Rail hunting refers to the tendency for a train to assume a correlated side-to-side (sinusoidal) movement as the locomotive and wagon wheels alternately make and break contact with the inside of the rail track. In extreme cases, this sinusoidal movement can become excessive and lead to track damage and/or derailment. Rail hunting is speed dependent for a given train and track combination, and is one factor in determining the maximum rated travel speed of the train.
The client wanted to determine whether rail hunting may have been a factor in a major derailment incident on a critical section of the national rail network. They approached ICON Technologies to develop a portable rail hunting detection system that could be deployed to different field locations as required.
Data was acquired from five scanning lasers mounted inside the track over a 20m section of interest. Sensors mounted ca. 150m ahead and behind the instrumented zone were used to initiate a measurement cycle as a train approached, and terminate it as the train receded. The lasers measured the distance of every wheel from the inside head of the track at 2 kSa/s per laser, with measurement cycles extending up to 3 mins depending upon the length and speed of the train. Data files were pushed to the client’s Perth office as they were acquired.
The test site was challenging – a remote desert location with no on-site power and functional but intermittent 3G communications. The system needed to be available 24/7 for projected deployments over many months, with trains passing at irregular intervals both day and night. The change in light background from the pitch black of a desert night, to full-noon sun, is a particular challenge to making reliable measurements from a scanning laser. Finally, the amount of data captured, at up to 8 MB per train, was vastly larger than is commonly managed in such a remote location.
The system used an NI cRIO controller to acquire the high-speed laser data and manage the fail-safe transmission of the large data files to the client’s office. We deployed the system in an all-weather enclosure with battery-backed solar power. We implemented a custom smart power management regime in both hardware and software to meet the significant peak power requirements of the controller and lasers. And we developed hardware installation protocols and software data analysis routines to ensure that data on the changing wheel position could be reliably and precisely extracted from signals acquired in the presence of vastly different levels of background light.
The system ran 24/7 for an initial deployment of 12 months and reliably captured data from all trains. It was able to capture reproducible evidence of anomalous motion patterns associated with specific rolling stock that traversed the test site multiple times over the deployment period.
A three-man team was able to setup and validate the system within a two period on site, with pulldown less than one day.
NI CompactRIO hardware
NI LabVIEW software
Micro-Epsilon Lasers
Custom power management
