Publications

2000

• Simulation expérimentale du risque
  
  • Luong Minh-Phong
  , 2000, pp.77-106. Cet Atelier a pour but de présenter quelques méthodologies capables de fournir des informations réalistes en vue d’une évaluation quantitative de la vulnérabilité des ouvrages de génie civil face aux risques de dysfonctionnement, de dommage ou de destruction. Cette synthèse résume les diverses contributions et réflexions du groupe de travail en insistant plus particulièrement sur la problématique du risque décrit, l’approche expérimentale proposée, les résultats obtenus sur des exemples précis et les perspectives.
• Simulation numérique d'une clarinette par projection modale
  
  • Facchinetti Matteo Luca
  • Boutillon Xavier
  • Constantinescu Andrei
  , 2000.
• Exploiting partial or complete geometrical symmetry in boundary integral equation formulations of elastodynamic problems
  
  • Bonnet Marc
  , 2000. Procedures based on group representation theory, allowing the exploitation of geometrical symmetry in symmetric Galerkin BEM formulations of 3D elastodynamic problems, are developed. They are applicable for both commutative and noncommutative finite symmetry groups and to partial geometrical symmetry, where the boundary has two disconnected components, one of which is symmetric. (10.1007/978-3-642-56288-4_18)
  DOI : 10.1007/978-3-642-56288-4_18
• Intelligent'' optimal design of materials and structures: what does it mean?
  
  • Zarka Joseph
  , 2000 (DETC2000/DAC-14492). The engineers have to face very important problems in the design, the test, the survey and the maintenance of their structures. These problems did not yet get full answer even from the best people in the world. Usually in these problems (such as no satisfactory constitutive modeling of materials, no real control of the accuracy of the numerical simulations, no real definition of the initial state and/or the effective loading of the structure), there is no solution and the experts do not understand the problem in its whole. Moreover, the available data may be not statistically representative (i.e. are in limited number), fuzzy, qualitative and missing in part. We propose a practical solution the « Intelligent Optimal Design of Materials and Structures » where the actual best knowledges of the researchers/experts are intelligently mixed to the results of experiments or real returns. Several examples of applications are given in this serial set· of papers to explain the real meaning and power of this approach.
• “Intelligent” Optimal Design of Woven Composites
  
  • Zarka Joseph
  • Doux Thierry
  , 2000, pp.275-277. By the coupling of the existing knowledge of experts, the numerical results and the experimental data with the help of some special automatic learning and optimization techniques, it is now possible to propose a very simple tool to design materials. The purpose of this paper is to show its particular application to the very difficult design of woven composites where it is necessary to adjust the mechanical properties, the electromagnetic properties and taking care of the price too. (10.1115/DETC2000/DAC-14494)
  DOI : 10.1115/DETC2000/DAC-14494
• Intelligent modeling of woven materials
  
  • Zarka Joseph
  • Doux Thierry
  , 2000. This paper shows the particular application of the general approach "Intelligent Optimal Design" to the design of a woven composite material used o make radomes and for which the mechanical and electromagnetical properties have to satisfy special properties. Although the design of such a material needs the control of almost 50 design parameters, with a very limited number of examples of materials in the data base, but with a special description of each material, it is possible to extract the rules necessary to predict the behaviour of any new material with a rather good accuracy and thereafter, to reach the optimal design (minimum cost) associated to any new requirements. (10.2514/6.2000-4942)
  DOI : 10.2514/6.2000-4942
• Méthode de Schwarz additive avec solveur grossier pour problèmes non symétriques
  
  • Alart Pierre
  • Barboteu Mikaël
  • Le Tallec Patrick
  • Vidrascu Marina
  Comptes Rendus de l'Académie des Sciences - Series I - Mathematics, Elsevier, 2000, 331 (5), pp.399-404. Dans cette Note, nous présentons la généralisation à la résolution de systèmes non symétriques d'une méthode de décomposition de domaines extensible de type Neumann–Neumann. Cette généralisation a pu être effectuée en utilisant le cadre théorique des méthodes de décomposition en sous-espaces de Schwarz additifs avec introduction d'un espace grossier correctement adapté au cas non symétrique. La méthode développée est testée sur des problèmes non symétriques engendrés par la formulation mixte de problèmes de contact avec frottement. (10.1016/S0764-4442(00)01667-0)
  DOI : 10.1016/S0764-4442(00)01667-0
• Microextensométrie locale par corrélation d'images numériques. Application aux études micromécaniques par microscopie électronique à balayage.
  
