Plasma-sprayed coatings: Identification of plastic properties using macro-indentation and an inverse Levenberg–Marquardt method

•We identify the plastic properties of porous plasma-sprayed coatings by means of macro-indentation and a Levenberg–Marquardt optimization.•The plastic properties can be represented by the Gurson–Tvergaard plasticity criterion.•The geometry of the residual indent retrieved after indentation suffices to extract plastic properties.•We could identify 5 material parameters that take on realistic values where σy0=1620MPa, K=33,340MPa, q1=3.29, q2=4.60 and f=4.2%. Plasma-sprayed coatings are widely used for thermal protection and wear stability of structural components. These coatings feature an anisotropic porous structure as a result of the thermal spraying process. Although current literature provides methods for the identification of elastic properties of these materials, to our knowledge little research is dedicated to describing their plastic behavior, especially when these coatings are subjected to macro-scale contacts encountered in industrial applications. In this work we present a novel inverse method for the identification of plastic properties of thick plasma-sprayed coatings by means of macro-indentation and finite element simulations coupled to a Levenberg–Marquardt optimization. This optimization aims to fit numerically generated residual indentation profiles to the experimentally obtained ones. For the description of the coatings’ plastic behavior we made use of the Gurson–Tvergaard plasticity criterion coupled to a linear isotropic work hardening of the matrix. This criterion is appropriate for ductile porous solids as it takes into account the hydrostatic pressure, and it is readily implementable in commercial finite element software. The constitutive parameters to be identified include the yield strength σy0 and the work hardening coefficient K of the solid matrix as well as two dimensionless fitting parameters q1 and q2 and the void fraction f. For the given plasma-sprayed coating we could show that the proposed method is capable of identifying these parameters after as little as three iterations where σy0=1624MPa, K=33,340MPa, q1=3.29, q2=4.60 and f=4.2%.