Sensitivity-matrix-based Identification of Anisotropic Elastic Constants and Microstructure Features in Plasma Sprayed Coatings Utilizing Ultrasonic Back-Reflection

Luoming Sun,Li Lin,Yijia Chen,Zhiyuan Ma,Xiaopeng Zhu,Mingkai Lei

DOI: https://doi.org/10.1016/j.mtla.2023.101897

2023-01-01

Materialia

Abstract:With the unique lamellar microstructure, macroscopic elasticity of plasma sprayed coatings shows either positive or inverse anisotropy due to complex microstructure of inter-layer pores and intra-layer cracks. It is extremely difficult to nondestructively evaluate anisotropic elastic constants of these coatings with thin thickness. Moreover, it is an absolute challenge to reveal the elastic anisotropy transitions between positive and inverse. It is also difficult to clarify the nonlinear relationship between macroscopic elasticity and microstructure, which is necessary for microstructural identification. In present work, a sensitivity-matrix-based method was developed to determine five independent elastic constants of plasma sprayed coating by ultrasonic back-reflection method, with the relative errors of inverted elastic constants less than 1.33 %. Then, an optimized physical-descriptorbased microstructure characterization and reconstruction method was proposed to model "pore and crack" coatings with different microstructural features, and relative errors of microstructural features between the reconstructed and preset values were less than 4.35 %. Based on above, the dependences between six microstructural features and elastic constants were established. The inverse elastic anisotropy was explained systematically as the increase of vertical crack content, the "roundness" of horizontal pore and its "weakening" preferred orientation. Finally, the sensitivity-matrix-based method was used to simultaneously identify the porosity, crack content, pore average aspect ratio, pore orientation factor, and crack orientation factor of plasma sprayed Al2O3 coating based on elastic constants inverted. The relative errors in these identifications were less than 14.85 %. This work provides a new sensitivity-matrix-based solving strategy for multi-parameter integrated quantitative identification of complex microstructural features in materials.