enhanced strength, uniformity, dimensional consistency, process ability and diversity in member sizes. It can be used as a timber substitute and is competitive with commonly used building materials, whilst offering renewable environmentally friendly characteristics [10–13]. Previous research on engineered bamboo was primarily focused on their flexural and compressive performances [14–22]. Due to the pretreatment [23] and further compound procedures, basic characteristics of engineered bamboo products were significantly improved; Sharma et al. [24] reported that tension strength of layered laminated bamboo was up to 191 MPa, whilst compression strength parallel to fibre was 64 MPa; similarly, a mean tension stress of 210 MPa was revealed by Verma et al. [25], which indicated somewhat comparable characteristic to steel. Since PBSL is a new-type of composite material, relevant testing standards and criterion are not established. Limited research was reported on the experimental column behaviour of PBSL under axial and eccentric compression [26–28]. Reported test results demonstrate that engineered bamboo products can be used for structural applications but significant research is required to ensure efficient use of this renewable material in construction. Compression behaviour of PBSL with multiple configurations under quasi-static (QS) tests and Drop-weight (DC) impact tests are investigated as part of the current study to develop better understanding of failure mechanisms under considered load cases. Impact loading is quite common in all types of structures but there is no evidence of such studies for PBSL. Impact loading is typically defined as a load or shock applied over a short time period, and loading at velocities below 10 m/s is generally accepted as a lowvelocity impact [29,30]. Like all other construction materials, PBSL will inevitably be subjected to different impact loads such as tool drops, hailstorms etc. during its service life [29,31]. Thus, understanding of low-velocity impact responses including deformation and failure mechanism are crucial for the improvement of materials [32]. Nevertheless, it is worth noting that no existing literature presented a thorough investigation on the failure and deformation mechanism of PBSL under drop-weight impact. Square specimens with three different heights were used in the current study to observe variation, if any, in compressive response due to fibre length and orientations. Obtained experimental results were analysed to characterise typical trends observed in load-deformation behaviour. Obtained experimental results were discussed to highlight material responses as well as deformation and failure mechanisms to pave way for further research in the field.

2. Experimental methodology

line of the supplier and no special instructions were provided to achieve higher density. All specimens were cut into a 25 mm square cross section with varying height ranging from 25 mm to 100 mm as illustrated in Fig. 1. The manufacturing tolerance was controlled within ±1 mm. Since PBSL is an anisotropic material, where fibre orientation could play a significant role on mechanical properties, specimens were classified into two subgroups such as parallel to the fibre (PL) and perpendicular to the fibre (PR). For each type, three specimens were tested to justify reliability of the obtained results. 2.2. Quasi static testing A SHIMADZU 100 kN universal testing machine was used for quasi-static testing of the bamboo specimens. As there is no standard procedure for testing bamboo specimens, the loading procedure was taken according to ASTM D143 [33]. Different loading rates were used based on the height of specimens to achieve a uniform rate of deformation as shown in Table 1. Designation for specimens adopted in the current study refers to material name (PBSL) – height of specimen – fibre orientation (PL/PR) and test type (QS/ DC-impact level). Deformation was captured by using SHIMADZU TRViewX Non-Contact Digital Video Extensometers; the overall setup used for quasi-static test is shown in Fig. 2. 2.3. Drop-weight compression impact testing An INSTRON CEAST 9350 drop tower equipped with a 40-mm diameter cylindrical striker with 5.641 kg of total weight dropping from various heights was used to conduct impact tests. All specimens were carefully placed concentrically with the drop-weight to ensure the impact position and to reduce any potential error. A Phantom V12 high speed camera with a frame rate of 2000 per second was used to capture the whole impact history and required illumination was provided by using a plasma lamp. A schematic diagram for the impact testing is shown in Fig. 3. The impact levels conducted in the current study were 105.6 J and 26.4 J, and relative equivalent impact velocities were 6 m/s and 3 m/s respectively, which are specified in sample designations shown in Table 2.

3. Results and discussion A total number of 18 specimens with varying heights and fibre orientations were tested under QS tests, and an additional 24 specimens were subjected to drop-weight low-velocity rebound impacts.

2.1. Material Materials used in this study were manufactured from commercial PBSL supplier using Moso bamboo filament bundles (Phyllostachys pubescens from Fengxi, Jiangxi Province, China) and phenolic glue (Danning 4I glue with 43.2% solids content, pH = 9.1 and 70 MPa.s viscosity) as raw materials. Bamboo strands and glue were placed into a predesigned mould with predetermined loading to form the desired structural shapes. Transverse compression of 90 MPa was applied during the pressing process under a pressure temperature of 140 °C [26]. Based on previous research [26–28], batch of PBSL gave a range of the bulk density between 990 kg/m3 to 1030 kg/m3 when the moisture content changed from 7.22% to 8.96%. However, in this study, the density of PBSL was 1260 kg/m3 with a coefficient of variation (COV) of 0.026; a higher density may have been obtained through additional pressure and enhanced properties of raw material. It is worth noting that the PBSL specimens were taken from typical production

Fig. 1. PBSL specimens showing different height and fibre orientations.

