Flow behaviour and aeroacoustic characteristics of a simplified high-speed train bogie

TL;DR: In this paper, the aerodynamic and aeroacoustic behavior of the flow past a simplified high-speed train bogie at scale 1:10 is studied using a two-stage hybrid method comprising computational fluid dynamics and acoustic analogy.

Abstract: Aerodynamic noise becomes significant for high-speed trains and its prediction in an industrial context is difficult to achieve. The aerodynamic and aeroacoustic behaviour of the flow past a simplified high-speed train bogie at scale 1:10 is studied using a two-stage hybrid method comprising computational fluid dynamics and acoustic analogy. The near-field unsteady flow is obtained by solving the Navier-Stokes equations numerically with the delayed detached-eddy model and the results are used to predict the far-field noise through the Ffowcs Williams-Hawkings method. The sound radiated from the same scaled bogie model is measured in an anechoic open-jet wind tunnel. The aeroacoustic characteristics of tandem wheelsets are also investigated for comparison. It is found that the unsteady flow past the bogie is characterized by coherently alternating vortex shedding from the axles and more randomly distributed vortices of various scales and orientations from the wheels and frame. The vortices formed behind the upstream geometries are convected downstream and impinge on the downstream bodies, generating a highly turbulent wake behind the bogie. The noise predictions correspond fairly well with the experimental measurements for the dominant frequency of tonal noise and the shape of spectra. Vortex shedding from the axles generates the tonal noise with the dominant peak corresponding to the vortex shedding frequency. The directivity exhibits a dipole shape for the noise radiated from the bogie. Compared to the wheelsets of the bogie, the noise contribution from the bogie frame is relatively weaker.

Summary

1. Introduction

• Over the last few decades, researches have been conducted regarding the source mechanisms of flow-induced noise, particularly in aerospace engineering for landing gears and airframes [1,2].
• In contrast, modelling numerically some simplified geometries can reveal more details of the flow behaviour and the corresponding aeroacoustic mechanisms for some main noise-generating components of high-speed trains.
• 4     DDES is an extension of the detached-eddy simulation (DES) method which combines the large-eddy simulation (LES) in the main flow region with the Reynolds-averaged Navier-Stokes (RANS) approach in the boundary layer region close to the solid objects.
• The cell size on the axle surface is implemented as 0.42 mm around the perimeter and 0.88 mm in the spanwise direction.

4.1. Flow field

• Fig. 3 visualizes the iso-surfaces of the second invariant of the velocity gradient 𝑄 to get an overview of the unsteady flow developed around the bogie.
• Distinct features are observed in different regions of the flow field.
• The flow separates from the upstream wheel front edges and interferes with the flow separated on the wheel tread; therefore, the coherent vortex shedding, seen behind the front axle, cannot be formed behind the front wheel and the wake developed there becomes fully three-dimensional.
• For the point one axle radius above and behind the front axle in the mid-plane between the wheel inner surface and axle mid-span, a tonal peak appears in the spectrum at 324 Hz, as seen in Fig. 5(a).
• A peak appears in the drag coefficient of the bogie and the front wheelset at 641 Hz, which is twice the frequency of the tonal peak in the lift coefficient while at a much lower amplitude.

4.3. Wall pressure fluctuations

• Fig. 8 displays the wall fluctuating pressure level in decibels (𝐿! = 10log 𝑝!"/𝑝!"#!  , where 𝑝!"  is mean-square fluctuating pressure and  𝑝!"#  is reference acoustic pressure 20𝜇𝑃𝑎) on the bogie surface, which can be used to identify the potentially significant noise source regions.
• This also indicates that the massive vortex shedding generated from the front axle may potentially be a major contributor to the noise radiated from the bogie.
• Furthermore, the high pressure fluctuations can be seen around the downstream wheelset due to the flow impingement by the incoming vortex convected from the upstream geometry as well as the flow separation developed from the rear wheel front edges and the vortex shedding formed behind the rear axle.
• Based on the near-field unsteady flow data obtained from the CFD calculations, the far-field noise signals can be predicted by the FW-H acoustic analogy using equivalent acoustic sources.
• Additionally, equivalent circular-shaped receiver positions are defined in the horizontal x-z plane (the coordinates referred to Fig. 1).

5.1. Acoustic spectra computation

• Flow statistics on lift and drag coefficients in Section 4.2 suggest that the flow transient is washed out after 0.1 s.
• Moreover, compared with the experimental data, the tonal peak has a higher amplitude from the calculations in both cases.
• The directivity of the noise radiated to far-field is calculated based on the OASPL determined from the PSD in the frequency range below 2 kHz.
• Note that the similar directivity pattern of sound radiation occurs from the two cases with the slight difference of noise amplitudes between them, which also demonstrates that the wheelsets are the dominant noise sources of the bogie and the noise contribution from the bogie frame is relatively small.

6. Conclusions

• The aerodynamic and aeroacoustic behaviour of the flow past a simplified bogie has been studied using the DDES model and FW-H acoustic analogy.
• It is found that both streamwise and spanwise vortices are generated due to flow separation and vortex shedding around the bogie.
• Furthermore, a vertical dipole pattern of noise radiation is predicted for the upstream wheelset; whereas the downstream wheelset has a multi-directional directivity pattern due to the lift and drag dipoles being aligned perpendicular to each other and its sound generation is relatively weaker.
• These findings are helpful to understand the aerodynamic noise generating mechanisms from the bogie at full scale.
• The turbulent inflow and the complex geometry will lead to complex flow structures and these will also affect the noise generation.

Figures

Fig. 9. Sketch of receiver locations

Fig. 2. Structured mesh topology around the bogie

Fig. 5. Power spectral densities of pressure at bogie wake positions

Fig. 16. Noise directivity of whole geometry in vertical x-y plane

Fig. 15. Three-dimensional noise directivity for simplified bogie

Fig. 7. Power spectral densities of drag coefficients

Fig. 17. Spectra comparisons of the radiated noise from a circular cylinder case

Fig. 12. Comparisons of far-field noise spectra between simulation and experiment 5.3. Acoustic directivity

Fig. 6. Power spectral densities of lift coefficients

Fig. 10. Spectra of acoustic pressure on far-field receivers (𝑈!=30 m/s)

Fig. 13. Three-dimensional noise directivity for front and rear wheelsets of simplified bogie

Table 2. Overview of the mesh size and timestep size for grid independence study

Table 1. Root-mean-square and mean values of lift and drag coefficients

Fig. 3. Iso-surface of the instantaneous normalized 𝑄-criterion (at value of 25)

Fig. 14. Noise directivity of front and rear half-bogies in vertical z-y plane

Fig. 4. Contours of the instantaneous spanwise vorticity field in a vertical plane (side views)

Fig. 11. Experimental setup in the anechoic chamber

Fig. 1. Simplified bogie model (1:10 scale, dimensions in millimetres)

Fig. 8. Wall pressure fluctuation level of the simplified bogie

Frequently asked questions

Q1. What have the authors contributed in "Flow behaviour and aeroacoustic characteristics of a simplified high-speed train bogie" ?

The aerodynamic and aeroacoustic behaviour of the flow past a simplified high-speed train bogie at scale 1:10 is studied using a two-stage hybrid method comprising computational fluid dynamics and acoustic analogy. 

Q2. What future works have the authors mentioned in the paper "Flow behaviour and aeroacoustic characteristics of a simplified high-speed train bogie" ?

Such factors need to be accounted for in future work.