Nonlocal plasticity models for localized failure

In quasi-brittle materials such as concrete, failure is preceded by highly localized deformation patterns. Localized failure is caused by strain softening, which cannot be described by classical local models, because they lack a length scale. Mathematically, the use of such models leads to ill-posed boundary value problems. If properly formulated, nonlocal models can provide a sound description of strain softening and model the complete failure process of quasi-brittle materials. A bifurcation analysis is performed for a number of nonlocal plasticity models of the gradient and integral type, and the load-displacement response up to complete failure is computed. The nonlocal models are constructed mainly as enhancements of the standard smallstrain, rate-independent, isotropic local plasticity model. Most nonlocal models respond in a physically appropriate manner at initial bifurcation, but for some models, the plastic zone widens at later stages of the loading process, and stresses remain at a spuriously high level. For the integral format of two nonlocal models selected for further development, the VermeerBrinkgreve model and the ductile damage model, the bifurcation analysis is extended to hardening and the two-dimensional, plane strain setting. For both models, the enhancement is based on a modified hardening law, which depends in a specific way on the local hardening variable and its nonlocal counterpart. The analysis determines the range of material parameters for which the solution is regular and provides an analytical expression for the initial distribution of the rate of plastic strain. The use of the Vermeer-Brinkgreve model is restricted to softening, while the ductile damage model can be used for hardening and softening. An efficient and robust iterative numerical algorithm for the stress return of integral-type nonlocal models that use the local and nonlocal hardening variable in the yield function is developed and implemented. Due to weighted averaging, the stress return is a coupled nonlinear complementarity problem. It is solved by a Jacobi-like iterative technique, derived in a rigorous manner for a special case and generalized based on a physical interpretation of the resulting equations. In view of the numerical modeling of concrete, the nonlocal enhancement is applied to plasticity models with pressure-dependent yield functions and non-associated flow rules. Simulations of laboratory tests demonstrate that the nonlocal model regularizes localization due to softening and non-associated plastic flow. Neither the load-displacement response nor the propagation of the process zone depend in a spurious manner on the computational grid. For eccentric compression of a prismatic column under plane strain conditions, the trend for the size effect on structural strength and post-peak response is correctly predicted.

Analyse de l'état de contrainte et de l'endommagement d'assemblages sertis en matériau mixte métal-composite sollicités en traction ou en flexion

Alain Michel Prenleloup

The constant improvement of polymer composite materials has allowed their integration in many industrial fields where their mechanical and chemical properties are apparently attractive. Mixed metal-composite insulators are now increasingly used for high voltage networks instead of conventional porcelain structures because of their high strength and reduced maintenance requirements due to their strong resistance against chemical degradation and dynamic loading. The composite insulators investigated in this study are made of an epoxy rod reinforced with glass fibres, two steel fittings being strongly crimped to both ends of the rod to transfer loads between the suspended electrical components and the fixation points. Due to the plastic deformation of the steel induced by the crimping process, a purely mechanical link results from the residual radial stresses at the interface between the composite rod and the metallic fitting. The strength of the joints plays an important role during the tensile and bending loading of the assemblies (clamping of the high-voltage line on the electrical tower and guiding of the electrical transformation substation). Experimental and numerical studies on the crimped joints are performed simultaneously in order to determine the strength of the structures and to characterise the damage and fracture mechanisms. The traction and bending tests are achieved to highlight the failure modes appearing for very high loads. In parallel, the numerical analysis provides complementary information on the stress state and the sensitivity of some influent parameters. The proposed dual approach allows the identification of the critical loads corresponding to the failure initiation and the possible mechanisms associated. The results for the tensile tests performed on structures subjected to a crimping load between 290 and 460 kN exhibit a nonlinear increase from 110 to 160 kN of the maximum pull-out load, and a transition of the failure mode of the link from a sliding of the rod out of the end-fitting for weakly crimped insulators and a delamination of the composite for strongly crimped ones. For still higher crimping conditions, the maximum tensile load decreases due to the damage introduced before tensile loading. In parallel to the experimental study, the numerical model developed for the crimping and traction steps provides the stress state in the joint during the solicitations. The comparison between the numerical predictions and the experimental observations highlights the strength limit of the bond by delamination of the outer layers of the composite. The use in the numerical model of a tensor damage factor depending on the stress state in the composite provides additional information on the failure initiation resulting from the stress combination and allows the characterisation of the critical crimping and loading conditions. The analysis of the crimped joints under tensile loading is extended to the study of insulators subjected to bending loading. Experimental tests are performed for two different sizes of insulators showing different final failure modes. For 51mm-thick insulators, a collapse of the compressive part of the rod occurs while for the 63mm-thick assemblies a shear crack propagation on the mid-plane of the rod can be observed. In addition to the fracture behaviour, the tests show a progressive embrittlement of the joint due to the damage onset appearing at 60 and 50% of the maximum mechanical loading of 15.7 and 21 kN found for the 51mm- and 63mm-thick rods respectively. Moreover, the numerical simulations contribute to the understanding of the failure zones and allow the identification of the critical loads and the stress state associated with the damage observed. Owing to the numerical and experimental analysis, structural modifications are proposed for improving the current bearing capacities of the crimped structures in order to shift the damage initiation and to increase the maximal strength of these assemblies under industrial use. In the tensile loading case, an optimal contact area between the rod and the end-fitting ensures an increase of 50% of the maximum mechanical load for the 18.7mm-thick insulator. For the bending case, the damage initiation at loading rates which are much lower than the bearing capacity can be postponed by changing the conical form of the metal end-fitting. However, the maximum bending load cannot be improved since it is directly limited by the diameter of the rod.

