- Research area
Push the Frontiers of Offshore Wind Technology
University of Sheffield
- Research project
Digital twins for health monitoring and fault detection in wind generators and converters
- Lead supervisor
- PhD Student
- Supervisory Team
Dr Xiao Chen (Lecturer in Electrical Machines - Department of Electronic and Electrical Engineering, University of Sheffield)
This Research Project is part of the Aura CDT’s Reliability and Health Monitoring Cluster.
Reliability is of paramount importance for the offshore wind industry as the cost of maintenance, downtime and repair can markedly affect the business case for adopting new and innovative technologies.
To increase availability without increasing maintenance and associated downtime, condition and health monitoring to support fault detection and predictive maintenance are essential in offshore wind. Although many CHM tools are being investigated for the structural elements of a wind generator, little has been done for the electrical generators and power electronics converters which are at the heart of the energy conversion system.
The digital twin concept, based on an accurate real-time simulation of the real system, has emerged as a powerful tool for condition monitoring and predictive maintenance.
Digital twin concepts are gaining interest in the research and industrial communities as effective tools to support smart manufacturing, operation and monitoring of industrial processes and equipment.
Digital twins (DT) are multiphysics, multiscale, high-fidelity simulation that emulate in real-time the state of a corresponding physical twin based on historical and real-time sensors data. The comparison of physical and virtual data throughout product lifecycle can provide valuable information on the state-of-health of a physical structure. While extensive research is being undertaken on DTs for structural health monitoring in OW, there is little if any application of the DT concept to electrical equipment. This is mainly due to the difficulties of multi-time scale modelling in the electrical domain where dynamics can range from sub-milliseconds transients following a power electronics switching transient to thermal and mechanical induced gradual ageing and degradation taking place over the lifetime of the machine.
Using a combination of high-fidelity analytical models, model order reduction techniques and machine learning, this project will develop and validate a multi-time scale digital twin concept for advanced condition monitoring and maintenance of direct-drive permanent magnet generators and converters for offshore wind. The proposed digital twin, will be able to accurately model all electrical transients in the electric drive train, ranging from the sub-millisecond time-scale of the switching converter to long-term degradation over the lifetime of the machine. Comparison of the digital-twin output and the real-time measurements from a range of sensors, combined with advanced signal-processing tools, will be used to demonstrate the ability to detect both gradual degradation and faults in the machine.