13th International Conference on Fracture June 16–21, 2013, Beijing, China -2- ZrO2 with an addition of 6 à 8% Y2O3. . The purpose of the bond coat is to improve the adhesion and to prevent the oxidation by itself forming a dense oxide with as good mechanical properties as possible.. In the typical gas-turbine application, such components will be loaded by the thermomechanical start/stop cycle, i.e., a start, followed by a full-load high-temperature period of from a couple of hours up to several days, finally followed by a shut-down. The design life requirements between inspections for a land-based gas turbine can typically be 3 000 cycles and 30 000 full-load high-temperature hours, after which the TBC must still be functional. This document deals with air-plasma-sprayed (APS) TBC, which is the type most commonly used in stationary gas turbines. 2. TBC fatigue life Thermomechanical fatigue (TMF) aspects. Experience shows that with time (and, consequently, accumulated load cycles) TBCs are prone to failing by spalling (flaking). The most dangerous spalling mechanisms have been shown to appear at the end of the shutdown after a long high-temperature full-load period. At the end of the high-temperature full-load period, the stress state in the interface-near region will be low due to high-temperature creep, and as the top coat (ceramic) has a much lower thermal expansion than the metallic layers below, the top coat will be in strong compression at the end of the shut-down (at ‘room temperature’). This leads to the three main spalling mechanisms illustrated in Fig. 1. Of the spalling mechanisms shown in the figure, most research has been concentrated on the explanation and analysis of the spalling from flat surfaces. TBC life models: history and present status. Early research (see, for instance, [1]) showed that the geometry of the interface (which has a pattern of repeated ridges and valleys caused by the plasma-spray process) leads to tensile stress normal to the interface during the stop cycle of the gas turbine, so that the initiation and growth of cracks in the interface is made possible. Similar analyses on an improved 3-D FE model in [2] confirmed this mechanism. This insight led us to the set-up of a fracture-mechanicallly based life model 3. Fracture-mechanical model of a TBC Figure 2. Measured damage evolution in APS TBC Figure 3. Idealised interface/crack geometry for the FM model
RkJQdWJsaXNoZXIy MjM0NDE=