Wind Turbine Generator Fault Detection via Current Resampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional fault detection techniques for wind turbine generators face challenges in accurately identifying faults due to varying shaft rotating frequencies and low signal-to-noise ratios in current measurements, making it difficult to extract fault signatures and detect imbalances and aerodynamic asymmetries effectively.
Innovation Solution
The method involves frequency and amplitude demodulation of current data, resampling to convert variable frequencies into constant values, and performing frequency spectrum analysis to identify excitations at fault characteristic frequencies, using techniques like upsampling, downsampling, and impulse detection to generate alerts for fault detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional frequency spectrum analysis methods are used on current measurements, then fault detection can be performed, but the varying shaft rotating frequency causes difficulty in extracting fault signatures
Solution Approach 1:
The patent transforms the varying frequency problem by changing the reference frame through resampling at a fixed rate, converting variable frequency fault signatures into constant frequency components that can be reliably detected using traditional spectrum analysis methods
Solution Approach 2:
The patent introduces an intermediary resampling step that acts as a bridge between the variable frequency current measurements and the fixed frequency spectrum analysis, enabling fault signature extraction despite operating condition variations
2Reliability
If sensors are mounted on wind turbine components for monitoring, then fault detection capability is improved, but sensor failure due to poor working conditions causes additional reliability problems
Solution Approach 1:
The patent enables the generator to monitor its own health by analyzing its own current measurements, eliminating the need for separate sensors exposed to harsh environments. The generator's electrical signals serve as both operational data and diagnostic indicators
Solution Approach 2:
The patent replaces mechanical vibration sensors with electrical current analysis, substituting a sensor-based mechanical monitoring system with an electrical signal processing approach that uses the generator's own operational currents for fault detection
3Productivity
If current measurements are used for fault detection, then online monitoring is enabled, but the low signal to noise ratio makes fault detection difficult
Solution Approach 1:
The patent extracts fault-related information from the noisy current measurements by using resampling to isolate and concentrate fault signatures at specific frequency components, separating them from the dominant fundamental frequency and noise
Solution Approach 2:
The patent applies spectral analysis techniques that identify characteristic frequency patterns analogous to vibration analysis, detecting fault signatures through frequency domain characteristics even in the presence of noise and dominant fundamental components
Data Source
AI summary
A wind turbine generator fault detection method is described. The method includes acquiring current data from a wind turbine generator during operation, determining frequency demodulated signals and amplitude demodulated signals by frequency demodulating and amplitude demodulating the current data, resampling the frequency and amplitude demodulated signals corresponding to the current data, monitoring a frequency spectra of the resampled frequency and amplitude demodulated signals corresponding to the current data to identify one or more excitations in the frequency spectra. In response to identifying one or more excitations in the frequency spectra at one or more of the variable fault characteristic frequencies, the method includes generating and transmitting an alert that indicates that a wind turbine generator fault is detected.