| Popis: |
The expected growth of new offshore wind turbine installations necessitates effective Operation and Maintenance strategies to ensure wind farm reliability. Key to these strategies is optimizing operational uptime for Crew Transfer Vessels (CTVs), which transport technicians to wind farms. However, CTV operations are often cancelled due to harsh weather, creating a critical need for continuous operation during favourable conditions. This demand frequently leads to corrective maintenance of the CTV, which increases the risk of unexpected breakdowns. Transitioning to condition-based maintenance can help mitigate the risk of unexpected breakdowns while ensuring uninterrupted operation. At the core of this is the technology of Structural Health Monitoring (SHM), aiming to identify deviations from a normal, undamaged condition of the structure. Essentially, if the structural characteristics remain constant, deviations from this normal condition can be interpreted as structural damage. However, under time-variant operating conditions, the structural characteristics vary, complicating the identification of deviations caused by damages. This study addresses this challenge by separating the operation of a CTV into operational modes, each defined by unique structural load conditions. Then, a Random Forest model using easily accessible input features is implemented to identify the active mode during operation. Hence, deviations from the normal condition of the specific mode can be identified. While the Random Forest model demonstrates high accuracy in this regard, future potential limitations emerged, prompting considerations for future improvements. |
