Matters

The past 100 years have been based on an ideal of material simplification by homogeneity, repeatability, and isothermal conditions in buildings. This has resulted in a severe misalignment with the planetary limits. We propose to address these problems by developing and implementing deep learning models, by studying fast-growing, lightweight tree structures, focusing on varied and temporal material densities. The density of wood is a key indicator of thermal, acoustic, and structural capacities, which we will analyze and activate through deep learning to create accurate and efficient analyses and propositions for bespoke wood structures. The research will be based on the analysis of wood samples by use of 2D and 3D imaging, thermal imaging, and acoustic measurements. The investigations will contribute with a new multimodal dataset, training of a deep learning model, and a new evolutionary design model that can propose novel material structures. In Matters deep learning models and evolutionary design models of thermal and acoustic properties will be combined to develop new knowledge and methods for composite, biogenic, and sustainable material structures for our built environment.