Keywords: detached eddy simulation, aerodynamic train model, CRH1, proper orthogonal decomposition, slipstream, train aerodynamics.

When designing a high-speed train, numerous aerodynamic effects have to be considered. One of these is slipstream, which is the flow that the train pulls with it arising from the friction between the train surface and the surrounding air. High slipstream velocities are a safety concern, for instance, for passengers waiting on platforms and trackside workers. Understanding the flow structures and the flow in the wake is important for slipstream, since it is in this region that the largest slipstream velocities occur.

Two different train geometries are studied in this paper, the aerodynamic train model (ATM) and the CRH1. This study is carried out to investigate if there are similarities in the wake flow structures behind the trains. However, it is found that since the separation is different at the tail of the trains, so also are the flow structures in the wake.

The flow fields are simulated using the hybrid model delayed-detached eddy simulation, in order to have time accurate solutions in the wake of the trains. The computed flow fields are decomposed into modes, using proper orthogonal decomposition (POD). The modes are interpreted as coherent flow structures and are analysed in depth.

Three different grids with different grid resolutions in the wake are used to show that the results are grid independent on the CRH1. Investigating velocity profiles in the wake and on the side of the train, shows that the intermediate grid resolution is appropriate for further studies.

The mean flow in the wake is different for the ATM and the CRH1. For the ATM the separated region is only small and located close to the side edges of the tail. While for the CRH1, the separated region covers the whole tail and forms a separation bubble.

It is found that the two dominant fluctuating flow structures behind the CRH1 model act on each of the counter-rotating vortices, respectively. Each flow structure twists the vortex around itself and leaves the other vortex straight. The twisting is propagated downstream. For the ATM the two dominant fluctuating flow structures have a different effect on the counter-rotating vortices. The first alternates between the left and right vortex, making one stronger and the other weaker. The second bends the vortices in the spanwise direction.

It is important to use enough snapshots when computing the POD modes. It is found that approximately the same number of snapshots of the flow is needed behind both train geometries. Even though the modes are different, this indicates that the time scales of the wake structures are similar for both geometries.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £110 +P&P)