A Fast Wave-Based Hybrid Method for Interactive Acoustic Simulation in Large and Complex Environments

Abstract

We propose to develop a hybrid wave-based method combining two efficient methods: the ECT and PSTD, for acoustic simulation in large urban environments. In this report, we have developed an efficient scheme for our simulation method to treat arbitrarily complex geometries. Our method can correctly capture the complex shapes to produce accurate results. The complex geometries are usually represented by some CAD formats. We have developed a mesh generation to convert the STL format to the mesh that our method can directly use. We have also explored some techniques to improve our computational performance. The use of vectorization makes our code almost 3 times faster even with the use of one core. The use of blocked range of Intel TBB gives another speedup of 2.5 to 3 times with the use of 4 cores. We also show an efficient hybridization strategy to significantly improve the efficiency of our hybrid method. In the conventional strategy, the total computational domain is partitioned into boundary subdomains with ECT methods and internal subdomains with PSTD methods. This strategy requires a lot of interfacing for any pairs of adjacent subdomains. In the new strategy, we regard the whole domain as just a single PSTD domain, but with embedded boundaries inside. Boundary techniques, such as ECT or others, can be then applied locally to correct the field near the boundaries.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 24, 2012
Accession Number
ADA558086

Entities

People

  • Tian Xiao

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Boundaries
  • Boundary Layer
  • Cartesian Coordinates
  • Computational Fluid Dynamics
  • Computational Science
  • Computations
  • Computer-Aided Design
  • Coordinate Systems
  • Demographic Cohorts
  • Efficiency
  • Embedding
  • Environment
  • Geometry
  • Grids
  • Runge Kutta Method
  • Simulations
  • Triangles

Fields of Study

  • Computer science

Readers

  • Computational Fluid Dynamics (CFD)
  • Distributed Systems and Data Platform Development