How can the ever-increasing human understanding of material properties and behaviours be used in an applied manner to design and construct innovative and pioneering structures? How can the scientific understanding of a material’s properties be used as an active force driving the structural formation of its physical implementation? How can the unique data of a specific material inform an architecture harnessing that material’s distinct properties?
The COW project aims to investigate and develop new tools and methods for achieving precisely this within the confined research space of timber structures. An innovative parametric tool that uses genetic algorithms to search for optimal geometries within a fitness landscape defined by existing site conditions and other data specific to an architecture scheme (including but not restricted to the properties of different timber-based building systems), as well as a wide range of proj- ect-specific desires, COW is a practical tool for optimising timber structures in accordance with a collection of pre-defined objectives.
The most successful attempts to define relations between such material data and its implied structural geometries, in order to turn them into architectural strategies, can be found in quite recent trajectories within architecture theory: the parametric systems that have allowed for a veritable revolution within the field’s digital realms, exemplified most (in)famously by architect Patrik Schumacher’s 2008 manifesto, Parametricism as Style.
COW uses existing site data and a wide range of project-specific desires to go beyond Schumacher’s rather broadly defined style. Instead, it uses evolutionary logics as a productive strategy for achieving optimal compromises between a scheme’s different driving forces. This position paper discusses how the COW system is situated within contemporary architectural theory and practice in general, and shows how it can begin to be used to promote timber architecture in the urban context.
Keywords: timber architecture, multi-objective optimisation, COW, genetic algorithms, material performance
Authors
Larsson M.
KTH Royal Institute of Technology, KTH Building Materials, EnWoBio – Engineered Wood and Biobased Building Materials Laboratory, Stockholm, Sweden
Login to download the PDF