Rail Wheel Defect Detection Using Median Signal Reference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting wheel defects, such as flat wheels, in rail vehicles are unreliable due to signal peaks from minor spikes affecting average values, leading to false negatives or false positives.
Innovation Solution
A measurement system using sensors to detect force effects on the rail, processing median data samples to establish a reference value for the normal wheel part, comparing data to this reference, and informing users of deviations exceeding predetermined thresholds, with optional derivative and integral filtering for enhanced accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If average value of measurement signal is used for comparison, then measurement method is simple, but reliability of defect detection deteriorates due to signal peaks from minor spikes affecting the average value
Solution Approach 1:
The patent changes the statistical parameter used for comparison from mean (average) to median. This parameter change makes the reference value robust against outliers and signal peaks caused by minor spikes, thereby improving defect detection reliability while maintaining method simplicity. The median inherently filters the effect of extreme values without requiring complex preprocessing.
2Reliability
If median data sample is used to establish reference value, then reliability of defect detection is improved, but data processing complexity increases
Solution Approach 1:
The patent extracts only the necessary characteristic (median value) from the measurement signal for comparison purposes. By focusing on this single robust statistical parameter rather than processing the entire signal complexly, the method achieves high reliability while keeping data processing relatively simple. The median extraction is computationally efficient compared to more complex signal processing techniques.
3Measurement precision
If measurement zone covers entire wheel circumference, then complete wheel inspection is achieved, but measurement time increases
Solution Approach 1:
The patent establishes a reference value (median) from the measurement signal before performing the actual defect comparison. This preliminary action allows the system to quickly compare subsequent measurement points against the pre-established reference, reducing the time required for complete wheel inspection while maintaining comprehensive coverage. The reference value is computed once and reused for multiple comparisons.
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 approach provides reliable and accurate detection of wheel defects, minimizing false alarms and ensuring timely identification of broken wheels, thereby improving safety and reducing damage to rail and vehicle structures.
Implementation Method 1
sensors adapted to measure force effects induced by the wheel to the rail
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The present invention relates to a solution for measuring forces of wheels travelling on a rail, and detecting defects on the wheels. In prior art solutions the sensor signals of defected wheels are compared to values of other, normal wheels, or to an average value of the defected wheel under measurement. This may lead to unreliable results. In the present invention a reference value is used for describing the force of a normal part of the wheel under measurement. The reference value is achieved by selecting a reference sample from the sensor signal of the wheel under measurement. The selected sample is preferably a median sample of the signal of the wheel under measurement, whereby a median value of the sensor signal is used as the reference value. Derivation, high-pass filtering and integration of the signal may further be used in the analysis of the sensor signals.