  • Doumalin Pascal
  , 2000. Ce travail a vise dans un premier temps au developpement d'une technique de microextensometrie locale par l'adaptation des methodes de correlation d'images numeriques, utilisees depuis plusieurs annees en imagerie optique macroscopique pour determiner des champs de deplacements locaux dans des structures, aux images numeriques obtenues, a echelle bien plus fine, par microscopie electronique a balayage (meb). L'etude a permis de reveler les difficultes specifiques rencontrees pour atteindre la precision theorique de ces techniques avec des images de meb et a conduit a proposer un certain nombre de procedures pour les traiter. On s'est en particulier attache a analyser les defauts d'images induits par les imperfections de balayage du meb, par de tres faibles variations de grandissement ou par la derive du faisceau, a etudier les consequences des reglages non optimaux du meb, conduisant a des niveaux de bruit trop eleves, ou encore l'effet de forme du repere pour marquer les echantillons sur la precision de mesure. Cette technique de microextensometrie a ensuite ete appliquee a l'investigation des mecanismes locaux de deformation dans des materiaux heterogenes, notamment au cours du tout premier stade de l'ecoulement plastique. On s'est interesse plus particulierement a la mise en evidence, la caracterisation, la quantification et enfin la modelisation du phenomene de localisation precoce de la deformation sous forme de bandes dans des materiaux elastoplastiques. Un moyen original pour quantifier statistiquement la portee des bandes a ete propose en appliquant sur le champ de deformation lui-meme les methodes statistiques classiquement utilisees pour caracteriser la distribution des phases.
• Une utilisation de Castem 2000
  
  • Constantinescu Andrei
  , 2000.
• An affine formulation for the prediction of the effective properties of nonlinear composites and polycrystals
  
  • Masson Renaud
  • Bornert Michel
  • Suquet Pierre
  • Zaoui André
  Journal of the Mechanics and Physics of Solids, Elsevier, 2000, 48 (6-7), pp.1203-1227. Variational approaches for nonlinear elasticity show that Hill’s incremental formulation for the prediction of the overall behaviour of heterogeneous materials yields estimates which are too stiff and may even violate rigorous bounds. This paper aims at proposing an alternative ‘affine’ formulation, based on a linear thermoelastic comparison medium, which could yield softer estimates. It is first described for nonlinear elasticity and specified by making use of Hashin–Shtrikman estimates for the linear comparison composite; the associated affine self-consistent predictions are satisfactorily compared with incremental and tangent ones for power-law creeping polycrystals. Comparison is then made with the second-order procedure and some limitations of the affine method are pointed out; explicit comparisons between different procedures are performed for isotropic, two-phase materials. Finally, the affine formulation is extended to history-dependent behaviours; application to the self-consistent modelling of the elastoplastic behaviour of polycrystals shows that it offers an improved alternative to Hill’s incremental formulation. (10.1016/S0022-5096(99)00071-X)
  DOI : 10.1016/S0022-5096(99)00071-X
• Quantifying Drilling Efficiency and Disruption: Field Data vs. Theoretical Model
  
  • Putot Claude
  • Guesnon Jean
  • Perreau Philippe
  • Constantinescu Andrei
  SPE Drilling and Completion, Society of Petroleum Engineers, 2000, 15 (02), pp.118-125. Mechanical and kinematics data at the bit have been collected during drilling operations and organized to identify optimal drilling efficiency. Transitional conditions towards abnormal behavior have been distinguished. Normal drilling is usually characterized by regular linear relationships between global control variables at the bit, which approximately correspond to an upscaling of a local Mohr-Coulomb-type failure mechanism of the rock. Most of our experimental data confirm this classical disposition. However, considerable departure from these relationships is noted, especially when there is bottomhole assembly (BHA) resonance or when there is partial balling at the bit. The field tests presented here allow us to verify classical theory and to draw up new guidelines for behavior in nonoptimal or dangerous, evolving operating conditions, such as those that occur under poor cleaning conditions. The main features of a mechanical structuring model involving transition towards low-efficiency conditions are presented, and it is shown how this model confirms our interpretation of field data. The model couples the drillability of rock and the general dynamics of the bit in an original and synthetic way. It makes it possible to understand and test the sensitivity of bit response to variables such as "weight on hook," torque on the drillstring, and flow rate of the fluid. (10.2118/64004-PA)
  DOI : 10.2118/64004-PA
• Représentations non structurées en optimisation topologique de formes par algorithmes évolutionnaires
  
