Optimization of a diffuser geometry using parametric modeling tools and finite-difference time-domain simulations

Julie Meyer, Tapio Lokki

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientific

3 Citations (Scopus)

Abstract

This paper presents an approach for optimizing a diffuser geometry using parametric modeling tools and finite-difference time-domain (FDTD) acoustic simulations. Geometries are automatically generated based on a set of parameters using the modeling software Rhinoceros and its tightly integrated graphical algorithm editor Grasshopper. Scattered polar responses are then predicted using room acoustic FDTD simulations. The simulation results are analyzed in terms of the uniformity of the scattering produced by the diffusing surface. Finally, an optimization algorithm is used to define a new set of geometry parameters at each iteration. The entire process iterates until reaching the most uniform scattered polar distribution across all receiver and source positions. Such an approach can be used to facilitate and speed up the design process of diffusers that can provide a good trade-off between performance and relatively simple geometry.
Original languageEnglish
Title of host publicationAuditorium Acoustics 2018
PublisherInstitute of Acoustics
Pages638-645
ISBN (Electronic)9781906913311
Publication statusPublished - 5 Oct 2018
MoE publication typeB3 Non-refereed conference publication
EventAuditorium Acoustics - Elbphilharmonie, Hamburg, Germany
Duration: 4 Oct 20186 Oct 2018
https://www.ioa.org.uk/civicrm/event/info?id=294

Publication series

NameProceedings of the Institute of Acoustics
PublisherInstitute of Acoustics
Number3
Volume40
ISSN (Print)1478-6095
ISSN (Electronic)1478-6095

Conference

ConferenceAuditorium Acoustics
Abbreviated titleIOA AA
Country/TerritoryGermany
CityHamburg
Period04/10/201806/10/2018
Internet address

Fingerprint

Dive into the research topics of 'Optimization of a diffuser geometry using parametric modeling tools and finite-difference time-domain simulations'. Together they form a unique fingerprint.

Cite this