Wind turbine multi-source heterogeneous voiceprint sensor data synchronization and fusion method and system
By combining high-precision hardware synchronization and software correction with a deep learning fusion model based on attention mechanisms, the problem of data synchronization and fusion in wind turbine acoustic signature monitoring has been solved, achieving high-precision fault diagnosis and improving the accuracy and reliability of wind turbine condition monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUANENG TONGLIAO WIND POWER CO LTD
- Filing Date
- 2026-02-28
- Publication Date
- 2026-06-12
AI Technical Summary
Existing wind turbine acoustic signature monitoring technology suffers from insufficient accuracy in data synchronization and fusion, weak multi-source heterogeneous data processing capabilities, and insufficient resistance to strong background noise interference, resulting in the inability of fault diagnosis results to meet the needs of engineering applications in terms of accuracy and reliability.
A method combining high-precision hardware synchronous acquisition with software-level synchronous signal correction is adopted to correct and align multi-source heterogeneous data. Feature extraction and fusion are performed through a deep learning fusion model based on attention mechanism. Spatial information of sensor array is used to suppress environmental noise and achieve adaptive weighted fusion of multimodal information.
It achieves high-precision time synchronization from microsecond level to sample level, effectively solves the clock drift problem, maximizes the complementary advantages of multimodal information, improves the accuracy and reliability of fault diagnosis, significantly improves the signal-to-noise ratio of wind turbine fault characteristics, and makes early diagnosis of weak faults possible.
Smart Images
Figure CN122196881A_ABST