Determination of train speeds on railway lines is an important task for both railway operators andresearchers working in the area of noise and vibration assessment. The objective of this paper isto present a new technique capable of automatically calculating train speed from vibration sensorsplaced at distance from the track structure. The calculation method combines three separate speedcalculation techniques to provide high precision speed estimates for arbitrary train speeds. Thefirst estimation technique involves the use of cepstral analysis to isolate key harmonics for usewith speed calculation. The second method is similar however the combination of a running rootmean square and a previously developed “dominant frequency method” are used. The third methoduses an analytical vibration frequency prediction model in combination with regression analysisto calculate train speed. All three methods are combined into one calculation procedure, resultingin high accuracy prediction capability. To show the robustness and ability of the new methodto calculate a wide range of train speeds, it is used to predict intercity and high speed rail trainpassage velocities generated from a previously validated vibration prediction numerical model.More importantly, it is used to predict train speeds during field trials performed on operationalrailway lines in Belgium. The new method is shown to have offer high performance for severaltrain types and track setups.
展开▼