3

X. Li et al. / Construction and Building Materials 228 (2019) 116724 Table 1 Details of specimens for quasi static testing. Specimen Designation

Geometric Dimensions (mm)

Number of Specimen

Fibre Orientation

Loading Rate (mm/min)

PBSL PBSL PBSL PBSL PBSL PBSL

25  25  25

3 3 3 3 3 3

PL PR PL PR PL PR

0.25 0.25 0.50 0.50 1.00 1.00

25 PL QS 25 PR QS 50 PL QS 50 PR QS 100 PL QS 100 PR QS

25  25  50 25  25  100

Fig. 2. Test set-up used for PBSL specimens under quasi-static loading.

3.1. Quasi static tests 3.1.1. Failure mechanism (1) Failure mode Load-deformation curves obtained from quasi-static testing are shown in Figs. 4–6 for different specimen heights considered in the current study. Relevant images captured by camera correspond to the onset of loading, the propagation of cracks due to gradual increase in loading, and the end of test. All loading histories exhibit similar tendency up to the ultimate load, while distinct differences were observed for different specimens after the onset of failure, especially for PL specimens. For PL specimens, in which load was applied parallel to the grains, failure triggered due to local bending and dislocation of bamboo culms. This dislocation propagated along a 45° plane as can be clearly observed. A splitting sound was clearly heard once the applied force exceeded the ultimate capacity. At the end of an observed plastic plateau, the bent area crossed the overall width and the test specimen buckled causing collapse of the cross-section which eventually initiated unloading. The unloading process in 100 mm specimens occurred smoothly yet relatively quickly until reaching the residual strength stage at about 29.5% Pu, where Pu is the ultimate compression resistance. 50 mm specimens showed some fluctuations in load reduction but stabilized at around 30.8% Pu. The unloading process in 25 mm PL specimens, however, demonstrated a distinctly different path – once the load reduced to almost 50% of Pu, a roughly linear increase was observed in load carrying capacity up to about 91.2% Pu before the load finally dropped again. The load carrying capacities of shorter specimens were higher than those for longer ones but the initial stiffness

Fig. 3. Schematic diagram of Dynamic Compression (impact) test.

showed somewhat inverse trend as can be seen in Fig. 7 (a). It is worth noting that all specimens showed considerable ductility prior to failure. For PR specimens, in which loading was applied perpendicular to the grains, failure began with the occurrence of cracks at side of the specimen. The main cracks expanded continuously until reaching the bearing capacity; the crack formation initiated the separation between bamboo culm and adhesive. At the end of test, the specimens were broken into pieces, which is quite different to PL specimens. Shorter specimens showed reduced initial stiffness (Fig. 7 (b)) and the descent phase flattened out gradually for shorter specimens, until reaching their residual strength at about 25.5% Pu, 30.8%Pu, and 46.1%Pu respectively. It is also worth noting that the loading period and axial displacements were larger than the corresponding PL specimens, especially in 25 mm specimens. All specimens showed considerable ductility prior to failure. (2) Failure influenced region Once the mechanical tests were complete, failure influenced region in each specimen was carefully investigated to identify important failure characteristics. Results from PL specimens indicated that complete bending and dislocation occurred within a specific area; approximately 25 mm along the height of specimen. This may correspond to the 25 mm cross-section in this study with a failure plane of 45 degree, as shown in Figs. 4–6 (a). In 25 mm

4

X. Li et al. / Construction and Building Materials 228 (2019) 116724

Table 2 Details of specimens (DC). Specimen Designation

Geometric Dimensions (mm)

Number of Specimen

Fibre Orientation

Impact level (J)

Equivalent velocity (m/s)

PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL

25  25  25

3 3 3 3 3 3 3 3

PL PL PR PR PL PL PR PR

26.4 105.6 26.4 105.6 26.4 105.6 26.4 105.6

3.00 6.00 3.00 6.00 3.00 6.00 3.00 6.00

25 25 25 25 50 50 50 50

PL DC-3 PL DC-6 PR DC-3 PR DC-6 PL DC-3 PL DC-6 PR DC-3 PR DC-6

25  25  50

(a) Failure mechanism observed in PL specimens (compressed parallel to the grain) (left) (b) Failure mechanism observed in PR specimens (compressed perpendicular to the grain) (right)

(c) Load vs deformation for 100 mm specimens under quasi-static compression Fig. 4. Load-deformation behaviour of 100 mm high specimens under quasi-static compression loading.

high specimens, the obvious densification effect allowed a significant rebound in compression resistance prior to failure. In PR specimens, major cracks extended from edge towards the centre of the height, as shown in Fig. 4 (b) and Fig. 5 (b). (3) Failure procedure Fig. 7 shows stress-strain behaviours of all tested configurations, where 7 (a) compares PL specimens and 7 (b) shows comparison for PR specimens. After careful observation, it is proposed

herein that the typical failure procedure of PBSL under quasistatic compression loading may be divided into 5 stages. The initial loading stage was characterised by elastic behaviour up to a specific load, which followed a nonlinear behaviour up to the ultimate load. A brief plastic plateau was observed for most of the specimens, which is followed by gradual unloading before residual strength reaches a stable state prior to failure. This typical failure path is shown in Fig. 7 (a), which closely correspond to previously reported research on laminated bamboo specimens under compression [21].

X. Li et al. / Construction and Building Materials 228 (2019) 116724

5

(a) Failure mechanism observed in PL specimens (compressed parallel to the grain)

(b) Failure mechanism observed in PR specimens (compressed perpendicular to the grain)

(c) Load vs deformation for 50 mm specimens under quasi-static compression Fig. 5. Load-deformation behaviour of 50 mm high specimens under quasi-static compression loading.

