Tamping effectiveness prediction using supervised machine learning techniques

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearch

Abstract

Railway maintenance planning is critical in maintaining track assets. Tamping is a common railway maintenance procedure and is often used when geometrical issues are first identified. Tamping repacks ballast particles under sleepers to restore the correct geometrical position of ballasted tracks. However, historical data shows that tamping is not always effective in restoring track to a satisfactory condition. Furthermore, ineffective, or unnecessary tamping tends to reduce the lifetime of existing track. An intuitive way of preventing ineffective tamping is to predict the likely tamping effectiveness. This work aims to predict the likely tamping effectiveness ahead of time using supervised machine learning techniques. Supervised machine learning techniques predict an outcome using labelled training data. In this case, the training database consists of multivariate sensor data from instrumented revenue vehicles (IRVs). The data between the previous and current tamping dates are used. This forms a time series database labelled with the tamping effectiveness of each track location based on the responses recorded from the IRVs before and after tamping. The labelled time series database is then used to train a time series classifier for prediction. This work uses the state of the art time series classification algorithm, k-nearest neighbour (k-NN) extended to the case of multivariate time series. k-NN is a non-parametric algorithm that does not make assumptions on the underlying model of the training data. With a sufficiently large training database, non-parametric algorithms can outperform parametric algorithms. Using k-NN, the tamping effectiveness of a potential tamping location that is not in the training database, or locations in the next tamping cycle, is predicted using the expected tamping effectiveness from a location in the training database that is the most similar to the target. This allows the algorithm to effectively to identify locations where tamping is likely to be ineffective. This work achieves high accuracy in the prediction of tamping effectiveness even at 12 weeks before tamping. It is hoped that the methodology will help in assisting decision making for maintenance planning activities.

Original languageEnglish
Title of host publicationICRT 2017 - Railway Development, Operations, and Maintenance
Subtitle of host publicationProceedings of the First International Conference on Rail Transportation 2017 - July 10–12, 2017 Chengdu, Sichuan Province, China
EditorsWanming Zhai, Kelvin C. P. Wang
Place of PublicationReston Virginia USA
PublisherAmerican Society of Civil Engineers
Pages1010-1023
Number of pages14
ISBN (Electronic)9780784481257
DOIs
Publication statusPublished - 2018
EventIEEE International Conference on Intelligent Rail Transportation (ICIRT) 2017 - Chengdu Sichuan Province, China
Duration: 10 Jul 201712 Jul 2017
Conference number: 1st

Conference

ConferenceIEEE International Conference on Intelligent Rail Transportation (ICIRT) 2017
Abbreviated titleICIRT 2017
CountryChina
CityChengdu Sichuan Province
Period10/07/1712/07/17

Keywords

  • Dynamic time warping
  • Machine learning
  • Railway maintenance
  • Tamping effectiveness
  • Time series classification

