Separating Periodic Vibration Peaks Using Kurtosis Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods require experienced analysts to identify periodic and non-periodic amplitude peaks in machine vibration data, leading to noisy results due to the estimation of Percent Periodic Energy, which complicates fault prediction in machinery.
Innovation Solution
A statistical method using kurtosis to separate periodic peaks from noise in autocorrelation spectra by sequentially removing large peaks until the kurtosis reaches a threshold, allowing for the creation of a cleaner Periodic Information Plot (PIP) that accurately identifies periodic peaks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the Percent Periodic Energy is estimated using traditional methods, then the PIP can be generated, but the result contains excessive noise
Solution Approach 1:
The patent segments the amplitude peaks into periodic and non-periodic components by analyzing the autocorrelation waveform. The periodic peaks are identified through statistical analysis of the autocorrelation function, while non-periodic peaks (noise) are separated out. This segmentation allows the PIP to contain only periodic peaks, eliminating the noise problem while maintaining accurate periodic peak identification.
2Measurement precision
If experienced vibration analysts are used to identify periodic patterns, then accurate fault identification can be achieved, but the process requires specialized expertise and is time-consuming
Solution Approach 1:
The patent implements an automated system that performs fault identification without requiring experienced vibration analysts. The system uses computational algorithms including autocorrelation analysis, statistical measures, and pattern recognition to automatically identify periodic peaks and diagnose machine faults. This self-service approach eliminates the need for specialized human expertise while maintaining high accuracy, and significantly reduces the time required for analysis.
3Loss of information
If all amplitude peaks are included in the PIP, then comprehensive vibration information is captured, but the plot becomes noisy and difficult to interpret
Solution Approach 1:
The patent extracts only the periodic peaks from the complete set of amplitude peaks by analyzing the autocorrelation waveform. The statistical analysis identifies which peaks exhibit periodic characteristics, and these are extracted to form the PIP. Non-periodic peaks (noise) are excluded from the PIP. This extraction process maintains all relevant vibration information while eliminating noise, making the plot clear and interpretable.
Data Source
AI summary
A statistical method is used to separate periodic from non-periodic vibration peaks in autocorrelation spectra. Generally, an autocorrelation spectrum is not normally distributed because the amplitudes of periodic peaks are significantly large and random relative to the generally Gaussian noise. In a normally distributed signal, the statistical parameter kurtosis has a value of 3. The method sequentially removes each largest amplitude peak from the peaks in the spectrum until the kurtosis is 3 or less. The removed peaks, which are all considered to be periodic, are placed into a set. The total energy of the peaks in the set is considered to be the total periodic energy of the spectrum. As the process of building the peak set proceeds, if its total energy becomes greater than or equal to a predefined energy threshold before its kurtosis reaches 3 or less, the process stops and the periodic peak set is defined.