Research aims to develop predictive maintenance tool

Written by Stuart Chirls, Senior Editor
image description

Televic Rail, Belgium, is participating in a research project aiming to identify new ways of mining data retrieved from sensors installed onboard trains with the objective of developing a predictive maintenance tool.

With up to 50 sensors now installed on vehicles, measuring parameters such as the temperature of components, vibrations, speed, acceleration, and load, Televic says investigating how to collect and use this information effectively is becoming increasingly time-consuming and resource intensive.

Televic is working alongside Renson, Cumul.io and research group IDLab on the Dynamic Visualization, Adaptive Analysis and Scalability for Mining Sensor Data with Integrated Feedback (DyVerSIFy) project. The aim of the research is to develop software components and methodologies in the fields of dynamic visualisation, adaptive anomaly detection and scalability to drive dynamic, adaptive and scalable sensor analytics. The new tool will attempt to present this data in a user-friendly way.

The two-year imec.icon research projects combine academic with industry and/or non-profit partners, and have a track-record of producing digital solutions which have become commercial products for the participating partners.

Televic has participated in three of the research initiatives in the past: Train Applications over an Advanced Communication Network (Track) (2009-2012); Railway Applications Integration and Long-term Networks (Rails) (2012-2014); and Train Passenger Interfaces for Smart Travel (TraPist) (2014-15).

These projects have led to the development of iSync which is used by Eurostar to handle applications and data communication between dispatchers and the train; customised IP network and information screens and audio communication for San Francisco Municipal Transportation Agency (SFMTA); and a personalised passenger information system for the British market.

DyVerSIFy will conclude in September 2019.

Categories: Analytics Tags: