For monitoring the conditions of railway infrastructures, axle box acceleration (ABA) measurements on board of trains is used. In this paper, the focus is on the early detection of short surface defects called squats. Different classes of squats are classified based on the response in the frequency domain of the ABA signal, using the wavelet power spectrum. For the investigated Dutch tracks, the power spectrum in the frequencies between 1060–1160Hz and around 300Hz indicate existence of a squat and also provide information of whether a squat is light, moderate or severe. The detection procedure is then validated relying on real-life measurements of ABA signals from measuring trains, and data of severity and location of squats obtained via a visual inspection of the tracks. Based on the real-life tests in the Netherlands, the hit rate of the system for light squats is higher than 78%, with a false alarm rate of 15%. In the case of severe squats the hit rate was 100% and zero false alarms.
Monitoring the railway infrastructure: Detection of surface defects using wavelets
2013-10-01
1108397 byte
Conference paper
Electronic Resource
English
Application of Wavelets and Fuzzy Sets to the Detection of Head-Checking Defects in Railway Rails
British Library Conference Proceedings | 2010
|Application of Wavelets and Fuzzy Sets to the Detection of Head-Checking Defects in Railway Rails
Springer Verlag | 2010
|Railway infrastructure monitoring by drones
IEEE | 2016
|