Oxnard Stephen
Department of Computing and Technology, Anglia Ruskin University, East Road, Cambridge, CB1 1PT, United Kingdom
J Acoust Soc Am. 2018 Apr;143(4):EL266. doi: 10.1121/1.5030917.
Numerical acoustic modeling enables simulation of sound propagation through bounded space. Recent research directed to refining Finite Difference Time Domain solutions for acoustic prediction has focused on emulating sound wave-surface interaction. Locally reacting surface properties are a popular choice for deriving boundary conditions that incorporate surface absorption properties. However, implementation of these boundary conditions, using the methods described in prevalent literature, is demonstrated here as unstable for complex room geometries. This work presents a reformulated implementation of frequency-dependent locally reacting surface boundary conditions for Finite Difference Time Domain simulations that is empirically demonstrated to be robust against simulation instabilities.
数值声学建模能够模拟声音在有限空间中的传播。最近针对改进用于声学预测的时域有限差分法的研究主要集中在模拟声波与表面的相互作用。局部反应表面特性是推导包含表面吸收特性的边界条件时常用的选择。然而，本文证明，使用现有文献中描述的方法来实现这些边界条件，对于复杂的房间几何形状而言是不稳定的。本文提出了一种重新制定的用于时域有限差分模拟的频率相关局部反应表面边界条件的实现方法，经实验证明该方法对模拟不稳定性具有鲁棒性。
