Search

Despite recent advances in machine learning, simulation driven design optimization using high-fidelity simulations may still be prohibitively expensive for practical applications. This paper investigates improvements in multi-fidelity surrogate-based hydrodynamic optimization, which are intended to make the process faster and more efficient. Specific innovations are: a) the use of a reduced initial dataset with only one data point for all fidelity levels except the lowest, to reduce the computational cost of surrogate-model initialization; b) accounting for noise variance in the selection of the fidelity level to sample, to avoid oversampling well-resolved but noisy fidelity levels, c) improving the automatic mesh adaptation protocol for the CFD simulations, to further optimize mesh adaptation; and d) restarting high fidelity simulations from converged low-fidelity results, to improve the overall efficiency of the design optimization process. These methodological advancements are demonstrated for an analytical test problem, as well as the shape optimization of a NACA 4-digit airfoil and the DTMB 5415 for calm-water resistance. The results show that the reduced dataset drastically reduces the computational cost of initialization and favors efficient low-fidelity exploration of the design space. The noise-corrected fidelity selection encourages the selection of higher fidelities, to effectively determine the true optimum. Finally, the CFD solver advancements make high-fidelity simulations faster, up to eight times. Compared with previous work, the solution of the DTMB 5415 problem exhibits a more robust training process, providing a slightly improved optimum, in less than half the computational time.

International audience; A generalized multi-fidelity (MF) metamodel of CFD (computational fluid dynamics) computations is presented for design-and operational-space exploration, based on machine learning from an arbitrary number of fidelity levels. The method is based on stochastic radial basis functions (RBF) with least squares regression and in-the-loop optimization of RBF parameters to deal with noisy data. The method is intended to accurately predict ship performance while reducing the computational effort required by simulation-based optimization (SBDO) and/or uncertainty quantification problems. The present formulation here exploits the potential of simulation methods that naturally produce results spanning a range of fidelity levels through adaptive grid refinement and/or multi-grid resolution (i.e. varying the grid resolution). The performance of the method is assessed for one analytical test and three SBDO problems based on CFD simulations, namely a NACA hydrofoil, the DTMB 5415 model, and a roll-on/roll-off passenger ferry in calm water. Under the assumption of a limited budget of function evaluations, the proposed MF method shows better performance in comparison with its single-fidelity counterpart. The method also shows very promising results in dealing with and learning from noisy CFD data.

International audience; The flow around the well-known KRISO Very Large Crude Carrier no. 2 (KVLCC2) is characterised by boundary layer thickening, flow separation, and vortex creation on the aft body of the ship, leading to a very nonhomogeneous velocity field in the propeller plane (figure 1). It is known since the 1990s that the numerical simulation of such flows is highly sensitive to the treatment of the turbulence. For RANS equations, standard turbulence models perform poorly; measures like the anisotropic treatment of turbulence production are necessary in order to compute correctly the most famous aspect of the flow, the hook-shaped low-velocity zone in the main longitudinal vortex. The case is therefore justly famous as a test of turbulence models for marine flow simulation. The focus of most studies that use the KVLCC2, such as Larsson et al. [3], is the comparison of different flow solvers or different turbulence models. To keep such studies compact and accessible, comparisons are mostly done for a few physical features only. The propeller plane with its hook shape in the axial velocity is often studied. The objective of this paper, on the contrary, is to use a single numerical simulation of good quality to visualise the entire flow field around the aft body of the KVLCC2.We then attempt a physical analysis of this flow, searching explanations of less-studied features such as the outer boundary layer shape and the flow behind the propeller hub cap.We hope that better knowledge of the flow will allow, in the future, to understand more precisely the behaviour of different turbulence models, which may ultimately lead to better models. The paper starts with a definition of the test case and the computation (section 2). Then section 3 presents the visualisation and analysis of the numerical solution. To assess the validity of the analysis, selected aspects of the solution are compared with experiments in section 4. The conclusion (section 5) addresses the challenges for turbulence modelling which the KVLCC2 test case still presents.

International audience; This paper studies if adaptive grid refinement combined with finite-volume simulation of the incompressible RANS equations can be used to obtain grid-independent solutions of realistic flow problems. It is shown that grid adaptation based on metric tensors can generate series of meshes for grid convergence studies in a straightforward way. For a two-dimensional airfoil and the flow around a tanker ship, the grid convergence of the observed forces is sufficiently smooth for numerical uncertainty estimation. Grid refinement captures the details of the local flow in the wake, which is shown to be grid converged on reasonably-sized meshes. Thus, grid convergence studies using automatic refinement are suitable for high-Reynolds incompressible flows.

International audience

International audience

For the simulation of water flow with waves, grid refinement must be anisotropic to limit the total grid size. For these flows, the grid has to be refined at the water surface, to resolve the conservation law which indicates the surface position, and below the surface to resolve the water flow. A combined criterion is presented, based on the free-surface position and on the Hessian of the pressure. It is shown that this criterion creates suitable grids for two- and three-dimensional flows.

International audience; L'objectif de cet article est de d'écrire l'état des travaux aussi bien expérimentaux que numériques réalisés au LMF pour étudier l'écoulement autour des palettes d'aviron. La démarche adoptée, l'influence des paramètres physiques prépondérants ainsi que les étapes vers la modélisation d'un coup d'aviron réel avec la précision adéquate sont ainsi présentés.

Colloque avec actes et comité de lecture. Internationale.; International audience; Nous présenterons une méthode de raffinement automatique de maillage pour la simulation des écoulements visqueux incompressibles, en calcul parallèle, sur des maillages non-structurés. La méthode est appliquée aux écoulements à surface libre stationnaires et non-stationnaires. Pour ces simulations, une forte augmentation de la qualité des solutions est obtenue.

A novel multigrid method for the solution of the steady Reynolds-averaged Navier–Stokes equations is presented, that gives convergence speeds similar to laminar flow multigrid solvers. The method is applied to Menter’s one-equation turbulence model. New aspects of the method are the combination of nonlinear Gauss–Seidel smoothing on the finest grid with linear coarse-grid corrections, and local damping in the initial stages of the computation, to keep the solution stable; the damping needed is estimated with the nonlinear smoother. Efficiency on the finest grid is increased with full multigrid, second-order accuracy is obtained with defect correction. Tests on boundary layers and airfoil flows show the efficiency of the method.