Rail Corrugation Recognition Using SVM and Wavelet Noise Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing intelligent rail corrugation recognition methods have low recognition accuracy and efficiency, leading to inefficient maintenance and increased operational and maintenance costs.
Innovation Solution
A rail corrugation recognition method and apparatus based on a support vector machine, which extracts characteristic vectors from wheel-rail noise information and constructs a binary classification model for real-time detection of rail corrugation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If planned maintenance based on specified operation time is used for rail grinding, then maintenance scheduling is simple, but maintenance efficiency is low and operation and maintenance costs increase due to excessive or insufficient grinding
Solution Approach 1:
The patent replaces the traditional mechanical time-based maintenance scheduling system with an intelligent recognition system based on support vector machine and wavelet packet transformation. This system analyzes wheel-rail noise signals to automatically detect rail corrugation states, substituting manual planning with automated intelligent diagnosis, thereby improving maintenance efficiency while reducing scheduling complexity
Solution Approach 2:
The system enables the rail maintenance system to self-diagnose its own state by using noise signals from normal operation. The support vector machine model automatically identifies rail corrugation conditions without external intervention, allowing the system to serve itself in detecting maintenance needs, thus eliminating the need for complex external scheduling
2Measurement precision
If existing intelligent rail corrugation recognition methods are used, then real-time detection is achieved, but recognition accuracy and efficiency are low
Solution Approach 1:
The patent transforms the recognition approach by changing the parameter extraction method from traditional features to wavelet packet transformation coefficients. This parameter transformation enables more accurate representation of noise signals in different frequency bands, significantly improving recognition accuracy while maintaining real-time processing capability through efficient wavelet decomposition
Solution Approach 2:
The patent applies segmentation by dividing the wheel-rail noise signal into different frequency bands through wavelet packet transformation. This segmentation allows the support vector machine to analyze specific frequency components separately, improving recognition accuracy for different types of rail corrugation while maintaining computational efficiency through targeted feature analysis
Data Source
AI summary
The present disclosure discloses a rail corrugation recognition method and apparatus based on a support vector machine, a device, and a medium. The method includes: obtaining wheel-rail noise signals in different time periods, and obtaining wheel-rail noise time domain information; dividing the wheel-rail noise time domain information into segmented wheel-rail noise time domain information corresponding to each of the different time periods; preprocessing each piece of segmented wheel-rail noise time domain information, and extracting a time domain statistical characteristic quantity and frequency domain eigenmode energy of each piece of segmented wheel-rail noise time domain information, to obtain a multi-dimensional wheel-rail noise characteristic vector; constructing a rail corrugation state recognition model based on a support vector machine, and training the rail corrugation state recognition model; and recognizing to-be-recognized wheel-rail noise data by using the rail corrugation state recognition model based on a support vector machine, to obtain a rail corrugation state.