Ikaskuntza automatikoko ikuspegi bat, ur-zutabe oszilatzailearen uhin-sorgailuen pronostikorako
- Izaskun Garrido Hernandez
- Jon Lecube Garagarza
- Fares Mzoughi
- Payam Aboutalebi
- Irfan Ahmad
- Aitor J. Garrido Hernández
- Salvador Cayuela Padilla 1
-
1
Universidad del País Vasco/Euskal Herriko Unibertsitatea
info
Universidad del País Vasco/Euskal Herriko Unibertsitatea
Lejona, España
- Aitor J. Garrido Hernández (ed. lit.)
- Matilde Santos Peñas (ed. lit.)
- Izaskun Garrido Hernandez (ed. lit.)
Editorial: Servicio Editorial = Argitalpen Zerbitzua ; Universidad del País Vasco = Euskal Herriko Unibertsitatea
ISBN: 978-84-09-58971-5
Año de publicación: 2024
Páginas: 25-30
Congreso: Jornada Internacional de Energía Eólica y Marina (5. 2023. null)
Tipo: Aportación congreso
Resumen
Wave-induced excitations lead to structural vibrations in Oscillating Water Columns (OWC), resulting in decreased power generation and a shortened lifespan. This article addresses the issue of generator degradation in the Mutriku MOWC plant through a machine learning-oriented strategy for prognosis and fault characterization. Specifically, the utilization of k-Nearest Neighbors (kNN) models has been suggested to forecast the time until OWC generator failure. The assessment relies on data gathered from sensors monitoring various operational parameters of the turbines. The results indicate that the proposed kNN model stands out as an effective solution for cost reduction in maintenance by allowing advanced scheduling months ahead. The high accuracy in predicting generator failures facilitates timely and cost-efficient maintenance practices, avoiding expensive breakdowns and enhancing turbine efficiency. These outcomes underscore the potential of machine learning approaches in tackling maintenance issues within the energy sector, emphasizing the relevance of proactive strategies to minimize operational expenses and optimize energy production.