Pierre Ferrant

Two-dimensional SPH simulations of wedge water entries

This paper presents a study based on the smoothed particles hydrodynamics (SPH) method, aiming at an accurate numerical simulation of solid-fluid coupling in a free surface flow context. The SPH scheme is first described and discussed through its formulations. Then a new technic based on a particle sampling method, and designed to evaluate fluid pressure on solid boundaries is introduced. This method is then extended to the capture of freely moving body dynamics in a fluid/solid coupling approach. This study involves a spatially varying resolution, based on the so-called variable smoothing length technique, for which a new formulation of the equations is proposed. Two distinct test cases of wedge water entry are presented in order to validate this new method. Pressure prediction is first compared with analytical and experimental results, evolution in time of the body dynamics is compared with experimental results in both cases, and the pressure field on the solid boundaries is studied and discussed on the first impact case.

SPH simulation of green water and ship flooding scenarios

Flooding of a ship’s deck (greenwater) or within its internal compartments can severely restrict the operational ability of the vessel, and the safety of its cargo. In severe circumstances such as those produced by freak waves or hull damage, the vessel can become unstable causing it to sink and/or capsize. The flows produced by such events tend to be highly dynamic, with large amounts of free surface deformation. For this reason, SPH is a valuable method for predicting the physics of such flows. In this paper, SPH is used to predict fluid behaviour for two different flooding scenarios. The first is the interaction between a vessel (represented by a rigid body) and undulating travelling waves. The predicted water heights on the deck are compared to experimental results in [1]. The second is the transient flooding behaviour that occurs during, and immediately after a side collision between two vessels. Water heights are measured close to the point of impact within the vessel. The measurements are compared to experimental results in [2].

Nonlinear Higher Order Spectral Solution for a Two-Dimensional Moving Load on Ice

We calculate the nonlinear response of an infinite ice sheet to a moving load in the time domain in two dimensions, using a higher-order spectral method. The nonlinearity is due to the moving boundary, as well as the nonlinear term in Bernoullis equation and the elastic plate equation. We compare the nonlinear solution with the linear solution and with the nonlinear solution found by Parau & Dias (J. Fluid Mech., vol. 460, 2002, pp. 281-305). We find good agreement with both solutions (with the correction of an error in the Parau & Dias 2002 results) in the appropriate regimes. We also derive a solitary wavelike expression for the linear solution - close to but below the critical speed at which the phase speed has a minimum. Our model is carefully validated and used to investigate nonlinear effects. We focus in detail on the solution at a critical speed at which the linear response is infinite, and we show that the nonlinear solution remains bounded. We also establish that the inclusion of nonlinearities leads to significant new behaviour, which is not observed in the linear solution.

The Transient and Progressive Flooding Stages of Damaged Ro-Ro Vessels: A Systematic Review of Entailed Factors

Roll-on/roll-off vessels appear to be sensitive to rapid capsizing due to an abrupt ingress of water caused by maritime accidents. As a result of the damage creation, the flooded ship can experience intermediate stages, which might be more devastating than the final condition, as the sudden loading could significantly alter the ship stability characteristics. Far from a probabilistic analysis, the paper under study presents the state-of-the-art in regards to flooding physics by treating some relevant important topics. It sheds light on the transient and progressive flooding stages, focuses on relevant factors, and suggests combinations between factors that strongly affect the flooding before the steady state is reached. Furthermore, the authors comment on some points, which remain difficult to take into consideration either numerically or experimentally, and propose, where found necessary, recommendations for a more reliable assessment of the flooding process. This review shows that the intermediate flooding phase depends upon many factors, and its assessment could be adequate in calm water condition. The effects and interdependency between these factors still require further investigation. Therefore, we recommend carrying out a wide range parametric investigation into these factors, which consider their interdependency and encourage the application of the design of experiments methodology.

Run-up on a body in waves and current. Fully nonlinear and finite-order calculations

Run-up on a large fixed body in waves and current have been calculated using both a fully nonlinear time-domain boundary element model and a finite-order time-domain boundary element model, the latter being correct to second order in the wave steepness and to first-order in the current strength. The results from the two models agree well in the low Froude number and low wave steepness regime. This serves as a cross-validation of the two boundary element models. Furthermore, the two sets of data provide an excellent method for examining the domain of validity for the second-order method. Such limits are, for the case studied, given in terms of maximum Froude number and maximum wave steepness.

