Starting point: 

How does the digitalisation of Architecture, Engineering, and Construction (AEC) enhance decarbonisation through intelligent algorithms in the context of Life Cycle Assessment (LCA)?

The topic may include: Building Lifecycle Assessment (BLCA), Design for Manufacturing and Assembly (DfMA), Artificial Neural Networks (ANNs), and KPI developments aligned with related Building Codes.

Project Summary: 

AI-Integrated Solutions for Decarbonizing AEC

Making well-informed decisions is essential but challenging in architecture, engineering, and construction (AEC) since everything is interconnected. This difficulty arises from the lack of integrated solutions for managing dynamic data flows and information throughout related procedures. Each sector involves actions requiring analysis, sorting, integration, and providing a true data structure for each design decision. However, the absence of a centralised database is the principalproblem in studying the integration of variant factors that can lead to informed decisions.

This research proposes the implementation of intelligent algorithms like Geometry Deep Learning (GDL) and/or Graph-based Neural Networks (GNNs) to address the challenge described. These algorithms are well-suited for providing each building with a study because they can check functional paradigms among impactful parameters within graphs and geometrical data analysis. Every building can be represented as a multi-faceted geometrical graph comprising subparts that encapsulate distinct factors such as environmental impact, structural integrity, and geometrical considerations. By having graphs, data flows from diverse and varied factors can be studied graphically and analytically. Thisidea is highly comprehensive for AEC; however, this PhD thesis focuses on providing a reliable framework for building industry decarbonisation. 

The methodology involves developing an algorithm using GNNs to address complex AEC decisions focusing on decarbonisation strategies. The model will be trained with relevant data and presents each building design as a graph that dynamically updates with unique features the designer considers. Therefore, the source identification, leakage management, and verification of raw materials are studied in each decision. The research outcome will demonstrate the necessity of robust data and building data structure management using ANNs for decarbonisation, which will help researchers and practitioners understand the network of impactful data behind their decisions and strategies.