Data driven design in sustainability

Abstract

The construction sector contributes to a significant portion of global energy consumption, CO2 emissions and the use of material resources. As climate change escalates with rapidly increasing temperatures, more extreme weather patterns and rising sea levels, an urge to reduce the environmental impact is evident.

A discrepancy is visible between early structural design and the required information for performing an environmental assessment. The first phase of design is an excellent time to steer the design to be more sustainable. However, today's procedures, involving LCA, only provides feedback at a later phase. This is due to the required highly detailed information for performing this complex and time consuming assessment. An assessment tool that supports the conceptual design would help to reduce the environmental impact.

The research project aims to show the feasibility of sustainability feedback using analytical methods on stored building data in early design. It provides LCA feedback in the conceptual design phase. A proof-of-concept is developed on industrial warehouses showcasing the functionality of the framework. Lacking the required building data and LCA scoring, a fully automated script is created with the parametric plugin Grasshopper, to generate the warehouse data. Due to time constrains and performance limitations, the scope was limited to the structural system of beams, columns, purlins and braces. The script randomly iterates over its parameter domain, automatically changing the geometry, performing a structural optimisation and calculating the LCA. This assessment is limited to a cradle-to-gate system boundary. Data is automatically written to a web-hosted database on the Packhunt.io platform of White Lioness technologies. The feedback is based on retrieving data from the database, applying a performance based algorithm, and finding the parameter that decreases the LCA score the most. Results are visualised to the user.

The result is the Interactive Design Assist (IDA) prototype that suggests parameters to the engineer in reducing the LCA score. This information provides guidance in the conceptual design phase on structural design of industrial warehouses. It demonstrates the large potential of reusing 'knowledge' inside the design process to improve building designs.