Numerical study on the repeatability and reproducibility of laboratory building acoustic measurements

Abstract

An important issue in building acoustics is the significant variability in laboratory test results that numerous round robin tests have indicated. The current wish to include the frequency bands 50-80 Hz in the procedures to determine single-number quantities has prompted new discussions. In this paper, wave based models are used to numerically investigate the fundamental repeatability and reproducibility. Regarding sound insulation measurements, both the pressure method (ISO 10140-2) and the intensity method (ISO 15186-1 and ISO 15186-3) are investigated in the frequency range 50-200 Hz. Flanking transmission measurements (ISO 10848) are also studied. The investigation includes the repeatability of the different measurement procedures, which depends on the influence of the source and receiver positions. The reproducibility in different test facilities is studied by looking at the influence of geometrical parameters like room and plate dimensions. Increasing the number of source or receiver positions has little effect on the overall uncertainty as the reproducibility uncertainty is generally much larger than the repeatability uncertainty. For small-sized test elements, the reproducibility of the intensity method is better. For heavy walls and lightweight double constructions, however, the predicted uncertainty is similar for the three measurement methods.