Assessment of accuracy and hydrodynamic behavior of different types of no-slip on the curved boundaries in the lattice Boltzmann method

This paper examines the various methods of applying no slip boundary condition on a fixed and rotary cylinder in the lattice Boltzmann framework. For this purpose, five methods of bounce back, LYMLS, QYMLS, LBFL and QBFL are chosen. The main challenge in all of these methods is how to calculate and interpolate the unknown distribution functions at the points around the boundary points. Results show that in the stable conditions (Re=20 and Re=40), the maximum error of calculation of the separation angle is 6.7 % and it is related to the bounce back method, while in the stable conditions, a significant difference cannot be seen between the bounce back and other methods. Also, the LBFL method has the most error in calculating the separation length (6% for Re=20 and 8.82 % for Re=40). By increasing the Reynolds number and increasing the rotational velocity, the bounce back method differs in the prediction of the lift and drag coefficients respect to other methods; so that there is a difference in the lift coefficient in the early times, t*> 7.78 for the conditions of k=0.2 and Re=200, between the bounce back and other methods, however with increasing time, this difference reduces, whereas the three methods of LYMLS, LBFL and QYMLS continue to produce similar results. Investigations show that the choice of the single release time τ has a serious effect on the convergence of the stated methods, so that the two methods of the QYMLS and QBFL have a smaller convergence limit.