[en] Graphite shows an increase in strength with an increase in temperature, and this characteristic makes graphite a prime candidate for structural applications in the high temperature gas-cooled reactors (HTGR). The core support structure in the HTGR uses high strength graphite. Its ultimate strength under a combined stress state is important for design use. A test program was therefore initiated to determine the multiaxial behavior of two specific types of graphites, in particular the biaxial failure surface which is to be correlated with various failure theories. The two candidate graphites are a fine-grain 2020 (of Stackpole Carbon Company) and a coarse-grain PGX (of Union Carbide Corp). Tubular specimens are used and are subjected axial loading and internal or external pressure. Various ratios of biaxial stresses are examined but emphasis is placed in the first and fourth stress quadrants. The tensile fracture pattern reveals a cleavage failure mode. Compressive failure occurs when a limiting value (equivalent to cohesion and internal friction) on any given plane is reached. The maximum stresses at the time of failure are calculated using the thick-wall cylinder formula. A modified Coulomb-Mohr theory with tension-cutoffs proposed by B. Paul (1961) fits the biaxial test data very well. The theory is consistent with the observed fracture patterns and mechanism. The theory appears to give a better correlation with the biaxial test data than the maximum normal stress, maximum shear stress, maximum distortion energy and Tsai and Wu's strength tensor theories. The failure surfaces obtained are for thin tubular specimen configuration in which the stress state is relatively uniform. The volume and grain size effects are nearly the same. In the application of the failure surfaces to the core support component design, these factors, non-uniform stress state, the volume and grain size effects should be included. (orig.)

[en] The paper describes the numerical analysis of creep problems of orthotropic shells of revolution under asymmetrical loads with application to a cylindrical shell. In this paper the authors study the creep problem of rotationally axisymmetrical shells made of a homogeneous and orthotropic material under asymmetrical load. The equations of equilibrium and the relations between the strains and the displacements are derived from the Sanders theory for thin shells. In the theory of creep it is assumed that in a given increment of time the total strain increments are composed of an elastic part and a part due to creep. The elastic strains are proportional to the stress by Hooke's law for orthotropic materials. The constitutive equations in the creep range are based on the orthotropic creep theory derived from the orthotropic plastic theory by Hill, and the effective creep strain is related to the effective stress by Norton-Vailey equation. The basic differential equations for the incremental values with respect to time are numerically solved by a finite difference method and the solutions at any time are obtained by summation of the incremental values. Resultant forces and resultant moments are given from numerical integration by Simpson's 1/3 rule. As a numerical example the creep deformation of simply supported cylindrical shell subject to asymmetrical, locally distributed load is treated. The numerical computations have been carried out for four cases of anisotropy. The effect of anisotropy has been studied. It is observed that the stress resultant and the displacement distributions are significantly affected by the anisotropy of the material. The analysis of shells in steady creep becomes a subcase of the more general transient creep process described here. (orig./HP)

[en] Progressive deformation due to ratchetting may lead to the failure of structure in two different manners: by local fracture of the material when the strain is greater than an admissible limit or by a global destabilization of the loaded structure. This second mecanism will be considered here and corresponds to the principal interaction between buckling and progressive deformation. Under additional loading over a given dead load P, a structure initially in equilibrium will have an associated displacement u(t). Its failure by instability may occur if this displacement is great enough so that the response attains an instable branch of the load - displacement curve. Under cyclic additional loading of thermal or mechanical nature, this situation happens if there is ratchetting. The progressive plastic deformation which increases after each cycle, will produce greater and greater displacement and then perhaps failure by instability before local fracture by excessive local strain. In this paper, we present a theoretical and numerical analysis of this phenomena in function of the material characteristics and of the amplitude of the loads. (orig./RW)

[en] The paper considers elastic-perfectly plastic discrete structures subjected to dynamic excitations, whose exact time sequence is supposed to be unknown. A general bounding principle is formulated, which provides bounds on several kinds of dynamic plastic deformations, such as plastic strains, residual displacements and plastic work produced within a single or more structural elements. (orig.)

[en] This paper deals with a method of implementation of strain-rate dependence effect of elastic-plastic materials into a finite element code. The strain-rate dependence of yield surface is analyzed based on the viscoplasticity theory, in which two kinds of yield surfaces are used. Among the two surfaces, one is quasi-static yield surface, which expands its radius due to classical plasticity theory, and the other one is dynamic yield surface, which is the expansion or shrinkage of the corresponding quasi-static yield surface. The amount of the change from the quasi-static yield surface depends on the magnitude of the strain-rate at that current time. Strain-rate is calculated from the velocity at the node. This method enables us to deal with the phenomenon of the shrinkage of yield surface due to the sudden drop of strain-rate within the framework of plasticity theory. In this analysis, the strain-rate history effect is neglected. The analysis is also confined currently within the range of small strain. This method is implemented into the implicit and multipurpose finite element code ADINA. Two examples of calculated results are given, one is an impact of a bar against the rigid wall, and the other is an expansion of a pipe due to the explosive pressure. (orig.)

