Predictive maintenance using tree-based classification techniques: A case of railway switches

Zaharah Allah Bukhsh (Corresponding Author), Aaqib Saeed, Irina Stipanovic, Andre G. Doree

Research output: Contribution to journalArticleAcademicpeer-review

2 Downloads (Pure)

Abstract

With growing service demands, rapid deterioration due to extensive usage, and limited maintenance due to budget cuts, the railway infrastructure is in a critical state and require continuous maintenance. The infrastructure managers have to come up with smart maintenance decisions in order to improve the assets’ condition, spend optimal cost and keep the network available. Currently, the infrastructure managers lack the tools and decision support models that could assist them in taking (un) planned maintenance decisions effectively and efficiently. Recently, many literature studies have proposed to employ the machine learning techniques to estimate the performance state of an asset, predict the maintenance need, possible failure modes, and such similar aspects in advance. Most of these studies have utilised additional data collection measures to record the assets’ behaviour. Though useful for experimentation, it is expensive and impractical to mount monitoring devices on multiple assets across the network. Therefore, the objective of this study is to develop predictive models that utilise existing data from a railway agency and yield interpretable results. We propose to leverage the tree-based classification techniques of machine learning in order to predict maintenance need, activity type and trigger's status of railway switches. Using the data from an in-use business process, predictive models based on the decision tree, random forest, and gradient boosted trees are developed. Moreover, to facilitate in models interpretability, we provided a detail explanation of models’ predictions by features importance analysis and instance level details. Our solution approach of predictive models development and their results explanation have wider applicability and can be used for other asset types and different (maintenance) planning scenarios.

Original languageEnglish
Pages (from-to)35-54
Number of pages20
JournalTransportation Research Part C: Emerging Technologies
Volume101
DOIs
Publication statusPublished - 1 Apr 2019

Fingerprint

German Federal Railways
Switches
assets
predictive model
infrastructure
Learning systems
Managers
manager
Decision trees
Failure modes
learning
Deterioration
budget
monitoring
scenario
Planning
planning
Monitoring
lack
costs

Keywords

  • Classification
  • Data-driven
  • Decision support
  • LIME
  • Machine learning
  • Predictive maintenance
  • Railway infrastructure
  • Switches and crossings

Cite this

@article{3a0bf56f2f5d436b83ec05e490525614,
title = "Predictive maintenance using tree-based classification techniques: A case of railway switches",
abstract = "With growing service demands, rapid deterioration due to extensive usage, and limited maintenance due to budget cuts, the railway infrastructure is in a critical state and require continuous maintenance. The infrastructure managers have to come up with smart maintenance decisions in order to improve the assets’ condition, spend optimal cost and keep the network available. Currently, the infrastructure managers lack the tools and decision support models that could assist them in taking (un) planned maintenance decisions effectively and efficiently. Recently, many literature studies have proposed to employ the machine learning techniques to estimate the performance state of an asset, predict the maintenance need, possible failure modes, and such similar aspects in advance. Most of these studies have utilised additional data collection measures to record the assets’ behaviour. Though useful for experimentation, it is expensive and impractical to mount monitoring devices on multiple assets across the network. Therefore, the objective of this study is to develop predictive models that utilise existing data from a railway agency and yield interpretable results. We propose to leverage the tree-based classification techniques of machine learning in order to predict maintenance need, activity type and trigger's status of railway switches. Using the data from an in-use business process, predictive models based on the decision tree, random forest, and gradient boosted trees are developed. Moreover, to facilitate in models interpretability, we provided a detail explanation of models’ predictions by features importance analysis and instance level details. Our solution approach of predictive models development and their results explanation have wider applicability and can be used for other asset types and different (maintenance) planning scenarios.",
keywords = "Classification, Data-driven, Decision support, LIME, Machine learning, Predictive maintenance, Railway infrastructure, Switches and crossings",
author = "{Allah Bukhsh}, Zaharah and Aaqib Saeed and Irina Stipanovic and Doree, {Andre G.}",
year = "2019",
month = "4",
day = "1",
doi = "10.1016/j.trc.2019.02.001",
language = "English",
volume = "101",
pages = "35--54",
journal = "Transportation research. Part C: Emerging technologies",
issn = "0968-090X",
publisher = "Elsevier",

