Details

Autor(en) / Beteiligte

• Brill, Laura C.
• Strong, John T.

Titel

Optimizing acoustic diffusion of an architectural feature using finite-difference time-domain

Ist Teil von

• The Journal of the Acoustical Society of America, 2018-09, Vol.144 (3), p.1791-1791

Erscheinungsjahr

2018

Link zum Volltext

Quelle

American Institute of Physics

Beschreibungen/Notizen

• Parametric design is gaining in popularity as a design tool; however, rigorous analysis of the acoustic characteristics of parametrically designed surfaces typically requires testing physical mockups of a small sample of candidate designs. A computational method using a combination of finite-difference time-domain (FDTD) analysis and a genetic optimization algorithm has been developed to analyze a large number of candidate designs and select the best performers. In this case study, an architect’s aesthetic vision was translated into geometric parameters used by a genetic algorithm to optimize acoustic diffusion performance. FDTD was used to simulate an impulse response and analyze the frequency response of reflective/diffusive geometry with a resolution of up to 11 kHz. The process utilized GPU computing clusters available through Amazon Web Services (AWS) to accelerate the analysis. The current availability and relative cost-effectiveness of cloud computing resources makes wave-based acoustic analysis a more viable and efficient option. The methods and resources discussed in this paper allow acoustic performance to be integrated into the architectural optimization process thereby making acoustic performance an active consideration in an iterative design approach.