Friday, 14. December 2018

Aerodynamic loads on track side objects

Guide-lines are required to guarantee the safety of persons standing near the track, e.g. track side workers or passengers waiting on platforms in railway stations, when exposed to aerodynamic loads generated by passing trains. In principal, the flow over a train can be split into three regions: The first is associated with the pressure wave generated by the head of the train, the second is characterized by the boundary layer next to the train and the last region is the wake behind the train.

The goal of the present work is to achieve fundamental physical knowledge of induced forces on track side objects. The question is how the unsteady flow field generated by the moving train interacts with objects of different shapes and sizes.

In the head region, friction and boundary layer effects can be neglected, hence the head induced pressure wave can be described theoretically using potential flow theory. This is done here for the ICE-3 train using three point sources in the head and respective sinks in the tail. As first tests, the aerodynamic forces on a sphere were considered. The sphere is modeled with a 3D-dipole with unsteady source strengths. The train passage is simulated by moving the sphere along the train in a quasi-static manner. The pressure and aerodynamic loads are calculated using the unsteady Bernoulli equation.

Parallel to the theoretical work, the unsteady aerodynamic forces are measured in model experiments. These were carried out in the Tunnel Simulation Facility (TSG) using different train models scaled 1:25 passing polystyrene spheres of different sizes between 15mm and 60mm, which corresponds to 375mm and 1500mm in full scale. The spheres were mounted on a cantilever balance to measure the unsteady forces in the milli-Newton range. Tests were carried out for different lateral distances of the object and different heights. In Figure 1 a typical test setup is shown. The balance is connected to a frame, which is mounted to the ceiling by steel wires to reduce disturbances caused by external vibrations propagating from the catapult through the floor into the mounting.

Figure 2 shows the train models used in the tests. The ICE 3 model, the Next-Generation-Train (NGT) and a generic model based on potential flow theoretical calculations. The latter is called potential train model in the following. 
First measurements revealed, that the force induced by the train head on a sphere in lateral direction can be well predicted by potential flow theory (see Figure 3), although not all details of the ICE 3 geometry can be modeled using only three sources in the train head. Interestingly, the main force fluctuations induced by the train tail fit qualitatively to the theoretically predicted loads, even though the flow around the train in this Region is dominated by turbulent effects and at the train tail the real flow field differs significantly from the potential flow assumed in the theoretical model. Additionally, the experimental curves show the effects of the inter-car gaps where partial fairing was installed at the first gap.

Figure 4 shows the loads normalized using the trains' cross sectional area for the three different train geometries. The normalized loads induced by the potential train model are similar to those induced by the ICE 3 model, although the nose length of the ICE 3 is more than twice as large as the one of the generic train.
The normalized loads induced by the NGT are smaller than those induced by the ICE 3 and the generic train and the duration of the force fluctuations is longer. This is caused by the extended nose length compared to both of the other trains. But altogether the results show that the nose length is not the only decisive factor for the strength of the forces. Other details of the geometry of the nose also seem to be essential for the loads acting on a track side object.

Figure 1: Test setup for unsteady force measurement in the TSG
Figure 2: Used train models in scale 1:25: 1) ICE 3, 2) NGT, 3) Potential train
Figure 3: Comparison between experimental data and theoretically calculated unsteady lateral forces for a normal distance of 11cm (equals 2.75m in full scale)
Figure 4: Comparison of the loads in travel direction for the three different train geometries normalized with the trains’ cross sectional area.


Dr. Klaus Ehrenfried
German Aerospace Center (DLR)
Institute of Aerodynamics and Flow Technology, Department Ground Vehicles
Phone: +49 551 709-2813

German Aerospace Center (DLR), Institute of Aerodynamics and Flow Technology, SCART
Bunsenstraße 10, 37075 Göttingen, Germany