Search

In this study, we present different numerical methods for the simulation of liquid-vapor phase change (boiling). We use a Level Set formalism to capture the liquid-vapor interface. Such a formalism...

International audience; A Smoothed Particle Hydrodynamics axisymmetric solver was developed in order to simulate the collapse of a single cavitation bubble close to an elastic-plastic material and study plasticity formation and hence material erosion. Findings indicate the relative importance of the material deformation due to the impact of the micro-jet and the shock wave that develop during collapse. A shock-wave dominated impact has a much higher material erosion ability compared to a micro-jet impact. Strain rate is found to have a significant effect on plastic deformation, with an overestimation of the plastic deformation up to 60% if strain rate effects are neglected in the case of stainless steel A2205. We also demonstrate that, although the impact pressure is maximum just below the collapsing bubble, maximum plastic strain occurs at a radial offset from the symmetry axis. This is the result of inertial effects that have an impact on both the magnitude and the position of the plastic domain in the material. A new non-dimensional parameter called effective pressure is introduced that can predict plastic strain location accurately for higher stand-off ratios. Alternatively, a characteristic time analysis also shows that it can be used for prediction of plastic strain zone in the solid for detached cavities.

This research is devoted to understanding the physical mechanism of cavitation erosion in liquid flows at the fundamental scale of cavitation bubble collapse. Cavitation bubbles form in a liquid when the pressure of the liquid decreases locally below the saturated vapor pressure p v s a t . The bubbles grow due to low ambient pressure and collapse when the surrounding liquid pressure increases again above p v s a t . Bubble collapse near solid walls can result in high velocity liquid jet and shock wave emission that cause high pressure loads on the wall. These pressure loads are responsible for the erosive damages on solid surfaces, as observed in applications like liquid fuel injection, hydrodynamic power generation and marine propulsion. On the other hand, the pressure loads from collapsing bubbles are exploited for applications like shock wave lithotripsy, drug delivery and cleaning surfaces. In this work, we follow a numerical approach, which begins with the development of a compressible solver capable of resolving the cavitation bubbles in the finite-volume code YALES2 1 employing a simplified homogeneous mixture model. The solid material response to cavitation loads is resolved with the finite element code Cast3M 2 . A one–way coupling approach for fluid–structure interaction (FSI) simulation between the fluid and solid domains is pursued. In this simple approach, the pressure field computed by the fluid solver at the fluid–solid interface is communicated to the solid solver which computes the deformation induced in the material. In the end, the dynamical events responsible for surface deformation are highlighted from 2D vapor bubble collapse dynamics and associated pressure loads on the solid wall are estimated. The response of different materials to bubbles collapsing at different distances from the solid wall is discussed.

International audience; A meshless Smoothed Particle Hydrodynamics solver is used to simulate the collapse of a cavitation bubble near a solid material taking into account the complex fluid-structure interaction. A parametric study has been performed to study the effect of stand-off ratio, bubble size, driving pressure and strain rate on the material response. We focus on plastic strain magnitudes and plastic strain energy dissipation to compare different cases and their ability to cause material erosion. Findings indicate that, in the case of repeated collapses, cavities attached to the solid have an ability to initiate damage quicker but exhibit lower erosion rate compared to the cavities detached from the solid. The incubation time does not depend on the size of the bubbles, unlike the erosion rate which is strongly affected by the bubble size. It is also found that the amount of cumulated plasticity is overestimated by more than 150% when the strain-rate sensitivity is not taken into account in the material modelling which suggests that using an appropriate plasticity model that includes strain-rate sensitivity is mandatory while studying the phenomenon of cavitation erosion.

International audience; A meshless Smoothed Particle Hydrodynamics solid solver is developed in order to study fluid-structure interactions and to predict cavitation erosion. The solid solver is developed in-house in an axisymmetric configuration. The existing SPH methods dedicated to solid materials do not allow a consistent treatment of particles close to the symmetry axis, so a density correction scheme is proposed here to derive new density and momentum equations for solid mechanics in axisymmetric SPH formulation. The new SPH equations are coded in the solver and the SPH solid solver is then validated against FEM results which shows excellent agreement.

Accurate numerical simulation of moving contact lines on complex boundaries with surface wettability effect remains a challenging problem. In this paper, we introduce a robust and accurate method to perform 3D contact line simulations on unstructured meshes with an imposed contact angle. The contact angle is imposed through a sub-grid scale curvature term and the contact line motion is enabled thanks to partial slip on the wall. Moreover, an original strategy has been designed to improve normal and curvature computation at contact line from the level-set field. The whole method is validated on 2D and 3D test cases and shows good mass conservation properties. The drop detachment from a horizontal fiber due to gravity or surface tension is then investigated. For this purpose, dynamic mesh adaptation is used to keep high resolution around the interface with moderate number of cells. These realistic cases demonstrate the ability of the numerical method to handle surface wettability effects on resolved complex geometries.

