Les divers composants d'un réacteur d'avion sont soumis à des chargements de fatigue mécaniques et thermiques fortement variables. Afin de prédire au mieux leurs durées de vie, il est alors nécessaire de prendre en compte l'ensemble de ces variations. Pour ce faire, un modèle représentant de façon incrémentale l'évolution de la plasticité en pointe de fissure a été développé. Celui-ci permet de prédire la vitesse de propagation d'une fissure tout en prenant en compte les effets d'histoire provenant de la plasticité produite lors de surcharges ou de sous-charges. Dans cette étude, ce modèle condensé de plasticité a été exprimé plus simplement à partir du facteur d'intensité des contraintes et l'identification automatisée de ses divers paramètres a été redéveloppée. Dans ce but, un essai de fissuration stable a été conçu pour déterminer simplement le seuil de non-propagation du matériau. Par ailleurs, les éventuelles fissures se propageant dans un composant sont en général amorcées en surface à partir de chocs ou de rayures. Ces fissures passent la majeure partie de leur vie dans un régime de propagation de fissure courte, différant de celui d'une fissure longue. En effet, à facteur d'intensité des contraintes équivalent, une fissure courte se propagera plus rapidement qu'une fissure longue. La prise en compte de ce comportement dans la prédiction de durée de vie est donc primordiale. En outre, il a été observé qu'une fissure longue subissant une contrainte T négative se propage de façon analogue à une fissure courte. Il a donc été choisi de développer un nouveau modèle adapté aux fissures courtes en prenant en compte l'influence de la contrainte T sur la vitesse de propagation. Enfin, à partir des essais de fissuration stable, un protocole a été développé pour réaliser des éprouvettes comportant fissure courte. Une campagne d'essais multiaxiaux pour différentes valeurs de contrainte T a été menée en régime permanent et transitoire afin d'étudier les similitudes entre effet de fissure courte et influence de la contrainte T. / The various components of an aircraft engine undergo strong changes in mechanical and thermal fatigue loadings. All these variations must be taken into account in order to anticipate the components' total fatigue life to the best. An incremental model showing the changes in plasticity at the crack tip has therefore been developed. It allows us to foresee the crack growth rate with history effects by modelling the plastic behaviour produced when overloads or underloads occur. In this study, the plasticity condensed model has been represented in an easier way, using the stress intensity factor and the automated identification of its parameters has been redeveloped. A stable crack growth test has been designed to determine easily the fatigue crack growth threshold of the material. Cracks usually grow from scratches or impacts on the surface of the component. These cracks spend much of their fatigue life growing with a short crack behaviour, differing from the long crack one. As a matter of fact, a short crack grows faster than a long crack for a same stress intensity factor. This behaviour must therefore be taken into account while modelling the crack growth to accurately predict the total fatigue life of the component. It appeared furthermore that a long crack undergoing a negative T-stress grows the same way as a short crack. It has consequently been chosen to develop a short crack growth model using the influence of T-stress on the crack growth rate. Finally, using the stable crack growth specimens, an experimental protocol has been designed to produce new specimens containing a short crack. A multiaxial tests campaign was carried out for various values of T-stress in permanent and transitional regimes in order to compare short crack and T-stress effects.

This thesis investigates fatigue crack growth rate behavior in the threshold and near-threshold regimes for a Ti-6Al-4V (STOA) alloy using two proposed ASTM procedures- (1) load-shedding (LS) using a larger load-shed rate than the current ASTM Standard E647 load-reduction (LR) test procedure, and (2) compression pre-cracking constant-amplitude (CPCA) or load-increasing (CPLI) and load-shedding (CPLS). Tests were conducted at a low stress ratio (R = 0.1) on compact C(T) specimens of two different widths (W = 51 and 76 mm) and threshold fatigue crack growth rates were generated. These test data were compared to previous test data produced from the same batch of material using the current LR and the CPCA test procedure. While no test procedure provided an exact representation of the threshold value (?Kth), the compression pre-cracking (CP) procedures were the most promising. The LR, LS, and CPLS test procedures were influenced by prior loading-history and various crack-closure mechanisms, leading to higher ?Kth values and slower crack growths in the threshold regime. The LS tests (at shed-rates of -0.08,-0.32, and -0.95 mm-1) generated ?Kth values that were 15% to 32% higher than the estimated threshold stress-intensity factor range (?*Kth)R=0.1. The CP test procedures are a more accurate alternative for developing near-threshold and threshold fatigue crack growth rates. The CPLS test procedure produced a ?Kth value that was 10% higher than (?*Kth)R=0.1. LR and LS tests produced different ?Kth values as a function of the specimen width for the given load ratio. The CP test procedures produced consistent crack growth rates over the same range of ?K values examined, independent of the specimen width. Further research is required for developing test procedure(s) capable of providing a more definitive representation of the ?Kth value and closureree fatigue crack growth rates in the threshold regime.

