Modélisation multiéchelle du phénomène de fretting dans le contact aube-disque
Résumé
Blade–disc fixings in the fan of aeroplane turbo–engines are highly loaded form–closed connections that allow micrometer size relative movement between blade and disc. This loading type is called fretting, the encountered damage mechanisms are wear and cracking. To reduce damage, palliative coatings are applied. To assure the reliability of the engine these coatings need to be reapplied during regular maintenance operations. The prediction of wear and fissuration in the fixings is a challenging task. Better tools for damage prediction will
allow an optimisation of the maintenance intervals and the design of more damage resistant blade–disc fixings.
The goal of this work is the construction of a finite element (FE) model that is suitable for an accurate prediction of fretting damage in blade–disc fixations. With this goal in mind, methods are elaborated to take (i) the change of the friction coefficient, (ii) cyclic plastic
material deformation, (iii) crack initiation, (iv) wear and related contact geometry change and (v) wear–fissuration interaction into account. Integrating some of the models into a finite element analysis of the blade–disc contact, an attempt to predict wear and fatigue in a blade–disc fixation is made. To describe the change of the apparent friction coefficient during the wear of the coatings, analytical and FE models are made. In these the apparent friction coefficient is regarded
to be the mean value of a micro–structurally heterogeneous contact. For the description
of the cyclic plastic deformation of Ti6Al4V a multikinematic von Mises material model and a polycrystal plasticity material model is used. Both models are suitable for the description of ratchetting, additionally the polycrystal plasticity model takes micro–scale plastic deformation, crystallographic texture and mean stress relaxation into account. Crack
initiation is computed using the Dang Van high cycle fatigue criterion. For the use with polycrystal plasticity models the criterion is reassessed. As a wear criterion the dissipated energy approach is used. Contact geometry change by wear is taken into account in FE
computations by iteratively updating the mesh by the computed wear. An attempt to describe the interaction of wear and fatigue is made using polycrystal plasticity and the Dang Van criterion in FE fretting computations with a cylinder–plate contact pair. The computations are compared to corresponding fretting experiments provided by partners from LTDS – Ecole Centrale de Lyon. Finally geometry change by wear, cyclic plastic deformation and fissuration of the blade–disc assembly are calculated in 2D and 3D FE computations. A main result is the circumstance that wear, fatigue and their interaction can be qualitatively described using a polycrystal plasticity material model and the Dang Van
criterion without the use of a classical wear model. The fact that this is not possible with von Mises material models shows that replacing macro–scale phenomenological models by more
physically based micro–scale models increases predictive power. In all FE computations with the blade–disc fixing including geometry change by wear and von Mises plasticity, wear increases fatigue life, cyclic plastic deformation can have the opposite effect. It is shown that a 2D simplification of the blade–disc model yields unrealistic results. This means that 3D
modelling is indispensable. A macro–scale model for the prediction of wear and cracking in the blade–disc fixing has been constructed in this thesis. An integration of the micro–scale models developed in this thesis into the blade–disc fretting model will further increase the predictive power.
allow an optimisation of the maintenance intervals and the design of more damage resistant blade–disc fixings.
The goal of this work is the construction of a finite element (FE) model that is suitable for an accurate prediction of fretting damage in blade–disc fixations. With this goal in mind, methods are elaborated to take (i) the change of the friction coefficient, (ii) cyclic plastic
material deformation, (iii) crack initiation, (iv) wear and related contact geometry change and (v) wear–fissuration interaction into account. Integrating some of the models into a finite element analysis of the blade–disc contact, an attempt to predict wear and fatigue in a blade–disc fixation is made. To describe the change of the apparent friction coefficient during the wear of the coatings, analytical and FE models are made. In these the apparent friction coefficient is regarded
to be the mean value of a micro–structurally heterogeneous contact. For the description
of the cyclic plastic deformation of Ti6Al4V a multikinematic von Mises material model and a polycrystal plasticity material model is used. Both models are suitable for the description of ratchetting, additionally the polycrystal plasticity model takes micro–scale plastic deformation, crystallographic texture and mean stress relaxation into account. Crack
initiation is computed using the Dang Van high cycle fatigue criterion. For the use with polycrystal plasticity models the criterion is reassessed. As a wear criterion the dissipated energy approach is used. Contact geometry change by wear is taken into account in FE
computations by iteratively updating the mesh by the computed wear. An attempt to describe the interaction of wear and fatigue is made using polycrystal plasticity and the Dang Van criterion in FE fretting computations with a cylinder–plate contact pair. The computations are compared to corresponding fretting experiments provided by partners from LTDS – Ecole Centrale de Lyon. Finally geometry change by wear, cyclic plastic deformation and fissuration of the blade–disc assembly are calculated in 2D and 3D FE computations. A main result is the circumstance that wear, fatigue and their interaction can be qualitatively described using a polycrystal plasticity material model and the Dang Van
criterion without the use of a classical wear model. The fact that this is not possible with von Mises material models shows that replacing macro–scale phenomenological models by more
physically based micro–scale models increases predictive power. In all FE computations with the blade–disc fixing including geometry change by wear and von Mises plasticity, wear increases fatigue life, cyclic plastic deformation can have the opposite effect. It is shown that a 2D simplification of the blade–disc model yields unrealistic results. This means that 3D
modelling is indispensable. A macro–scale model for the prediction of wear and cracking in the blade–disc fixing has been constructed in this thesis. An integration of the micro–scale models developed in this thesis into the blade–disc fretting model will further increase the predictive power.
xxx