Cite this

Tan, C. W., Webb, G. I., Petitjean, F., & Reichl, P. (2018). Tamping effectiveness prediction using supervised machine learning techniques. In W. Zhai, & K. C. P. Wang (Eds.), ICRT 2017 - Railway Development, Operations, and Maintenance: Proceedings of the First International Conference on Rail Transportation 2017 - July 10–12, 2017 Chengdu, Sichuan Province, China (pp. 1010-1023). Reston Virginia USA: American Society of Civil Engineers. https://doi.org/10.1061/9780784481257.101
Tan, Chang Wei ; Webb, Geoffrey I. ; Petitjean, Francois ; Reichl, Paul. / Tamping effectiveness prediction using supervised machine learning techniques. ICRT 2017 - Railway Development, Operations, and Maintenance: Proceedings of the First International Conference on Rail Transportation 2017 - July 10–12, 2017 Chengdu, Sichuan Province, China. editor / Wanming Zhai ; Kelvin C. P. Wang. Reston Virginia USA : American Society of Civil Engineers, 2018. pp. 1010-1023
@inproceedings{70baabb9bcf940f4a6afc17cc7a70163,
title = "Tamping effectiveness prediction using supervised machine learning techniques",
abstract = "Railway maintenance planning is critical in maintaining track assets. Tamping is a common railway maintenance procedure and is often used when geometrical issues are first identified. Tamping repacks ballast particles under sleepers to restore the correct geometrical position of ballasted tracks. However, historical data shows that tamping is not always effective in restoring track to a satisfactory condition. Furthermore, ineffective, or unnecessary tamping tends to reduce the lifetime of existing track. An intuitive way of preventing ineffective tamping is to predict the likely tamping effectiveness. This work aims to predict the likely tamping effectiveness ahead of time using supervised machine learning techniques. Supervised machine learning techniques predict an outcome using labelled training data. In this case, the training database consists of multivariate sensor data from instrumented revenue vehicles (IRVs). The data between the previous and current tamping dates are used. This forms a time series database labelled with the tamping effectiveness of each track location based on the responses recorded from the IRVs before and after tamping. The labelled time series database is then used to train a time series classifier for prediction. This work uses the state of the art time series classification algorithm, k-nearest neighbour (k-NN) extended to the case of multivariate time series. k-NN is a non-parametric algorithm that does not make assumptions on the underlying model of the training data. With a sufficiently large training database, non-parametric algorithms can outperform parametric algorithms. Using k-NN, the tamping effectiveness of a potential tamping location that is not in the training database, or locations in the next tamping cycle, is predicted using the expected tamping effectiveness from a location in the training database that is the most similar to the target. This allows the algorithm to effectively to identify locations where tamping is likely to be ineffective. This work achieves high accuracy in the prediction of tamping effectiveness even at 12 weeks before tamping. It is hoped that the methodology will help in assisting decision making for maintenance planning activities.",
keywords = "Dynamic time warping, Machine learning, Railway maintenance, Tamping effectiveness, Time series classification",
author = "Tan, {Chang Wei} and Webb, {Geoffrey I.} and Francois Petitjean and Paul Reichl",
year = "2018",
doi = "10.1061/9780784481257.101",
language = "English",
pages = "1010--1023",
editor = "Wanming Zhai and Wang, {Kelvin C. P.}",
booktitle = "ICRT 2017 - Railway Development, Operations, and Maintenance",
publisher = "American Society of Civil Engineers",
address = "United States of America",

}

Tan, CW, Webb, GI, Petitjean, F & Reichl, P 2018, Tamping effectiveness prediction using supervised machine learning techniques. in W Zhai & KCP Wang (eds), ICRT 2017 - Railway Development, Operations, and Maintenance: Proceedings of the First International Conference on Rail Transportation 2017 - July 10–12, 2017 Chengdu, Sichuan Province, China. American Society of Civil Engineers, Reston Virginia USA, pp. 1010-1023, IEEE International Conference on Intelligent Rail Transportation (ICIRT) 2017, Chengdu Sichuan Province, China, 10/07/17. https://doi.org/10.1061/9780784481257.101

Tamping effectiveness prediction using supervised machine learning techniques. / Tan, Chang Wei; Webb, Geoffrey I.; Petitjean, Francois; Reichl, Paul.

ICRT 2017 - Railway Development, Operations, and Maintenance: Proceedings of the First International Conference on Rail Transportation 2017 - July 10–12, 2017 Chengdu, Sichuan Province, China. ed. / Wanming Zhai; Kelvin C. P. Wang. Reston Virginia USA : American Society of Civil Engineers, 2018. p. 1010-1023.

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearch

TY - GEN

T1 - Tamping effectiveness prediction using supervised machine learning techniques

AU - Tan, Chang Wei

AU - Webb, Geoffrey I.