[en] Many constitutive models for frictional materials such as concrete and soils are so complex that the number of parameters is too large to allow the models to be used in general purpose stress analysis codes. Alternately, if simplified models are used, then one or more of the essential features such as shear enhanced compaction, dilatation, strain hardening and softening, and cracking cannot be included. For engineering applications, these two aspects must be balanced. A formulation based on the theory of plasticity has been developed which requires only the first and third invariants of stress and inelastic strain. The flow surface evolves to a limit surface as the path length of the inelastic strain increases. Here the assumption is made that this path length increases less rapidly if the mean pressure increases. The result is that considerably more inelastic deformation is displayed for those paths associated with large values of mean pressure. Comparisons of predictions with experimental data for a weak concrete support the assumption. (orig.)

[en] This paper discusses the present status of the simplified method which was proposed a few years ago by J. ZARKA and J. CASIER to obtain a straightforward evaluation of the total inelastic strains which accumulate in a structure under cyclic loadings when a linear kinematic hardening behaviour is assumed for the materials involved. An iterative procedure has been developped within the past year to handle plastic shakedown under non radial anisothermal cyclic loadings. The theoretical basis of this iterative procedure is presented. An application example is then given which highlights the efficiency of the method. (orig.)

[en] Inelastic analysis may be required to evaluate the behavior of perforated plates used at elevated temperature such as steam generator tube sheets of fast breeder reactors. We have presented in previous papers a method of inelastic analysis utilizing the concept of the equivalent solid plate, where the inelastic material constants of the equivalent solid plate are determined on the basis of the inelastic behavior of the perforated plate subjected to in-plane loadings. The applicability of the method to the case of out-of-plane loading was still to be examined. In order to investigate into this problem, we have performed experiments to measure the elastic-plastic-creep responses of perforated plates when subjected to four point bending and compared them with the calculated results utilizing the concept of the equivalent solid plate. (orig./RW)

[en] Contact between two bodies often results in very high local stress concentrations, which causes local plastic deformation and permanent deformations after unloading. This permanent deformation can be at high importance for the functionality of a construction, e g static indentation in bearing races. In this paper a method for analysis for elastoplastic bodies in contact with the finite element method will be presented. The procedure is based on the general finite element system ASKA, which permits access to all available facilities, e g a highly improved substructure technique. This permits that only the interesting areas will be handled as elastoplastic, which reduces the computer time and the amount of data on mass storage. The analysis is separated into two different parts. (1) Calculation of the yield load. To obtain the magnitude of applied external forces and the stress state at incipient yield for the following elastoplastic analysis an iterative procedure is needed due to the nonlinearity introduced from the contact conditions. (2) Elastoplastic analysis. This part of the analysis is performed in a combined incremental and iterative algorithm calculating the load increments corresponding with the next discrete change of the contact surface automatically. The paper includes presentation of results from yield load and elastoplastic analysis of a cylinder in contact with an infinite half plane and a smooth circular disc in an infinite plate with hole. (orig.)

[en] Applications of the boundary element method to plate bending problems are quite sparse. The usual approach has been to treat same specific problems or types of problems. However, many practically important problems have been left without consideration. Plates on elastic foundations is one of these practically important areas and, moreover, it seems to be especially suitable to be treated with boundary element method. The analysis presented in this paper is based on the Kirchhoff plate bending theory and a fundamental singular solution is a displacement field caused by a unit lateral load acting at a point of an infinite plate resting on a linearly elastic foundation, which can be either of Winkler or Pasternak types or an elastic half space or, more generally, a foundation for which an axisymmetric external stress will result in an axisymmetric state of deformation. The derived integral equations base on the so-called direct formulation of the boundary element method. The two solution equations are formulated in terms of displacement rotation, moment and resultant boundary shear in every boundary nodal point. From these four unknown variables in every boundary node at least two has to be prescribed. The considered boundary conditions are free, simply supported and rotationally restrained edges. The developed computer code was designed to include the singular solutions for plates on Winkler, Pasternak and elastic half space foundations. These solutions were plotted and compared with each other and with available solutions in literature. The developed method has the advantage that the treatment of plates with finite dimensions and varying boundary conditions will be greatly facilitated. The analytic solution for these types of problems are rare and, moreover, the evaluating of numerical results from these solutions is a cumbersome task. (orig.)