The slipstream and wake structure of high-speed trains

2017-02-24T00:12:07Z (GMT) by Bell, James Robin
This work investigates the slipstream, the airflow induced by a vehicles movement, of high-speed trains (HSTs). Such flows can be hazardous to waiting commuters at platforms, track-side workers and infrastructure. Scaled wind-tunnel and moving-model experimental methodologies were developed to assess the slipstream risk of HSTs in their prototype design phase. The slipstream of an Inter-City-Express 3 (ICE3) HST was measured and compared to full-scale experimental data. The scaled methodologies replicated the full-scale slipstream profiles and gust results when measured at low positions. However, they appeared less accurate at high measurement positions. This difficulty is proposed to be caused by the presence of ambient wind in the full-scale results and the higher Reynolds numbers, which reduces the overall coherence of the wake topology responsible for the peak slipstream velocities. Analysis of the scaled moving-model and full-scale results provided insight into the causal flow physics of the peak slipstream velocities. These results indicated that the fluctuation and resulting peak instantaneous velocity are a result of alternating `types' of individual slipstream runs that could be explained by capturing a periodic wake at different phases. A model of the salient time-averaged and unsteady features of the wake has been developed from numerous experimental wind-tunnel techniques and analysis. The wake features a pair of streamwise counter-rotating vortices that move downwards and outwards causing the largest slipstream velocities. The unsteady wake was found to contain vortex shedding from the train's sides and weaker vortex shedding from the roof that interact and merge with the streamwise vortices. This results in an anti-symmetric spanwise oscillation of the streamwise vortices with additional vertical translation at a frequency of St= 0.2. These dynamics cause the slipstream velocity to oscillate and results in the maximum instantaneous slipstream velocities measured, as the near-wake oscillated towards a stationary observer. The effect of modelling a reduced length train on the slipstream and wake was also investigated. Increasing L/H resulted in a thicker surface boundary layer that increased slipstream velocity above the streamwise vortices. The slipstream at low measurement positions and the time-average and unsteady wake appeared to be insensitive to the L/H modelled.