PhD OWI-LAB PHD POSITION ON CONDITION MONITORING AND AI ON REMOTE EDGE DEVICES

The Faculty of Engineering, Department Industriële ingenieurswetenschappen, is looking for a PhD-student with a doctoral grant. 

More concretely your work package, for the preparation of a doctorate, contains: 

The VUB Acoustics and Vibrations Research group has a core focus on wind energy as a member of OWI-lab. Our vision is to bring methodological advancements all the way to application in industry. This is achieved by working closely together with the different wind farm operators within the Belgian offshore zone, e.g., Parkwind, Norther, Otary. The main focus of the department is on performance monitoring and health monitoring of the different machine components. To this end, multiple dedicated measurement campaigns have been performed throughout the Belgian offshore zone, resulting in a large in-house database of experimental real-world data enabling large-scale validation of developed algorithms.

Wind turbine drivetrains are critical components, and their failures can lead to significant downtime and operational costs. Traditional condition monitoring approaches often face challenges in accurately detecting early-stage faults, especially in the presence of highly impulsive signals characteristic of damaged gears or bearings. Furthermore, the increasing decentralization of wind farms necessitates robust and efficient on-site data processing capabilities.

This PhD project will address these challenges by exploring and developing advanced methodologies for wind turbine drivetrain condition monitoring.

The successful candidate will primarily focus on:

• Signal processing techniques: Developing and implementing signal processing algorithms specifically tailored to analyze signals that contain interfering impulsive content, often encountered in data coming from main and pitch bearings.

• Machine learning for anomaly Detection and diagnostics: Leveraging state-of-the-art machine learning and deep learning models for automated fault detection, classification, and time-till-failure prediction. This will involve exploring various architectures and unsupervised learning techniques to identify anomalies and diagnose specific fault types based on processed sensor data (e.g., vibrations, currents).

• Edge device deployment for remote applications: A crucial aspect of this project will be the optimization and deployment of developed signal processing and machine learning algorithms onto resource-constrained edge devices. This will involve investigating techniques for model compression and efficient inference to enable on-board condition monitoring directly at the wind turbine, reducing data transmission requirements, central computing needs, and enabling rapid responses to potential issues in remote locations. The aim is to develop robust and scalable software solutions suitable for practical industrial deployment on a large scale.

The project will involve working with real-world wind turbine data, collaborating with industry partners, and conducting experimental validation. The candidate will be expected to publish their research in leading international journals and present at conferences. Given the group’s international relations with other universities, it is also highly recommended to do a joint PhD with one of these universities through an extended research stay abroad to further strengthen the candidate’s profile.

For this function, our Brussels Humanities, Sciences & Engineering Campus (Elsene) will serve as your home base.