Dulshan Costa
Doctor of Philosophy candidate
Construction
Biography
Dulshan Costa is a PhD candidate in the Faculty of Architecture, Building and Planning at the University of Melbourne. He holds a Bachelor of Science Honours degree in Quantity Surveying with First Class Honours from the University of Moratuwa, Sri Lanka. Prior to commencing his doctoral studies, he worked as a Temporary Lecturer at the University of Moratuwa, where he contributed to teaching, research, and academic activities in the fields of Quantity Surveying and Construction Management.
His doctoral research focuses on developing and validating a risk forecasting system that integrates building-specific deconstruction knowledge to support safer and more efficient circular economy practices. Through the application of artificial intelligence, data-driven modelling, and digital technologies, he investigates how health and safety hazards can be forecast during building deconstruction to improve safety outcomes, enhance resource recovery, and maximise the value of recovered materials within the construction industry.
Thesis
Forecasting Deconstruction Risks to Enable Safer and Higher-Value Material Recovery
Deconstruction is a sustainable alternative to demolition that supports Circular Economy objectives by enabling the recovery and reuse of building materials. However, deconstruction is highly labour-intensive and relies heavily on manual dismantling processes. As building components are progressively removed, structural stability continuously changes, creating unpredictable working conditions and increasing workers' exposure to hazards. These risks can lead to accidents, project delays, additional costs, and reduced recovery of valuable building components, ultimately discouraging stakeholders from adopting deconstruction practices.
This research challenges the common assumption that health and safety (H&S) risks in deconstruction are too unpredictable to forecast. It aims to develop and validate a risk forecasting system that integrates expert knowledge, historical accident records, and building-specific digital data to predict H&S risks before they occur. By analysing these sources, the study seeks to identify recurring risk patterns, underlying causes, and context-specific factors, demonstrating that deconstruction hazards follow identifiable and modellable relationships rather than occurring entirely at random. The resulting knowledge base will provide a structured understanding of deconstruction risks, their consequences, and their severity.
The study adopts the Design Science Research (DSR) methodology to design, develop, and evaluate the proposed forecasting system. A pilot system will first be developed to examine the feasibility of risk prediction. Subsequently, data collected through literature reviews, expert interviews, accident records, and Revit-based Building Information Models (BIM) will be used to develop the final system. Following the DSR cycle of problem investigation, treatment design, and validation, the system will be evaluated through expert validation and benchmark testing. Ultimately, the research seeks to determine whether accurate forecasting of H&S risks in inherently uncertain deconstruction environments is achievable, thereby enhancing worker safety, improving project efficiency, and increasing industry confidence in deconstruction as a viable Circular Economy strategy.
Contact
- Email millaniyaged@student.unimelb.edu.au
- LinkedIn https://www.linkedin.com/in/dulshancosta/
- ORCID Profile ORCID