0
Research Papers: Gas Turbines: Turbomachinery

# Numerical Prediction of the Sources and the Modal Content of the Acoustic Field in a Radial Compressor Outflow

[+] Author and Article Information
Marius C. Banica

ABB Turbocharging,
Bruggerstrasse 71a,
e-mail: marius.banica@ch.abb.com

Peter Limacher

ABB Turbocharging,
Bruggerstrasse 71a,
e-mail: peter.limacher@ch.abb.com

Heinz-Jürgen Feld

ABB Turbocharging,
Bruggerstrasse 71a,
e-mail: heinz-juergen.feld@ch.abb.com

Carsten Spinder

ABB Turbocharging,
Bruggerstrasse 71a,
e-mail: carsten.spinder@ch.abb.com

Contributed by the Turbomachinery Committee of ASME for publication in the JOURNAL OF ENGINEERING FOR GAS TURBINES AND POWER. Manuscript received October 4, 2016; final manuscript received February 21, 2017; published online April 19, 2017. Assoc. Editor: Riccardo Da Soghe.

J. Eng. Gas Turbines Power 139(9), 092605 (Apr 19, 2017) (17 pages) Paper No: GTP-16-1483; doi: 10.1115/1.4036284 History: Received October 04, 2016; Revised February 21, 2017

## Abstract

In large modern turbochargers, transonic compressors often constitute the main source of noise, with a frequency spectrum typically dominated by tonal noise at the blade passing frequency (BPF) and its harmonics. Inflow BPF noise is mainly generated by rotor locked shock fronts. Outflow noise, while also dominated by BPF tones, is linked to more complex source mechanisms. Its modal structure and the relationships between sources and modal sound pressure levels (SPL) are less well understood, and its numerical analysis is, in general, significantly more complex than for compressor inflows. To shed some light on the outflow acoustic characteristics of radial machines, transient simulations of a 360 deg model of a radial compressor stage, including its vaned diffuser and volute, were carried out. Four increasingly finer grids were used for this purpose. On all grids, numerical damping had detrimental effects on prediction quality. A simple and mathematically sound method is proposed to account for this damping. With it, the global outflow acoustic power level ($PWLg$) is predicted to within an accuracy of 2 dB of the experimental result on the finest grid. This shows that satisfactory accuracy can be obtained with state-of-the-art computational fluid dynamics (CFD) codes if care is taken with the simulation setup. The simulations are further validated with experimental data from 17 transient wall pressure sensors.

<>

## Figures

Fig. 1

Schematic representation of the compressor operating map with the design point (on the operating line) and current operating point (next to the operating line)

Fig. 2

Computational domain with details of the compressor, diffuser, volute, and location of pressure taps Dn. Symbols: Pressure taps on hub in light color and on shroud in black; squares and circles indicate differences >4 dB and <4 dB in WPFL at the first BPF, respectively. x = 0 is located at the volute exit, u is the mean flow velocity in x-direction, EP = experimental plane, and US = upstream plane. Also shown are details of grids 3 and 4 around the leading edge of vane 11.

Fig. 3

Simple test case for the investigation of various numerical issues

Fig. 4

p′ on the wall for a (1,0) mode in the test case from Fig. 3, analytical (solid) and numerical (dashed) solution

Fig. 5

Experimental probe rack showing the rotatable pressure transducers in the EP (from Ref. [27])

Fig. 6

Distribution of axial Mach number Mx in the EP at the end of the simulation on grid 4

Fig. 7

Distribution of circumferential Mach number Mϕ in the EP at the end of the simulation on grid 4

Fig. 8

Variation of the mass flow-averaged value for Mx as a function of time in the EP for the final 16 revolutions on grid 2

Fig. 9

Frequency spectrum for a random sample point in the EP, grid 4

Fig. 10

PWL as a function of distance traveled in the propagation domain for the simple test case from Fig. 3

Fig. 11

Magnitude of Mach number at 50% span on grid 4: (a) full view, tongue region at (b) t = 0, (c) t = T/4, (d) t = T/2, and (e) t = 3T/4. Symbols: Pressure taps on hub (light-colored) and shroud (black); squares and circles indicate differences >4 dB and <4 dB in WPFL at the first BPF, respectively.

Fig. 12

Instantaneous static pressure distribution, p, in the diffuser exit at the end of the run on grid 4 on a plane with constant radius r = 1.02RDE. The vertical dashed line in the plot marks the location of the trailing edge (TE) of vane 12.

Fig. 13

Logarithm of the density gradient (see Eq. (22)) at 50% span at t = 0. Symbols: Pressure taps on hub (light colored) and shroud (black); squares and circles indicate differences >4 dB and <4 dB in WPFL at the first BPF, respectively.

Fig. 14

WPFL as a function of sample length in terms of impeller revolutions and equivalent blade passes for D4 and D5

Fig. 15

Experimental and numerical WPFL at the first BPF for the diffuser pressure taps

Fig. 16

Experimental and numerical WPFL at the second BPF for the diffuser pressure taps

Fig. 17

Spectral content of experimental (solid) and numerical (dashed, grid 4) data for four transient wall pressure taps

Fig. 18

PWLg and PWLgc as a function of upstream distance on grid 4

Fig. 19

PWLmn in the EP at the first BPF for (a) experiment, ((b) and (c)) grid 4, ((d) and (e)) grid 3, and ((f) and (g)) grid 1. Uncorrected data: left column. Corrected data: right column.

Fig. 20

WPFL distribution at the (a) first and (b) second BPF in the tongue region and at the (c) first and (d) second BPF opposite the tongue, grid 4

## Discussions

Some tools below are only available to our subscribers or users with an online account.

### Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related Proceedings Articles
Related eBook Content
Topic Collections