[Back]

C. Junger:
"Computational aeroacoustics for the characterization of noise sources in rotating systems";
Supervisor, Reviewer: M. Kaltenbacher, S. Becker; Mechanik und Menchatronik, 2019; oral examination: 2019-10-17.

The objective of computational aeroacoustics is the prediction of sound generated by turbulent flow. Its utilization in design processes demands the prediction, and consequently the manipulation of sound. Therefore, not only the radiated sound is of interest, but also the flow mechanisms generating the noise sources have to be understood. A large area where sound prediction
is applied - from aircraft engines down to vacuum cleaners - includes rotating systems.
The aeroacoustic investigations in this work are applied to a low pressure axial fan benchmark, to be able to compare the results with measurements. The first aim of this work is the investigation
of the prerequisites a simulation needs to fulfill to predict the sound correctly, as the rotation yields special challenges on the flow computation and the aeroacoustic methods.
The second aim is the comparison with different prediction methods, and the third aim is to investigate the obtained information about noise sources in the rotating system.
Applied and compared are different methods, from empirical formulas to state of the art computational aeroacoustic prediction methods. In the computational prediction, detailed information about the flow field is needed. To obtain the information, a transient flow simulation
was performed. For the validation of the flow simulation, flow velocity, pressure rise and wall pressure spectra were compared with the measurement results from the benchmark. From the sound prediction with the perturbed convective wave equation it can be seen that at least five revolutions of the fan have to be computed before tonal components in the acoustic spectrum can be predicted. The interpolation of acoustic sources as well as the spatial discretization have little influence on the acoustic result but large influence on the computational effort.
Compared to that, blending of the acoustic source term can have more influence on the result of the aeroacoustic prediction. The over all sound power level is predicted with only a deviation of 0.6 dB compared to the measurements but the spectral prediction of the first subharmonic peak was not sufficient. The Ffowcs-Williams and Hawkings analogy is able to predict the first subharmonic peak but has a deviation of 2.1 dB in the over all sound pressure level compared
to the measurements. Semi empirical methods predict the over all sound power level to an accuracy below 2 dB. The stochastic noise prediction methods, predict the sound sources at the leading edges of the fan blades and in the flow of the boundary layer and tip flow. From
the flow simulation, it can be seen that the tip flow leads to interactions with the blades. The investigations of the aeroacoustic sources show that the tip flow results in noise sources at the outer diameter of the fan, but the strength of the sources varies over the simulation time.
Further strong sources occur at the leading edges of the blades. For higher frequencies acoustic sources become more and more compact.

https://publik.tuwien.ac.at/files/publik_282258.pdf