Modeling Effects of Material Properties and Three-Dimensional Surface Roughness on Thermal Barrier Coatings

Abstract

Thermal barrier coatings (TBCs) are known to spall as a result of the residual stresses that develop during thermal cycling. TBC's are multi-layered coatings comprised of a metallic bond coat, thermally grown oxide and the ceramic top coat, all on top of a Ni-base superalloy substrate. The development of residual stresses is related to the generation of thermal, elastic and plastic strains in each of the layers. The focus of the current study is the development of a finite element analysis (FEA) that will model the development of residual stresses in these layers. Both interfacial roughness and material parameters (e.g., modulus of elasticity, coefficient of thermal expansion and stress relaxation of the bond coat) play a significant role in the development of residual stresses. The FEA developed in this work incorporates both of these effects and will be used to study the consequence of interface roughness, as measured in SEM micrographs, and material properties, that are being measured in a parallel project, on the development of these stresses. In this paper, the effect of an idealized three-dimensional surface roughness is compared to residual stresses resulting from a grooved surface formed by revolving a sinusoidal wave about an axis of symmetry. It is shown that cylindrical and flat button models give similar results, while the 3-D model results in stresses that are significantly larger than the stresses predicted in 2-D.

Considerations for the incorporation of measured surfaces in finite element models

“…This approach was used by Glynn et al . (2001) and Tang et al (2006). However, visual inspection of the surfaces presented by Glynn et al (2001) and Caminha et al (2004) reveals that surface layers may have variable thicknesses and that the interface may have substantially a different topography than the surface.…”

“…A brief overview of the hemispherical asperity assumption in early analytical models of thermal contact resistance can be found in Madhusudana (1996). Other asperity shapes in the literature include half cylinders (Heichal and Chandra 2005; Komvopoulos and Gong 2007), cones (Kumar and Ramamurthi 2001), squares (Salti and Laraqi 1999), saw tooth patterns (Prasher 2001), and sinusoids (Glynn et al 2001; Poon and Sayles 1994). In many cases, symmetry or similarity arguments have been used to extend the behavior of a pair of interacting geometric asperities (or a single asperity and an ideal surface) to describe a pair of interacting surfaces covered in asperities (Komvopoulos and Gong 2007; Madhusudana 1996; Marotta et al 2001; Shai and Santo 1982).…”

“…It is becoming increasingly common to see FE models that use probabilistic surface data. Glynn et al (2001) performed 2D and 3D FE analyses on four‐layer thermal barrier coatings using probabilistic sinusoidal surface roughness and symmetry arguments. Hyun et al (2004) and Pei et al (2005) used probabilistic self‐affine fractal surfaces in a custom FE model to establish “general relationship between contact properties and key material parameters.” Tang et al (2006) performed a FE analysis of a hot strip rolling using an oxide‐coated steel sheet with a probabilistic rough surface.…”

A comparison of methods to evaluate the behavior of finite element models with rough surfaces

“…The analyst may also knowingly include assumptions and simplifications that increase the model error in an attempt to reduce the model size, complexity and solution time. These assumptions include idealized geometry (Glynn et al , 2001; Komvopoulos and Gong, 2007; Kong et al , 2009), reduced model order using symmetry arguments (Walter et al , 2007; Walter and Mitterer, 2009), plane strain assumptions (Tang et al , 2006), reduced model size (Walter and Mitterer, 2009) and reduced model resolution (Thompson, 2007; Groning et al , 2009). Although these assumptions are often clearly described in the literature, the quality of those assumptions, and thus the quality of the models themselves and the results that they produce, is often unclear.…”

Damage evolution of a thermal barrier coating system with 3-dimensional periodic interface roughness: Effects of roughness depth, substrate creep strength and pre-oxidation

“…Furthermore, the effect of interface roughness on the stress state was studied numerically by finite element method (FEM) simulations [8,[17][18][19][20][21][22][23][24][25][26]. For model development a periodic TBC-substrate interface roughness was assumed in these studies to reduce calculation complexity.…”