  • Hamda Hatem
  • Jouve François
  • Lutton Evelyne
  • Schoenauer Marc
  • Sebag Michèle
  , 2000, Canum 2000 : Actes du 32e Congrès national d'analyse numérique, pp.153-179. Les algorithmes évolutionnaires sont des méthodes d'optimisation stochastiques inspirées -- grossièrement -- de l'évolution naturelle des populations. Méthodes globales d'ordre 0, leur robustesse et leur souplesse leur permettent d'attaquer la résolution numérique de problèmes difficiles à résoudre autrement. Mais c'est leur capacité à travailler sur des espaces de recherche non standards qui leur offre les perspectives les plus originales. Dans le domaine de l'optimisation topologique de formes, les résultats obtenus il y a quelques années montraient la faisabilité de l'approche évolutionnaire, mais étaient limités par le fait que la complexité de l'espace de recherche était liée à celle du maillage utilisé lors de la simulation numérique. Ce papier introduit un ensemble de représentations compactes et non structurées dont la complexité n'est pas fixe, mais est ajustée par l'algorithme lui-même. Les résultats présentés dans cet article montrent que leur utilisation permet de repousser les limites de l'optimisation topologique de formes évolutionnaire. Des résultats sur des problèmes test simples tentent ensuite de comparer les mérites des diverses approches proposées.
• On the simulation of large viscoplastic structures under anisothermal cyclic loadings
  
  • Verger Laetitia
  • Constantinescu Andrei
  • Charkaluk Eric
  , 2001, pp.341-350. The optimal design of parts submitted to thermomechanical loadings is a key issue for the safety and quality assessments of structures. In this context, the present paper discusses the choice of a behavior model and the determination of its parameters. For cast iron and aluminum alloy applications, two constitutive laws are compared : one is based on a classical unified viscoplastic model and the other is based on a two-layer plastic-viscous rheological model. The following discussion shows that the use of anisothermal experiments is very important for the identification of material parameters and that simple models can be successfully applied for the predictive lifetime assessments of large 3D structures. (10.1007/978-94-015-9628-2_33)
  DOI : 10.1007/978-94-015-9628-2_33
• Une transformation du problème d'élasticité linéaire en vue d'application au problème de l'inclusion et aux fonctions de Green
  
  • Pouya Ahmad
  Comptes Rendus de l'Académie des Sciences - Series III - Sciences de la Vie, Elsevier, 2000, t 328 (Série II b), pp.437-443. Nous présentons une transformation simple du problème de structure élastique linéaire. Le problème transformé se présente sous forme d'un nouveau problème de structure élastique linéaire avec un comportement, une géométrie et des données de forces et de déplacements imposés différents. Le problème transformé peut être plus simple à traiter ou correspondre à des cas de solutions analytiques connues. A l'aide de cette transformation, on peut ramener le problème de l'inclusion ellipsoïdale en un problème d'inclusion sphérique, étendre les résultats connus pour le tenseur d'Eshelby pour les cas de matrice isotrope ou d'isotropie transverse, à des cas plus larges, et enfin, établir l'expression analytique de la fonction de Green du milieu infini pour des cas de comportements orthotropes plus riches que l'isotropie transverse, ainsi que des cas non orthotropes.
• Numerical lifetime assessment of engine parts submitted to thermomechanical fatigue
  
