I had read a lot of papers about large eddy simulation at high Reynolds number, but when I put it into my LBM code, I couldn' t get a realiable believable result.
Dear Dr. Tao,
The Lattice Boltzmann (LB) approach is used mainly as the direct method without any assumptions concerning the dissipation. The smallest scale represented in the calculations is thus determined by the lattice length. For this reason the method can be applied for simulation of the rather low Reynolds number flows. Numerous studies indicate conclusively that using the LB method without modeling of the unresolved scales leads to instability. Your observation is therefore correct.
There are two main approaches to extend the LB scheme to the case of large Reynolds number flow. The first one is based on the combination of the LBM with the traditional lattice gas model. The second method relies on a simpler solution where the traditional subgrid-scale model (Smagorinsky type) is used within the LB method. The later solution is discussed in section 3 of the classic paper by Hou et al. (1994),
https://arxiv.org/abs/comp-gas/9401004
My recommendation is to use the above technique in your LB model. It is simple to implement and easy to evaluate.
Sincerely,
Janusz
Dear Dr. Tao,
The Lattice Boltzmann (LB) approach is used mainly as the direct method without any assumptions concerning the dissipation. The smallest scale represented in the calculations is thus determined by the lattice length. For this reason the method can be applied for simulation of the rather low Reynolds number flows. Numerous studies indicate conclusively that using the LB method without modeling of the unresolved scales leads to instability. Your observation is therefore correct.
There are two main approaches to extend the LB scheme to the case of large Reynolds number flow. The first one is based on the combination of the LBM with the traditional lattice gas model. The second method relies on a simpler solution where the traditional subgrid-scale model (Smagorinsky type) is used within the LB method. The later solution is discussed in section 3 of the classic paper by Hou et al. (1994),
https://arxiv.org/abs/comp-gas/9401004
My recommendation is to use the above technique in your LB model. It is simple to implement and easy to evaluate.
Sincerely,
Janusz
Dear Dr. Tao,
Please, be prepared to work on potential extensions of the fundamental high-order discretized equations and implement high-order boundary conditions in combination with the right number of lattices and sizes. Step-by-step you will increase the Reynolds and validate the model as you increase the resolution and accuracy of the boundary conditions. Try not to follow the approach of traditional CFD models, i.e., postpone adding a subgrid scale model unless you can document that it is absoluttely necessary.
Best wishes with developing your model,
Jannette Frandsen
- Badges
- Science topic
- Similar topics
- Engineering
- Thermal Engineering
- Thermofluid
- Fluid Mechanics
- Fluid Dynamics