International audience

A numerical method for the simulation of multiphase flows with phase change on unstructured grids is presented. Based on the work of Tanguy et al. [31] using a Level Set/Ghost Fluid Method coupling for two-dimensional axisymmetric cartesian grids, we extend the method to two- and three-dimensional unstructured grids with the aim of taking a step towards realistic boiling simulations in industrial context. The mass transfer rate at the interface accounting for phase change is computed as a function of the liquid and vapor heat fluxes at the interface and of the latent heat of the fluid by means of a new framework improving the accuracy of differential operators. The mass transfer rate is then used in the projection method to solve Navier-Stokes equations and in the advection equation of the Level Set function to account for the interface movement due to phase change. An immersed Dirichlet boundary condition is imposed at the interface to ensure that boiling always occurs at saturation temperature. We demonstrate the accuracy of our method first in the case of a static growing bubble with a fixed constant mass transfer rate, and second with a mass transfer rate computed from the thermal fluxes at the interface. In both cases, the bubble radius at final time converges with grid refinement towards the theoretical value.

International audience; This numerical work focuses on the collapse of a single air bubble close to a deformable wall generated by the impact of an incident high-pressure wave. A CFD code was fully coupled to a FEM solid code in order to compute the bubble collapse and the subsequent plastic deformation in the material. The CFD code FSI4CAV was developed from the numerical model of Johnsen et al. [1]. A mobile mesh capability was added in order to account for the displacement of the fluid-structure interface. This has been done through the implementation of an ALE (Arbitrary Lagrangian Eulerian) method. The solid response to the bubble collapse was computed with the FEM software CAST3M [2] assuming an elastic-plastic constitutive law for the material. The communication between the two codes is achieved through the MPI library. The paper presents a detailed analysis of the two-dimensional bubble collapse dynamics and the corresponding solid behaviour. The damping of the impact pressure with respect to a perfectly rigid wall was computed as well as the plastic deformation developed in the material. The comparison of the simulation results with a non-coupled approach where the pressure applied to the solid is computed from CFD simulations definitely shows the importance of including the fluid-structure interaction into the simulation of cavitation erosion.

International audience; The motivation behind this research lies in understanding the physical mechanism of cavitation erosion in compressible liquid flows, with applications in the field of aerospace, hydrodynamics, diesel injectors etc. As a consequence of collapsing vapor cavities in cavitating flow near solid boundaries, high pressure impact loads are generated. These pressure loads are believed to be responsible for the erosive damages on solid surface observed in most applications. For our investigation, the initial geometry is a single vapor bubble near a solid boundary collapsing due to the pressure difference between the bubble and surrounding liquid. The numerical approach employs a simplified homogenous mixture or 'single fluid' model with barotropic assumption in a fully compressible finite-volume fluid solver. The numerical method is validated against the well-known Rayleigh collapse of a pure 3D vapour bubble. It is then used for the simulation of a 2D vapour bubble collapsing in the proximity of a solid boundary placed at a specified distance from the centre of the bubble. The pressure loads are computed from the evolving dynamics of collapsing bubble near a solid boundary which can be used to determine the resulting surface deformation. The developed compressible cavitation solver in the CFD code Y ALES2 can efficiently model small and large scale cavitating structures in a fully resolved three dimensional flow.

This study report clinical course, etiology, management, and long-term outcome of children who developed toxic epidermal necrolysis-like reaction (TEN-LR) after allogeneic hematopoietic stem cell transplantation (allo-HSCT). We retrospectively collected children with TEN-LR occurring after allo-HSCT performed in 2 pediatric bone marrow units between 2005 and 2014. We identified 6 cases of TEN-LR of 322 patients (1.8%). Possible triggers of TEN included antibiotics, antiepileptics, antimycotics, and Mycoplasma infection. In 3 patients TEN-LR occurred concurrently with severe multiorgan acute graft versus host disease. The management of TEN included administration of high doses of intravenous immunoglobulins and steroids (n=6), anti-tumor necrosis factor (n=3), and plasmapheresis (n=3) and whenever possible, discontinuation of the potentially causative drugs. Four patients (66%) reached a complete clinical response of TEN-LR after a median of 11.2 days. Two children (34%) are presently alive, 1 with long-term ocular sequelae. TEN-LR is a potentially lethal complication that may occur after HSCT also in pediatric patients. In our experience, TEN-LR and acute graft versus host disease probably coexisted and an overlap between the 2 forms is suggested. The multidisciplinary approaches involving specialized nurses, hematologists, dermatologists, burn surgeons, and infectious disease specialists is crucial to treat these patients.

International audience; We perform particle scale simulations of suspensions submitted to shear reversal. The simulations are based on the Force Coupling Method, adapted to account for short range lubrication interactions together with direct contact forces between particles, including surface roughness, contact elasticity and solid friction. After shear reversal, three consecutive steps are identified in the viscosity transient: an instantaneous variation, followed by a rapid contact force relaxation, and finally a long time evolution. The separated contributions of hydrodynamics and contact forces to the viscosity are investigated during the transient, allowing a qualitative understanding of each step. In addition, the influence of the contact law parameters (surface roughness height and friction coefficient) on the transient are evaluated. Concerning the long time transient, the difference between the steady viscosity and minimum viscosity is shown to be proportional to the contact contribution to the steady viscosity, allowing in principle easy determination of the latter in experiments. The short time evolution is studied as well. After the shear reversal, the contact forces vanish over a strain that is very short compared to the typical strain of the long time transient, allowing to define an apparent step between the viscosity before shear reversal and after contact force relaxation. This step is shown to be an increasing function of the friction coefficient between particles. Two regimes are identified as a function of the volume fraction. At low volume fraction, the step is small compared to the steady contact viscosity, in agreement with a particle pair model. As the volume fraction increases, the value of the viscosity step increases faster than the steady contact viscosity, and, depending on the friction coefficient, may approach it.

International audience; This paper is devoted to cavitation erosion modeling. It presents some recent numerical developments made in the code initially developed in collaboration with E.Johnsen and collaborators at University of Michigan [1] in order to account for fluid-structure interaction. The considered test case is that of a single air bubble collapsing near a wall due to an incident shock wave in the surrounding water. In our investigation, we focus on the code implementation and optimization and the bubble implosion mechanism. The paper is focused on the various events occurring during bubble collapse and the computation of the time evolution of the pressure distribution. The influence of the amplitude of the incident wave and the distance of the bubble to the wall are investigated.

Two dimensional numerical simulations of sets of vesicles in a Poiseuille flow are presented. Vesicles are a simple model to describe the dynamics of red cells in blood flow. At the scale of vesicles, the hydrodynamics is well described by the Stokes equation, whose linearity allows the use of Green's functions via the boundary integral method. This is coupled with the fast multipole method to acheive optimal scaling with respect to the number of discretization points. Results are presented for sets of different number of vesicles, showing their spatial organization. Vesicles assume a parachute-like shape and align one to the other in the centre of the parabolic profile. The relative distances depend on the total number of vesicles and on the position in the set.

The rheology of a dilute two-dimensional suspension of vesicles (closed bags of a lipid bilayer membrane) is studied by numerical simulations. The numerical methods used are based on the boundary integral formulation (Green's function technique) and the phase field approach, which has become a quite popular and powerful tool for the numerical study of free-boundary problems. The imposed flow is an unbounded linear shear. The goal of the present study is to elucidate the link between the rheology of vesicle suspensions and the microscopic dynamics of the constituent particles (tank-treading and tumbling motions). A comparison with emulsion rheology reveals the central role played by the membrane. In particular, at low viscosity ratio λ (defined as the viscosity of the internal fluid over that of the ambient one), the effective viscosity decreases with λ, while the opposite trend is exhibited by emulsions, according to the classical Taylor result. This fact is explained by considering the velocity field of the ambient fluid. The area-incompressibility of the vesicle membrane modifies the surrounding velocity field in a quite different manner than what a drop does. The overall numerical results in two dimensions are in reasonable agreement with the three-dimensional analytical theory derived recently in the small deformation limit (quasi-spherical shapes). The finding that the simulations in two dimensions capture the essential features of the three-dimensional rheology opens the way for extensive and large-scale simulations for semi-dilute and concentrated vesicle suspensions. We discuss some peculiar effects exhibited by the instantaneous viscosity in the tumbling regime of vesicles. Finally, the rheology is found to be relatively insensitive to shear rate.

We study the dynamics and rheology of a single two-dimensional vesicle embedded in a linear shear flow by means of numerical simulations based on the boundary integral method. The viscosities inside and outside the vesicle are supposed to be identical. We explore the rheology by varying the reduced area, i.e. we consider more and more deflated vesicles. Effective viscosity and normal stress differences are computed and discussed in detail, together with the inclination angle and the lateral membrane velocity (tank-treading velocity). The angle is found, surprisingly, to reach a zero value (flow alignment) at a critical reduced area even in the absence of viscosity contrast. A Fast Multipole Method is presented that enables to run efficiently simulations with a large number of vesicles. This method prevails over the direct summation for a number of mesh points beyond a value of about 10 3 . This offers an interesting perspective for simulation of semi-dilute and concentrated suspensions.

Coalescing drops spontaneously jump out of plane on a variety of biological and synthetic superhydrophobic surfaces, with potential applications ranging from self-cleaning materials to self-sustained condensers. To investigate the mechanism of self-propelled jumping, we report three-dimensional phase-field simulations of two identical spherical drops coalescing on a flat surface with a contact angle of 180°. The numerical simulations capture the spontaneous jumping process, which follows the capillary–inertial scaling. The out-of-plane directionality is shown to result from the counter-action of the substrate to the impingement of the liquid bridge between the coalescing drops. A viscous cutoff to the capillary–inertial velocity scaling is identified when the Ohnesorge number of the initial drops is around 0.1, but the corresponding viscous cutoff radius is too small to be tested experimentally. Compared to experiments on both superhydrophobic and Leidenfrost surfaces, our simulations accurately predict the nearly constant jumping velocity of around 0.2 when scaled by the capillary–inertial velocity. By comparing the simulated drop coalescence processes with and without the substrate, we attribute this low non-dimensional velocity to the substrate intercepting only a small fraction of the expanding liquid bridge.

Self-propelled jumping upon drop coalescence has been observed on a variety of textured superhydrophobic surfaces, where the jumping motion follows the capillary–inertial velocity scaling as long as the drop radius is above a threshold. In this paper, we report an experimental study of the self-propelled jumping on a Leidenfrost surface, where the heated substrate gives rise to a vapour layer on which liquid drops float. For the coalescence of identical water drops, we have tested initial drop radii ranging from 20 to $\def \xmlpi #1{}\def \mathsfbi #1{\boldsymbol {\mathsf {#1}}}\let \le =\leqslant \let \leq =\leqslant \let \ge =\geqslant \let \geq =\geqslant \def \Pr {\mathit {Pr}}\def \Fr {\mathit {Fr}}\def \Rey {\mathit {Re}}500\ \mu \mathrm{m}$, where the lower bound is related to the spontaneous takeoff of individual drops and the upper bound to gravitational effects. Regardless of the approaching velocity prior to coalescence, the measured jumping velocity is around 0.2 when scaled by the capillary–inertial velocity. This constant non-dimensional velocity holds for the experimentally accessible range of drop radii, and we have found no cutoff radius for the scaling, in contrast to prior experiments on textured superhydrophobic surfaces. The Leidenfrost experiments quantitatively agree with our numerical simulations of drop coalescence on a flat surface with a contact angle of 180°, suggesting that the cutoff is likely to be due to drop–surface interactions unique to the textured superhydrophobic surfaces.

8-Methoxypsoralen (8-MOP) plus UVA irradiation (PUVA therapy) has been used for the treatment of psoriasis. PUVA therapy has been associated with an increased risk of developing skin squamous cell carcinoma (SCC). In order to determine the PUVA-induced p53 mutation spectrum, a yeast expression vector harbouring a human wild-type p53 cDNA was incubated with 8-MOP, and UVA irradiated in vitro. PUVA-damaged and undamaged DNA was transfected into a yeast strain containing the ADE2 gene regulated by a p53-responsive promoter. An 8-MOP concentration-dependent decrease in survival and increase in mutant frequency were observed. At a fixed 8-MOP concentration, survival decreased and mutant frequency increased as UVA irradiation increased. Eleven mutant clones contained 11 mutations: 10 were single base pair substitutions, the remaining one being a complex mutation. All eight T:A-targeted mutations were at 5'-TpA sites, hallmark mutations of PUVA mutagenesis. Through a rigorous statistical test, the PUVA-induced p53 mutation spectrum appears to differ significantly (P < 0.0002) from that observed in SCC in PUVA-treated patients. The present work demonstrates that a specific PUVA-induced mutational fingerprint could be obtained and recognized on human p53 cDNA. This result may suggest that PUVA therapy can be a risk factor for the development of SCC in psoriasis patients through a mechanism not involving the induction of p53 mutations.

International audience; The behaviour of red blood cells (RBCs), modelled as vesicles, in Poiseuille flow, mimicking the microvasculature, is studied with numerical simulations in two dimensions. RBCs moving in the centre of the Poiseuille flow (as in blood capillaries) are shown to attract each other and form clusters only due to hydrodynamic interactions, provided that their distance at a given time is below a certain critical value. This distance depends on physical parameters, such as the flow strength. Our simulations reveal that clusters are unstable above a threshold value in the number of forming RBCs, beyond which one or few cells escape the pack by a self-regulating mechanism that select the marginally stable size. This size selection depends on the flow strength as well as on the RBC swelling ratio. The results are interpreted via the analysis of the perturbation of the flow field induced by the vesicles and the interplay with bending and tension forces. This sheds a novel light on the process of collective motion of RBCs observed in vivo.

We report an experimental study on shearing a monolayer of monodisperse bubbles floating on liquid in a narrow-gap Couette device. The bubbles in such a "bubble raft" coalesce only if the shear rate exceeds a threshold value. This is in contrast to the conventional wisdom that bubbles and drops coalesce for gentler collisions, at shear rates below a critical value. Furthermore, the threshold shear rate increases with the bubble size and the viscosity of the suspending liquid, contravening reasoning based on capillary number. Through visualization and scaling arguments, we investigate several plausible mechanisms for the anomalous coalescence. None explains all aspects of the observations. The most promising model is one based on inertial forces that compress the bubbles radially inward and accelerate film drainage.

International audience; Red blood cells (RBC) flowing in microcapillaries tend to associate into clusters, i.e., small trains of cells separated from each other by a distance comparable to cell size. This process is usually attributed to slower RBCs acting to create a sequence of trailing cells. Here, based on the first systematic investigation of collective RBC flow behavior in microcapillaries in vitro by high-speed video microscopy and numerical simulations, we show that RBC size polydispersity within the physiological range does not affect cluster stability. Lower applied pressure drops and longer residence times favor larger RBC clusters. A limiting cluster length, depending on the number of cells in a cluster, is found by increasing the applied pressure drop. The insight on the mechanism of RBC clustering provided by this work can be applied to further our understanding of RBC aggregability, which is a key parameter implicated in clotting and thrombus formation.

Cross-streamline migration of deformable entities is essential in many problems such as industrial particulate flows, DNA sorting, and blood rheology. Using two-dimensional numerical experiments, we have discovered that vesicles suspended in a flow with curved flow lines migrate towards regions of high flowline curvature, which are regions of high shear rates. The migration velocity of a vesicle is found to be a universal function of the normal stress difference and the flow curvature. This finding quantitatively demonstrates a direct coupling between a microscopic quantity (migration) and a macroscopic one (normal stress difference). Furthermore, simulations with multiple vesicles revealed a self-organization, which corresponds to segregation, in a rim closer to the inner cylinder, resulting from a subtle interaction among vesicles. Such segregation effects could have a significant impact on the rheology of vesicle flows.

We report direct numerical simulations of liquid filaments breaking up into droplets on partially wetting substrates. It is motivated by recent experiments, linear stability analyses, and lubrication-based calculations. The fluid flow is governed by the Stokes equations and the contact line motion is handled by a phase-field model, which also serves to capture the interfacial motion. The coupled Stokes and Cahn-Hilliard equations are solved using a finite-element algorithm in three dimensions. This avoids additional approximations of the fluid flow or contact line motion, and allows us to compute arbitrary contact angles on the substrate. We simulate both the breakup of infinite liquid filaments via growing capillary waves and that of finite liquid filaments with drops pinching off from the ends, with a focus on the effect of the wetting angle. In both cases, substrate hydrophobicity promotes breakup of the thread, and decreases the spacing of the daughter drops. The results show the differences in the two processes and in the final drop size and spacing. The development of capillary waves agrees well with prior linear analysis and the end-pinching results offer new insights into this poorly understood phenomenon.

