Modellierung und Auralisation von Körperschall in Gebäuden
Final Report Abstract
Structure-borne sound sources in buildings are difficult to characterise because of the complexity involved in dealing with six degrees-of-freedom (three translational and three rotational components) at every contact point between a source and receiving structure. Especially for the practical source characterisation in building acoustics some simplification is desired. It is often argued that the reduction of the degrees-of-freedom to the normal components leads to substantial errors. It is shown in this work that this is not the case for the prediction of the sound pressure out of the normal interaction forces. The normal components can predict the sound radiation from structure-borne sound sources quite accurately, even in the case of firm contact through medium-sized contact areas with a diameter of 17 mm. The common rubber interlayer reduces the likelihood of moment excitation even more. A case study of a washing machine on a wooden floor was investigated to prove this point. The prediction of the interaction force from independently measured source and receiver quantities showed to involve larger errors especially at low frequencies. The errors in the prediction are introduced on the one hand by the coupling-invariance of the source activity and on the other hand by errors in the calculation of the coupling. The coupling-invariance can be dealt with by obtaining the quantities in the coupled state. The errors in the calculation of the coupling are dominant at low frequencies and cannot be solved in a simple and practical way. They are caused by the assumption that in-plane shear forces can be neglected. A solution would be to include more degrees-of-freedom but this is probably illusory for the practical source characterisation in building acoustics. The prediction of the interaction forces has to be further investigated through case studies that incorporate more degrees-of-freedom alongside with FEM calculations to verify the results. The importance of the moments has been treated in detail but it is still not sufficient to predict the coupling completely. The current project performs a case study of a washing machine in which the moments are eliminated by using very small contact areas. The prediction still fails due to the assumption that the in-plane shear forces can be neglected. This is partially verified with an ideal source on rubber feet but further work is necessary. Especially the characterisation of rubber interlayers will have to be integrated. The importance of the in-plane shear forces was also observed by Moorhouse et al.. Once there is a detailed understanding of the importance of the degrees-of-freedom of structure-borne sound sources found in practice it is possible to think about simplifications. Those results can then be used to define a measurement procedure for the EN 12354-5 standard that deals with structure-borne sound sources in buildings. Currently only high-mobility sources (e.g. a washing machine on a very heavy concrete floor) are treated in this standard. The coupling between a source and a receiver structure with similar mobilities is not possible yet. The reception plate method is used in the EN 12354-5 standard as it is an appealing technique because of its simplicity. The source is positioned on a standardised plate and only the average velocity on the plate has to be measured. This certainly involves errors compared to a precise prediction. However, those errors are not known yet because the results of the reception plate have never been compared to a precise predicition. The project defines a transfer path method in the coupled state (source and receiver stay together). The methodology is very similar to the operational transfer path analysis (OTPA) commonly used in NVH applications but it has the advantage that it is based on accurately measured frequency response functions with magnitude and phase information. OTPA uses estimates based on the operational velocity and sound pressure records over time which leads to errors in the case of correlation between the contact points. As this is usually the case in mechanical coupling, OTPA results are usually unreliable. It would be an improvement to combine OTPA with coupled transfer paths.
Publications
- Prediction of the sound radiation from a plate excited by a structure-borne sound source. Proceedings of NOVEM, 2009
Matthias Lievens and Pascal Dietrich
- Investigation into the importance of the degrees of freedom for the characterisation of structure-borne sound sources. Acta Acustica united with Acustica, 96:899-904, 2010
Matthias Lievens