A weight function based modified strip-yield model is adapted to include the mode I effects of roughness induced crack closure. The weight function methodology is verified for arbitrary mixed mode loading conditions under load levels experienced in fatigue crack growth. A rough crack geometry is simulated as a sine wave fitted to a sawtooth defined by an asperity angle and period. Additional mode I crack closure due to mode II sliding of crack faces is calculated and incorporated into the strip-yield model by lengthening and shortening the crack face elements. Simulations of fatigue crack growth using the newly developed model are examined via a design and analysis of computer experiments. The effects of model parameters are identified. Combined roughness and plasticity induced crack closure for long cracks is studied.

The first part of this dissertation employs a three-dimensional elastic-plastic finite element model of straight-through crack growth to correlate four well-known methodologies characterizing fatigue crack closure. The compliance offset and the adjusted compliance ratio (ACR) are experimental methods, whereas the node displacement and the contact stress methods are numerical approaches. Evolutions of crack closure from all four methodologies are compared for a numerical model of a single edge-cracked tension specimen subjected to different levels of constant amplitude cyclic loading. In the second part, a detailed two dimensional stress analysis is conducted for a single pin-joint under plane stress conditions. This study investigates the influence of material nonlinearity, friction, and pre-existing residual stresses from cold-working process on the local radial and hoop stress levels around the pin-loaded hole. Next, the beneficial influence of cold working process is quantified by computing the Mode I stress intensity factors style='mso-bidiont-style:normal'>KI for a single radial crack eanating from a side of a loaded hole. Two different loading configurations are considered: (a) an open hole in tension, (b) a pin-loaded hole. The stress intensity factors are computed using the J integral solutions and the weight functions specific to the crack configuration. The reductions in KI values due to different levels of cold-working process are presented for a range of crack lengths. The final part of the research involves a numerical investigation of an on-line crack compliance technique that is used for experimental measurements of residual stress fields along the crack growth path. A finite rectangular sheet is considered with a single crack emanating from a side of a central hole. The residual stress field is introduced around the hole by cold-working simulation. As part of validation, the normalized residual stress intensity factors computed using the on-line crack compliance technique are compared with those from the style='mso-bidiont-style:normal'>J-integral approach for the case of elastic crack growth. The influence of crack tip plasticity on the performance of the on-line crack compliance technique is studied by comparing the solutions of the elastic and elastic-plastic crack growth models.

Fatigue and crack-growth tests were conducted on 7050-T7451 aluminum alloy under a wide range of loading conditions. Crack-growth tests were conducted on compact, C(T), specimens under constant-amplitude loading, single-spike overloads, and a simulated aircraft spectrum loading. Fatigue tests were also conducted on single-edge-notch bend, SEN(B), specimens under constant-amplitude loading and three aircraft load spectra. The FASTRAN, life-prediction code, was used to make crack-growth predictions on the C(T) specimens; and to make fatigue-life calculations using a 12-micrometer initial flaw size at the center of the edge-notch on the SEN(B) specimens. The predictions agreed fairly well with most of the tests, except the model was unconservative on the single-spike overload tests and the severe spectrum Mini-TWIST+ Level 1 tests. The discrepancy was suspected to be caused by a low constraint factor and/or crack paths meandering around overload plastic zones. A roughness- and plasticity-induced crack-closure model would be needed to improve the model.

No description available.

No description available.

No description available.

No description available.

No description available.