AU - Petitjean, Francois

AU - Reichl, Paul

PY - 2018

Y1 - 2018

N2 - Railway maintenance planning is critical in maintaining track assets. Tamping is a common railway maintenance procedure and is often used when geometrical issues are first identified. Tamping repacks ballast particles under sleepers to restore the correct geometrical position of ballasted tracks. However, historical data shows that tamping is not always effective in restoring track to a satisfactory condition. Furthermore, ineffective, or unnecessary tamping tends to reduce the lifetime of existing track. An intuitive way of preventing ineffective tamping is to predict the likely tamping effectiveness. This work aims to predict the likely tamping effectiveness ahead of time using supervised machine learning techniques. Supervised machine learning techniques predict an outcome using labelled training data. In this case, the training database consists of multivariate sensor data from instrumented revenue vehicles (IRVs). The data between the previous and current tamping dates are used. This forms a time series database labelled with the tamping effectiveness of each track location based on the responses recorded from the IRVs before and after tamping. The labelled time series database is then used to train a time series classifier for prediction. This work uses the state of the art time series classification algorithm, k-nearest neighbour (k-NN) extended to the case of multivariate time series. k-NN is a non-parametric algorithm that does not make assumptions on the underlying model of the training data. With a sufficiently large training database, non-parametric algorithms can outperform parametric algorithms. Using k-NN, the tamping effectiveness of a potential tamping location that is not in the training database, or locations in the next tamping cycle, is predicted using the expected tamping effectiveness from a location in the training database that is the most similar to the target. This allows the algorithm to effectively to identify locations where tamping is likely to be ineffective. This work achieves high accuracy in the prediction of tamping effectiveness even at 12 weeks before tamping. It is hoped that the methodology will help in assisting decision making for maintenance planning activities.

AB - Railway maintenance planning is critical in maintaining track assets. Tamping is a common railway maintenance procedure and is often used when geometrical issues are first identified. Tamping repacks ballast particles under sleepers to restore the correct geometrical position of ballasted tracks. However, historical data shows that tamping is not always effective in restoring track to a satisfactory condition. Furthermore, ineffective, or unnecessary tamping tends to reduce the lifetime of existing track. An intuitive way of preventing ineffective tamping is to predict the likely tamping effectiveness. This work aims to predict the likely tamping effectiveness ahead of time using supervised machine learning techniques. Supervised machine learning techniques predict an outcome using labelled training data. In this case, the training database consists of multivariate sensor data from instrumented revenue vehicles (IRVs). The data between the previous and current tamping dates are used. This forms a time series database labelled with the tamping effectiveness of each track location based on the responses recorded from the IRVs before and after tamping. The labelled time series database is then used to train a time series classifier for prediction. This work uses the state of the art time series classification algorithm, k-nearest neighbour (k-NN) extended to the case of multivariate time series. k-NN is a non-parametric algorithm that does not make assumptions on the underlying model of the training data. With a sufficiently large training database, non-parametric algorithms can outperform parametric algorithms. Using k-NN, the tamping effectiveness of a potential tamping location that is not in the training database, or locations in the next tamping cycle, is predicted using the expected tamping effectiveness from a location in the training database that is the most similar to the target. This allows the algorithm to effectively to identify locations where tamping is likely to be ineffective. This work achieves high accuracy in the prediction of tamping effectiveness even at 12 weeks before tamping. It is hoped that the methodology will help in assisting decision making for maintenance planning activities.

KW - Dynamic time warping

KW - Machine learning

KW - Railway maintenance

KW - Tamping effectiveness

KW - Time series classification

UR - http://www.scopus.com/inward/record.url?scp=85048707427&partnerID=8YFLogxK

U2 - 10.1061/9780784481257.101

DO - 10.1061/9780784481257.101

M3 - Conference Paper

SP - 1010

EP - 1023

BT - ICRT 2017 - Railway Development, Operations, and Maintenance

A2 - Zhai, Wanming

A2 - Wang, Kelvin C. P.

PB - American Society of Civil Engineers

CY - Reston Virginia USA

ER -

Tan CW, Webb GI, Petitjean F, Reichl P. Tamping effectiveness prediction using supervised machine learning techniques. In Zhai W, Wang KCP, editors, ICRT 2017 - Railway Development, Operations, and Maintenance: Proceedings of the First International Conference on Rail Transportation 2017 - July 10–12, 2017 Chengdu, Sichuan Province, China. Reston Virginia USA: American Society of Civil Engineers. 2018. p. 1010-1023 https://doi.org/10.1061/9780784481257.101