Please follow this post in this discussion you ill the the summery of lattice boltzmann and Navier - stokes equation comparison and i also attached one pdf that ill help you regarding your query.

https://www.researchgate.net/post/Lattice_Boltzmann_equation_vs_Navier-Stokes_equations2

Certainly you can. Be aware that does not matter which technique you planning to us, you need high number of grid points, lattice numbers, control volume, etc. to get high resolution. However, LBM is more stable.

Many thanks for your time and reply.

I want to know that which procedure has lower simulation time with same accuracy?

Of course, it depend two thing:

1. The high resolution synchrotron tomography image data.

2. At low scale, before you use tomography image induced volume model to simulate the extended flow problem, you should define constructive relationship by yourself.

E.g For thermal-chemical degradation problem, the governing equation which you used is totally different with macro scale standard fluid problem

As in my view LES will be more accurate as well as high computing cost as compare to LBM. And there is a one key things which i want to put in your consideration that for high resolution you need to expanse high computation cost as well as high simulation time in LES. Although in the LBM ill be relatively low simulation time and computing cost . Since your problem is incompressible multi-phase flow so i would like to suggest if accuracy is most important parameter then go with LES or you have an option to chose between among them, then LBM is a good option.

Many thanks for your comments.

High resolution is always related to high resolution numerical scheme, such as sixth order compact scheme, spectral element methods. Standard LBM is nearly 1.5 order in precision. For high Reynolds number flow, LBM is more time-cost than the RANS. As far as i'm concerned, LBM is used near the wall region and RANS is used in the region far from the wall.