}

TY - JOUR

T1 - Predictive maintenance using tree-based classification techniques

T2 - A case of railway switches

AU - Allah Bukhsh, Zaharah

AU - Saeed, Aaqib

AU - Stipanovic, Irina

AU - Doree, Andre G.

PY - 2019/4/1

Y1 - 2019/4/1

N2 - With growing service demands, rapid deterioration due to extensive usage, and limited maintenance due to budget cuts, the railway infrastructure is in a critical state and require continuous maintenance. The infrastructure managers have to come up with smart maintenance decisions in order to improve the assets’ condition, spend optimal cost and keep the network available. Currently, the infrastructure managers lack the tools and decision support models that could assist them in taking (un) planned maintenance decisions effectively and efficiently. Recently, many literature studies have proposed to employ the machine learning techniques to estimate the performance state of an asset, predict the maintenance need, possible failure modes, and such similar aspects in advance. Most of these studies have utilised additional data collection measures to record the assets’ behaviour. Though useful for experimentation, it is expensive and impractical to mount monitoring devices on multiple assets across the network. Therefore, the objective of this study is to develop predictive models that utilise existing data from a railway agency and yield interpretable results. We propose to leverage the tree-based classification techniques of machine learning in order to predict maintenance need, activity type and trigger's status of railway switches. Using the data from an in-use business process, predictive models based on the decision tree, random forest, and gradient boosted trees are developed. Moreover, to facilitate in models interpretability, we provided a detail explanation of models’ predictions by features importance analysis and instance level details. Our solution approach of predictive models development and their results explanation have wider applicability and can be used for other asset types and different (maintenance) planning scenarios.

AB - With growing service demands, rapid deterioration due to extensive usage, and limited maintenance due to budget cuts, the railway infrastructure is in a critical state and require continuous maintenance. The infrastructure managers have to come up with smart maintenance decisions in order to improve the assets’ condition, spend optimal cost and keep the network available. Currently, the infrastructure managers lack the tools and decision support models that could assist them in taking (un) planned maintenance decisions effectively and efficiently. Recently, many literature studies have proposed to employ the machine learning techniques to estimate the performance state of an asset, predict the maintenance need, possible failure modes, and such similar aspects in advance. Most of these studies have utilised additional data collection measures to record the assets’ behaviour. Though useful for experimentation, it is expensive and impractical to mount monitoring devices on multiple assets across the network. Therefore, the objective of this study is to develop predictive models that utilise existing data from a railway agency and yield interpretable results. We propose to leverage the tree-based classification techniques of machine learning in order to predict maintenance need, activity type and trigger's status of railway switches. Using the data from an in-use business process, predictive models based on the decision tree, random forest, and gradient boosted trees are developed. Moreover, to facilitate in models interpretability, we provided a detail explanation of models’ predictions by features importance analysis and instance level details. Our solution approach of predictive models development and their results explanation have wider applicability and can be used for other asset types and different (maintenance) planning scenarios.

KW - Classification

KW - Data-driven

KW - Decision support

KW - LIME

KW - Machine learning

KW - Predictive maintenance

KW - Railway infrastructure

KW - Switches and crossings

UR - http://www.scopus.com/inward/record.url?scp=85061435024&partnerID=8YFLogxK

U2 - 10.1016/j.trc.2019.02.001

DO - 10.1016/j.trc.2019.02.001

M3 - Article

VL - 101

SP - 35

EP - 54

JO - Transportation research. Part C: Emerging technologies

JF - Transportation research. Part C: Emerging technologies

SN - 0968-090X

ER -