Testing on the rails
A team of UK scientists are developing a non-contact railhead scanning device that can pinpoint minute cracks on a rail line at speeds of 125 miles per hour (mph). The ultrasonic detector identifies rolling contact fatigue or gauge corner cracking, which occur on heavily-used rails and caused the Hatfield, UK, train derailment in October 2000, killing four people.
The system, invented by researchers at the University of Warwick, uses electromagnetic acoustic transducers (EMATs). These generate broadband ultrasonic waves which travel along the track and return to the EMAT at a rate of 3,000m/s.
‘Existing test methods use a contacting probe usually based in a wheel on an axle,’ explains Lead Investigator Professor Steve Dixon. ‘Using those systems, you more or less have to stay at the same point during the time it takes a sound to propagate from the transducer down into the rail, off the defect and back to the transducer. We use [EMATs], which generate sound in the railhead without touching the rail.’ This allows the detector to move at a much faster rate than current devices, which usually operate at around 25mph. Conventional approaches can also miss deep cracks if they are screened by smaller ones.
The technique can determine the extent of cracks at a deeper depth. The sound waves’ lower frequencies penetrate up to 15mm into the railhead surface. By examing the change in the frequency content of the wave that is able to pass under the crack, its exact depth can be determined.
The goal is to attach the device to commuter trains, allowing tracks to be checked more regularly, says Dixon. This endeavour, however, will cause complications for data analysis. ‘It would be no good to collect all the data and analyse it later, because you can never be 100% sure of where you were when you found the defect, plus the volume of data would be astronomical. So we need computer hardware that can capture and process the data on the fly.’
The group is working with data capture and analysis equipment creators to devise a field-programmable gate array processor that can operate at the required speeds. Thus far, researchers have run trials at 25mph on simulated rails. But they hope to begin tests on real tracks within the next month, working towards higher speeds.
Dixon notes that the device will not be a total solution, as it will only examine the surface of the railhead. ‘There are a plethora of other defects that can exist within rail. We see this as a complementary method to help with safety mangement.’
He says work will continue on the device ‘until we see the first go on a train’, and adds that the group is looking for a commercial partner to take it forward.