Broken Rail Prediction and Detection Using Wavelets and Artificial Neural Networks

2011 ◽  
Author(s):  
Brad M. Hopkins ◽  
Saied Taheri
Keyword(s):  
Neural Networks ◽  
Defect Detection ◽  
Classification Algorithm ◽  
Vertical Acceleration ◽  
Acceleration Signal ◽  
Acceleration Data ◽  
Frequency Domains ◽  
Artificial Neural ◽  
The Fourier Transform ◽  
Signal Processing Methods

Current track health monitoring requires time consuming use of railway monitoring vehicles. This paper presents a rail defect detection and classification algorithm that could potentially be used with bogie side frame vertical acceleration data from a data acquisition system located onboard a train car during daily operation. The algorithm uses wavelets to process the vertical acceleration data and detect irregularities in the signal. Wavelets have proven themselves to be useful in event detection and other applications where localization is needed in both the time and frequency domains. Traditional signal processing methods may use the Fourier transform which is limited to localization only in the frequency domain. Wavelets provide a solution for recognizing rail defects and determining their location. The wavelet-processed data is fed into an artificial neural network for defect classification. Neural networks can be a powerful tool in pattern recognition and classification because of their ability to be taught. The network in this algorithm has been trained to recognize impending breaks and breaks in a rail from the original vertical acceleration signal and the first four scales of the wavelet transformed signal. This paper presents an offline analysis of a set of collected data using the proposed defect detection and classification algorithm.

Rail Track Defect Detection Using Derivative Wavelet Transform

2012 ◽  
Author(s):  
Shahyar Taheri ◽  
Saied Taheri
Keyword(s):  
Wavelet Transform ◽  
Defect Detection ◽  
Vertical Acceleration ◽  
High Frequencies ◽  
Acceleration Signal ◽  
Detection Systems ◽  
Limit Value ◽  
Accelerometer Signal ◽  
Acceleration Data ◽  
Track Monitoring

Railroad track monitoring systems are used for finding rail defects that may lead to a derailment of the train. The classical limit-value based defect detection systems are simple but are limited in their capability to detect small defects. As a cutting-edge supervision method, signal derivative filters can help to reveal information in the acceleration signal collected while the train is moving on the rail. The derivative filters are designed based on the required performance of the application. However, their design should be done with caution because they can greatly amplify the noise in the data, especially in high frequencies. Derivative filters can be implemented in the sample domain of space or time. The derivative filters in time domain are not always sufficient to study all the features of a signal. To explore the signal content, wavelet transformation was chosen, because it gives accurate description of the frequency contents according to their position in time. It should be noted that the wavelet transform that gives the derivative of a signal, has the properties of smoothing and differentiation. The proposed algorithm processes the data using continuous and discrete derivative wavelets filters, and is able to locate defects and provide information that may help to distinguish between various types of rail and wheel defects, including rail cracks, squats, corrugation, and wheel out-of-rounds. The wavelet-based algorithm developed was applied to a sample accelerometer signal and the results show the potential of this algorithm to locate and diagnose defects from limited bogie vertical acceleration data.

Evaluation and Modeling of Passenger Vehicle Celeration Performance Based on Artificial Neural Networks

2015 ◽  
Vol 744-746 ◽  
pp. 1938-1942
Author(s):  
Yi He ◽  
Duan Feng Chu
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Comfort Level ◽  
Subjective Rating ◽  
Acceleration Data ◽  
Acceleration And Deceleration ◽  
The Subject ◽  
Artificial Neural ◽  
The Relationship ◽  
Subjective Sensation

As the siginificant factors influence passengers comfort, the vehicle celebration performance may easy to cause accidents, such as hard acceleration and deceleration performance. In order to find the relationship between passengers comfort and celebration performance, 35 passengers and three professional drivers were recruited in the field experiment. The passengers’ comfort feelings were analysed by subject questionnaires, the acceleration and deceleration data were received by CAN bus.The Artificial Neural Networks (ANNs) model was elaborated to estimate and predict the passengers comfort level of driver unsafe acceleration behavior situations. Therefore, the subject views of the passengers could be compared to object acceleration data. An ANN is applied to interconnect output data (subjective rating) with input data (objective parameters). Finally, it is found the investigatioin have demonstrated that the objective values are efficiently correlated with the subjective sensation. Thus, the presented approach can be effectively applied to support the drive train development of bus.

Track Health Monitoring Using Wavelets

2010 ◽  
Author(s):  
Brad M. Hopkins ◽  
Saied Taheri
Keyword(s):  
Signal Processing ◽  
Fourier Transform ◽  
Defect Detection ◽  
Health Monitoring ◽  
Finite Energy ◽  
Detection Algorithm ◽  
Transient Signal ◽  
Vertical Acceleration ◽  
Typical Day ◽  
The Fourier Transform

This paper presents a defect detection algorithm for rail health monitoring that could potentially be used with limited bogie. Current wheel and track monitoring requires expensive track instrumentation and/or time consuming operation of railway monitoring vehicles. The proposed health monitoring algorithm can potentially be used with a portable data acquisition system that can be relocated from train to train to monitor and diagnose the conditions of the track as a train is driven during typical day-to-day operation. The algorithm processes the data using wavelets and is able to locate defects and provide information that may help to distinguish between various types of rail defects. In recent years, wavelets have been used extensively in signal processing because of their ability to analyze a signal simultaneously in the time and frequency domains. The Fourier transform has been used traditionally in signal processing to locate dominant frequencies in a signal, but it is unable to provide time localization of those frequencies. Unlike the Fourier transform, the wavelet transform uses a set of basis functions with finite energy, which is advantageous for detecting the irregular events that may show up in a transient signal. The wavelets used in the proposed signal processing routine were chosen for optimal signal decomposition through consideration of the signals that are likely to be generated from common rail and wheel defects, including rail cracks, squats, corrugation, and, wheel out-of-rounds. A sample accelerometer signal was generated from information found in existing literature and was then processed using the proposed defect detection algorithm. Results show the potential of this algorithm to locate and diagnose defects from limited bogie vertical acceleration data. This study is intended to present a proof-of-concept for the proposed defect detection algorithm, providing a basis for which a more comprehensive defect detection and diagnosis algorithm can be developed.

Sign in / Sign up

Export Citation Format

Share Document