Machine learning and ontology supported comprehensive building comfort framework design

Bie, Sisi 2023. Machine learning and ontology supported comprehensive building comfort framework design. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF (Thesis) - Accepted Post-Print Version Download (15MB) | Preview
	PDF (Cardiff University Electronic Publication Form) Restricted to Repository staff only Download (1MB)

Abstract

This dissertation presents a comprehensive study integrating Building Information Modelling (BIM), machine learning, and ontology to enhance building comfort and energy efficiency. It explores the current state of BIM, the advancements in machine learning, and the critical dimensions of building comfort, establishing a comprehensive comfort framework that addresses the multifaceted nature of designing sustainable and comfortable buildings. The study is motivated by the need for a holistic approach to building design that efficiently balances energy consumption with occupant comfort while leveraging the potential of advanced technologies like ontology for structured knowledge representation and dynamic reasoning. The methodology adopted in this research involves a structured approach to data generation, leveraging BIM's rich repository of building data and the predictive capabilities of machine learning. The study develops a comprehensive comfort framework through quantitative and qualitative methods, normalizing and standardizing various comfort dimensions into a unified metric that can be used to assess and compare the comfort levels of different building designs. This metric moves beyond traditional assessments, typically focused on thermal comfort, to include factors such as acoustic, visual comfort, and air quality, thereby providing a more holistic view of occupant comfort. The dissertation further incorporates ontology to create a dynamic and adaptable comfort assessment system. By integrating Semantic Web Rule Language (SWRL) and ontology-based reasoning, the study enhances the model’s ability to evaluate real-time comfort data, predict future conditions, and suggest system optimizations. This ontology-based framework also enables the customization of comfort profiles according to user preferences, allowing for a more personalized assessment and adjustment of comfort parameters. Additionally, the integration of machine learning with BIM and ontology is explored to revolutionize traditional building design and performance analysis. Machine learning algorithms, such as Linear Regression (LR), Artificial Neural Networks (ANN), and Random Forests (RF), are utilized to predict building performance and optimize comfort and energy efficiency strategies. This approach not only enhances the accuracy and efficiency of performance predictions but also reduces reliance on time-consuming and often biased traditional methods. The research addresses a significant gap in the literature by proposing and validating a BIM-based machine learning and ontology engine. This engine provides a robust, dynamic, and comprehensive analysis of building data, leading to more informed decision-making in the design phase. The proposed system is tested and validated through a series of case studies, demonstrating its potential to transform building design and management practices. The dissertation concludes with a reflection on the research findings, discussing the implications for the construction industry and outlining future research directions. It emphasizes the need for continuous improvement and innovation in the integration of BIM, machine learning, and ontology, advocating for their adoption in creating more sustainable, comfortable, and efficient buildings. This study contributes to the field by offering a novel approach to building design, emphasizing the importance of holistic comfort and energy efficiency through the integration of advanced technologies.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Engineering
Uncontrolled Keywords:	1). Comprehensive Comfort Framework 2). Machine Learning 3). BIM Model 4). Building Energy Analysis 5). Building Performance 6). Comfort Level 7). Artificial Neuron Network 8). Ontology
Date of First Compliant Deposit:	21 January 2025
Last Modified:	21 Jan 2025 16:43
URI:	https://orca.cardiff.ac.uk/id/eprint/175419

Actions (repository staff only)

Edit Item