The goal of the study project was to design a structure using bamboo collected from the botanical garden during the winter semester of 24/25.

The focus was on two areas of research: the physical simulation of the structure and the design-oriented grid generation.

Form-active support systems are “results of structural mechanics”; their form is largely determined by the underlying load case. The limits of this thesis are to be tested with the help of different grid typologies—is it possible to generate a grid that enables or promotes a certain deformation?

The design methodology (computational workflow) consists of five interdependent sub-areas. Using the geometry data obtained for the target geometry, a corresponding grid is generated, which is then simulated under a specific load case in the next step. Depending on the deviation from the target geometry, this grid can be further refined if necessary. A static estimate shows the load in relation to the normal forces and bending moments around x, y, and z.

The project goal consists of two areas of research: physical simulation and static evaluation of the construction, and design-oriented grid generation and manipulation. To what extent can design and form be controlled through grid manipulation?

3D scanning enables precise documentation of the natural location and facilitates the planning of support points. Bamboo stalks and dead and living tree trunks are used for this purpose.

Two curved modules are designed to capture visitors' interest and together form the prototype. The curved design language, the overlapping elements, and the lightness reinforce the character of the construction.

The elevation profiles of the models show which areas require greater deformation. Points are projected onto these areas, and their elevation values are later accumulated to form the parameters for grid generation.

The aim of grid generation is to control local stiffness via the density of the grid.

With the help of another prototype, three different load cases and the corresponding deformation were scanned and documented. This basis is used for simulation calibration and validation.

Overlay of the point cloud of the scanned, deformed grid structure and the simulated beam axes (red).

Once the two grid variants have been generated, the focus shifts to the simulation setup. This requires the alignment of the modules, the parameterization of the load case, and the determination of the material properties of the bamboo, which have already been approximately determined.

Next, the deformed grid is further refined depending on the deviation from the target geometry in order to further maximize local stiffness differences.

A comparison of the surface curvatures shows that the parametric, post-compacted grid offers an improvement of 13.11% compared to a conventional square grid.

1. In this case, the parametric, redensified grid favors the desired deformation (+ 13.11%). Even if the structure is subject to natural deformation, the change in local stiffness can lead to more favorable deformation.

2. In the non-linear, biaxial physical simulation, bamboo was abstracted and considered as an isotropic material; nevertheless, an approximate result was achieved, particularly in the simulation calibration. Determining the exact utilization of the material is only possible to a limited extent due to its strong anisotropy, as well as its strong natural growth, artisanal processing, and heterogeneity.

Simulated grid structure with corresponding beam load in the context of the scanned location.

Scanned, constructed grid structure in the botanical garden

Acknowledgement

Supervision

Prof. Dr. -Ing. Oliver Tessmann

Sandro Siefert M.Sc.

Botanical Garden

Prof. Dr. Simon Poppinga (Wiss. Leiter Bot. Garten)

Dirk Hoyer (Gartenbautechniker)

Workshop

Dilek Tagit

Danijela Mitrovic

Giana Varma

Sinem Özdemir

Alessandro Garruto