The PhD defence will take place in Kristen Nygaards sal (5370), Ole-Johan Dahls hus.
Ex auditorio questions: the chair of the defence will invite the attending audience at Kristen Nygaards sal to ask ex auditorio questions.
Trial lecture
“Trustworthiness, explainability, and accountability in the use of ML and AI tools in geotechnical practice”
Time and place: April 7, 2025 11:15 AM. Kristen Nygaards sal (5370), Ole-Johan Dahls hus.
Main research findings
Machine learning can transform how we build tunnels in rock.
The need for efficient, sustainable underground construction grows with urbanisation, resource demands, and climate challenges.
Current tunnelling methods are often constrained by suboptimal decision-making, driven by insufficient point-based data and a reliance on subjective human judgement. Methods for evaluating the rock mass ahead of the tunnel face and around the tunnel profile remain limited. Although extensive spatial datasets, such as Measure While Drilling (MWD) data, are increasingly available, they are not fully leveraged to enhance decision-making.
My research explores the application of all branches of machine learning (ML) in rock engineering, aiming to improve tunnelling efficiency. It focuses on ML models that automate the interpretation of MWD data, translating it into actionable insights for real-time decision-making. These models provide reliable decision support ahead of the tunnel face, aiding planning and operational efficiency. Furthermore, the thesis investigates reinforcement learning (RL) for optimising complex tunnelling processes, such as predictive cutter disk maintenance in Tunnel Boring Machine (TBM) tunnelling and decision-making for rock support and blasting design in Drill & Blast tunnelling.
The findings demonstrate that ML can improve the objectivity, efficiency, and safety of tunnelling operations by providing automated, data-driven decision support.
Predicting rock mass type and quality parameters from MWD-data ahead of the tunnelling face
For further information please see:
Adjudication committee
- Professor Davide Elmo, University of British Columbia, Canada
- Associate Professor Jelena Ninic, University of Birmingham, UK
- Professor Erik Velldal, University of Oslo, Norway
Supervisors
-
Professor Jim Tørresen, Department of Informatics, University of Oslo, Norway
-
Senior Researcher, PhD Zhongqiang Liu, The Norwegian Geotechnical Institute (NGI), Norway
Chair of defence
- Professor Xing Cai, Department of Informatics, University of Oslo
Contact information at Department: Mozhdeh Sheibani Harat