International audience; In the present study a meshless Smoothed Particle Hydrodynamics cavitation solver is developed. The fluid bubble collapse solver is validated against analytical Rayleigh-Plesset equation and shows good agreement. The solid solver capable of solving elastic-plastic deformation and material damage is developed and is validated against FEM results. A fluid structure interaction solver capable of solving cavitation erosion is presented. A single bubble collapse is demonstrated in the paper.

International audience

The objective of this thesis is the numerical simulation of the boiling henomenon on unstructured grids. Boiling is the phase change of fluid particles from the liquid phase to the vapor phase under the action of thermal fluxes at the interface separating the two phases. Boiling is thus encountered in two-phase flows and driven by the mass transfer rate at the interface. This mass transfer rate is computed from the thermal fluxes on both sides of the interface. Consequently, a highly accurate numerical method is needed to locate the interface throughout the simulation. The Navier-Stokes equations are then coupled to the heat equation by means of the mass transfer rate at the interface. Such simulations have been performed by Tanguy et al. (J. Comput.Phys., 2014) on two-dimensional axisymmetric cartesian grids. In this thesis, weextend this methodology to three-dimensional unstructured grids (composed ofirregular tetrahedra, useful to describe complex geometries). We then developed a specific solver in the YALES2 code (finite-volume-based code for simulations of two-phase flows on 3D unstructured grids). The interface motion is captured by the Level Set method. Phase change implies velocity and pressure discontinuities at the interface which especially depend on the mass transfer rate. These discontinuities are taken into account by the Ghost Fluid Method, with two velocity fields and two temperature fields. This methodology being already well established for structured cartesian grids, the contribution of this thesis relies on the ability to simulate phase change by boiling on three-dimensional unstructured grids. The particularities of unstructured grids have demanded numerous developments for the reinitialization of the Level Set function after advection, as well as the use of high-order operators forthe computation of the mass transfer rate at the interface. The proposed developments are finally validated on unstructured grids against the analytical test-case of a 3D bubble expanding inside a superheated quiescent liquid., Cette thèse a pour objectif la simulation numérique du phénomène d'ébullition sur maillages non structurés. L'ébullition est le changement de phase des particules fluides de la phase liquide vers la phase vapeur sous l'action des flux thermiques à l'interface séparant les deux phases. Il s'agit donc d'un phénomène rencontré au sein d'écoulements diphasiques et piloté par le taux de transfert de masse à l'interface. Ce taux de transfert de masse est calculé à partir des flux thermiques de part et d'autre de l'interface. Cela implique donc la nécessité d'adopter une méthode de suivi d'interface très précise pour localiser l'interface à tout instant de la simulation. Les équations de Navier-Stokes sont alors couplées à l'équation de la chaleur par l'intermédiaire du taux de transfert de masse à l'interface. De telles simulations ont été menées par Tanguy et al. (J. Comput. Phys., 2014) sur des maillages cartésiens axisymétriques en deux dimensions. Dans cette thèse, nous étendons cette méthodologie à des maillages non structurés en trois dimensions (maillages composés de tétraèdres non réguliers utiles pour décrire des géométries complexes). Pour ce faire, nous avons développé un solveur spécifique dans le code YALES2 (code diphasique basé sur la méthode des volumes finis pour des maillages 3D non structurés). Le suivi de l'interface est assuré par la méthode Level Set. Le changement de phase engendre des discontinuités de vitesse et de pression à l'interface qui dépendent notamment du taux de transfert de masse. Ces discontinuités sont prises en compte par la méthode Ghost Fluid à l'aide de deux champs de vitesse et deux champs de température. Cette méthodologie étant déjà bien établie pour des maillages structurés cartésiens, l'apport de cette thèse réside dans la possibilité de simuler le changement de phase par ébullition sur des maillages non structurés en trois dimensions. Les spécificités des maillages non structurés ont nécessité de nombreux développements pour la réinitialisation de la fonction Level Set après advection, ainsi que l'utilisation d'opérateurs d'ordres élevés pour le calcul du taux de transfert de masse à l'interface. L'ensemble des développements proposés est finalement validé sur maillages non structurés à l'aide du cas-test analytique d'une bulle 3D en expansion dans un liquide surchauffé au repos.