EPFL2008

Finite element modeling and experimental study of brittle fracture in tempered martensitic steels for thermonuclear fusion applications

Pablo Federico Mueller

In this work we have studied brittle fracture in high-chromium reduced activation tempered martensitic steels foreseen as structural materials for thermonuclear fusion reactors. Developing the adequate materials that can withstand the severe irradiation conditions of the burning plasma in a future fusion reactor is one of the major challenges to be solved in order to make profit from the great advantages of thermonuclear fusion as an energy source. High-chromium tempered martensitic steels such as F82H and the most advanced version Eurofer97 are among the main candidate materials for structural applications in future fusion power plants due to low irradiation-induced swelling, good mechanical and thermal properties, and reasonably fast radioactive decay. The most concerning drawback of these kind of steels is irradiation embrittlement, which is manifested by a ductile-to-brittle transition temperature shift to higher temperatures after irradiation whose amplitude depends on the irradiation conditions (temperature, neutron flux, neutron fluence, etc). The aim of this work was to study and model brittle fracture in the ductile-brittle transition region of this kind of steels in the as-received unirradiated conditions. It is necessary to be able to transfer laboratory specimen fracture data to real components and structures in order to assess the performance of these steels in the different operating and transient conditions they could find during the operation life of a fusion reactor. In order to do so, the specimen geometry effects and specimen size effects on measured fracture toughness need to be properly understood, taken into account and predicted with an appropriate model. In particular, specimen size effect on measured toughness is a major concern for the nuclear materials research community owing to the limited irradiation volume in current and planed materials irradiation facilities. The main results of this PhD work are summarized below. The microstructure of Eurofer97 and F82H has been characterized and compared by means of optical microscopy, scanning electron microscopy, transmission electron microscopy and energy-dispersive X-ray spectroscopy in order to identify microstructural features that could play a role in the measured fracture toughness. Both steels have similar but slightly different chemical composition and final heat-treatments but the prior austenitic grain size measured in F82H is approximately 8 times larger than in Eurofer97. It was shown that the alloying element Tantalum, added to stabilize the austenite grain size, played a different role in both steels. After a careful analysis of the particles present in both steels, it was found that Tantalum in Eurofer97 formed carbides of an average size around 100 nanometers. In contrast in F82H it did not form small carbides but formed big oxide inclusions with a size up to 30 µm. These large particles do not effectively pin the grain boundaries. The different behavior of Tantalum in these steels is believed to be mainly a consequence of the larger content of Oxygen present and the smaller amount of Aluminum in F82H compared to Eurofer97. The Master-Curve ASTM-E1921 standard is a method initially developed to determine the ductile-to-brittle transition reference temperature, T0, in fission reactor ferritic steels from a small number of experiments. In this work the applicability of the Master-Curve method to reduced activation tempered martensitic steels such as Eurofer97 and F82H was studied in detail. Fracture tests with pre-cracked sub-sized compact tension specimens (three different sizes, 0.18T, 0.35T and 0.87T of Eurofer97) were carried out in the temperature range [-196 °C, -40 °C]. The toughness-temperature behavior and scatter were shown to deviate from the ASTM-E1921 standard predictions near the lower shelf. Using the method of maximum likelihood, the athermal component of the Master-Curve was calculated to better fit our fracture toughness data from the lower to the middle transition region. We showed that these Master-Curve adjustments are necessary to make the T0 values obtained near the lower shelf with 0.35T size C(T) specimens consistent with those obtained in the middle transition region with 0.87T C(T) specimens. The ASTM-E1921 specimen size limitations, setting the maximum toughness measurable with a given specimen size, were found to be too lenient for this kind of steels. This problem was especially evident in fracture toughness data of Eurofer97 and F82H obtained in the upper transition temperature range with two different specimen sizes. Thus a more stringent specimen size requirement was proposed to avoid inconsistent transition temperature determinations. A promising local fracture model with the potential of predicting cleavage fracture toughness was studied. Finite element simulations were undertaken for compact specimens, notched specimens and tensile specimens of Eurofer97 steel tested from 20 °C down to -197 °C. Three and two dimensional as well as axisymmetric simulations were run in order to calculate the stress and strain fields at the onset of brittle fracture. For each tested temperature, the calculated load-displacement curves were found to reproduce very well those of the experiments. A local approach fracture criterion was studied. This criterion states that when the maximum principal stress is larger than a critical stress within a critical volume ahead of the crack tip, notch or neck, cleavage is triggered leading to macroscopic fracture of the specimen. It was shown that this model is able to predict the minimum fracture load of the notched specimens with the same values of critical stress and critical volume that were calibrated to predict the lower bound of fracture in compact specimens. This local approach model was also successfully used to predict the strong size effect observed experimentally in pre-cracked compact tension specimens in the upper transition region. The critical fracture stress determined in the standard tensile specimens was found higher than that of the fracture specimens. It was suggested that this difference stems from a significant difference in the stress state between the different specimens (triaxiality). Finally, the comparison between the overall fracture behavior in the transition between F82H and Eurofer97 steels indicated that these two materials are quite similar. A difference in the reference temperature T0 of about 20 °C was found, with a nominal value of -100 °C and -80 °C for F82H and Eurofer97 respectively. However, some excessive scatter in the toughness data was found in F82H with more data points than expected lying below the 1% tolerance bound; this was not observed in Eurofer97. From an engineering point of view, the entire fracture databases of these two steels are encompassed practically by the same lower tolerance bound.

