An End-to-End, Real-Time Solution for Condition Monitoring of Wind Turbine Generators
Data-driven wind generator condition monitoring systems largely rely on multi-stage processing involving feature selection and extraction followed by supervised learning. These stages require expert analysis, are potentially error-prone and do not generalize well between applications. In this paper,...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-09-01
|
| Series: | Energies |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1996-1073/13/18/4817 |
| _version_ | 1797553670975389696 |
|---|---|
| author | Adrian Stetco Juan Melecio Ramirez Anees Mohammed Siniša Djurović Goran Nenadic John Keane |
| author_facet | Adrian Stetco Juan Melecio Ramirez Anees Mohammed Siniša Djurović Goran Nenadic John Keane |
| author_sort | Adrian Stetco |
| collection | DOAJ |
| description | Data-driven wind generator condition monitoring systems largely rely on multi-stage processing involving feature selection and extraction followed by supervised learning. These stages require expert analysis, are potentially error-prone and do not generalize well between applications. In this paper, we introduce a collection of end-to-end Convolutional Neural Networks for advanced condition monitoring of wind turbine generators. End-to-end models have the benefit of utilizing raw, unstructured signals to make predictions about the parameters of interest. This feature makes it easier to scale an existing collection of models to new predictive tasks (e.g., new failure types) since feature extracting steps are not required. These automated models achieve low Mean Squared Errors in predicting the generator operational state (40.85 for Speed and 0.0018 for Load) and high accuracy in diagnosing rotor demagnetization failures (99.67%) by utilizing only raw current signals. We show how to create, deploy and run the collection of proposed models in a real-time setting using a laptop connected to a test rig via a data acquisition card. Based on a sampling rate of 5 kHz, predictions are stored in an efficient time series database and monitored using a dynamic visualization framework. We further discuss existing options for understanding the decision process behind the predictions made by the models. |
| first_indexed | 2024-03-10T16:19:54Z |
| format | Article |
| id | doaj.art-bd1b6b1281f6457dbc20821eeb682b37 |
| institution | Directory Open Access Journal |
| issn | 1996-1073 |
| language | English |
| last_indexed | 2024-03-10T16:19:54Z |
| publishDate | 2020-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Energies |
| spelling | doaj.art-bd1b6b1281f6457dbc20821eeb682b372023-11-20T13:47:51ZengMDPI AGEnergies1996-10732020-09-011318481710.3390/en13184817An End-to-End, Real-Time Solution for Condition Monitoring of Wind Turbine GeneratorsAdrian Stetco0Juan Melecio Ramirez1Anees Mohammed2Siniša Djurović3Goran Nenadic4John Keane5Department of Computer Science, University of Manchester, Manchester M13 9PL, UKDepartment of Electrical and Electronic Engineering, University of Manchester, Manchester M1 3BB, UKDepartment of Electrical and Electronic Engineering, University of Manchester, Manchester M1 3BB, UKDepartment of Electrical and Electronic Engineering, University of Manchester, Manchester M1 3BB, UKDepartment of Computer Science, University of Manchester, Manchester M13 9PL, UKDepartment of Computer Science, University of Manchester, Manchester M13 9PL, UKData-driven wind generator condition monitoring systems largely rely on multi-stage processing involving feature selection and extraction followed by supervised learning. These stages require expert analysis, are potentially error-prone and do not generalize well between applications. In this paper, we introduce a collection of end-to-end Convolutional Neural Networks for advanced condition monitoring of wind turbine generators. End-to-end models have the benefit of utilizing raw, unstructured signals to make predictions about the parameters of interest. This feature makes it easier to scale an existing collection of models to new predictive tasks (e.g., new failure types) since feature extracting steps are not required. These automated models achieve low Mean Squared Errors in predicting the generator operational state (40.85 for Speed and 0.0018 for Load) and high accuracy in diagnosing rotor demagnetization failures (99.67%) by utilizing only raw current signals. We show how to create, deploy and run the collection of proposed models in a real-time setting using a laptop connected to a test rig via a data acquisition card. Based on a sampling rate of 5 kHz, predictions are stored in an efficient time series database and monitored using a dynamic visualization framework. We further discuss existing options for understanding the decision process behind the predictions made by the models.https://www.mdpi.com/1996-1073/13/18/4817wind turbinereal-time diagnosticgeneratorconvolutional neural networkscondition monitoring |
| spellingShingle | Adrian Stetco Juan Melecio Ramirez Anees Mohammed Siniša Djurović Goran Nenadic John Keane An End-to-End, Real-Time Solution for Condition Monitoring of Wind Turbine Generators Energies wind turbine real-time diagnostic generator convolutional neural networks condition monitoring |
| title | An End-to-End, Real-Time Solution for Condition Monitoring of Wind Turbine Generators |
| title_full | An End-to-End, Real-Time Solution for Condition Monitoring of Wind Turbine Generators |
| title_fullStr | An End-to-End, Real-Time Solution for Condition Monitoring of Wind Turbine Generators |
| title_full_unstemmed | An End-to-End, Real-Time Solution for Condition Monitoring of Wind Turbine Generators |
| title_short | An End-to-End, Real-Time Solution for Condition Monitoring of Wind Turbine Generators |
| title_sort | end to end real time solution for condition monitoring of wind turbine generators |
| topic | wind turbine real-time diagnostic generator convolutional neural networks condition monitoring |
| url | https://www.mdpi.com/1996-1073/13/18/4817 |
| work_keys_str_mv | AT adrianstetco anendtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT juanmelecioramirez anendtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT aneesmohammed anendtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT sinisadjurovic anendtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT gorannenadic anendtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT johnkeane anendtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT adrianstetco endtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT juanmelecioramirez endtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT aneesmohammed endtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT sinisadjurovic endtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT gorannenadic endtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators AT johnkeane endtoendrealtimesolutionforconditionmonitoringofwindturbinegenerators |