A study of microphone arrays for the location of vibrational sound sources

The original objective of the work was to develop an acoustic imaging technique, Nearfield Acoustic Holography (NAH), into a reasonably af-fordable practical system for in-situ applications in an industrial environment. In order to place NAH in the general context of source identification techniques, the thesis summarizes theoretical considerations about sound sources and sound fields, and the principles of different sound source location techniques used in practical situations. The development of NAH theory for planar arrays is the central point of this summary, and the different systems that apply this theory are discussed. Theoretical research and computer simulations show that, at the present state of the art, sound source reconstructions using NAH can only work in certain cases which have to be analyzed individually. The limitations of the theory are better understood when the source reconstruction is studied using the proper methodology for the inherent ill-posed inverse problem. Such a study allowed us to improve the theoretical framework, and to obtain stable source reconstructions. A reliable system for field measurements of different types of sources was not found to be feasible at present. However, it is shown that the inverse problem theory allows us to overcome some of the limitations of the current theory for developing a practical system for field measurements.The difficulties encountered with NAH, and the study of the inverse problem, led us to attempt to develop an alternative, simpler, system which could minimize the inverse problem of source reconstruction. The system considered was a hemispherical array with analogue signal processing. Applying this principle transforms the source reconstruction problem into a re-modelling problem, reducing the inverse problem to the solution of a stable direct problem. Although the antenna could not be tested to its full potential, the initial results were not applicable for source reconstruction, and showed that the signal processing requires to be done by computer software, to take full advantage of the capabilities of the array.