-
Indexed by:
Journal paper
-
Document Code:
21486
-
Journal:
JOURNAL OF TONGJI UNIVERSITY（NATURAL SCIENCE）
-
Included Journals:
EI
-
Place of Publication:
上海
-
Discipline:
Engineering
-
First-Level Discipline:
Architecture
-
Document Type:
J
-
Volume:
50
-
Issue:
6
-
Page Number:
793-801
-
ISSN No.:
0253-374X
-
Key Words:
wind engineering； lattice Boltzmann method； time steps； compressibility errors； over relaxation；large-eddy simulation
-
CN No.:
31-1267/N
-
DOI number:
10.11908/j.issn.0253-374x.21486
-
Date of Publication:
2022-06-01
-
Abstract:
Lattice Boltzmann method-based large-eddy simulation （LBM-LES） is a new method to solve turbulence problems in recent decades. However， improper time step settings may affect the simulation accuracy of LBM-LES. This paper first analyzed and summarized the impact of time step δt on the results of LBM-LES， theoretically. An oversized δt will cause compressibility error in the velocity field， while a too small δt can lead to the over-relaxation colliding mode， causing the numerical oscillation of velocity field. Subsequently， LBM-LES simulations of an isothermal indoor airflow case were conducted to discuss these errors quantitatively. The results show that a large δt leads to a sharp density change， and the velocity field in the regions where the Mach number （M） in the lattice Boltzmann unit exceeds 0.3 showing that there are obvious compressibility errors. Meanwhile， a too-small δt causes apparent numerical oscillations of both time-averaged and fluctuating velocities. This phenomenon is more significant when the grid resolution is higher. Therefore，it is suggested that δt should be small enough to ensure M<0.3 in the maximum velocity regions， based on which， a larger δt should be utilized to prevent numerical oscillations.
-
Links to published journals:
