Wind Turbine Bearing Fault Detection via Frequency Segmentation

The patent applies preliminary action by calculating the characteristic frequency of the bearing defect before performing spectrum analysis. This allows the system to pre-identify the frequency range of interest and apply targeted filtering or enhancement only in that range, avoiding the need to increase overall frequency resolution which would introduce phantom components. The characteristic frequency is calculated based on bearing geometry and operating conditions, enabling focused detection without global resolution increases.

The patent introduces an intermediary approach by using measured vibration signals to empirically determine characteristic frequencies rather than relying solely on theoretical calculations that require precise contact angle measurements. The system uses signal processing techniques to identify actual operating frequencies from the vibration data itself, acting as an intermediary between theoretical models and actual fault detection. This avoids the complexity of directly measuring or calculating contact angles while still achieving accurate frequency identification.

The patent applies segmentation by dividing the vibration spectrum into multiple frequency bands or channels, each centered on characteristic defect frequencies and their harmonics. Instead of using a single global indicator, the system calculates separate indicators for each frequency channel, allowing sensitive detection of low-energy faults in specific frequency ranges while maintaining operational simplicity. This segmented approach enables targeted monitoring of bearing defect frequencies without requiring complex multi-parameter analysis.