### Refine

#### Keywords

- Strömungsakustik (4)
- Aerodynamik (2)
- Akustik (2)
- Schallabstrahlung (2)
- Turbulente Strömung (2)
- aeroacoustics (2)
- axial fan (2)
- fan noise (2)
- Axialventilator (1)
- Druckstoß (1)

#### Institute

The sound emission of low-pressure axial fans is substantially influenced by the fan blade geometry and the inflow conditions, induced by the fan installation system. However, the combined impact of these parameters has not yet been comprehensively investigated or understood. Hence the motivation for this thesis was to undertake a compact, systematic experimental study on the sound emission of axial fans with different blade geometry parameters under distorted inflow conditions.
The impact of the fan blade design was investigated on the basis of nine fans with different blade loading distributions and different fan blade skew, but otherwise identical geometric parameters. Additionally, the effectiveness of leading edge serrations in reducing axial fan noise was examined with a parametric study of a generic flat-plate fan and eleven sets of fan blades with leading edge modifications that included single-sine (sinusoidal), double-sine and random amplitude leading edge serrations.
The inflow conditions were altered to incorporate either an increased inflow turbulence intensity or an inhomogeneous inlet velocity profile. These conditions were realised by grids - three turbulence grids and two velocity gradient grids - that were mounted upstream of the fans.
A large increase in the sound radiation with elevated tonal and broadband components was observed with the modified inflow conditions. Thereby, the sound emission of the forward-skewed fans showed a greater susceptibility for distorted inflow conditions than that of the backward-skewed fans. It was found that owing to the fan blade shape, the forward-skewed fans are prone to higher unsteady blade forces and increased pressure fluctuations on the fan blade leading edges. These factors determine the tonal and broadband sound radiation. For the backward-skewed and unskewed fans, dominant subharmonic components, originating from flow phenomena in the tip region, occurred under both free and distorted inflow conditions.
All types of leading edge serrations achieved a sound reduction compared with the reference fan with straight leading edges - for both free and distorted inflow conditions. The greatest reduction was observed for the sinusoidal leading edges, followed by the double-sine and random amplitude leading edges. The reduction was mainly dependent on the serration wavelength rather than the serration amplitude: the smaller the leading edge wavelength, the greater was the overall sound reduction.
Overall, the investigations showed that the sound emission of low-pressure axial fans is highly dependent on the combined impact of the fan blade geometry and the inflow conditions. The findings contribute to a better understanding of the sound generation mechanisms in axial fans and show means for designing low-noise fans in complex cooling or ventilation systems.

Induced cross-draft and natural ambient winds affect cooling fan performance in air-cooled condensers, which is a considerable detrimental factor to power plants in arid regions. Against this background, this work tries to comprehensively examine the various effects of uniform ambient flow fields on single axial fans. Deviating from the common approach in pertinent research to study multiple fans operating in an array, this work's research examines a single fan inside a uniform flow field under varying conditions. Investigated fan parameters include operating point, blade shape, installation conditions, and orientation in flow field. A specific fan test rig was designed and used to assess ambient flow field influence on integral fan performance in a wind tunnel environment. Detailed information on the resultant flow field is investigated using Laser Doppler Anemometry and numerical methods. Fan blade vibration is captured with Tracking Laser Scanning Vibrometry. Finally, this work presents experimental and numerical investigations of the cross-flow influence on fan noise.

