Issue 29

A. De Rosis et alii, Frattura ed Integrità Strutturale, 29 (2014) 343-350; DOI: 10.3221/IGF-ESIS.29.30 345 of the interface lattice sites. Since the part composed of the lower density is not affected by the solution of the LB equation, the missing values of the particle distribution functions should be computed by a pressure boundary, as sketched in Fig. 1. The interested reader can refer to [9, 10] for further discussion about this approach. Figure 1 : Handling of the missing distribution functions. Red dotted arrows represent the distribution that need to be reconstructed. The present method, as a whole, is briefly described in the following steps: 1. initialize fluid and solid domain; 2. solve the fluid system, Eq. (1) (LB); 3. compute mass fluxes (VOF) and reconstruct a valid set of distribution functions at the interface (pressure boundary); 4. compute the fluid macroscopic variables  and v , Eq. (2); 5. enforce the no-slip condition at the fluid-structure interface via the iterative IB algorithm (compute g i ( x ,t) and compute the forces on the structure); 6. correct the populations with the term g i ( x ,t) and compute, again, the fluid macroscopic variables  and v , Eq. (2); 7. perform structure solution; 8. update nodes fill levels and re-initialize lattice state (i.e. emptied interface node become gas node, etc.); 9. advance in time going to step 2. R ESULTS AND DISCUSSION n this section, two cases are investigated. First, the dam break phenomenon is simulated. Then, the dynamics of two slender deformable beams invested by a flow is examined. Dam break Making reference to Fig. 2, a water column of height H =0.3m and width B =0.6m is placed in the right part of a tank of length L =1.61m, height D =6m. At t =0 the wall placed at x =0 is removed. The system described has been discretized using a relatively coarse grid, composed of 400 lattice nodes in x-direction and 150 in z-direction. During the test, the height of the water column is recorded at 4 checkpoints, H1, H2, H3 and H4. In the following, findings from a LB simulation are compared to the experimental data given in [23- 25]. In Fig. 3, the time evolution of the simulated and experimental free-surface profiles are compared. Moreover, in Tab. 1, the simulated and experimental time instants at which the water column crosses the checkpoints H2, H3 and H4 and impacts the boundary are reported. As it is possible to observe, a very close agreement is experienced. Finally, in Fig. 4, the present findings in terms of water height are compared with literature results [23-25] at the above mentioned checkpoints. As it can be noted, also in this case a good agreement is obtained. I