  • Lederer Guy
  • Charkaluk Eric
  • Verger Laetitia
  • Constantinescu Andrei
  , 2000. In this work a numerical method for the design of components submitted to severe cyclic thermomechanical loading is developed. This tool is based on a Finite Element (FE) analysis. In a first part the temperature distribution is obtained and used in the second part for the mechanical computation. The analyses use the description of the geometry of the part, a precise information of the thermal properties, an appropriate behavior of the material at low and high temperature and a good assessment of the boundary conditions (heat transfer coefficients, contact, …). This method is applied to asses the low cycle fatigue design of a diesel turbo-charged exhaust manifold in cast iron. These calculations, failure location and lifetime estimation, obtained on two versions of this component, are compared to experimental data. The results show a good agreement in terms of critical zones location and of lifetime comparison in both versions and permit thus to classify the versions. (10.4271/2000-01-0789)
  DOI : 10.4271/2000-01-0789
• On discontinuous stresses at the interface of a bimaterial thermo-elasto-plastic body
  
  • Bui Huy Duong
  • Taheri Said
  • Dang Van Ky
  Advances in Natural Sciences, 2000, 1, pp.79-84.
• The dynamics of discrete mechanical systems with perfect unilateral constraints
  
  • Ballard Patrick
  Archive for Rational Mechanics and Analysis, Springer Verlag, 2000, 154 (3), pp.199-274. The dynamics of discrete mechanical systems with perfect unilateral constraints is formulated in a very general setting. The well-posedness of the resulting evolution problem is studied. It is proved that existence and uniqueness of a maximal solution is ensured provided strong assumptions are made on the regularity of the data: they are supposed to be analytic. Simple examples show that this regularity assumption may not be relaxed. Sufficient conditions to ensure that the maximal solution is defined for all time are supplied. The continuous dependence of the solution on initial conditions is also studied and the numerical computation of the solution is discussed. (10.1007/s002050000105)
  DOI : 10.1007/s002050000105
• An energetic approach in thermomechanical fatigue for silicium molybden cast iron
  
  • Charkaluk E.
  • Constantinescu Andreï
  Materials at High Temperatures, Taylor and Francis, 2000, 17 (3), pp.373-380. The purpose of this paper is to define a low cycle fatigue criterion in order to predict the failure of engineering structures. The major problem in defining a predictive fatigue criterion is that it should be applicable for structures submitted to complex multiaxial thermo-mechanical loadings but should be identifiable from simple experiments on specimens. After a short critical review of the principal criteria used in low cycle fatigue it will be shown that the dissipated energy per cycle permits a correlation of isothermal and anisothermal results obtained on silicon molybdenum cast iron in the case of specimens and also on structures. (10.1179/mht.2000.17.3.001)
  DOI : 10.1179/mht.2000.17.3.001
• Plan/plan sliding contact modelling by the boundary element method
  
  • Serre Ingrid
  • Bonnet Marc
  • Pradeilles-Duval Rachel Marie
  , 2000.
• Hétérogénéité des contraintes au sein d'un matériau lamellaire. Application à un alliage intermétallique TiAl
  
  • Gélébart Lionel
  • Crépin Jérôme
  • Bretheau Thierry
  • Bornert Michel
  , 2000.
• A plasticity model for discontinua
  
  • Leroy Yves
  • Sassi William
  , 2000, pp.77-107. This article is concerned with the development and application of a simple continuum theory for rocks that may contain both randomly as well as preferentially oriented plane discontinuity surfaces. The theory stipulates that displacement discontinuities are independently activated on these surfaces as soon as an appropriate yield criterion is fulfilled; these displacement jumps account for the irreversible, ‘plastic’ part of the bulk deformation. In stress space, the critical conditions for the activation of discontinuous slip or opening displacements define an overall yield envelope that could be initially anisotropic, reflecting for example a weakness of certain orientations due to pre-existing joint sets. For the yield conditions studied in this paper, essentially a Coulomb-type friction law and a simple fracture opening condition, the inferred stress-strain response under typical triaxial loading conditions reveals the sensitivity of the two discontinuous deformation modes to the confining pressure. The incipient growth of a geological fold in such a material is modelled as a problem of plate bending. The slip- and opening-modes of deformation are found to be activated typically in the fold intrados and extrados, respectively. Under certain conditions, both modes will be activated simultaneously at the same locality and contribute to the total deformation. Field observations on a well exposed sandstone anticline are reported here, which support this conclusion. The present plasticity model for discontinua can clearly be explored in more detail for realistic distributions of faults and joints taken from field observations. It could also be improved in various ways in its description of the underlying deformation mechanisms. Apart from its interest as a mechanical constitutive model, it can also serve as a point of departure for studies of stress-sensitive, anisotropic permeability distributions in fractured formations. (10.1007/978-3-642-59617-9_5)
  DOI : 10.1007/978-3-642-59617-9_5
• An affine formulation for the prediction of the effective properties of nonlinear composites and polycrystals
  