Fast transient events such as water hammer are potentially dangerous for any piping system. In some cases, an accurate calculation must include fluid-structure interaction (fsi), because it can have a high impact on the amplitudes and frequencies of the hydraulic and structural systems. The important question is: when is fsi important and has it to be taken into account?
Therefore, fsi effects in an industry-oriented piping system have been investigated and quantified. Only a part of the piping system, a U-shaped bend, was able to move. Consequently, all high-amplitude interaction effects were localised at this bend.
In order to vary the bend frequency, the support position was shifted, while the length of the piping system and hence the fluid frequency remained constant. The structural frequency was changed in a range around the water hammer frequency.
Two-way junction coupling effects, such as beats in the time signals and correlations in the frequency spectra of the bend displacement and the pressure, were resolved only in the short system with the medium and long bends, where the differences originated from the bend geometry. The longer the distance between the parallel levers of the bend, the stronger was the coupling between the bend and the water column. The coupling is independent of the excitation. Characteristics of the pressure spectra were identified to quantify the strength of the coupling. Using an analogy with the two-mass oscillator, the coupled fluid mass was calculated.
Simple and partly more detailed coupled simulations showed that basically the strength of the coupling is a function of the amount of fluid mass contained in the vertical pipe section between the parallel levers. Furthermore, the fact that only a part of the fluid column is coupled with the bend oscillation caused a reduction of the strength of the coupling. Using the simulation data and the analogy with the two-mass oscillator, it is possible to predict the strength of the coupling on the basis of the geometry and the mass of the bend. The generation of a detailed model on the basis of a coupled code package used in the industry was time-consuming, but reproduced the coupling effects observed in the measurements.
Unsteady friction, which is defined by the unsteady wall shear stresses in transient flows and is associated with friction coupling and fluid damping, was analysed in oscillating pipe flow and water hammer flow. The good agreement between the wall shear stresses of the analytical solution, CFD (computational fluid dynamics) simulations and the measurements verified the numerical model and the measurement method. In the case of water hammer flow, differences between the wall shear stress distributions of the CFD simulations (laminar and turbulent approaches) and the measurements originated from the limited frequency resolution of the hot film probe, where simultaneously the mean flow velocities agreed well.
Finally, the results of a CFD simulation of water hammer in a pipe geometry were compared with the results of one-dimensional calculations using a quasi-steady friction model and unsteady friction models such as the Brunone model and the Diesselhorst model. No model was able to reproduce the fluid damping occurring in the CFD simulation because it originated from dissipation close to the wall and not in loss of momentum caused by friction at the wall.

The turbulent flow over flexible surfaces is normally associated with relevant acoustic
sound radiation. A mechanism is the sound generation due to turbulent pressure
fluctuations in the flow field. This is denoted as aeroacoustic sound radiation. Another
mechanism is the sound radiation induced by vibrating mechanical structures. This
is denoted as vibroacoustic sound. The excitation of mechanical structures can be
caused by hydrodynamical wall pressure fluctuations from turbulent eddy structures
in the flow field as well as from pressure forces induced by sound waves which have
spatial dimensions matching to the wave length of the structural eigenmodes. In case
of vibroacoustic sound radiation, the current work is focused on the excitation of plate
like structures due to forces generated by hydrodynamical wall pressure fluctuations
in order to compute the vibroacoustic sound radiation. A method for the numerical
calculation of the fluid–structure–acoustics interaction is introduced.
First principal investigations in this work deal with an aerodynamic test case consisting
of a quadratic wall–mounted square cylinder in conjunction with a flexible flat
plate placed in the turbulent wake of the cylinder obstacle. The aeroacoustic sound
radiation from the turbulent flow field was computed based on acoustic source terms
with help of Lighthill’s acoustic analogy. The turbulent flow field for the calculation
of acoustic source terms and for the determination of the hydrodynamical wall forces
was analysed by LES. Measurements of the square cylinder’s sound radiation in an
acoustic wind tunnel showed good comparability to the numerical results. The vibroacoustic
sound radiation was computed based on the acoustic particle velocity of
the vibrating flexible plate’s surface. The two–sided coupled computation of the fluid
load induced structural vibrations and the vibroacoustic sound radiation was compared
with a one–sided approach. It is shown, that due to the consideration of the influence
of the acoustic medium on the vibrating plate, a significant reduction of the plate’s
sound radiation occurs. As a result of the omitted back–coupling of the structural
displacements on the flow field, the introduced approach is limited to applications with
negligible small mechanical displacements.
Motivated from performance loss issues of sonar systems caused by flow induced
vibrations of hydrophone hulls, the fluid–structure–acoustics interaction methodology
was applied on the computation of the vibroacoustic sound radiation into a water filled
cavity in the interior of a towed underwater vehicle with turbulent outer flow. The
towed underwater vehicle was equipped with an acoustic window in form of a thin steel
plate on one of its planar side areas. Based on numerical investigations of the outer flow field of the underwater vehicle, the turbulent boundary layer flow in the vicinity of the
acoustic window was abstracted as a turbulent periodic half–channel flow and analysed
with help of LES. Comparison of the turbulent wall pressure spectra at the acoustic
window between simulation and experiments showed good comparability. The pressure
field in the cavity induced by the vibration of the acoustic window was investigated with
a two–sided coupled simulation to consider the interaction between the pressure field
and the vibrations of the acoustic window. The comparison of the pressure levels in the
interior of the cavity between measurements and simulation showed comparable values
for low frequencies. For higher frequencies, differences due to the absence of damping
effects and spatial resolution limits in terms of the hydrodynamical force distribution
in the simulation were evident. The wavenumber–frequency analysis of the pressure
field yielded a separation between the acoustic pressure and the pressure induced by
the eigenmodes of the vibrating plate. The change of the dispersion relation of the
acoustic window plate caused by the coupling to water was reproduced correctly.
Conclusively, in the current work, a fluid–structure–acoustics interaction approach
for wall–bounded turbulent flows and structural vibrations with small amplitudes is
introduced. The amount of modelling for the computation of the participated physical
phenomena was minimized.

