Enhancing the Practicality of Tools to Estimate the Whole Life Embodied Carbon of Building Structures via Machine Learning Models

Abstract

The construction and operation of buildings account for significant environmental impacts, including greenhouse gas (GHG) emissions, energy demand, resource consumption and waste generation. While the operation of buildings is fairly well regulated and globally considered in the pathways to net-zero mid-century targets, a different picture emerges when looking at the other life cycle stages, which incur the so-called embodied impacts. These cover raw material extraction and product manufacturing through to construction and end of life activities. Only a handful of examples exist where such embodied carbon (EC) emissions are enshrined in law with most of the ongoing debate still around estimating and understanding where such emissions occur and how to mitigate them. Building structures account for a significant share of a building’s embodied emissions and they also are the building element with the longest service life, thus presenting potential lock-in challenges for choices made today. To support the ongoing global effort to mitigate embodied carbon and equip engineers and designers worldwide with easy-to-use and robust calculation tools, we describe a real-time decision-support tool to aid building design that leverages machine learning (ML) methods from computer science to speed-up the computationally expensive process of finite element analysis (FEA) traditionally exploited in structural engineering. We demonstrate that replacing FEA calculations with a model learnt using ML from a large dataset offers real time decision support while guaranteeing the same level of confidence and accuracy that a traditional FEA-based method would offer at the design stage. The tool has been developed both as a standalone version and as a plugin for Trimble SketchUp to maximise its usability and diffusion. It offers results correlated with uncertainty analysis in the form of probability density functions to account for the inherent variability of input data that characterises early stages in the design process. This research contributes to the ongoing global efforts to decarbonising the built environment and offers an immediately implementable method and tool for doing so.

Artificial Intelligence and Digital Tools for Assisting Low-Carbon Architectural Design: Merging the Use of Machine Learning, Large Language Models, and Building Information Modeling for Life Cycle Assessment Tool Development

“…The discussion around incorporating AI, BIM, and LCA tools into architectural design underlines the potential benefits and existing challenges. While AI can aid in the decisionmaking process, for example, in LCA [32], improving design efficiency, the journey toward seamlessly integrating these technologies faces hurdles like data accessibility issues in the construction industry and the need for increased model accuracy to enhance the reliability of LCA studies derived from BIM models [18,21,22]. Developing industry-wide standards data formatting and interoperability between different BIM and LCA tools could also help support data accessibility.…”

“…AI has been identified as one of the answers to the problem of data accessibility and decision making with LCA early in the design process, allowing designers to perform analyses without a finished 3D model [32]. Machine learning-based solutions are based on either access to a pretrained model or to a dataset that can be used for training the model.…”

“…Similarly, Theodore Galanos showed how machine learning can drastically reduce the time needed for environmental analysis [29], while Zargar and Brown utilized neural networks for the static evaluation of architectural designs [30]. The application of AI extends to optimizing buildings' energy efficiency [31] or environmental footprint, as explored by Pomponi and others [32]. Furthermore, the technology was applied to predict the energy consumption of buildings in varying climates in work by Işeri et al, who leveraged neural networks in their predictions [33].…”

On the Potential of District-Scale Life Cycle Assessments of Buildings

“…D'Amico et al applied different ML algorithms on a parametrised non-residential building model to determine the energy demand and GWP, showing a high degree of approximation of the ML results by comparison to values obtained from granular simulations [92]. Pomponi et al applied different ML algorithms to determine the GWP of concrete and steel structures, leading to the development of a mass and carbon footprint estimation tool with statistical uncertainty statements in the form of probability density functions [93]. In a literature review on ML in LCA, Ghoroghi et al [94] infer that on the district and building level, ML is mostly employed at the LCI stage to predict missing data and to optimise the material selection in the design process.…”

A Framework for User Centric LCA Tool Development for Early Planning Stages of Buildings

“…Usually, the tools are published by researchers or consultants that have used them for their own work. Researchers develop the tools usually for a specific study or project, for example Hester et al (2018) have developed the Building Attribute to Impact Algorithm for guiding the design in early stages, Lobaccaro et al (2018) developed a tool to minimize the embodied GHG emissions in a zero emission building, Kiss and Szalay (2020) developed a workflow for multi-objective environmental optimization of buildings, and Pomponi et al (2021) applied machine learning to support structural design decision while Basic et al (2019) developed Bombyx mainly for teaching purposes. In some cases, the tools are published free or open source, e.g., Bombyx and Beetle.…”