  • Masson Renaud
  • Bornert Michel
  • Suquet Pierre
  • Zaoui André
  , 2000. Variational approaches for nonlinear elasticity show that Hill's incremental formulation for the prediction of the overall behaviour of heterogeneous materials yields estimates which are too stiff and may even violate rigorous bounds. This paper aims at proposing an alternative `affine' formulation, based on a linear thermoelastic comparison medium, which could yield softer estimates. It is first described for nonlinear elasticity and specified by making use of Hashin±Shtrikman estimates for the linear comparison composite; the associated affine self-consistent predictions are satisfactorily compared with incremental and tangent ones for power-law creeping polycrystals. Comparison is then made with the second-order procedure and some limitations of the affine method are pointed out; explicit comparisons between different procedures are performed for isotropic, two-phase materials. Finally, the affine formulation is extended to history-dependent behaviours; application to the self-consistent modelling of the elastoplastic behaviour of polycrystals shows that it others an improved alternative to Hill's incremental formulation. (10.1016/S0022-5096(99)00071-X)
  DOI : 10.1016/S0022-5096(99)00071-X
• Dynamic testing at high rates of strain
  
  • Gary G.
  , 2000.
• La programmation logique inductive à la lumière de la transition de phase
  
  • Giordana Attilio
  • Saitta Lorenza
  • Sebag Michèle
  • Botta Marco
  , 2000, pp.157-172. En logique du premier ordre, le test de couverture correspond à un problème de satisfaction de contraintes. Or, dans le domaine de la satisfaction de contraintes, les problèmes les plus difficiles à résoudre en moyenne se concentrent dans une étroite région, appelée transition de phase, qui marque la transition entre les problèmes presque surement satisfiables, et les problèmes presque surement insatisfiables. Pour étudier la complexité et la faisabilité de la Programmation Logique Inductive (PLI), nous avons généré quelques centaines de problèmes d’apprentissage, situés dans et hors de la transition de phase; ces problèmes ont été soumis à FOIL, SMART+ et G-Net. Ces expérimentations systématiques établissent deux résultats. Tout d’abord, la transition de phase constitue un attracteur pour l’apprentissage, dans le sens où toutes les hypothèses produites appartiennent à cette région. En second lieu, une "zone aveugle de l’apprentissage" apparaît: pour tout problème situé dans cette zone, les systèmes de PLI considérés échouent tous à apprendre quelque hypothèse pertinente que ce soit. Ces résultats sont interprétés, et la discussion conduit à remettre en cause les biais usuels de la PLI.
• Self-consistent estimates for nonlinear viscoelastic polycrystals: a simplified resolution of the affine formulation adapted to monotonic loading paths
  
  • Brenner Rénald
  • Castelnau Olivier
  • Masson Renaud
  • Bacroix Brigitte
  • Zaoui André
  , 2000, 177-180, pp.153-158. A simplified resolution of the self-consistent model for non-linear viscoelastic polycrystals is presented. The formulation is based on an affine linearization of the constitutive relations as recently proposed by Masson and Zaoui [4], which adapts Hill's conception of self-consistent modelling to nonlinear viscoelastic behaviours. The approximation of the inverse Laplace-Carson transform introduced here, which is believed to be well adapted to monotonic loading paths, makes the resolution of the whole self-consistent problem possible in the direct time space. This results in a largely simplified numerical algorithm. Comparisons with the predictions of a model treating accurately the inverse Laplace-Carson transform are presented, and very good results are obtained for creep of FCC and HCP polycrystals. (10.4028/www.scientific.net/KEM.177-180.153)
  DOI : 10.4028/www.scientific.net/KEM.177-180.153