3.1.2. Data analysis Table 3 shows key test results recorded for 25 mm, 50 mm and 100 mm specimens. It is worth noting that, as there is no standard for defining energy absorption for quasi-brittle material, amount of energy absorption were determined from the force-deformation curves up to the peak strength by using numerical integration methods. Specific geometric dimensions and fibre orientation not only influenced the failure mechanism, as discussed in the previous section, but also showed clear effects on load carrying capacity, deformability and energy absorption capacity of the specimens. Majority of test results from each group showed good consistency but some considerable variations for strain at ultimate stress, energy absorption and unit energy absorption (maximum COV of 0.23) were observed for PBSL 50 PL specimens. Natural fibres always possess some degree of variation in material properties and hence 25–30 specimens are usually suggested to achieve any statistical characteristic value. The current study also highlights the importance of such statement. The inherent nonhomogeneous nature of bamboo may have affected this variation in obtained results; more samples should be tested to obtain characteristic values for use in structural

design. However, the obtained results show some specific trends towards overall variation in material response due to difference in sample sizes, and will be used in the following analysis. (1) Load carrying capacity Size effect showed negligible influence on load-carrying capacity for PR specimens as shown in Fig. 8. The mean ultimate stress for PR specimens compressed perpendicular to the grain was 56.5 MPa with a COV of 0.06. In PL specimens, however, shorter specimens produced considerably higher mean stress values due to the observed densification effect; especially 25 mm specimens produced a mean ultimate stress of 109.02 MPa, whereas other specimens produced a mean of 91.76 MPa. The mean ultimate stress for all PL specimens was 97.62 MPa. PBSL is a known anisotropic material and previous studies indicated that PR/PL strength ratio was around 0.6. In this study, the PR/PL ratio showed some variation due to change in specimen length but the overall PR/PL ratio was 0.58, which is very much in line with previous studies. Aforementioned results demonstrate that PBSL offers significantly higher compressive resistance than other bio-composite

6

X. Li et al. / Construction and Building Materials 228 (2019) 116724

(a) Failure mechanism observed in PL specimens (compressed parallel to the grain)

(b) Failure mechanism observed in PR specimens (compressed perpendicular to the grain)

(c) Load vs deformation for 25 mm specimens under quasi-static compression Fig. 6. Load-deformation behaviour of 25 mm high specimens under quasi-static compression loading.

materials. Li et al. [17–21] reported that the compression strength of laminated bamboo lumber (LBL) varied between 53.50 MPa and 80.04 MPa, whilst that for commonly used European Glulam (GLT) and Cross Laminated Timber (CLT) were reported to be 24 MPa and 28 MPa, respectively [34–36]. PBSL clearly has an upper hand over other engineered products manufactured from bio-composite materials but would require comprehensive research for its wider use in construction. (2) Deformation capacity

(a) Stress-strain for PL specimens 100 mm, 50 mm and 25 mm

Specimen size effect showed a significant influence on its deformation capacity. Strain at ultimate stress increased from 0.037 to 0.069 as the specimen height was reduced from 100 mm to 25 mm. In terms of fibre orientation, PR specimens deformed somewhat more than the corresponding PL specimens when the specimen size was reduced to 25 mm as shown in Fig. 9. The deformation ratio of larger specimens, however, tended to be stable irrespective of fibre orientations. (3) Energy Absorption

(b) Stress-strain for PR specimens 100 mm, 50 mm and 25 mm Fig. 7. Stress-strain behaviours of PBSL under QS tests.

Size effect showed a significant influence on energy absorption capacities. As shown in Table 3, larger specimens absorbed more energy during the whole loading procedure due to their larger volume. However, unit energy absorbed by shorter specimens were significantly higher than those shown by longer 50 mm and 100 mm specimens; this is clearly obvious from Fig. 10. Significant densification effect observed in 25 mm PL specimens contributed to the highest unit energy absorption. PL specimens always absorbed more energy than their PR counterparts, which results

7

X. Li et al. / Construction and Building Materials 228 (2019) 116724 Table 3 Details of quasi-static tests results. Specimen Designation

Maximum Force, Pu (kN)

Maximum Stress, ru (MPa)

PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL

65.58 70.54 68.29 54.53 59.45 60.81 52.96 58.70 58.26 35.99 35.70 37.99 33.82 32.26 32.25 35.87 37.10 36.97

104.93 112.86 109.26 87.25 95.12 97.30 84.74 93.92 93.22 57.58 57.12 60.78 54.11 51.62 51.60 57.39 59.36 59.15

25 PL-1 QS 25 PL-2 QS 25 PL-3 QS 50 PL-1 QS 50 PL-2 QS 50 PL-3 QS 100 PL-1 QS 100 PL-2 QS 100 PL-3 QS 25 PR-1 QS 25 PR-2 QS 25 PR-3 QS 50 PR-1 QS 50 PR-2 QS 50 PR-3 QS 100 PR-1 QS 100 PR-2 QS 100 PR-3 QS

Mean Stress,

ru,mean (MPa)

97.62

56.52

Strain at ultimate stress, ePu

Energy Absorption (J)

Unit Energy Absorption (J/mm3*e-6)

0.056 0.052 0.062 0.059 0.039 0.045 0.032 0.039 0.039 0.083 0.062 0.063 0.045 0.044 0.055 0.043 0.035 0.037

