In the recent years train axle failures have caused incidents with major loss of life. The aim of this project is to research the feasibility of detecting axle cracks on trains using on-board self-powered monitoring. The vision is to use continuous live monitoring of axles with low cost self-powered wireless systems, that are easy to install, to replace expensive and disruptive NDT methods in maintenance sheds.
We designed the MONAXLE rig to allow research and development of in-situ axle cracks detecting systems.
The train carriage is carried on two bogies, each of which contained two axles.
Each axle is subjected to a 4 point rotating bend motion. This creates cyclic loading which can propagate any crack which may form within the axle.
The MONAXLE rig is designed to test 1:2.5 scaled axle under the same 4-point bend motion observed in real life.
TWI initiated and propagated cracks into several 1:2.5 scaled axles using a 3 point bend (non rotating) arrangement.
We, at the UoS, tested several axles, with known crack lengths, using the MONAXLE rig. Multiple sensors, including accelerometers, thermocouples, displacement and rotation measurement devices, were employed to record the axle behaviour during the tests.
Perpetuum employed the information from the MONAXLE rig to find relationship between crack length and vibrations frequencies. The graph shows the repeatability of the data generated by the rig.
The MONAXLE rig
We designed and built this bespoke rig with the objective of exploring the use of accelerometers to detect the natural frequency of axles and to monitor how this changes in the presence of cracks. We equipped the rig with the instrumentation required to detect crack growth and to apply load to simulate the condition of a real train.
The objective of this new technology is to enable axle inspection regimes to extend to major overhauls only; while currently axle inspections are required on every axle, every year, taking about 24 man-hours per a 4 car unit. This is expensive and disruptive to operations. This innovative technology will provide significant cost reductions for train operators and a greater certainty on safety.
The project developed advanced low power sensing, signal processing and analytics using Perpetuum’s energy harvester powered sensor platform, a test rig built at the University of Southampton and key expertise of TWI Ltd. in non-destructive testing (NDT) and fatigue cracks propagation.
We made progress in developing a sensor platform able to withstand a demanding rail freight vibration environment.
We carried out numerous tests on several scaled axles with and without cracks. The cracks were in different positions (e.g. middle of the axle or closer to the wheel) and presented different sizes (depth and length). The vibration data clearly showed that this new technology can detect the presence of cracks while the train is running. This image shows a phasogram of single axis accelerometer data collected for a cracked axle. The results will be presented at the 5th International conference of Railway Technology – RAILWAY 2021.