PhD Defence by Pierangelo Libianchi

Supervisors

• Principal supervisor: Finn T. Agerkvist, DTU Electro, Denmark.
• Co-supervisor: Associate Professor, Efren Fernandez Grande, DTU Electro, Denmark.
• Co-supervisor: Researcher Elena Shabalina, Head of Technology Center E-Textiles and Acoustics, Head of Acoustics, German Institutes of Textile and Fiber Research Denkendorf.
• Co-supervisor: Associate Professor Jonas Brunskog, DTU Electro, Denmark.

Assessment committee

• Associate Professor Vincente Cutanda Henriquez, DTU Electro, Denmark (chair).
• Associate Professor Jens Ahrens, Chalmers University, Sweden
• Professor Dick Botteldooren Ghent University, Belgium

Master of the ceremony

• Professor Frieder Lucklum, DTU Electro, Denmark.

Abstract
The exposure to noise has a detrimental effect to the health of an individual. Outdoor live events also are a source of noise since any undesired sound fall into this definition. It is important to limit the noise emissions from concerts, as one would do with any other noise source. Outdoor live events have particular characteristics: large content of low frequency noise that propagates over large distances and an audience area where the sound field should be left untouched. It is possible to limit the noise emissions by properly designing the sound reinforcement system. The only direction where the design does not help is in the direction of the audience. To reduce emissions in this direction we employ an array of subwoofers to create a secondary sound field that matches the field generate by the sources on stage but with opposite phase. In this way, the destructive interference between the two fields weakens the noise emissions where the two fields match. In this thesis, we split the problem in two parts: the design of the filters to apply to the control sources to produce the anti-field and characterization of outdoor sound propagation to enhance the range of the system.

For the first part of this thesis, we introduce a new method to generate the filters with the aim of simplifying the use of such a system. We introduce an iterative algorithm that allows controlling the radiation pattern of the control sources to avoid increasing the sound pressure level outside of the quiet area. Furthermore, it allows limiting the amplitude of the solution and avoiding distortion from the control sources, which would degrade the performance of the system. We further look at other approaches taken to deal with this problem, we provide a broad view of techniques, and challenges that to develop further this field and its applicability.

In the second part, we provide an overview of the main outdoor propagation models, highlighting the approximation to obtain the corresponding governing equations and thus, their limits. We also provide basics of micrometeorology to understand better the properties of the medium and how sensitive is the sound field to uncertainties in the wind profile. Finally, we look into the possibility of using a neural network as a surrogate model for the Helmholtz equation to recover the sound speed profile through a neural full waveform inversion approach.