Scholarlink

Large eddy simulationsLarge Eddy Simulation (LES) can provide accurate predictions of pressure exerted by the wind on a building. However, LES is a computationally overwhelming task because of its high resolution and storage demands. So far, its industrial use has been impracticable in the field of wind engineeringWind engineering. Recent developments on high-performance computingHigh-performance computing, relying on the decline in the GPU costs (relative to processing capability) and alternative computational fluid dynamicsComputational Fluid Dynamics (CFD) paradigms, might change this scenario. In this paper, we present an LESLarge Eddy Simulation (LES) solver based on the lattice Boltzmann methodLattice Boltzmann Method (LBM) (LBM) implemented in a GPU environment. The LBMLattice Boltzmann Method (LBM) has a second-order accuracy with respect to time and space discretizations, and its operations are mostly local, making it extremely efficient in parallel processing. With the current solver, engineering-level simulations of wind pressures on a building can be performed overnight. Two validationValidation cases with Reynolds number of \(10^5\) are analyzed, one of a low-rise standard building and other a high-rise CAARC building. The results are discussed in terms of the statistical properties of velocity and pressure signals. Great computational performance has been achieved, at least an order of magnitude faster than a traditional finite volume solver.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Viable Framework for Wind Pressure Assessments with Large Eddy Simulations

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Rodrigo S. Romanus,
  • Alan Lugarini,
  • Waine Oliveira,
  • Admilson T. Franco

摘要

Large eddy simulationsLarge Eddy Simulation (LES) can provide accurate predictions of pressure exerted by the wind on a building. However, LES is a computationally overwhelming task because of its high resolution and storage demands. So far, its industrial use has been impracticable in the field of wind engineeringWind engineering. Recent developments on high-performance computingHigh-performance computing, relying on the decline in the GPU costs (relative to processing capability) and alternative computational fluid dynamicsComputational Fluid Dynamics (CFD) paradigms, might change this scenario. In this paper, we present an LESLarge Eddy Simulation (LES) solver based on the lattice Boltzmann methodLattice Boltzmann Method (LBM) (LBM) implemented in a GPU environment. The LBMLattice Boltzmann Method (LBM) has a second-order accuracy with respect to time and space discretizations, and its operations are mostly local, making it extremely efficient in parallel processing. With the current solver, engineering-level simulations of wind pressures on a building can be performed overnight. Two validationValidation cases with Reynolds number of \(10^5\) are analyzed, one of a low-rise standard building and other a high-rise CAARC building. The results are discussed in terms of the statistical properties of velocity and pressure signals. Great computational performance has been achieved, at least an order of magnitude faster than a traditional finite volume solver.