63.25 62.36 70.84 132.17 83.66 100.91 170.60 182.01 180.89 53.01 37.76 38.74 54.32 48.83 57.84 104.19 85.57 91.42

4048.00 3991.04 4533.76 4229.44 2677.12 3229.12 2729.60 2912.16 2894.24 3392.64 2416.64 2479.36 1738.24 1562.56 1850.88 1667.04 1369.12 1462.72

Fig. 8. Comparative stress analysis for PBSL specimens.

from their significantly higher resistance under quasi-static loading. 3.2. Drop-weight impact tests on PBSL specimens A total of 24 specimens were tested under drop-weight impact loading. Following sub-sections present all necessary details of experimental procedures followed in this paper. 3.2.1. Impact levels for the drop-weight tests Impact level is one of the key parameters in dynamic compression test but there is no guideline available for PBSL. Impact levels considered in this study were based on the following considerations: (1) A complete material response was desired to appropriately understand the deformation mechanism of PBSL specimens under drop-weight rebound compression impact; (2) As observed in quasi-static (QS) failure procedure, cracks propagated rapidly once the applied load reached the ultimate limit. This can be considered as a reference point to assess the rupture energy;

(3) Material properties of composite may show some sensitivities to dynamic loads [37,38]. In absence of any relevant test evidence, PBSL was assumed to be strain rate sensitive and specimens will be tested under varying strain rates for evaluation of material’s performance. Thus, impact energy was assumed to be the absorbed energy from relevant QS test curves by using integral method up to ultimate load. Larger energy absorption was observed in 50 mm PL specimens with an average value of 105.6 J, and the smallest value was 43.2 J in 25 mm PR specimens. Therefore, 6 m/s impact rate (101.5 J) was used to investigate the elastic-plastic behaviour approaching to failure, whilst 3 m/s (25.4 J) was used to study the elastic responses of each specimen. Relevant details are used in specimen designations as all tests results are summarized in Table 4. 3.2.2. Material response All tests were finalised within 1 ms. Shorter samples seemed to be more sensitive to loading time, and the maximum time discrepancy was observed from the ultimate force for PBSL 25 specimens with a mean COV of 0.18, as shown in Table 4. However, initial

8

X. Li et al. / Construction and Building Materials 228 (2019) 116724

Fig. 9. Strain analysis for the tested PBSL specimens.

Fig. 10. Energy absorbed by the tested specimens.

effects on displacement responses were significantly reduced; absorbed energy amounts were determined from forcedeformation curves and hence were not affected by time. Following sub-sections elaborate on the results obtained from impact testing. (1) Loading procedure adopted in dynamic compression test The force history was recorded in situ by the load cell to analyse how contact force changed with time as shown in Figs. 11 and 12. All loading patterns showed similar trend, where the following 4 stages can be identified: stage (1) demonstrated an almost linear trend where material response was predominantly elastic; stage (2) refers to the onset of plastic deformation, commenced at approximately 90% of the peak load and followed a plateau; stage (3) is the beginning of descent with a steeper slope, which followed by the final unloading, and stage (4) represented a relatively mild

unloading stage that continued until the load unloaded completely. During the last two stages, the impactor started to rebound and moved away from the specimen, and no obvious damage occurred in the used specimens showing considerable resistance to impact loading. (2) Load-time response Typical force versus time curves of 50 mm specimens subjected to different impact levels are shown in Fig. 13; relevant data from Table 4 are summarized in Fig. 14. Size effects on dynamic response force varied with the impact level. Negligible effect was found under lower impact. These maximum response forces showed good consistency within each group of PL DC-3 and PR DC-3. In general, PL group was 14.8% larger than PR group. However, results showed some sensitivity to impact effects under higher impact level, especially PR specimens. Maximum response forces of group PL DC-6 was still steady, whilst

9

X. Li et al. / Construction and Building Materials 228 (2019) 116724 Table 4 Drop-weight compression impact test results. Specimen Designation

Maximum Force, Pu (N)

Time at ultimate force, tPu (ms)

Maximum Displacement, dmax (mm)

Time at ultimate displacement, tdmax (ms)

Plastic Deformation, dp (mm)

Absorbed Energy (J)

Average (J)

PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL

37,807 36,193 36,878 70,587 74,802 74,430 30,306 31,746 32,559 65,234 60,403 60,311 36,635 37,389 37,157 69,960 72,607 69,333 31,026 31,978 31,537 55,782 52,426 53,738

0.304 0.396 0.342 0.322 0.412 0.294 0.656 0.504 0.446 0.424 0.472 0.426 0.480 0.444 0.462 0.306 0.306 0.308 0.490 0.496 0.482 0.424 0.650 0.428

1.008 1.176 0.993 2.129 2.072 2.142 1.405 1.371 1.334 2.573 2.711 2.764 1.215 1.091 1.178 2.126 1.995 2.085 1.396 1.384 1.353 2.895 2.885 2.785

0.594 0.656 0.588 0.648 0.618 0.630 0.762 0.740 0.726 0.718 0.772 0.778 0.668 0.620 0.654 0.634 0.606 0.626 0.760 0.750 0.744 0.900 0.826 0.796

0.376 0.566 0.369 1.101 1.123 1.069 0.663 0.593 0.573 1.454 1.439 1.429 0.506 0.404 0.469 1.019 0.946 1.053 0.497 0.519 0.508 – 1.422 1.378

