Stock discretised structural timber elements

Abstract

The current sustainability crisis requires a shift in the way material availability is currently considered. The current trend is to dispose of material easily after it has served its first purpose, while often the material still has potential to be reused. The construction sector is a large contributor of the extraction of virgin resources, leading in the case of timber to deforestation. In The Netherlands alone around 1.740 kiloton waste wood is collected annually. 23% of this wood consists of solid non-glued or treated reusable timber, translating roughly to a waste stream of 450kton of reusable wood that is discarded. The current design approach requires shift, instead of design for manufacturing designers should focus on designing with what has already been manufactured. The design concept of discrete timber is well suited for this design approach as it involves building a structure or element out of smaller parts. This method allows for direct reuse and repurposing of a varying stock of discharged timber pieces, effectively enhancing the life-cycle of timber from a currently down-cycling scenario into a circular loop. Moreover a discrete system is not beholden to the traditional implication of uniform an rectangular cross-sections and a geometry can be easily optimised to the acting forces by leaving out parts on lower stressed area’s. Using programming and optimisation techniques, this thesis focusses on creating an efficient and adaptable structural system from a varying stock of reclaimed timber pieces while maximizing the future reuse potential of the used parts. The main computational problems addressed in this research are efficiently filling a design domain with available stock and minimizing cutting losses, and structurally optimising the design domain. An algorithm is created in the visual programming environment Grasshopper in combination with Python. The algorithm discretizes a given design space into the pieces found in the database by sequentially solving three combinatorial problems. The algorithm optimises the placement of the pieces so that higher strength grade pieces are placed in area with higher stress levels. The assembly is optimised by removing all non-vital structural parts resulting in a final efficient structure. The algorithm's performance is tested on stability of results, optimisation method, size influence of parts, filling rate of the design space, strength grade influence and buckling. This work serves as a prove of concept for designing with a highly versatile stock of reclaimed components.