Gear Fault Detection via Envelope Synchronous Averaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting mechanical faults in gears, such as those in aircraft or helicopter engines, are often inaccurate due to the strong meshing signal overpowering the modulating signals, and are not robust against interferences, making it difficult to determine the shape of the modulating signals effectively.
Innovation Solution
A method involving filtering the vibratory signal using an optimal filter, determining the square envelope of the signal through Hilbert transform, and calculating a synchronous average to attenuate spurious components, followed by normalization to obtain a monitoring indicator that reflects the state of the gear teeth, which is displayed in polar coordinates to indicate faults.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If classical spectrum monitoring methods are used to detect gear faults, then the meshing signal can be analyzed, but the fault signature is obscured by the strong meshing signal making detection inaccurate
Solution Approach 1:
The patent extracts the modulating signals containing fault information from the composite gear vibration signal by separating them from the dominant meshing signal through spectral analysis and envelope detection, enabling accurate fault detection despite the strong meshing interference
Solution Approach 2:
The patent introduces an intermediary processing approach using envelope detection and spectral kurtosis analysis to bridge the gap between the strong meshing signal and the weak fault signatures, allowing the extraction of fault information that would otherwise be obscured
2Measurement precision
If demodulation methods are used to restore modulating signal envelope, then fault information can be extracted, but performance degrades in the presence of periodic interferences and stationary colored noise
Solution Approach 1:
The patent changes the analysis parameters by using spectral kurtosis to identify optimal frequency bands and applying adaptive filtering techniques that adjust to the interference characteristics, maintaining reliable fault detection in noisy environments with periodic interferences and colored noise
Solution Approach 2:
The patent implements feedback through iterative refinement of the envelope detection process, where the spectral characteristics are continuously analyzed and the processing parameters are adjusted based on the detected signal properties, improving robustness against varying interference conditions
3Measurement precision
If source separation methods are used to restore modulating signals effectively, then fault information can be extracted, but the calculation cost becomes excessively high
Solution Approach 1:
The patent applies partial action by focusing computational resources only on the specific frequency bands containing fault information, identified through spectral kurtosis analysis, rather than processing the entire signal spectrum, thus achieving effective fault detection with reduced computational complexity
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method provides a simple, intuitive, and robust means to detect and characterize gear faults, distinguishing between local and distributed faults, while being resistant to mechanical and electromagnetic noise, with reasonable computational resources.
Implementation Method 1
determining, E3, the square envelope of said monitoring signal, defined by the squared absolute value of the Hilbert transform of the monitoring signal
Data Source
AI summary
The invention concerns a method for monitoring a gear system (1) comprising at least two wheels, each wheel having a characteristic frequency (f1, f2), a vibratory or acoustic signal representative of these vibrations having been acquired by a sensor (21), a vibratory or acoustic digital signal x(t) having been obtained, the method comprising, for each characteristic frequency of the system (1), the steps of filtering (E1) the digital signal by means of a filter in such a way as to obtain an image monitoring signal of at least one vibratory component of at least one defect; determining (E3) the square envelope of said monitoring signal, defined by the square absolute value of the Hilbert transform of the monitoring signal so as to extract an image of the modulating signals associated with the rotation of the wheels from the monitoring signal; determining (E4) the synchronous average of said square envelope relative to the period of rotation of a wheel of interest chosen from the wheels of the gear, said average helping attenuate spurious components in the envelope originating from other mechanical sources; determining (E5) a monitoring indicator dependent on said synchronous average, said monitoring indicator being defined by the square root of the synchronous average normalised by the median value of the synchronous average.