Robust and Predictive Fuzzy Key Performance Indicators for Condition-Based Treatment of Squats in Railway Infrastructures

Ali Jamshidi, Alfredo Núñez, Rolf Dollevoet, Zili Li
2017 Journal of Infrastructure Systems  
9 This paper presents a condition-based treatment methodology for a type of rail surface defect called "squat". 10 The proposed methodology is based on a set of robust and predictive fuzzy key performance indicators. A 11 fuzzy Takagi Sugeno interval model is used to predict squat evolution for different scenarios over a time 12 horizon. Models including the effects of maintenance to treat squats, via either grinding or replacement of 13 the rail, are also described. A railway track may contain
more » ... y track may contain a huge number of squats distributed in the rail 14 surface with different levels of severity. We propose to aggregate the local squat interval models into track-15 level performance indicators including the number and density of squats per track partition. To facilitate the 16 analysis of the overall condition, we propose a single fuzzy global performance indicator per track partition 17 based on a fuzzy expert system that combines all the scenarios and predictions over time. The proposed 18 methodology relies on the early detection of squats using Axle Box Acceleration measurements. We use real-19 life measurements from the track Meppel-Leeuwarden in the Dutch railway network to show the benefits of 20 the proposed methodology. The use of robust and predictive fuzzy performance indicators facilitates the 21 visualization of the track health condition and eases the maintenance decision process. 22 23 Keywords: Design of key performance indicators, railway track condition monitoring and maintenance, 24 interval fuzzy models. 25 26 Please cite as: A. Jamshidi, A. Núñez, R. Dollevoet, and Z. Li, "Robust and predictive fuzzy key performance indicators for conditionbased treatment of squats in railway infrastructures". Find published version at www.ascelibrary.org This material may be downloaded for personal use only. Any other use requires prior permission of the American Society of Civil Engineers 2 ( ) Asset label J t relates to the KPI for the health condition 40 of an asset called "Asset", uniquely labelled as "label" at time t. In The Netherlands, the assets in the railway 41 network includes more than 3,000 km of track, 388 stations, being one of the densest networks in Europe. In 42 this network, the design of an optimal maintenance plan for all its assets is a challenging problem. To 43 optimally design the maintenance plans, infrastructure manager requires to provide crucial information of 44 each asset , and maintenance decision making considering risk averse situations 45 (Rockafellar and Royset 2015) . Thus, the optimal maintenance plan is a necessity because of the high 46 demand from users and government for a better quality of service, and the need of keeping costs as low as 47 possible. 48 Maintenance Performance Indicators evaluate the system performance and can be used to guarantee that 49 these assets operate at an acceptable level of functionality and safety. In Parida and Chattopadhyay (2007) , a 50 Please cite as: A. Jamshidi, A. Núñez, R. Dollevoet, and Z. Li, "Robust and predictive fuzzy key performance indicators for conditionbased treatment of squats in railway infrastructures".
doi:10.1061/(asce)is.1943-555x.0000357 fatcat:7hwksx5rf5adhfhjc5tiwdzadi