Reliability analysis of detecting false alarms that employ neural networks: a real case study on wind turbines.
Operations and maintenance tasks are critical to the reliability of a wind turbine. The state-of-the-art demonstrates the effectiveness of reliability centred maintenance, but there are no research studies that consider false alarms to reliability of the wind turbines. This paper presents a novel ap...
| Autores: | , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2019 |
| País: | España |
| Institución: | Universidad de Castilla-La Mancha |
| Repositorio: | RUIdeRA. Repositorio Institucional de la UCLM |
| OAI Identifier: | oai:ruidera.uclm.es:10578/21480 |
| Acceso en línea: | https://doi.org/10.1016/j.ress.2019.106574 http://hdl.handle.net/10578/21480 |
| Access Level: | acceso abierto |
| Palabra clave: | Reliability centred maintenance Condition monitoring Artificial neural network Wind energy conversion systems False alarms |
| Sumario: | Operations and maintenance tasks are critical to the reliability of a wind turbine. The state-of-the-art demonstrates the effectiveness of reliability centred maintenance, but there are no research studies that consider false alarms to reliability of the wind turbines. This paper presents a novel approach based on artificial neural networks to reliability centred maintenance. The methodology is employed for false alarm detection and prioritization, training the artificial neural networks over the time to increase the system reliability. The approach is applied to a real dataset from a supervisory control and data acquisition system together with a vibration monitoring system of a wind turbine. The results accuracy is done by confusion matrices, studding real alarms with the estimations provided by the approach, and the results are validated with real false alarms and compared by the results given by a fuzzy logic model. The method provides accuracy results (over 90%). A novelty is to use a two real dataset from a wind turbine to create a redundant response to detect false alarms by artificial neural networks. |
|---|