Probabilistic Lifetime Prediction of Thermal Barrier Coating Systems Depending on Manufacturing Scatter

Abstract

The assessment of Bondcoat/Thermal barrier coating systems is an inherent part of the lifing process of gas turbine component. On the one hand, coatings are considered in the constitutive modelling — e.g. in the thermal model and for the prediction of eigenfrequencies of gas turbine blades. On the other hand, the influence of the coating system on the lifetime of the part (target cyclic life and target operation hours) needs to be assessed. This paper addresses the prediction of coating lifetime. Lifing models of Bondcoat/Thermal barrier coating systems (BC/TBC) are commonly built using isothermal furnace cyclic tests (FCT). The lifetime of the BC/TBC under such test conditions has been shown to depend on multiple coating parameters like TBC thickness, TBC porosity, BC thickness, BC roughness, and also on testing temperature. For example, the TBC life (defined as time to partial TBC spallation) is reduced with increasing temperature, with increasing TBC thickness and decreasing porosity and BC roughness. When operating in a gas turbine (GT), the TBC surface temperature and the BC temperature depend on engine operating conditions, heat transfer of combustion gas and cooling air, coating microstructure and thickness. For instance, a TBC with high porosity typically demonstrates a lower thermal conductivity than that with low porosity. For otherwise same boundary conditions, the BC temperature will decrease with increasing TBC porosity and increasing TBC thickness. The benefit of having a high coating porosity observed in FCT is further amplified by its impact on reducing the BC temperature in GT operation. To the contrary, the positive impact of a reduced TBC thickness observed in FCT is reduced by its negative impact on an increased BC temperature during GT operation. Taking these effects into account a probabilistic lifing model is proposed based on Monte Carlo simulations. Using this model the impact of the manufacturing scatter on the BC/TBC life can be assessed, and enables improved manufacturing by focusing on those parameters that are most critical for coating lifetime.

Kinetic Monte Carlo simulation of ZrO2 coating deposited by EB‐PVD

“…In contrast to homogeneous materials, the conductivity of the coating is not only a material property but is also a function of the microstructural parameters such as porosity and microcracks 58,59 . Obviously, TBCs with high porosity typically demonstrate a lower thermal conductivity than that with low porosity 60 …”

“…58,59 Obviously, TBCs with high porosity typically demonstrate a lower thermal conductivity than that with low porosity. 60 In addition, the high porosity in TBCs can also relax the thermal stress around crack tip. Therefore, we believe that lower porosity (higher preparation temperature and kinetic energy, lower deposition rate) will lead to lower coating life.…”