Uncertainty quantification of the sound transmission loss of building components at the design stage

摘要

The sound transmission of a building component does not only depend on its material properties, but also on the facility in which it is tested and on the test procedure, especially at low frequencies. This introduces uncertainty in a design situation where the test facility and measurement procedure are still unknown, i.e., when predicting the variability of the transmission loss across different facilities and procedures is of interest. In this work, a robust probabilistic framework is developed for quantifying the combined effect of the uncertain parameters involved in sound insulation prediction problems. The main idea is to construct the probability distribution of these parameters from the available information using the maximum entropy principle. The probability distribution of the predicted sound transmission loss, which is subsequently obtained through Monte Carlo simulation, is then fully compatible with the available information but otherwise maximally uncertain. The method is applied for quantifying the uncertainty of the transmission loss of single and double walls at the design stage, where the available information on the uncertain parameters is limited to e.g. standardized requirements and nominal damping values.