Ghost-fluid-based boundary treatment in lattice Boltzmann method and its extension to advancing boundary

A ghost-fluid-based boundary treatment in lattice Boltzmann method (GF-LBM) proposed by Tiwari and Vanka (2012) is one of the complex boundary treatment in fluid flow computation. In this paper, this treatment is developed to multi-relaxation-time LBM and further discussed with validation. It is found that, in case of stationary and sliding boundary such as Taylor-Couette flow, this treatment is superior to other bounce-back-based treatment. However, a defect in the density conservativeness is found due to the prediction of the density at image point under a non-penetration condition. Authors improve the treatment by using the normal moment relation followed by the validation. Next, the treatment is extended to other moving boundary problem such that the boundary advances with stepping over the computational grids. It is found that, this treatment is comparable to the second-order interpolated bounce-back scheme, and its precision depends on node numbers considered in each treatment. GF treatment is found to have high precision especially at curved boundary in motion.