This research is devoted to understanding the physical mechanism of cavitation erosion in compressible liquid flows on the fundamental scale of cavitation bubble collapse. As a consequence of collapsing bubbles near solid wall, high pressure impact loads are generated. These pressure loads are believed to be responsible for the erosive damages on solid surface observed in most applications. Our numerical approach begins with the development of a compressible solver capable of resolving the cavitation bubbles in the finite-volume solver YALES2 employing a simplified homogenous mixture model. The solver is extended to Arbitrary Lagrangian-Eulerian formulation to perform fluid structure interaction simulation with moving mesh capabilities. The material response is resolved with the finite element solver Cast3M, which allowed us to perform one-way and two-way coupled simulations between the fluid and solid domains. In the end, we draw comparisons between 2D and 3D vapor bubble collapse dynamics and compare them with experimental observations. The estimated pressure loads on the solid wall and different responses of materials for attached and detached bubble collapses are discussed. Finally, the damping of pressure loads by different materials is identified with two-way coupled fluid-structure interaction.; Ce travail de recherche est dédié à la compréhension des mécanismes physiques de l’érosion de cavitation dans un fluide compressible à l’échelle fondamentale de l’implosion d’une bulle de cavitation. Suite à l’implosion d’une bulle de vapeur à proximité d’une surface solide, des très hautes pressions sont générées. Ces pressions sont considérées responsables de l’endommagement (érosion) des surfaces solides observé dans la plupart des applications. Notre approche numérique démarre avec le développement d’un solveur compressible capable de résoudre les bulles de cavitation au sein du code volumes finis YALES2 en utilisant un simple modèle de mélange homogène des phases fluides. Le solveur est étendu à une approche ALE (Arbitraire Lagrangien Eulérien) dans le but de mener des simulations d’interaction fluide-structure sur un maillage mobile. La réponse du matériau solide est calculée avec le code de calcul éléments finis Cast3M, et nous a permis de mener des simulation avec un couplage d’abord monodirectionnel, ensuite bidirectionnel, entre le fluide et le solide. On compare des résultats obtenus à deux dimensions, puis à trois, avec des observations expérimentales. On discute les chargements de pression estimés, et les réponses de différents matériaux pour des implosions de bulle à des différentes distances de la surface. Enfin, à travers l’utilisation de simulations avec couplage bidirectionnel entre fluide et solide, on identifie l’amortissement des chargements de pression pour les différents matériaux.

This research is devoted to understanding the physical mechanism of cavitation erosion in compressible liquid flows on the fundamental scale of cavitation bubble collapse. As a consequence of collapsing bubbles near solid wall, high pressure impact loads are generated. These pressure loads are believed to be responsible for the erosive damages on solid surface observed in most applications. Our numerical approach begins with the development of a compressible solver capable of resolving the cavitation bubbles in the finite-volume solver YALES2 employing a simplified homogenous mixture model. The solver is extended to Arbitrary Lagrangian-Eulerian formulation to perform fluid structure interaction simulation with moving mesh capabilities. The material response is resolved with the finite element solver Cast3M, which allowed us to perform one-way and two-way coupled simulations between the fluid and solid domains. In the end, we draw comparisons between 2D and 3D vapor bubble collapse dynamics and compare them with experimental observations. The estimated pressure loads on the solid wall and different responses of materials for attached and detached bubble collapses are discussed. Finally, the damping of pressure loads by different materials is identified with two-way coupled fluid-structure interaction.; Ce travail de recherche est dédié à la compréhension des mécanismes physiques de l’érosion de cavitation dans un fluide compressible à l’échelle fondamentale de l’implosion d’une bulle de cavitation. Suite à l’implosion d’une bulle de vapeur à proximité d’une surface solide, des très hautes pressions sont générées. Ces pressions sont considérées responsables de l’endommagement (érosion) des surfaces solides observé dans la plupart des applications. Notre approche numérique démarre avec le développement d’un solveur compressible capable de résoudre les bulles de cavitation au sein du code volumes finis YALES2 en utilisant un simple modèle de mélange homogène des phases fluides. Le solveur est étendu à une approche ALE (Arbitraire Lagrangien Eulérien) dans le but de mener des simulations d’interaction fluide-structure sur un maillage mobile. La réponse du matériau solide est calculée avec le code de calcul éléments finis Cast3M, et nous a permis de mener des simulation avec un couplage d’abord monodirectionnel, ensuite bidirectionnel, entre le fluide et le solide. On compare des résultats obtenus à deux dimensions, puis à trois, avec des observations expérimentales. On discute les chargements de pression estimés, et les réponses de différents matériaux pour des implosions de bulle à des différentes distances de la surface. Enfin, à travers l’utilisation de simulations avec couplage bidirectionnel entre fluide et solide, on identifie l’amortissement des chargements de pression pour les différents matériaux.

