Bearing Defect Detection by Swept Speed Pattern Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional condition monitoring applications for bearing defects in trains are costly and prone to errors due to the need for accurate shaft speed measurements, which are often unreliable and require extensive manual analysis, especially when wheel diameters change.
Innovation Solution
A method using a computer system that processes vibration data from sensors to identify bearing defects without knowing the shaft speed, by sweeping patterns across a speed range and comparing them to predefined defect patterns, allowing for automatic detection of defects such as spalling, brinelling, and lubrication issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional condition monitoring applications use shaft speed sensors to identify bearing defects, then defect detection capability is improved, but system cost and complexity increase
Solution Approach 1:
The patent extracts the shaft speed sensor from the system entirely. Instead of using a physical speed sensor, the invention uses pattern sweeping algorithms that operate on vibration data across a range of speeds without requiring actual speed measurement, thereby eliminating the sensor hardware while maintaining defect detection capability
Solution Approach 2:
The patent replaces the mechanical shaft speed sensor with a computational approach. The pattern sweeping algorithm processes vibration spectra mathematically to identify bearing defects without mechanical speed sensing, substituting a mechanical measurement system with a signal processing method
2Measurement precision
If conventional condition monitoring applications require accurate shaft speed measurements within a few percent tolerance, then vibration spectrum frequency component identification is improved, but measurement precision requirements increase system cost and error proneness
Solution Approach 1:
The patent applies dynamics by sweeping patterns across a range of speeds rather than fixing them to a single measured speed. The pattern sweeping process dynamically adjusts frequency components across multiple speed values, allowing the system to identify defects without requiring precise knowledge of the actual shaft speed at any given moment
Solution Approach 2:
The patent changes the parameter approach by instead of measuring speed precisely and using that single value, the system varies speed parameters across a range during pattern sweeping. This parameter variation allows defect identification to be robust against speed measurement errors or uncertainties
3Adaptability or versatility
If conventional condition monitoring applications manually manage wheel diameter parameters affecting shaft speed calculations, then adaptability to wheel wear is improved, but productivity decreases due to extensive manual analysis
Solution Approach 1:
The patent enables the system to be self-sufficient by automatically compensating for wheel diameter changes through pattern sweeping. The algorithm inherently adapts to varying wheel conditions without requiring manual intervention to update parameters, as the sweeping process naturally accounts for speed variations caused by wheel wear
Solution Approach 2:
The patent performs preliminary pattern sweeping across the entire speed range before definitive defect identification. This preliminary action pre-computes the expected vibration patterns at various speeds, allowing the system to quickly identify defects without requiring manual parameter adjustments when wheel conditions change
Data Source
AI summary
A method and system is provided for performing speed and defect identification of a component such as, for example, a bearing. The method can be implemented by a computer, such that the computer receives from one or more sensors condition monitoring data. The computer sweeps patterns along a speed range against the condition monitoring data and multiplies each pattern component of the patterns by a matching environmental spectral component. The computer, then, adds the pattern components together to produce one or more results.