Documents

Eliminate Track Caused Derailments with V/TI Monitoring & Cluster Analysis

Vehicle/Track Interaction (V/TI) Monitor Cluster Analysis and Implementation

By Dwight Clark (Union Pacific), Thomas Toth (Union Pacific), Matthew Dick (ENSCO, Inc.), Rafael Maldonado (ENSCO, Inc.)

The Vehicle/Track Interaction (V/TI) Monitor is an autonomous system that measures accelerations caused from interaction events between a rail vehicle and the track. Traditionally, the V/TI Monitor utilizes single event exception criteria, meaning one exception triggers one field response. In 2010 derailment analysis was performed on several track caused derailments. In the analysis, the V/TI Monitor exceptions detected leading up to the derailment were reviewed. It was discovered that many of the track caused derailments had repeated low level V/TI exceptions.

Additionally the exceptions were a mix of different types (i.e. carbody acceleration + axle impact + ten foot profile mid-chord offset). Using the data associated with the derailments, an algorithm was developed to scan the database looking for similar patterns. When the algorithm detects the pattern, it is flagged and a notification is made to field personnel. In 2011 the algorithm was put into production in Union Pacific’s Northern Region. Since the implementation, an analysis was performed which found a significant decrease of track caused derailments following the implementation of the algorithm.

This paper outlines the development and implementation of the V/TI Cluster algorithm to help identify derailment risk locations. Additionally the paper will discuss feedback from field personnel and lessons learned from the project.



Download  View supplier



Subscribe to Railway-News
Get the latest contract alerts and Railway News emailed directly to your inbox.

We respect your privacy and take protecting it seriously.

Please tick to confirm you would like to receive emails we think may be of interest to you. We sometimes tailor the content you see to what we think you'll find interesting, and so monitor your interaction with our website and email content (find out more).