Fault Diagnosis and Data Reconstruction of Temperature Sensors for Wind Turbine Stator Winding
A proposed method for diagnosing sensor faults in the stator winding temperature sensor of a wind turbine is introduced, aiming to address the significant issues that arise when such malfunctions occur, affecting the overall stable operation of the turbine. The method combines supervisory control an...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2025-01-01
|
| Series: | Shock and Vibration |
| Online Access: | http://dx.doi.org/10.1155/vib/4713545 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | A proposed method for diagnosing sensor faults in the stator winding temperature sensor of a wind turbine is introduced, aiming to address the significant issues that arise when such malfunctions occur, affecting the overall stable operation of the turbine. The method combines supervisory control and data acquisition (SCADA) with an improved firefly sparrow search algorithm (FISSA)–optimized deep belief network (DBN). The input parameters of the model are determined via Pearson’s correlation coefficient method, whereas the exponentially weighted moving average (EWMA) is utilized to detect faults. To ensure system stability in the event of sensor failure, a multi-input single-output sensor fault data reconstruction model is proposed on the basis of long-term and short-term memory (LSTM) networks, which are further optimized using the seagull optimization algorithm (SOA-LSTM). An illustrative verification is subsequently conducted using actual data obtained from a wind farm located in Central China. The results of the verification demonstrate that the proposed method can provide an alarm time more than 4 days in advance compared with the actual recording time. It is beneficial for wind farm staff to replace equipment in advance and improve the economic benefits of wind farm operation. |
|---|---|
| ISSN: | 1875-9203 |