17.400 17.816 17.620 75.835 79.567 74.891 16.259 16.033 15.139 73.920 75.029 73.551 16.388 16.304 16.273 75.032 76.175 75.875 14.542 14.284 14.897 – 70.048 68.912

17.612

25 25 25 25 25 25 25 25 25 25 25 25 50 50 50 50 50 50 50 50 50 50 50 50

PL-1 DC-3 PL-2 DC-3 PL-3 DC-3 PL-1 DC-6 PL-2 DC-6 PL-3 DC-6 PR-1 DC-3 PR-2 DC-3 PR-3 DC-3 PR-1 DC-6 PR-2 DC-6 PR-3 DC-6 PL-1 DC-3 PL-2 DC-3 PL-3 DC-3 PL-1 DC-6 PL-2 DC-6 PL-3 DC-6 PR-1 DC-3 PR-2 DC-3 PR-3 DC-3 PR-1 DC-6 PR-2 DC-6 PE-3 DC-6

those for group PR DC-6 were a bit scattered. Force level for PL group was 19.4% higher than PR group. Average force levels were approximately 1.90 times higher when impact level was changed from 3 m/s to 6 m/s, irrespective of the fibre orientation and specimen size, as shown in Fig. 14. Peak load levels were achieved earlier under higher impact level. Overall, the impact effect increased with increasing impact level as considered in the current test. Difference in specimen heights did not show any significant effect, especially under lower velocity impact. In addition, the difference between maximum response forces for different fibre orientations was highly reduced with a mean PR/PL ratio of 0.83.

76.765

15.810

74.167

16.322

75.694

14.574

69.480

3.2.3. Deformation pattern The deformation history of a specimen can be classified into two stages, typically expressed as elastic stage and plastic stage. To reveal the deformation mechanism from real material responses, the overall deformation should be controlled within the plastic limit. Thus, drop-weight rebound compression impact test is a useful technique to understand deformation characteristics under impact loading. Obtained experimental evidences are analysed and reported in the following sections. Deformation versus time response can be obtained by using a double integration of force versus time curve, expressed as:

Fig. 11. Impact force versus time curves of 50 mm specimens.

10

X. Li et al. / Construction and Building Materials 228 (2019) 116724

Fig. 12. Impact force versus time curves of 25 mm specimens.

Fig. 13. Typical force versus time curves of 50 mm specimens.

ZZ D ¼

FðtÞ  Mg 2 d t M

ð1Þ

where D is the deformation, F(t) is force obtained from the load cell, M is the total weight including the striker (5.641 kg), g is the acceleration of gravity taken as 9.8 m/s2. (1) Maximum deformation-time response Typical deformation versus time curves for 50 mm specimens subjected to different impact levels and the corresponding captured images are shown in Fig. 15 and 16; relevant maximum deformations from each curve and mean deformation rates K are graphically presented in Fig. 17. Size effect showed almost negligible effect on deformation capacity; the COVs of maximum deformation from PL DC-3, PR DC-3, PL DC-6 and PR DC-6 groups are 0.065, 0.004, 0.015 and

0.039 respectively. In terms of fibre orientation, maximum deformations in PR specimens were marginally higher and the trend became more obvious for increased impact level. It is worth noting that all specimens under same impact level presented a perfect linear fit regardless of specimen height and fibre orientation. Mean maximum deformation was 1.95 times higher when impact velocities was changed from 3 m/s to 6 m/s. (2) Plastic deformation-time response Fig. 18 illustrates the plastic deformation versus time data obtained from the conducted tests. Size effect was negligible when specimens with same fibre orientation were subjected to the same impact level showing consistent amplitude for plastic deformation. Average plastic deformation was 2.45 times higher when impact level was changed

X. Li et al. / Construction and Building Materials 228 (2019) 116724

11

Fig. 14. Force versus time subjected to different specimens.

Fig. 15. Typical deformation versus time curves for 50 mm specimens.

from 3 m/s to 6 m/s. It is also worth noting that plastic deformation of PR specimens were always larger than those of PL specimens, and the difference increased rapidly with increasing impact level. (3) Load-deformation response Typical deformation versus force curves recorded for 50 mm specimens are illustrated in Fig. 19. As obvious in Fig. 19, deformation-force curves can be divided into 4 stages: The first stage is similar to what was observed in force versus time curves, demonstrating a linear increase up to approximately 90% of the maximum force defined as recoverable elastic deformation. The second stage commenced from the onset of plastic deformation up to the maxima. During this stage, force remained steady near the peak with some fluctuations, whilst continuing to accumulate plastic deformation. The third stage initiated with the onset of unloading; the force dropped rapidly with essentially unchanged deformation, and the last stage demonstrated a full-recovery of elastic deformation. PL specimens showed better performance in resisting elastic deformation as reflected by the dynamic stiffness, which is expressed by the slope of the force ver-

sus deformation curves, and its magnitude increased for all specimens as the impact level was increased. PR specimens took longer time to deform, which was obvious at higher impact level, and this allowed PR specimens to absorb more energy during the plastic stage. (4) Deformation mechanism in dynamic compression The current study shows that the dynamic deformation performance of PBSL varies with fibre orientations. This section elaborates on the observed material deformation mechanism in regards to the aforementioned differences in mean deformation ratio, plastic deformation and dynamic stiffness. Fig. 20 (a) shows the bamboo culm at cross-section under microscope. The typical bamboo vascular bundle consists of several fibrous sheaths made from continuously distributed sclerenchymatous cells; these tube-like sclerenchymatous fibre bundles play a key role on material properties [39]. PBSL manufacturing procedures make the high-strength adhesives to fill up the gap between dried cells, especially these tube-like fibre

