Computational aeroacoustics (CAA) is the study of the generation, propagation, and scattering of sound waves in gases, particularly air, using numerical methods and computer simulations. High fidelity schemes are needed to control numerical dispersion and disipation. The figure above shows acoustic waves in a meridional plane across a ducted rotor.
CAA ducted rotor
The widespread use of drones for future delivery of packages and transportation is inevitable as a recent Morgan Stanley market research update suggests it could be a $9 trillion dollar industry by 2050. One of the main issues will be the noise produced by eVTOL and drones.
In order to understand the noise generation and propagation, a high-fidelity compressible solver is needed. Furthermore, numerical simulation must accurately represent intricate geometric details. One of the advantages of LBM is that the entire geometry can be mapped with an octree mesh. Being massively parallel, it can simulate flows at high Reynolds number. Because it has low dissipation and dispersion errors, Low Mach number flows can be simulated with high accuracy.
Why Mach Matched ?
Matching Mach and Reynolds numbers are essential for high fidelity simulation in aeroacoustics. While the importance of Reynolds number matched simulation is well-known in the aerodynamic community, the effect of Mach number is more subtle (see my publication for details JFM_aeroacoustics ).
In aeroacoustics the Mach number essentially sets the ratio between the hydrodynamic to acoustic wavelength. Disturbances at Low mach number are so low that double precision accuracy is required to estimate noise levels that are below 20 decibels. As such low disipation schemes are required. Furthermore, Mach number also sets the CFL condition in explicit schemes, as such low Mach number flows can be extremly expensive to simulate.
In sum LBM provides:
High fidelity aeroacoustic simulation tool, especially at Low mach and high Reynolds numbers, for futuristic design of eVTOL's and drones.
Direct estimation of far-field noise.
Several orders of magnitude faster than Navier-stokes solvers.
My upcoming publication highlights these points, stay tuned!