Analysis of the relationship between geometric parameters of railway track and twist failure by using data mining techniques. (January 2023)

• Record Type: Journal Article
• Authors: Izadi Yazdan Abadi, Ehsan; Khadem Sameni, Melody; Yaghini, Masoud
• Is Part Of: Engineering failure analysis. Volume 143:Part A(2023)
• Publication Date: 2023-01
• Subjects: Fault prediction -- Track twist -- Chain failures -- Geometric parameters -- Data mining; System failures (Engineering) -- Periodicals; Fracture mechanics -- Periodicals; Reliability (Engineering) -- Periodicals; Pannes -- Périodiques; Rupture, Mécanique de la -- Périodiques; Fiabilité -- Périodiques; Fracture mechanics; Reliability (Engineering); System failures (Engineering); Periodicals; Electronic journals; 620.112
• DOI: 10.1016/j.engfailanal.2022.106862
• Languages: English
• ISSNs: 1350-6307

Abstract

Highlights: The effect of 4 other track parameters on twist is examined for the first time. Alignment of the rail (AL) and the super elevation (XLV) have the greatest impact on twist. Polynomial regression and association rules have been used to analyze failures. Heavy and expensive machinery can be utilized less frequently for track inspections. This study paves the way toward condition-based maintenance. Abstract: Maintaining railway tracks is capital intensive, time consuming and safety-critical. Novel maintenance methods can decrease these costs and improve efficiency by analysis of collected data from tracks. Twist is one of the common track failures, which poses the risk of derailment, fatalities, injuries and financial loss. In this paper, track parameters are studied for part of a major railway route in Iran. Polynomial regression and association rules, which are popular data mining approaches are used to discover relationship between twist failure with failures of other track parameters for the period between 2018 and 2020. The results show that alignment and super elevation have the highest impact on twist and most of the times these failures occur simultaneously. By adopting this approach twist failure can be identified in order to avoid chain failure and move toward condition-based maintenance.