EPFL2009

Deformation mechanism in nanocrystalline FCC metals studied by atomistic simulations

Christian Brandl

Polycrystalline materials with crystallite diameters below hundred nanometer exhibit extraordinary strength which goes along with a decrease in ductility. In order to tailor tough materials, which combine strength and ductility, the underlying deformation mechanisms have to be understood. Molecular dynamics suggested partial dislocation mediated plasticity in nanocrystalline face-centred cubic (FCC) metals in terms of 1) dislocation nucleation at the grain boundaries, 2) dislocation propagation through the nano-grain, and 3) eventual dislocation absorption in the neighbouring grain boundary network. The partial dislocation nucleation sites were seen to exhibit stress concentrations such as in triple junctions. Moreover, it has been shown that the partial dislocation nucleation is energetically related to the so-called general stacking fault energies (GSFEs). Transmission electron microscopy, in-situ X-ray diffraction, and deformation experiments can be interpreted in terms of the aforementioned dislocation processes. On basis of this knowledge this thesis utilizes atomistic simulation methods to extend the understanding of the fundamental deformation mechanisms which mediate the plasticity in bulk nanocrystalline FCC metals. In the first part of this thesis, ab-initio simulations are exploited to investigate the effect of the superimposed strains in Ni, Al, and Cu on the GSFEs, which describe the energy penalty for rigidly shearing of the crystal. The energy curves are discussed in terms of a) the second and third order elastic constants b) the energy difference between hexagonal closed packed and face centred cubic structures, c) the GSFE / unstable-twinning-fault-energy interdependence and c) the implication on dislocation nucleation. The second part explores the strain rate, temperature, applied stress, and grain size dependence of nanocrystalline Al utilizing classical molecular dynamics simulations. The deformation behaviour is examined on the basis of the underlying deformation mechanism. The higher the strain rate the more the onset of dislocation mediated plasticity is delayed. This can manifest itself into a stress overshoot that diminishes as the strain rate is lowered. The associated critical resolved shear stresses are seen to cause the strain rate sensitivity in the flow stress. In conjunction with the rate dependence of the critical resolved shear stresses, the temperature dependence of the flow stresses supports the picture of thermally activated dislocation processes which control the deformation of nanocrystalline Al in molecular dynamics. Moreover, the strain rate simulations evidence that a subset of dislocation processes operate at the athermal limit which indicates, once more, the difficulty to extrapolate from the high strain rate regime of molecular dynamics towards common experimental conditions. Furthermore, using an iterative strain/relaxation method, athermal critical stresses are calculated for an isolated nano-grain with a nucleated dislocation therein. Dislocation cross-slip is seen to enable a dislocation to by-pass stress concentration, which can hinder the dislocation propagation through the grain, and allows the dislocation to be deposited in preferred grain regions as e.g. triple junctions. This rationalized the observed lack of strain hardening in the molecular dynamics study. Additionally dislocation cross-slip is more pronounced at lower strain rates which indicate the growing importance of the grain boundary structure for dislocation motion as the stresses are lowered. Slip transfer through a high angle grain boundary is observed and is understood in terms of the grain boundary structure, which exhibits regions of preferred dislocation transmission sites. Such structural features are discussed by means of the implication on mesoscopic models of dislocation transmission. Moreover, the grain size study shows discrete stress-drops, which are associated with dislocation mediated processes, in the so-called "inverse" Hall-Petch regime. The deformation behaviour of nanocrystalline Al in molecular dynamics is rationalized in terms of thermally activated processes, and therein the implications of the grain size dependencies are qualitatively discussed.

EPFL2010