Page 66 - Winter2018

Computational Acoustics
formation of a new TSG, that would be a suitable and suf- ficient justification for forming one. Given the increasingly computational nature of acoustics and the sciences in gen- eral, and the interest of many members of the ASA in recent developments in this area, it is important that the ASA deep- en its support for CA-related activities. Particularly worth noting is that many of the signatories of the TSG petition were relatively young ASA members, which likely reflects the emphasis of their research projects and employment in- terests. Hence the CA TSG should help the ASA to encour- age and retain younger members. Many petition signatories also hail from countries other than the United States and Canada; thus the TSG can help expand the reach of the ASA to researchers in other countries.
Although most of the feedback we received regarding for- mation of a CA TSG was enthusiastic, some thoughtful ar- guments were also made against the idea. The most frequent was that computational acoustics is a tool that should be discussed within the scope of the existing TCs rather than a proper research topic warranting focused discussions in a separate, specialized group. In response, it can be argued that computer science is now widely accepted as a legitimate academic discipline, and in many other fields (e.g., fluid dy- namics, physics, and biology), computational techniques are regarded as an important and rapidly expanding area of inquiry, distinct enough to be the subject of specialized groups, meetings, and journals. Thus the formation of a CA TSG is really just a recognition of the important status of computation across the spectrum of modern science.
Currently, CA topics are often discussed independently within many of the TCs. This lack of interaction can be a detriment to scientific progress because many computa- tional approaches have multiple applications. One pertinent example is the formulation of computational methods for sound refraction and scattering, such as the parabolic equa- tion, that have applications both underwater (e.g., Jensen et al., 2011) and in the atmosphere (e.g., Salomons, 2012). Another example is finite-element methods for calculating sound fields in interior spaces, which are important in struc- tural, engineering, and architectural acoustics (Thompson, 2006; Marburg and Nolte, 2008; Vorländer, 2013). The CA TSG thus provides a forum for researchers to discuss recent advances in these and other topics crossing the existing TC boundaries.
\
The technical scope envisioned for the TSG includes the fol- lowing topics:
• Numerical methods for acoustic wave propagation, scat- tering, interactions with structures and boundaries, ra- diation, and other acoustically related phenomena
• Practical utilization of acoustical computations for en- gineering and noise control, and integration into other simulations
• Optimization, parallelization, and acceleration of compu- tational algorithms
• Validation, benchmarking, and uncertainty analysis in computational models
• Computational learning methods, data analytics, and vi- sualization
The first of these topics includes various numerical methods for solving differential and integral equations, for example, finite-difference methods, boundary-element methods, finite-element methods, parabolic equations, wavenumber integration, and ray tracing. It lies at the heart of how most researchers view computational acoustics. Some particular problems that have received strong interest in recent years include methods to efficiently handle complex boundaries and irregular meshes (e.g., Thompson, 2006; Marburg and Nolte, 2008), time-domain formulations for attenuation and impedance (e.g., Tam and Auriault, 1996), three-dimension- al solutions for sound propagating in the ocean and atmo- sphere (e.g., Castor and Sturm, 2008), wave propagation and scattering in moving and inhomogeneous media (e.g., Osta- shev and Wilson, 2016), and nonlinear wave propagation (sonic booms and explosions; e.g., Blanc-Benon et al., 2002).
Regarding the practical utilization of acoustical computa- tions for engineering and noise control, a vital research area is how to efficiently capture complex physical phenomena with limited computational resources. A prime example is model- ing noise in complex urban environments where phenomena such as reflections, shadowing by buildings, scattering and ab- sorption from facades and balconies, and distributed sound sources are all important for accurate sound-level prediction. Modeling of such phenomena in large urban spaces through direct numerical methods (e.g., finite differences or bound- ary elements) can be prohibitive even with supercomputers, although such models can be used to calibrate less intensive empirical and heuristic models.
64 | Acoustics Today | Winter 2018