A Probabilistic Method for Predicting Particle-Blade Impact Kinematics in Turbomachinery

Abstract The trajectories of dust particles ingested into gas turbine engines are dictated by their inertial properties, which influence the likelihood of collision with solid surfaces along the gas path, including compressor blades and casing. If a particle impacts a surface, it will either bounce, stick, or shatter. The fate is dependent on the kinematics of the interaction, which vary widely across the blade span and chord, and across the range of particle size and composition that a gas engine may typically ingest. This makes the prediction of compressor fouling or erosion using traditional deterministic methods computationally expensive. In this contribution, a new reduced-order model is proposed that can rapidly predict the probability of particle interaction with the blade and casing, and the probability of particle escape from the rotor outlet as a function of a dimensionless parameter, the generalized centrifugal Stokes number. Numerical simulations of particle-laden flow through a single-stage axial compressor rotor, the NASA Rotor 37, are performed to generate a dataset of particle-surface interaction kinematics for three minerals, each with a wide size distribution (0.3–135 μm), at three different blade rotation speeds. We also introduce a new dimensionless parameter, the particle kinetic energy ratio, to enable a non-dimensional particle impact kinematic analysis. Our results show that particle impact kinetic energy may be estimated on the basis of Stokes number alone.