TWINRAIL – Railway Maintenance 4.0 Based on the Virtual Connection of Mobile Sensors

Project Description:

This project aims to develop an algorithm capable of determining the degree of ballast deterioration during the tamping process, facilitating the identification of areas where ballast should be replaced without additional costs beyond routine maintenance.

Project Objectives:

Research sensors and instrumentation for gathering information based on the physical magnitudes of the proposed solutions.
Analyze variables that provide information on the condition of the equipment and infrastructure.
Select the most appropriate signals from the collected data.
Research analysis techniques using Artificial Intelligence.
Develop a tool for collecting a basic pattern dictionary of the machine, based on SAX.
Determine optimal operation and generate warnings or alarms regarding anomalous functioning.
Investigate the operation pattern using algorithms to measure the repetition of patterns (Jmotif) and the distance between repetitive patterns (Discords).
Perform measurements in idle mode and on new or perfectly functioning infrastructure to obtain initial pattern dictionaries that will serve as benchmarks for correct situations without faults or anomalies.