12

X. Li et al. / Construction and Building Materials 228 (2019) 116724

If the load is applied perpendicular to fibres, its loading condition resembles the manufacturing process as shown in Fig. 20 (c). Bamboo culms and adhesive bear the load directly during the whole loading history; bamboo cells start being crushed followed by the compaction of adjacent adhesive once the deformation exceeds the elastic limit. The larger deformation-rate, longer deformation-time and smaller dynamic stiffness produce a better performance on plastic deformation and hence it absorbs more energy during plastic stage. Thus, under dynamic impact, the PL specimens were stiffer but the PR specimens showed better deformability. (5) Energy absorption Fig. 21 summarizes the absorbed energy from each specimen, which are calculated through an integration of force-deformation curve. The area enclosed by a curve is regarded as the absorbed energy of the tested specimen and was obtained by numerical integration following Eq. (2):

Z

Eabsorbed ¼

Fig. 16. Captured images at different loading state.

bundles. This increases the integrity of PBSL and makes it significantly stiffer, and the material demonstrates uniform deformation during loading history. This process strengthens the substrate material. If the load is applied parallel to fibres, the specimen deforms as an axial loaded column as shown in Fig. 20 (b). During the elastic stage, strengthened adhesive and fibre culms along the loading direction have a better dynamic stiffness and are able to resist the integral deformation of specimen, and absorb more energy. Meanwhile, its lesser amplitude of plastic deformation ensures a faster response after plastic deformation stage.

FðDÞdD

ð2Þ

where Eabsorbed is the absorbed energy from the specimen and F(D) is the force curve which is a function of deformation. Fig. 21 shows that the ability of total energy absorption was mainly controlled by the impact level; although the maximum differences within groups were 12% in DC-3 and 9% in DC-6, specimen height and fibre orientation did not play significant roles. Specimens within groups showed good agreement with each other. Overall, the effect of specimen height on impact response force, maximum deformation and plastic deformation were limited. Fibre orientation influenced the process of energy absorption, but the amplitudes of energy absorption were similar during the whole deformation history. PL specimens absorbed relatively more energy during elastic-deformation stage, while PR specimens absorbed more energy during plastic-deformation stage resulting in a good agreement in total energy absorption. Detailed parameter comparisons can be seen in Fig. 22. 3.3. Comparisons 3.3.1. Energy absorption comparison Energy absorption subjected to different test types are summarized in Table 5 and graphically presented in Fig. 23. Energy

Fig. 17. Maximum deformation for specimens versus time subjected to different impact levels.

X. Li et al. / Construction and Building Materials 228 (2019) 116724

13

Fig. 18. Plastic deformation versus time subjected to different specimens.

Fig. 19. Typical deformation versus force curves of 50 mm specimens.

absorption capacity of PBSL under QS test varies with the height of specimen and fibre orientation. However, dynamic energy absorption capacity under rebound impact test is primarily controlled by the impact level regardless of the height of specimen. Comparisons are summarized as: (1) DC design impact energy is an input value for the testing machine, and DC total energy is the true value obtained from real-time test. The systemic error is about 2.8% regardless of impact levels, as shown in Table 5. All DC tests are considered valid. (2) Comparing DC tests under the same impact level, the mean ratios (Mean energy to DC total energy) are 0.652 in DC-3 specimens and 0.750 in DC-6 specimens respectively, which indicated that nearly 35% and 25% energy dissipation occurred during the rebound procedure. It is consistent with our previous analysis that specimens conducted under higher level impact had a better capacity of total energy absorption.

(3) PBSL reveals a remarkable energy absorption ability under impact compression as absorbed energy increased dramatically with increase in impact levels. When compared with QS specimen, the mean amplitude of energy absorption in 25 mm PR specimens increased by 71.8% under higher impact level.

3.3.2. Carrying capacity comparison Fig. 24 summarizes the maximum force from aforementioned tests. All QS tested specimens failed, and the maximum force represented its load carrying capacity. However, no visible damage occurred in DC tests; this maximum force represented the response carrying capacity under relevant impact level. The essential difference lies in maintaining the integrity of DC specimens while showing greater carrying capacity under dynamic conditions. PBSL is confirmed, and in accordance with our previous assumption, as a strain-rate sensitive material.

14

X. Li et al. / Construction and Building Materials 228 (2019) 116724

(a) Bamboo culm at cross section

(b) PL deformation

(c) PR deformation

Fig. 20. Microstructure of bamboo culm and loading schematic diagrams.

Fig. 21. Energy absorption of each specimen.

