AI for Designing Low Carbon Buildings: Accelerating Achievement of Net Zero 2050 in our Built Environment

Aim The key aim of our work is to develop and evaluate Artificial Intelligence (AI) based decision support systems to promote low carbon buildings in the process of design, construction and operation. We will achieve this by describing novel AI-augmented workflows and tools for designers, focusing on their application to support low carbon decision making at key stages of the building design process. The Problem The built environment has a vital role to play in responding to the climate emergency - addressing carbon is a critical and urgent focus. Buildings are currently responsible for 39% of global energy related carbon emissions: 28% from operational emissions and the remaining 11% from embodied carbon materials and construction (World Green Building Council, WGBC, 2019). The built environment has been identified as a key sector to achieve the UK government's vision for a technology-driven transition to decarbonise the economy and reach net zero by 2050. AI-based processes have significant potential to accelerate the achievement of climate neutrality in the built environment, by leveraging alternative low carbon design solutions at key stages of designer decision making. (Yang et al., 2022). We will explore the problem by evaluating the integration of AI processes with the existing design workflows of the host organisation Grimshaw Architects, an internationally recognised sustainable design practice, using a wide range of new and retrofit eco-focused projects as test cases. Secondment Objectives Objective 1 | Evaluate the use of Grimshaw Architects' existing building information and performance data to establish an optimised dataset collection and classification system for low carbon AI model training. (WP1). Objective 2 | Disseminate AI-augmented low carbon best practice design workflows and tools across networks of architectural education, practice, and industry via establishment of the AI Low Carbon Building Network (AILCB) (WP2). Objective 3 | Reduce carbon expenditure by authoring novel AI-based low carbon design decision support software tools integrated with Building Information Modelling software (BIM) across five key stages of building design, construction, and use. (WP3). Objective 4 | Produce an AI for Low Carbon Building policy brief to establish clear recommendations for building designers, contractors and users on dataset capture, model training, ethical considerations, and the use of AI tools for carbon reduction. (WP4). Applications and Benefits Outcomes will be tested across a variety of project types and disseminated via engagement events to communicate the research and deliver scalability and wide application for the benefit of carbon reduction in practice, to: Set benchmarks in the industry for standardisation, classification, segmentation, and optimisation of building design information into datasets for low carbon AI-model training. (Objective.1) Support interdisciplinary networks and learning between academia, practice and industry through engagement events and the establishment of the AI Low Carbon Building Network (AILCB) (Objective.2) Reduce carbon expenditure across the UK construction sector through the sharing of novel low carbon AI-based decision support tools (Objective.3) Add value to the UK architecture sector through the support and upskilling of present and future architectural practitioners in the use of AI for low carbon design by describing novel best practice workflows via the AI for Low Carbon Building Policy Brief (Objective.4) The project will create a significant launchpad from which we can position future collaborative work at the forefront of global AI-based low carbon design in the construction sector.