Anomaly Detection in the Time Series Data from Fehn Pollux Ship with ECO Flettner Rotor
- An ECO Flettner rotor has been installed on board the vessel MV Fehn Pollux to reduce the vessel’s carbon emissions and save fuel. The extent of fuel-saving is assessed using recorded data of apparent wind speed, apparent wind angle, and rotor speed by the vessel’s data acquisition and storage system. However, the data contains anomalies caused by noise, vibration, or errors. Detecting anomalies could help to understand the reason for their occurrence, improve the calculation of energy savings, and increase the accuracy of the trained models. To detect anomalies in apparent wind speed, apparent wind angle, and rotor speed, three anomaly detection approaches are proposed. The paper describes the proposed anomaly detection concepts, and it gives an insight into their implementation process. Additionally, it evaluates proposed anomaly detection capabilities.
| Author: | Farzaneh Nourmohammadi, Allanazar Jumabayev, Elmar Wings |
|---|---|
| DOI: | https://doi.org/10.1109/INDIN45523.2021.9557422 |
| ISBN: | 978-1-7281-4395-8 |
| Parent Title (English): | IEEE 19th International Conference on Industrial Informatics (INDIN 2021), 21.07.-23.07.2021, Palma de Mallorca (Spain) |
| Publisher: | IEEE |
| Document Type: | Conference Proceeding |
| Language: | English |
| Year of Completion: | 2021 |
| Release Date: | 2025/03/07 |
| Tag: | Anomaly Detection; Deep Neural Networks; Flettner Rotor; Long-Short Term Memory; Moving Median; Recurrent Neural Networks |
| Pagenumber: | 6 |
| First Page: | 1 |
| Last Page: | 6 |
| Institute: | Fachbereich Technik |
| Research Focus Area: | Industrielle Informatik |