Due to the electrification of cars and the reduction of noise generated by the powertrain, the automotive industry has gained growing interest for computer-based acoustic design of HVAC systems. In order to implement tools for virtual noise reduction, it is necessary to develop simulation procedures that are able to reliably predict the aeroacoustic noise of HVAC systems and their components. For this purpose, a hybrid simulation technique was proposed and tested in this work: First of all, the non-stationary flow was computed by use of a common finite volume solver; subsequently, the aeroacoustic sources were evaluated; finally, the sound radiation was simulated by a finite element code.
The coupling procedure between these simulation steps is a crucial task to achieve excellent results, for which two aeroacoustic approaches were tested. Applying Lighthill's analogy was able to yield spectra in good agreement with experiments. However, the effort required for the implementation and acoustic simulation was high, while the possibilities for source analysis were limited. These problems did not occur when the acoustic perturbation equations (APE) in form of the perturbed convective wave equation (PCWE) were used. This approach provides insight into the pure sound source and propagation fields. However, it was found to be more sensitive regarding numerical noise and the truncation of sources.
Based on these findings, this work revealed how the proposed procedure can be used to successfully predict the far-field noise spectra of a generic air outlet, a real air outlet of a car, an HVAC unit, a centrifugal fan, and eventually a complete HVAC system.
Therefore, this work demonstrated the applicability of hybrid approaches for the prediction of HVAC noise. Furthermore, details on their correct usage are provided to enable the error diagnostics of aeroacoustic simulations also in other fields of application. These findings offer the basis for the development of computer-based acoustic design tools for HVAC systems and thus pave the way for further improvements regarding practical usage.

Die vorliegende Dissertation setzt sich mit der Zerstäubung von Flüssigkeiten auseinander. Das Hauptaugenmerk liegt auf der Entwicklung eines zeitlich modulierenden Zerstäubungsverfahrens für die Erzeugung sehr feiner Sprays bei kleinen Durchsätzen der zerstäubten Flüssigkeit. Das entwickelte Verfahren basiert auf der Erzeugung monodisperser Tropfenketten durch den kontrollierten Zerfall von Flüssigkeitsstrahlen. Diese Primärtropfen zerspritzen durch den Aufprall auf einen Prallkörper in deutlich kleinere Sekundärtröpfchen und bilden dadurch ein sehr feines Spray. Experimente zur Tropfenprallzerstäubung zeigten, dass diese im Vergleich mit anderen Zerstäubungsmechanismen, insbesondere bei kleinen Durchsätzen, die kleinsten Tröpfchen liefert.
Zusätzlich sind einige Grundlagenuntersuchungen auf dem Gebiet der Mehrphasenströmungen beschrieben. Es wurde die existierende Theorie für den ersten windinduzierten Zerfall von Flüssigkeitsstrahlen um die Berücksichtigung der axialsymmetrischen Grenzschicht in der Strahlumgebung erweitert. Für den pulsartig angeregten Strahlzerfall konnte ein Zusammenhang zwischen den mechanischen Schwingungsamplituden des Düsenhalters und der strömungsmechanischen Anfangsstörung des Strahlzerfalls hergestellt werden.