Structure Health Diagnosis of Metro Rail Track by Using Vibration Mappings and Machine Learning

EasyChair Preprint 9631

Abstract

This abstract presents the methodology for the evaluation of metro rail tracks which can be used to produce a vibration map, also it is very fundamental for track maintenance, where we predict vehicle body vibration based on deep learning, which represents one of the newest areas in the Artificial Intelligence field. In the entire world, track quality had been evaluated through established track geometry standards. However, due to the incapability of these types of standards to detect some abnormality of track geometry conditions, that can cause vehicle body vibration. These vibrations also shake up nearby residents; measurements are carried out to evaluate the risk for structural integrity. To design railway tracks using SolidWorks software and structure & health diagnosis using Ansys software integration with experimental data for feature extraction used EEMD and for classification apply artificial neural network (ANN) where a model is proposed to make an accurate and point-wise prediction, due to which we can achieve optimal performance. This case study is based on the Delhi Metro Rail Corporation (DMRC), for track health monitoring system was installed on several trains running on the green line in the underground which aims to improve the maintenance process. The early detection and surveillance of defects help to extend the service life of the tracks and it also reduced operating costs. A data acquisition system is used to analyze the continuously recorded measurements (vehicle body vibration), which consist of vertical bogie acceleration and surrounding noise, for example with a frequency of 22kHz. However, ANN- LSTM can predict vertical vehicle-body vibration below 10Hz and lateral vehicle-body vibration below 1Hz. The above analysis shows that the performance of metro rail tracks by using the vehicle body vibration method acts as a performance-based model which evaluated the index of track quality.

Keyphrases: Artificial Neural Network (ANN), Track quality evaluation, machine learning, track health diagnosis, vehicle-body vibration

@booklet{EasyChair:9631,
  author    = {Madhavendra Saxena and Parag Jain and Pankaj Dhiman and Priya Singh},
  title     = {Structure Health Diagnosis of Metro Rail Track by Using Vibration Mappings and Machine Learning},
  howpublished = {EasyChair Preprint 9631},
  year      = {EasyChair, 2023}}