A non-linear wave decomposition model for efficient wave-structure interaction. Part A

This paper deals with the development of an enhanced model for solving wave-wave and wave-structure interaction problems. We describe the application of a non-linear splitting method originally suggested by Di Mascio et al. 1], to the high-order finite difference model developed by Bingham et al. 2] and extended by Engsig-Karup et al. 3,4]. The enhanced strategy is based on splitting all solution variables into incident and scattered fields, where the incident field is assumed to be known and only the scattered field needs to be computed by the numerical model. Although this splitting technique has been applied to both potential flow and Navier-Stokes solvers in the past, it has not been thoroughly described and analyzed, nor has it been presented in widely read journals. Here we describe the method in detail and carefully analyze its performance using several 2D linear and non-linear test cases. In particular, we consider the extreme case of non-linear waves up to the point of breaking reflecting from a vertical wall; and conclude that no limitations are imposed by adopting this splitting. The advantages of this strategy in terms of robustness, accuracy and efficiency are also demonstrated by comparison with the more common strategy of solving the incident and scattered fields together.

A Potential/RANSE Approach for Regular Water Wave Diffraction about 2-d Structures

Abstract A new formulation is proposed for the simulation of viscous flows around structures in waves. It consists in modifying the Reynolds-averaged Navier-Stokes equations: velocity, pressure or free-surface elevation fields are split into incident and diffracted fields to compute the diffracted flow only. The incident flow may be explicitly given by a stream function theory for non-linear regular waves, or by a spectral method for irregular waves. This method avoids classical problems (large CPU time, poor quality of generated wave) of numerical generation of waves in a viscous flow solver. The 2D flow around an immersed square in regular waves demonstrates the effectiveness of the method.

An experimental study on the flooding of a damaged passenger ship

Some recent marine incidents have shown that damaged Ro-Ro ferries can be extremely vulnerable to loss of stability. After an abrupt ingress of water caused by a maritime accident, flooding of the spaces below the car deck may endanger the ship and eventually lead to its sinking within a short period of time. The flooding stages depend upon many factors pertaining to the vessel, the accident, and the environment. Some of these factors interact during the flooding. An experimental investigation using an International Towing Tank Conference (ITTC) Ro-Ro ferry was carried out to provide a thorough insight into the flooding physics. Both transient and progressive phases are found to be highly dependent upon water and air behaviours. The damage area, the time of damage creation and the air ventilation level inside damaged compartments are key factors in determining the final ship state.

Comparison of Simulation and Tank Test Results of a Semi-submersible Floating Wind Turbine Under Wind and Wave Loads

A model of the Dutch Tri-floater semi-submersible platform equipped with the NREL 5MW wind turbine has been tested in the hydrodynamic and ocean engineering tank of Ecole Centrale Nantes under wind and wave loads. This paper aims at comparing the results obtained with numerical simulations with these experimental results. The numerical model is based on the FAST design code from NREL and a user defined platform load model for calculating hydrodynamic and mooring loads. This hydrodynamic model includes non linear hydrostatic and Froude-Krylov forces, diffraction/radiation forces obtained from linear potential theory and Morison forces to take into account viscous effects on the braces and heave plates.First the hydrodynamic model is calibrated against the results of free decay tests without wind. A good agreement is achieved by calibrating mooring properties and heave plates properties of the numerical model. Then a comparison of regular wave cases without wind is realised, and a fair agreement is observed for surge, heave and pitch motions of the floating system. Finally comparisons are realised for regular wave cases with a constant wind speed. A good agreement is observed for the steady state surge and pitch offset. Surge and heave motions also shown a good agreement, these degrees of freedom are not being strongly influenced by wind loading. For pitch motion, numerical simulations show differences around 0.4 rad/s wave frequency, for which model tests have shown a significant influence of wind loading on system motion.Copyright

Numerical simulation of a 3D viscous flow around a vertical cylinder in non-linear waves using an explicit incident wave model

The numerical simulation of 3D wave-body interaction problems in viscous fluid still leads to numerical difficulties, mainly due to large grid and CPU time requirements, especially for wave generation and propagation aspects. The correct treatment of open boundary conditions also remains problematic in direct approaches The method proposed here allows to avoid these problems by solving a modified problem for the diffracted flow only, the incident wave field being explicitly described by potential flow models. A set of results, both on 2D and 3D geometries is presented and commented, showing the capacity of the proposed scheme to solve problems of practical interest in ocean or offshore engineering in an efficient and accurate manner.Copyright