Yann Monerie has defended his HDR on 27th September 2010
Pierre GILORMINI, D.R., CNRS/Mechanics and Materials Engineering Processes Laboratory, referee
Ricardo LEBENSOHN, Senior researcher, Los Alamos National Laboratory (USA), referee
Jean-Jacques MARIGO, Professor, Ecole Polytechnique, referee
André CHRYSOCHOOS, Professor, Université Montpellier II, member
Sylvain LECLERCQ, Senior Scientist HDR, EdF R&D, Moret-sur-Loing, member
Anna PANDOLFI, Associate Professor, Politecnico di Milano (Italy), member
Farhang RADJAÏ, D.R., CNRS/Mechanics and Civil Engineering Laboratory, member
Pierre SUQUET, D.R., CNRS/Mechanics and Acoustics Laboratory, member
This work summarizes about fifteen years of research in the field of micromechanics of materials. Emphasis is placed on the most recent work carried out in the context of nuclear safety. Micromechanics finds a natural place there, aiming to predict the behavior of heterogeneous materials with an evolving microstructure. The applications concerned mainly involve the nuclear fuel and its tubular cladding. The uranium dioxide fuel is modeled, according to the scales under consideration, as a porous ceramic or a granular medium. The strongly irradiated Zircaloy claddings are identified with a composite medium with a metal matrix and a gradient of properties. The analysis of these classes of material is rich in problems of a more fundamental nature. Three main themes are discussed: 1/ Homogenization, 2/ cracking, rupture and fragmentation, 3/ discrete media and fluid-grain couplings.
The analytical scale change methods proposed aim to estimate or limit the linear and equivalent nonlinear behaviors of isotropic porous media and anisotropic composites with a metal matrix. The porous media under consideration are saturated or drained, with a compressible or incompressible matrix, and have one or two scales of spherical or ellipsoid pores, or cracks. The composites studied have a macroscopic anisotropy related to that of the matrix, and to the shape and spatial distribution of the inclusions. Thermoelastic, elastoplastic, and viscoplastic behaviors and ductile damage of these media are examined using different techniques: extensions of classic approaches, linear in particular, variational approaches and approaches using elliptical potentials with thermally activated elementary mechanisms. The models developed are validated on numerical finite element simulations, and their functional relevance is illustrated in comparison to experimental data obtained from the literature. The significant results obtained include a plasticity criterion for Gurson matrix cracked media.
The main aim of this topic is the numerical simulation of multiple cracking of strongly heterogeneous media from their sound state to their fractured state. A method called ‘Non Smooth Fracture Dynamics' is proposed. It is based on a cohesive-volume finite element model and on a non-regular dynamic multibody management (implicit scheme). The main theoretical and practical difficulties of the cohesive-volume method are discussed in detail: non-uniqueness of solutions, instabilities, dependence on the mesh system, local diversity, and experimental identification of the cohesive properties. By combining this method with analytical and numerical homogenization techniques, a two-scale volume and surface approach is developed for the cracking of media with a property gradient: the effect of the spatial distribution of weakening inclusions on the macroscopic fracture criteria and on the tortuosity of crack paths is revealed. An intermediate result of this work is the statistical characterization of the representative elementary volumes in cracking and fracture.
This more recent topic includes the numerical and stochastic analysis of discrete media in the presence or absence of a fluid phase. For the numerical analysis, the non-regular dynamic multi-body method is used. In the case of an interstitial or surrounding fluid, this method is coupled with two other classes of method according to the inertial regime and the size of the system considered: porous medium methods (homogeneous fluid equivalent) or fictitious domain type (direct numerical simulation). These methods are confirmed on fluidization and sedimentation tests. For the analysis, some results are obtained for gravity flows: blocking statistic in silo configuration, compaction effects during undersea avalanches.