Computationally efficient ship airwake simulations for rotorcraft shipboard operations using a GPU-accelerated Lattice-Boltzmann Solver

Abstract

It is desired to achieve computationally efficient mid-fidelity ship airwake predictions to be used in ship– rotorcraft interaction studies. The Lattice-Boltzmann Method (LBM) is a viable candidate to obtain flow field solutions rapidly, owing to its highly parallelizable nature. To demonstrate the LBM’s capabilities and to show the suitability of the method to solve the large computational problem of a ship airwake simulation in a timeefficient manner, a scaled model of the SFS-2 frigate shape was simulated in uniform inflow conditions using a GPU-accelerated Lattice-Boltzmann solver. A convergence study was conducted, and it was found that the LBM results converged already at relatively coarse resolutions, but only for the mean flow quantities. The obtained results were compared to published experimental data. The LBM results showed good correlation with velocity probe measurements on the starboard side of the landing deck, but they did not predict flow asymmetry seen in the measurements. Results were also compared with particle image velocimetry data, and the flow topology as well as the mean velocity magnitudes obtained by the LBM showed excellent agreement. Satisfactory agreement was found for the turbulence intensities that were mostly underpredicted, although the gradients were correct. The LBM model was able to obtain these results in three hours and 25 minutes on eight local GPUs. Considering its computational efficiency and good accuracy for quick turn-around solutions, the current LBM model was concluded to be a good mid-fidelity alternative to high-fidelity CFD methods that require orders of magnitude more time and computational resources for ship airwake simulations.