An improved CFD/experimental combined methodology for the calibration of empirical erosion models

A significant source of uncertainty in CFD-based erosion prediction is represented by the erosion model, used to turn the particle-wall impingement characteristics into an estimate of the removal of material. In engineering computations, frequent use is made of empirical erosion models obtained by fitting the experimental results of air-solid jet impingement tests. However, since the applicability conditions of these models are often unknown or unavailable to the users of CFD codes, they are frequently misapplied, producing not only uncertain, but at times also highly inaccurate wear estimates. As part of his PhD thesis defended at the University of Tulsa in 2016, Amir Mansouri proposed a methodology to calibrate the coefficients of an empirical erosion correlation by combining experimental data and CFD results of a slurry jet impingement test. This methodology provided an effective way to overcome the criticisms of using empirical erosion models taken from the literature without facing the complexity and the high computational burden of a multiscale, physically-based approach, in which the micro-scale mechanical interactions between the particles and the target surface are resolved according to the fundamental laws of damage mechanics. This paper presents an improved formulation of Mansouri's methodology, which not only increases the calibration accuracy of the empirical erosion equation, but also enhances the reliability of the CFD-based erosion prediction model as a whole. To this purpose, numerical simulations were supported by in-house slurry jet impingement experiments on an aluminum sample and curved Glass Reinforced Epoxy samples