Single-number quantities are widely used to characterise acoustic properties of building products and buildings. The uncertainty of these values is now investigated by analysing a large amount of round robin results. Methods are developed to calculate the uncertainty of the single number rating from third-octave band sound insulations and the associated uncertainties. The Monte-Carlo method is used to verify the calculation methods. It turns out that correlation effects between third-octave bands significantly influence the uncertainty of the single number rating. Whereas an upper limit for the uncertainty can be calculated by assuming a full, positive correlation between all third-octave bands, the uncorrelated case does not provide a lower limit due to the occurrence of negative correlation. Since correlation effects cannot be predicted it is recommended to use the averaged uncertainties as derived from round robin tests until a deeper understanding of the governing mechanisms enables a calculation of the uncertainties.
展开▼