4. Conclusion Parallel Bamboo Strand Lumber (PBSL) is a promising new material with the potential to be widely used in structural applications. The current study aimed at promoting its use by providing some reliable experimental evidence on the compression behaviour of PBSL under both quasi-static and impact loading. Experimental investigations on the deformation and failure mechanism of PBSL specimens under different types of compression loading have been presented with sufficient details in this paper. Tests were conducted on 25 mm square specimens with different heights to achieve a clear insight into failure characteristics. Fibre orientation was carefully considered as this has been reported to be a major factor in natural fibrous materials such as timber and bamboo. A typical 5-stage failure path was observed for all tested specimens, which was highly affected by the overall height of the specimens. Densification effects were obvious for 25 mm high

specimens, especially for specimens where load was applied parallel to the fibres (PL specimens), resulting in considerable changes in the post-ultimate loading pattern. The deformation capacity and energy absorption ratio for longer specimens were not highly affected by fibre orientations, and are hence considered more appropriate for testing of compression behaviour of PBSL specimens. Drop-weight impact specimens were tested using an instrumented drop tower for different impact levels, ranging from 25.4 J to 101.5 J. The considered 4-stage loading procedure was in accordance with relevant load-deformation response. Specimen height showed approximately negligible effect on the ultimate response force. The maximum deformation was strongly dependent on the impact velocity showing PBSL as a strain rate sensitive material, and dynamic deformation mechanism exhibited distinctly different deformation-paths for the considered specimens. Initial impact levels could be estimated by using integration

15

X. Li et al. / Construction and Building Materials 228 (2019) 116724

Fig. 22. Parameter analysis between (a) size effect, (b) fibre orientation and (c) impact level.

Table 5 Energy absorption subjected to different test types. Specimen Designation

QS Average Absorbed Energy (J)

DC Average Absorbed Energy (J)

Mean Energy (J)

DC Total Energy (J)

Design impact energy (J)

PBSL PBSL PBSL PBSL PBSL PBSL PBSL PBSL

65.490 43.170 105.580 53.660 65.490 43.170 105.580 53.660

17.612 15.810 16.322 14.574 76.765 74.167 75.694 69.480

16.080

24.669

25.385

74.027

98.676

101.538

25 25 50 50 25 25 50 50

PL DC-3 PR DC-3 PL DC-3 PR DC-3 PL DC-6 PR DC-6 PL DC-6 PR DC-6

Fig. 23. Energy absorption subjected to different test types.

method from quasi-static load-deformation curves. Total energy absorption capacity was affected by the impact level but not by the specimen height. Further experimental and numerical studies

are required to have a clear understanding of PBSL under compression and to subsequently develop design rules for practical applications.

16

X. Li et al. / Construction and Building Materials 228 (2019) 116724

Fig. 24. Maximum force subjected to different test types.

Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The study in this paper is based upon work supported by a Project Funded by the Natural Science Foundation of Jiang-su Province (No. BK20181402), the National Natural Science Foundation of China, China (51878354) and the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD). The writers gratefully acknowledge Ali Ameri, Zongjun Li, M.A. Islam, Bohan Jin and others from the UNSW at ADFA, the Nanjing Forestry University and Deakin University for helping with this work. References [1] E.A. Nurdiah, The potential of bamboo as building material in organic shaped buildings, Procedia-Social Behav. Sci. 216 (2016) 30–38. [2] D. Jayanetti, P. Follett, Bamboo in construction, in: Modern Bamboo Structures, CRC Press, 2008, pp. 35–44. [3] P. Van der Lugt, A. Van den Dobbelsteen, J. Janssen, An environmental, economic and practical assessment of bamboo as a building material for supporting structures, Constr. Build. Mater. 20 (9) (2006) 648–656. [4] J.J. Janssen, Designing and Building with Bamboo, International Network for Bamboo and Rattan Netherlands, 2000. [5] Q.Y. Jia, T. Fung, A.D. Ziegler, Carbon stocks in bamboo ecosystems worldwide: estimates and uncertainties, For. Ecol. Manage. 393 (2017) 113–138. [6] S. Lakkad, J. Patel, Mechanical properties of bamboo, a natural composite, Fibre Sci. Technol. 14 (4) (1981) 319–322. [7] S. Amada, Y. Ichikawa, T. Munekata, Y. Nagase, H. Shimizu, Fiber texture and mechanical graded structure of bamboo, Compos. B Eng. 28 (1–2) (1997) 13– 20. [8] M. Mahdavi, P.L. Clouston, S.R. Arwade, Development of laminated bamboo lumber: review of processing, performance, and economical considerations, J. Mater. Civ. Eng. 23 (7) (2011) 1036–1042. [9] Y. Li, B. Xu, Q. Zhang, S. Jiang, Present situation and the countermeasure analysis of bamboo timber processing industry in China, J. Forest. Eng. 1 (1) (2016) 2–7. [10] J.F. Correal, L.F. Lopez, Mechanical Properties of Colombian Glued Laminated Bamboo, Taylor & Francis, London, UK, 2008. [11] Lopez, L. and J. Correal, Exploratory study of the glued laminated bamboo Guadua angustifolia as a structural material. Maderas: Ciencia y Technologia, 2009. 11(3): p. 171–182. [12] N. Nugroho, N. Ando, Development of structural composite products made from bamboo II: fundamental properties of laminated bamboo lumber, J. Wood Sci. 47 (3) (2001) 237. [13] C. Verma, V. Chariar, Development of layered laminate bamboo composite and their mechanical properties, Compos. B Eng. 43 (3) (2012) 1063–1069.