The thesis is focused on development of a Smoothed Particle Hydrodynamics (SPH) Fluid-Structure Interaction (FSI) cavitation solver to understand the phenomenon of material deformation under cavitation load better. This summary presents a brief overview of the methodology used to solve a fluid-structure interaction simulation of a bubble collapse over a deformable solid medium. The fluid solver and the solid solver are validated against Rayleigh-Plesset spherical bubble collapse case and FEM solver respectively. The fluid solver is developed using an open source SPH code SPHYSICS_2D, the code is changed from 2D to 2D axisymmetric. The solid SPH solver is developed in-house in 2D axisymmetric, a novel scheme is derived to solve typical issues near symmetry axis in the solid axisymmetric SPH solver. The solid solver has the capability to solve for non-linear isotropic hardening with strain rate effects (commonly known as Johnson-Cook plasticity model).A case each for a detached and an attached cavity is simulated using the FSI solver, the results show that for the same magnitude of pressure wave initiating the collapse and the same size of the bubble, the micro-jet can produce twice the maximum plastic deformation compared to a shock wave, hence a micro-jet dominated impact would exhibit a smaller incubation time compared to the detached cavity. It is also observed that the volume of material that is plastically deformed in case of a micro-jet is miniscule compared to a shock wave impact (almost 800 times smaller). This would imply that even though the incubation time for material erosion might be lower for a micro jet collapse, the shock wave can plastify a much larger volume of material and so the erosion rate should be higher for a shock wave impact. Hence it could be inferred that the material erosion ability of a shock wave is much higher than a micro-jet.An important and novel finding in the present study is the response of the material for a detached cavity where plastic deformation does not occur at the center of collapse but at an offset from the center. The results show that even though the pressure experienced by the material is the highest at the center, it does not produce the maximum plastic deformation. This is for the first time that such a phenomenon is reported in cavitation studies, we find that the phenomenon is linked to inertial effects where the material does not respond to the load as the rate of loading and unloading is extremely high. The effect is linked to the high loading and unloading rate near the center of the collapse due to the flat geometry of the solid medium. The study clearly demonstrate that maximum pressure does not always correspond to the location of maximum plastic deformation or material erosion.Fluid-Structure Interaction simulations for different stand-off ratios, driving pressure and bubble radius have been computed. Results show that for varying stand-off ratio while keeping the bubble radius and driving pressure constant, the attached cavities (SR, La thèse est organisée autour du développement d'un modèle numérique d’interaction fluide - structure pour simuler la déformation induite dans un matériau solide lors de l’implosion de bulles de cavitation. Le solveur est développé à partir du code open source SPHYSICS_2D utilisant la méthode des Smoothed Particles Hydrodynamics (SPH). Dans cette thèse, le code 2D a été modifié pour traiter le cas de fluides en conditions 2D-axisymétrique. Le solveur solide SPH a été complètement développé en interne en 2D-axisymétrique avec un nouveau schéma pour résoudre les problèmes apparaissant à proximité de l'axe de symétrie. Une loi de comportement élasto-visco-plastique de type Johnson Cook est implémentée dans le solveur solide afin de prendre en compte l’effet de la vitesse de déformation sur l’écrouissage du matériau.Les applications du solveur FSI traitent le cas d’une bulle unique implosant au voisinage d’une surface solide. Deux cas sont envisagés : celui d’une bulle détachée de la surface solide pour laquelle l’effondrement génère une onde de choc ; et celui d’une bulle au contact de la surface pour lequel un micro-jet de liquide vient impacter la surface solide. Pour une taille de bulle donnée, les résultats montrent que le micro-jet peut produire deux fois solide pour laquelle l’effondrement génère une onde de choc ; et celui d’une bulle au contact de la surface pour lequel un micro-jet de liquide vient impacter la surface solide. Pour une taille de bulle donnée, les résultats montrent que le micro-jet peut produire deux fois plus de déformation plastique que l'onde de choc, réduisant ainsi le temps d'incubation. Par contre, le volume de matière déformée plastiquement dans le cas du micro-jet (cavité attachée) est 800 fois plus petit que celui déformé par l’impact d'une onde de choc (cavité détachée). Par conséquent, la capacité d'érosion d'une cavité détachée est beaucoup plus élevée que celle d'une cavité attachée. Un important résultat de cette étude concerne les cavités détachées où il est montré que la déformation plastique ne se produit pas au centre de l'effondrement mais à un décalage par rapport à l’axe de symétrie. Les résultats montrent également que même si la pression subie par le matériau est la plus élevée au niveau de l’axe de symétrie, la déformation plastique ne sera pas maximale à cet endroit mais dans une zone éloignée du centre. Nous Une étude paramétrique est menée pour quantifier les effets de la distance bulle/paroi, de la pression d’effondrement et du rayon de la bulle. Les résultats montrent que l'énergie totale absorbée et le taux d'érosion devraient être plus élevés pour une cavité détachée que pour une cavité attachée. La densité d'énergie absorbée (d'où le temps d'incubation) et l'énergie totale absorbée (d'où le taux d'érosion) augmentent avec la pression d’effondrement. Le changement du rayon de la bulle tout en gardant les autres paramètres constants n'affecte pas beaucoup l'amplitude de la déformation plastique ni la densité d'énergie absorbée, ce qui suggère que quelle que soit la taille de la bulle de cavitation, le temps d'incubation devrait rester similaire. Cependant, comme le volume de la zone déformée plastiquement varie presque linéairement avec la taille de la bulle, l'énergie totale absorbée ou le taux d'érosion augmente significativement avec la taille de la bulle.Dans le passé, les études similaires n'ont jamais pris en compte la sensibilité à la vitesse de déformation dans le modèle de plasticité. Nos simulations montrent que l'ampleur de la déformation plastique est alors surestimée d'environ 60% pour les cavités détachées présentées dans ce document et d'environ 200% pour les cavités attachées. Nous montrons ainsi que de telles études réductrices fondées sur des modèles de plasticité insensibles à la vitesse de déformation conduisent à une sous-estimation du temps d'incubation et à une surestimation du taux d'érosion.

