Advancements of Particle-Surface Interaction Studies through Novel Measurement Technique Development and Engineering Modelling
Files
TR Number
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Turbomachinery operating on aircraft are often exposed to dusty or sandy environments during typical service. Engines on commercial and military flights operating in desert regions such as the Middle East or even Phoenix, Arizona can become severely damaged by ingesting dirt, grit, sand, and dust. Due to the high speeds, pressures, and temperatures, ingested particles can inflict erosion upon the blades, stators, and other components within the operating turbomachinery. Left unchecked, this erosion can lead to an increase in surge and stall probability while also contributing to higher service frequency and maintenance cost. Historically, particle-induced erosion is thoroughly documented and has been studied extensively; however, the underlying physics that govern the particle-surface interactions present in turbomachinery have remained elusive. The work described in this dissertation aims to describe a novel experimental technique used to measure and quantify particle-surface interactions characteristic of those present in turbomachinery. Specifically, the technique captures fully time-resolved trajectories of microparticles rebounding off a flat surface. It has been developed to measure the coefficient of restitution for particles of various material composition and shape incident on various surface materials at differing speeds and angles of incidence. The coefficient of restitution is a kinetic energy conservation metric that characterizes the amount of kinetic energy lost by particle during impact with a static surface and can be related to erosion extent and erosion mode. Many key findings were made during the experimental campaign that focused on particle bounce. It is shown that measuring fully-time resolved trajectories of bouncing particles leads to the highest quality coefficient of restitution data. Specifically, obtaining fully-time resolved trajectories allows for the stochasticity present in particle bounce to be measured and for the uncertainty in the coefficient of restitution to be fully characterized. It is shown that particle shape is not only the key driver that contributes to the stochasticity present in particle rebound, but also an important factor for determining the amount of plastic deformation that occurs on the flat surface. These findings are underscored in a novel coefficient of restitution model that accounts for the jagged particle shape present on typical particles and the plastic deformation of the surface material. This novel model also provides an analytical prediction of some of the stochasticity, or spread, present in coefficient of restitution measurements caused by particle shape. The modeled particle bounce and surface deformation is compared with experimental results. It is demonstrated that the new model accurately captures the slope of normal coefficient of restitution vs. normal velocity while surface deformation measurements can be used as an auxiliary validation for particle bounce models. In addition to measuring the coefficient of restitution for particle bounce, a novel measurement technique has also been developed to directly measure particle breakage. Along with the breakage probability of a particle, both the number and speed of the fragments for each broken breakage are measured. As expected, the breakage probability generally scales with normal velocity. It is shown that the average rebounding angle distribution for broken fragments is identical to that of bouncing particles for identical impact conditions. Moreover, average fragment velocities were shown to be about the same as that of bouncing particles. Finally, it is demonstrated that automated breakage detection allows for a significantly higher number of breakage events to be measured. This allows for the accuracy of the breakage probability measurement to be directly estimated with an uncertainty estimate. Raw results from the experimental study along with the novel coefficient of restitution model can be used to develop models for erosion in turbomachinery. Specifically, the coefficient of restitution is typically implemented in computational fluid dynamics (CFD) simulations to predict particle paths and induced erosion in turbomachinery. Currently, CFD simulation results do not agree with real-world erosion findings. This implies that the underlying physics governing erosion are not fully understood. Higher accuracy models, such as the one developed in this dissertation, coupled with empirical data can be leveraged to increase the accuracy of CFD simulations to predict erosion. In the long term, if erosion can be predicted, new engine designs can be developed that will be erosion resistant. These engines may feature new geometry to aid in expelling particles from an engine along with different materials that may be more erosion resistant.