HomeEducationDoctorate (PhD & EngD)For current candidatesPhD infoUpcoming public defencesPhD Defence Rick Haasjes | Towards an active acoustic anechoic chamber

PhD Defence Rick Haasjes | Towards an active acoustic anechoic chamber

Towards an active acoustic anechoic chamber

The PhD defence of Rick Haasjes will take place in the Waaier Building of the University of Twente and can be followed by a live stream.
Live Stream

Rick Haasjes is a PhD student in the Department of Applied Mechanics & Data Analysis. (Co)Promotors are prof.dr.ir. B. Rosic and dr.ir. A.P. Berkoff from the Faculty of Engineering Technology.

An acoustic anechoic chamber is a space in which no reflections occur at the walls. The acoustic anechoic chamber therefore approaches free-field conditions, in which no reflections occur either. Acoustic free-field conditions allow measurements to be conducted without contamination from any boundary. The free-field conditions are realized with special laboratory facilities known as acoustic anechoic chambers. This allows for standardization, product testing, research and development and calibration, all under free-field conditions. The sound absorption in an acoustic anechoic chamber is typically realized by using absorption material, which is placed at the boundaries of the room. However, because the thickness of the absorption material is related to the wavelength of the acoustic waves, the performance of the absorption material deteriorates at lower frequencies. Consequently, a typical acoustic anechoic chamber has a lower cut-off frequency of up to 200 Hz, meaning that free-field conditions are not guaranteed below this frequency.

Active noise control is effective at lower frequencies, which therefore makes it a promising technique to complement the passive absorption and to lower the cut-off frequency of an acoustic anechoic chamber. To implement an active noise control system in an acoustic anechoic chamber, an accurate estimation of the reflected sound field is necessary. The solution to this problem is found in the application of the Kirchhoff-Helmholtz integral, which in two dimensions uses several microphones distributed along a circle. The microphones measure the acoustic pressure and the particle velocity, which are the input to the Kirchhoff-Helmholtz integral. The output of the Kirchhoff-Helmholtz integral is the reflected sound field due to any source located within the circle, while for any source located outside the circle it outputs the total sound field.

With increasing geometry sizes, more complex geometries and smaller wavelengths, the active noise control system needs to be up scaled. This means that a larger number of microphones is required to accurately measure the sound field, and a larger number of sources is required to accurately generate the secondary sound field. This results in a strong increase in computational complexity and memory consumption. Low computational complexity and memory consumption are important features of the desired algorithms. Three algorithms are derived. Two algorithms achieve the largest reduction in computational complexity and memory consumption by computations in the frequency domain. However, filters that are computed in the frequency domain are not necessarily causal, which is necessary for real-time implementation. Therefore, techniques to ensure causality are applied at the one of the algorithms, allowing the use of this algorithm for real-time systems. One of the algorithms is computed in the time domain and therefore has the highest computational complexity and memory consumption, but in comparison with the other two algorithms, relaxes some requirements to reduce frequency domain artifacts.

The algorithms, together with the Kirchhoff-Helmholtz integral, are used in two-dimensional numerical studies and two-dimensional experiments. In the numerical studies a small-scale and a large-scale chamber are simulated. The results of these numerical studies show that the Kirchhoff-Helmholtz integral, together with two of the derived algorithms is effective at the suppression of the reflections from the walls and allows the number of microphones and loudspeakers to be increased. To verify the real-time effectiveness, an experimental setup is built. The experiments are conducted on a setup which is limited in its height so that it is considered two-dimensional within the frequency range of interest. The setup is equipped with 12 sensors along a circle. Each sensor has three microphones which are distributed in radial direction, to measure the acoustic pressure and to approximate the particle velocity. The results of the real-time experiments show a significant reduction in reverberation time, which is an indication of effective suppression of reflections.