[14] W. Yang, X. Wang, L.I. Guofen, Mechanical properties test of bamboo scrimber flexural specimens reinforced with bars, Acta Mater. Comp. Sinica 31 (4) (2014) 1030–1036. [15] Y. Wei, W.U. Gang, Q. Zhang, S. Jiang, Theoretical analysis and experimental test of full-scale bamboo Scrimber flexural components, J. Civil Archit. Environ. Eng. (2012). [16] H. Li, G. Wu, Z. Xiong, I. Corbi, O. Corbi, X. Xiong, H. Zhang, Z. Qiu, Length and orientation direction effect on static bending properties of laminated Moso bamboo, Eur. J. Wood Wood Products 77 (4) (2019) 547–557. [17] H. Li, G. Wu, Q. Zhang, A.J. Deeks, J. Su, Ultimate bending capacity evaluation of laminated bamboo lumber beams, Constr. Build. Mater. 160 (2018) 365–375. [18] H.-t. Li, R. Liu, R. Lorenzo, G. Wu, L.-b. Wang, Eccentric compression properties of laminated bamboo columns with different slenderness ratios, Proc. Instit. Civil Eng. – Struct. Build. 172 (5) (2019) 315–326. [19] H. Li, G. Wu, Q. Zhang, J. Su, Mechanical evaluation for laminated bamboo lumber along two eccentric compression directions, J. Wood Sci. 62 (6) (2016) 503–517. [20] H. Li, G. Chen, Q. Zhang, M. Ashraf, B. Xu, Y. Li, Mechanical properties of laminated bamboo lumber column under radial eccentric compression, Constr. Build. Mater. 121 (2016) 644–652. [21] H. Li, J.-W. Su, Q.-S. Zhang, A.J. Deeks, D. Hui, Mechanical performance of laminated bamboo column under axial compression, Compos. B Eng. 79 (2015) 374–382. [22] J. Zhou, D. Huang, Z. Huang, Y. Shen, Experimental behavior of fabricated beam-column connection of bamboo frame under monotonic loading, J. Forestry Eng. 3 (3) (2018) 122–127. [23] M. Huang, X. Zhang, W. Yu, W. Li, X. Liu, W. Zhang, Mechanical properties and structure characterization of bamboo softened by high temperature steam, J. Forestry Eng. 1 (4) (2016) 64–68. [24] B. Sharma, A. Gatoo, M. Bock, H. Mulligan, M. Ramage, Engineered bamboo: state of the art, Proc. Instit. Civil Eng.-Constr. Mater. 168 (2) (2014) 57–67. [25] C. Verma, V. Chariar, R. Purohit, Tensile strength analysis of bamboo and layered laminate bamboo composites, Int. J. Eng. Res. Appl. 2 (2) (2012) 1253– 1264. [26] H. Zhang, H. Li, I. Corbi, O. Corbi, G. Wu, C. Zhao, T. Cao, AFRP influence on parallel bamboo strand lumber beams, Sensors 18 (9) (2018) 2854. [27] H. Li, Z. Qiu, G. Wu, O. Corbi, L. Wang, I. Corbi, C. Yuan, Slenderness ratio effect on eccentric compression performance of parallel strand bamboo lumber columns, J. Struct. Eng. 145 (8) (2019) 04019077. [28] H. Li, Z. Qiu, G. Wu, D. Wei, R. Lorenzo, C. Yuan, H. Zhang, R. Liu, Compression behaviors of parallel bamboo strand lumber under static loading, J. Renewable Mater. 7 (7) (2019) 583–600. [29] H. Wang, K.R. Ramakrishnan, K. Shankar, Experimental study of the medium velocity impact response of sandwich panels with different cores, Mater. Des. 99 (2016) 68–82. [30] J. Wang, A.M. Waas, H. Wang, Experimental and numerical study on the lowvelocity impact behavior of foam-core sandwich panels, Compos. Struct. 96 (4) (2013) 298–311. [31] K.R. Ramakrishnan, S. Guérard, P. Viot, K. Shankar, Effect of block copolymer nano-reinforcements on the low velocity impact response of sandwich structures, Compos. Struct. 110 (1) (2014) 174–182. [32] S. Abrate, Impact on laminated composite materials, Appl. Mech. Rev. 44 (4) (1991) 155–190. [33] A. Standard, D143-14, Standard methods of Testing Small Clear Specimens of Timber, ASTM International, West Conshohocken, PA, 2014. DOI: 10.1520/ D0143. [34] R. Brandner, G. Flatscher, A. Ringhofer, G. Schickhofer, A. Thiel, Cross laminated timber (CLT): overview and development, Eur. J. Wood Wood Prod. 74 (3) (2016) 331–351.

X. Li et al. / Construction and Building Materials 228 (2019) 116724 [35] M. Wallner-Novak, J. Koppelhuber, K. Pock, Cross-Laminated Timber Structural Design-Basic Design and Engineering Principles According to Eurocode, 2014. [36] B.J.B.S.I. EN, 14080: Timber Structures–Glued laminated timber and glued solid timber–Requirements, 2013. [37] W. Zhang, Y. Zhang, G. Zhang, Static, dynamic mechanical properties and microstructure characteristics of ultra-high performance cementitious composites, Sci. Eng. Comp. Mater. 19 (3) (2012).

17

[38] W. Zhang, Y. Zhang, Research on the static and dynamic compressive properties of high performance cementitious composite (HPCC) containing coarse aggregate, Arch. Civil Mech. Eng. 15 (3) (2015) 711–720. [39] M.A. Jian-Feng, W.Y. Chen, L. Zhao, D.H. Zhao, Bionic design of columnar structure based on microstructure of bamboo, J. Mach. Des. (2008).