Search

The withdrawal capacity – as determined by a modified ASTM D1761 test – of screws embedded in Phyllostachys edulis bamboo culm walls is presented, demonstrating capacities and behaviour suitable for structural load bearing applications. This study considers twelve screw type-size combinations and considers whether these are predrilled or the screws are inserted without predrilling, resulting in 20 screw type-size-predrill combinations. For the better-performing screws, average withdrawal parameters exceeding fax = 40 N/mm2 were achieved; twice that determined for three-ply plywood having comparable thickness. Comparison with a comparable dataset of screw withdrawal tests from Guadua angustifolia Kunth bamboo indicates that the value of fax is likely species-dependent. In order to mitigate splitting upon screw insertion into bamboo, screw diameters generally less than 6 mm were required unless the screws are inserted into predrilled holes. There was no advantage observed to using self-drilling screws.

Bamboo is often described as a functionally graded unidirectional fibre-reinforced composite material. As such, the Rule of Mixture (RoM) is often reported as an appropriate method of capturing longitudinal material and mechanical properties. However the dominant failure mode of bamboo is associated with longitudinal splitting of the culm and is therefore affected by transverse material properties. The efficacy of RoM in the longitudinal direction suggests its adoption in the transverse direction. It is concluded that the transverse behaviour of bamboo is not adequately captured using the often-adopted RoM approach. Experimental, morphological and numerical approaches are presented to better understand the reasons for bamboo not behaving as a unidirectional fibre reinforced material as is uniformly described in the literature. The approach has identified possible causes of the deviation of the observed behaviour – all which require further study.

Image analysis is used to quantify the distribution of fibre volume ratio, Vf, in strips of P. edulis bamboo obtained from two commercially available glue-laminated bamboo beam products. In total, 58 cross sections containing more than 3500 19 × 6 mm strips were analysed. Simple digital manipulation techniques were found to work well in establishing fibre volume data from the 1200 dpi source images. Total fibre volume for each strip was established and was found to vary in a linear manner through the strip thickness. The observations presented indicate significantly different bamboo source (feedstock) material used by the two manufacturers. Autocorrelation analysis was used to demonstrate that the orientation of the individual strips in each beam section was not random. The impact of a non-random distribution of strip orientation in a section subject to flexure is relatively small but does result in variation from analyses that assume a homogenous distribution of mechanical properties. Finally, nodes, a weak location in glued-laminated bamboo members were observed to represent 3–4% of all strips in a given cross section. The distribution of nodes appeared to be random.