Statistical approaches to study variations in thermal spray coatings

Thermal spray materials have a layered structure consisting of lamellae, pores, oxides, cracks, and unmelted particles which result in a material of anisotropic and heterogeneous character. The variability of these materials exists not only in different directions but throughout the coating thickness. The present work encompasses several experimental and theoretical expects to assess the variability in material properties of thermal spray materials.;Microhardness tests with different loads were used to assess how variability was related to the "test volume" of the material. A general trend was that the larger the load, the smaller the average microhardness and the larger the Weibull modulus. The microhardness distributions within the coatings also vary after heat treatments. High variability was found at high temperature.;The reliability of test results was assessed by a Monte Carlo method, where subsets of data were randomly selected from the parent data sets and analyzed statistically. The distribution of the mean fluctuates around the mean of the parent data sets. The coefficient of variation for the mean is ;Four point bend tests with in situ acoustic emission monitoring were performed on as-sprayed and heat-treated specimens in both cross section and in-plane orientations. The AE responses were analyzed with respect to time and displacement. For as-sprayed samples tested in the cross section orientation, microcracking before failure was observed. However, catastrophic failure was observed when tested in the in-plane orientation. Microcracking before failure was commonly observed after heat treatment.;The energy and amplitude distributions change after heat treatment but no significant difference can be noticed in different orientations. Macrocracks, characterized by their high energy and high amplitude, tend to occur when the event activity is low, dominate the cracking mechanism, and induce final failure. For as-sprayed and heat-treated samples, the AE responses differ not only in the percentage of macrocracks (these increase after heat treatment) but where the macrocracks occur during the tests (macrocracks can only be observed after a normalized displacement of 0.82).