| Popis: |
Bending Active Gridshells (BAGs) may be made by joining straight lengths to form a grid and then bending the grid into a curved, shell-like geometry. Therefore, the laths used must be flexible, section sizes are kept small and hence single layers have limited load-carrying capacity. As a result, BAGs (usually made of timber laths) require multiple layers to carry design loads. Each layer is curved independently, and the deformed layers are locked together at joints and using shear blocks to act as a composite unit. This research arose from prior research on double-layer timber BAGs carried out in UL (Collins, 2016). Collins’ BAG prototype was more flexible than anticipated. Preliminary investigations suggested that secondary bending could be contributing to the flexibility of Collins’ gridshell. This research, therefore, aims to quantify the relative contribution of secondary bending to the overall deflection of individual double-layer gridshell assemblies. It is assumed that the wider the joint and shear block spacing, the greater the likelihood of secondary bending happening between shear blocks and joints. This study uses both analytical and numerical models to generate predictions of primary and secondary deflection in double-layer shallow flat and curved assemblies. Robust sensitivity convergence testing of FEA models is employed together with simplified analytical models and physical xperimental testing. The model predictions are compared to each other and to experimental data. Timber BAGs are three-dimensional, and display geometric and material non- linearity, time-dependent deformation, complex connection behavior, and, potentially, inter-layer slip. However, as the first stage of a larger planned research program, this study addresses flat, bending inactive, double-layer, linear elastic individual assemblies analytically, numerically, and experimentally. Grillage assemblies are not treated in this research. All modelling is conducted within the limits of small deflection theory. Physical experiments use one exemplar geometry and are limited to simple supports and single points loads. An initial numerical study only of bending inactive curved assemblies is also made. The research shows that, except for very stiff, short cases, compact analytical expressions, based on commonly used modelssumptions for secondary bending, when compared with finite element and experimental data, predict the impact of secondary bending on actual deflections of flat single assemblies well. It is also shown that the deflection trends for curved assemblies are similar to flat assemblies in some respects. |
