Nguyen Hoang Nam Du has followed the PhD Programme at Faculty of Engineering and Science, Specialisation in Mechatronics.
- Trial lecture starts at 10:15
- Public defence starts at 12:15
Title of trial lecture: "Challenges in transferring fault diagnosis methods from laboratory environments to industrial applications"
Disputation chair: Paul Svennevig
Assessment committee:
- First opponent: Professor Jose Antonino-Daviu, Polytechnic University of Valencia, Spain
- Second opponent: Dr. Jenni Pippuri-Makelainen, VTT Technical Research Centre of Finland and Aalto University, Finland
- Chair of the assessment committee: Professor Joao Leal, UiA
Supervisors in the doctoral work:
- Main supervisor: Professor van Khang Hyunh, UiA
- Co-supervisor: Professor Kjell Gunnar Robbersmyr, UiA
Summary of thesis
Energy demand has increased significantly following the pandemic and the geopolitical instability worldwide. According to the IEA report, the growth of renewable sources reached 34.5% by the end of 2025. Although remarkable efforts are being made to develop and deploy renewable energy sources to meet the demand, they remain well short of the annual growth target of 8\%. Moreover, to accelerate annual energy growth, permanent magnet synchronous machines (PMSMs) are central to the sustainable development of large-scale energy production, energy transition, electrification, autonomous vehicles, and the aerospace industry.
PMSMs are now becoming the preferred choice for large-scale wind turbines due to their high efficiency, precise control, torque density, and ease of maintenance. Despite the rapid growth of diagnostic technologies enabled by artificial intelligence, PMSMs are subject to electrical faults and mechanical and thermal stresses under dynamic operating conditions. Additionally, the scarcity of historical fault data in practical applications limits the effectiveness of data-driven approaches, which typically require large amounts of labeled data. Therefore, robust fault detection and identification (FDI) based on data-driven approaches are key not only to deploying reliable and safe energy systems but also to reducing labor and maintenance costs.
This dissertation addresses these challenges through three key contributions. First, it improves FDI performance under low-severity faults and dynamic operations. Second, a framework is developed to integrate domain knowledge with data-driven approaches to enable explainable decision-making across different time scales in predictive maintenance. Third, an autonomous online fault diagnosis framework is proposed to provide traceable decision trajectories and physical reasoning during dynamic operations.
The proposed methods are described theoretically and validated through experimental setups. The results demonstrate that the models achieve robustness and explainability for each prediction in fault diagnosis, making them well-suited for deployment in predictive maintenance strategies for PMSMs. A brief summary of the models and their findings is presented in this dissertation, while detailed explanations and results are provided in the appendices.
What to do as an online audience member
The disputation is open to the public. To follow the trial lecture and the public defence online, register on Zoom.
We ask online audience members to join no earlier than 10 minutes in advance. After these times, you can leave and rejoin the meeting at any time.
Opponent ex auditorio:
Deadline for the public to pose questions is during the break between the two opponents. Questions ex auditorio can be submitted to Kristine Evensen Reinfjord.
Contact person
- kristine.reinfjord@uia.no
- Phone
- +47 37 23 30 14