TWI played a key role in an EU project that has developed methods of continuously monitoring the condition of active tidal energy generators.
Trials have now been completed in REMO, a collaborative research project looking into the feasibility of using vibration analysis (VA) and acoustic emission (AE) in combination to monitor the condition of in-service tidal stream generators. The developed prototype, now at the demonstration stage, assesses the structural and mechanical integrity of tidal systems to provide advance warning of the presence of faults and impending failures.
Harnessing the potential of tidal energy
Due to the predictable nature of tides, tidal energy is an environmentally attractive renewable energy source. However, considerable investment costs and expected costs of lifetime maintenance in hostile marine environments have hindered large-scale implementation. Operational availability has been shown to be as low as 25%. This availability needs to be increased substantially if tidal energy harvesting devices are to become commercially viable. Integrated condition monitoring (CM) can provide a reliable tool for assessing the real-time condition of critical components of tidal systems, enabling cost‑effective maintenance based on prediction rather than correction.
REMO aimed to create a novel CM system to help reduce the projected lifecycle maintenance costs of tidal stream energy by 50% and the generator downtime to a level comparable with wind turbines.
Details of the techniques
- An established inspection method for monitoring high-frequency rotating parts
- A qualitative technique providing information on the type of damage present.
- A technique gaining industrial acceptance for its ability to monitor low-speed rotating machinery
- Provides a quantitative measure of mechanism deterioration
- Requires validation of its ability to detect damage in service.
Using a combination of the two allows for the complementary analysis of waves through the whole frequency spectrum, enabling us to monitor components rotating at different speeds.
Validation of AE
To validate the use of AE, researchers undertook laboratory trials where a scaled test rig was set up to generate noise and vibration representative of a full‑scale tidal power system. A test plan was designed to replicate the kind of progressive gearbox damage that could lead to catastrophic failure. Clear spikes in an AE signal were observed when a defect was present – all the indicators extracted from the AE signal differed between the healthy and fault conditions. These results show the potential use of AE for fault detection in tidal turbine gearboxes.
Researchers developed and integrated software and hardware (Figure 2 ) for acquisition of VA and AE, providing remote access and supervisory control. The architecture is modular, scalable and flexible, so that the system can be adapted to different energy recovery systems.
Underwater trials at TWI’s diving tank facilities in Middlesbrough validated the system.
The team analysed different operational conditions relevant to increased damage levels to show damage detection and correlation capabilities of the REMO system. Figure 3 and Figure 4 show results of the tests corresponding to an increase of the feature extracted when the data acquired comes from a faulty gearbox using VA and AE respectively.
Researchers developed pattern recognition signal processing software based on similarity analysis and Euclidean distance that achieves the next targets:
- Determine the signature of a healthy turbine and evolution of this healthy signature throughout the system lifetime
- Identify deviations from the healthy signature
- Provide an automated warning of the presence of significant defects well before irreversible damage or failure arises.
Figure 5 depicts the capability of the software to identify deviations from a healthy situation based on the Euclidean distance dissimilarity.
Furthermore, the software is able to generate an alarm when the Euclidean distance moves outside of a knowledge-based reference using the traffic lights shown in Figure 6.
This project has shown that AE offers promising results and provides information about the damage level of the gearbox for tidal turbine gearbox monitoring. Adding VA data to this offers additional information about the prematurity and accuracy of defect detection including determination of the kind of defect.
For more information please visit remo-project.eu or please contact us.