Boundary conditions determined by experimentally measured surface properties for wave-based time-domain simulations

Abstract

Boundary conditions are essential and necessary when simulating sound wave propagation inside an enclosed space. In this work, complex-valued surface acoustical properties for the actual materials are obtained from impedance tube measurements. This work applies model-based Bayesian inference to estimate the exact parameters for arbitrary boundary conditions. The Bayesian approach presented in this work demonstrates the capability to predict the complex-valued surface boundary function of frequency in broadband ranges. This approach is equivalently applicable in different boundary models in time domain depending on wave-based numerical methods used for acoustic modelings, such as the spectral element method or finite-difference time-domain method. With the accurately estimated boundary models, the efficacy of the boundary conditions is explored within the time-domain approaches.