Finite element simulation of stress evolution in thermal barrier coating systems

摘要

Gas turbine materials exposed to extreme high temperature require protective coatings. To design reliable components, a better understanding of the coating failure mechanisms is required. Damage in Thermal Barrier Coating Systems (TBCs) is related to oxidation of the Bond Coat, sintering of the ceramic, thermal mismatch of the material constituents, complex shape of the BC/TGO/TBC interface, redistribution of stresses via creep and plastic deformation and crack resistance. In this work, experimental data of thermo-mechanical properties of CMSX-4, MCrAlY (Bond Coat) and APS-TBC (partially stabilized zirconia), were implemented into an FE-model in order to simulate the stress development at the metal/ceramic interface. The FE model reproduced the specimen geometry used in corresponding experiments. It comprises a periodic unit cell representing a slice of the cylindrical specimen, whereas the periodic length of the unit cell equals an idealized wavelength of the rough metal/ceramic interface. Experimental loading conditions in form of thermal cycling with a dwelltime at high temperature and consideration of continuous oxidation were simulated. By a stepwise consideration of various material properties and processes, a reference model was achieved which most realistically simulated the materials behavior. The influences of systematic parameter variations on the stress development and critical sites with respect to possible crack paths were shown. Additionally, crack initiation and propagation at the peak of asperity at BC/TGO interface was calculated. It can be concluded that a realistic modeling of stress development in TBCs requires at least reliable data of i) BC and TGO plasticity, ii) BC and TBC creep, iii) continuous oxidation including in particular lateral oxidation, and iv) critical energy release rate for interfaces (BC/TGO, TGO/TBC) and for each layer. The main results from the performed parametric studies of material property variations suggest that porosity in the TBC should be increased and sintering decreased, in order to prevent or hinder continuous paths of tensile stresses above the valleys in the TBC. It was shown that variations of creep rates in the BC influence marginaly stress values in TBCs. Therefore neither a positive nor a negative influence on the lifetime can be extrapolated. It was shown that higher creep rates in the TBC layer led to a lower stress level. The extreme variations of thermal expansion coefficient (±50%) help in better understanding of these variations on stress development. The creep of base material only slightly affects stress field development, under pure thermal cycling and can therefore be neglected in this case. As the tensile stresses increase with a relatively high fraction of lateral oxidation not only the out-of-plane oxidation kinetics, but also its lateral component should be low. The modification of amplitude and wavelength of the asperity showed that with increasing roughness a continuous radial tensile path in the TBC and partially in the TGO was formed already after 161 cycles. The variations of wavelength, amplitude and shapes improve the understanding of stress development. The large variety of parametric variations studied by the present work in a highly complex and rather realistic FE model contribute significantly to an enhanced understanding of TBCs. This is supported by the final conclusion, that the set of crucial parameters could be reduced to the time dependent deformation behavior of TBC and TGO, the oxidation kinetics, including lateral oxidation and the shape function of the interface asperity.