The thesis is focused on development of a Smoothed Particle Hydrodynamics (SPH) Fluid-Structure Interaction (FSI) cavitation solver to understand the phenomenon of material deformation under cavitation load better. This summary presents a brief overview of the methodology used to solve a fluid-structure interaction simulation of a bubble collapse over a deformable solid medium. The fluid solver and the solid solver are validated against Rayleigh-Plesset spherical bubble collapse case and FEM solver respectively. The fluid solver is developed using an open source SPH code SPHYSICS_2D, the code is changed from 2D to 2D axisymmetric. The solid SPH solver is developed in-house in 2D axisymmetric, a novel scheme is derived to solve typical issues near symmetry axis in the solid axisymmetric SPH solver. The solid solver has the capability to solve for non-linear isotropic hardening with strain rate effects (commonly known as Johnson-Cook plasticity model).A case each for a detached and an attached cavity is simulated using the FSI solver, the results show that for the same magnitude of pressure wave initiating the collapse and the same size of the bubble, the micro-jet can produce twice the maximum plastic deformation compared to a shock wave, hence a micro-jet dominated impact would exhibit a smaller incubation time compared to the detached cavity. It is also observed that the volume of material that is plastically deformed in case of a micro-jet is miniscule compared to a shock wave impact (almost 800 times smaller). This would imply that even though the incubation time for material erosion might be lower for a micro jet collapse, the shock wave can plastify a much larger volume of material and so the erosion rate should be higher for a shock wave impact. Hence it could be inferred that the material erosion ability of a shock wave is much higher than a micro-jet.An important and novel finding in the present study is the response of the material for a detached cavity where plastic deformation does not occur at the center of collapse but at an offset from the center. The results show that even though the pressure experienced by the material is the highest at the center, it does not produce the maximum plastic deformation. This is for the first time that such a phenomenon is reported in cavitation studies, we find that the phenomenon is linked to inertial effects where the material does not respond to the load as the rate of loading and unloading is extremely high. The effect is linked to the high loading and unloading rate near the center of the collapse due to the flat geometry of the solid medium. The study clearly demonstrate that maximum pressure does not always correspond to the location of maximum plastic deformation or material erosion.Fluid-Structure Interaction simulations for different stand-off ratios, driving pressure and bubble radius have been computed. Results show that for varying stand-off ratio while keeping the bubble radius and driving pressure constant, the attached cavities (SR

His numerical work focuses on the collapse of a single air bubble close to adeformable wall generated by the impact of an incident high-pressure wave. A CFD code was fully coupled to a FEM solid code in order to compute the bubble collapse and the subsequent plastic deformation in the material. The CFD code was developed from the numerical model of Johnsen et al. (University of Michigan). A mobile mesh capability was added in order to account for the displacement of the fluid-structure interface. An ALE (Arbitrary LagrangianEulerian) method was implemented to switch from an Eulerian description in a fixed mesh to an Eulerian description in a moving mesh. The solid response to bubble collapse was computed with the FEM software CAST3M (developed by CEA) assuming an elastic-plastic constitutive law for the material. The communication between the two codes is achieved through the MPI library.For the CFD code, bubble collapse dynamics was validated by comparison with two others codes: research software CaviFlow and commercial software Ansys Fluent. The coupling between the CFD and the FEM code was validated on the case of the impact of a wave on an elastic medium.The paper will present a detailed analysis in 2D of both the dynamics of bubble collapse and solid behaviour for various conditions. They were obtained by changing the amplitude of the incident shock wave, the standoff distance and the material properties. Special attention will be given to the shock wave that forms when the microjet hits the bubble interface and to the impact of this shock wave on the material surface. In particular, the damping of the impactpressure with respect to a perfectly rigid wall was computed as well as the plastic deformation of the material surface. These final data gave us crucial information about the requirement of taking account of fluid-structure interaction in numerical modeling of cavitation erosion.; Lorsqu’un liquide est soumis à de fortes dépressions, des bulles de gaz etde vapeur peuvent apparaître en son sein. Ce phénomène est appelé cavitation Ces bulles sont susceptibles d’imploser dans les zones de surpression etd’endommager les parois solides environnantes à travers la création d’écoulements violents. Dans les applications industrielles, ce phénomène d’érosionpar cavitation peut réduire la durée de vie des machines hydrauliques. Dans ledomaine biomédical, le phénomène d’érosion par onde de choc est utilisé pourl’ablation de tissus mous ou la destruction de calculs rénaux.Dans ce travail, à caractère numérique, nous nous intéressons à une bulled’air isolée implosant au voisinage d’une paroi déformable sous l’effet d’unesurpression générée par une onde de choc incidente. Le travail de thèse aconsisté à mettre en place un couplage fluide structure pour simuler le phénomène d’implosion de la bulle dans la partie fluide et la déformation plastiqueengendrée dans la partie solide. Pour cela, une stratégie en deux étapes a étémise en oeuvre. La première étape a consisté à construire un code fluide avecmaillage mobile pour gérer le déplacement de l’interface fluide/solide. Le codedéveloppé est dérivé du modèle numérique développé par Eric Johnsen et al.à University of Michigan dans lequel nous avons implémenté une méthodeALE (Arbitrary Lagrangian Eulerian) permettant de passer d’une descriptioneulérienne sur maillage fixe à une description eulérienne sur maillage mobile.La seconde étape a consisté à modéliser la déformation élasto-plastique de lapartie solide au cours de l’implosion de la bulle grâce à un couplage fort ducode fluide avec le logiciel CAST3M. La communication entre les deux codesutilise la bibliothèque MPI.A l’issue du couplage, nous avons été en mesure de simuler l’implosiond’une bulle d’air à proximité d’une paroi déformable. Nous avons pu calculerle niveau de déformation de la paroi solide ainsi que l’amortissement de la sur-pression en paroi pour plusieurs matériaux. L’outil numérique ainsi développépermettra à terme d’étudier le comportement de revêtements amortissants capables de protéger des structures industrielles de l’érosion de cavitation oud’optimiser les effets bénéfiques des traitements par onde de choc dans lesapplications biomédicales.