Numerical Investigation of Particle Breakup and Ingestion into an Axial Low Pressure Compressor at Engine Icing Operating Points

摘要

Three-dimensional simulations of the Honeywell ALF502 low pressure compressor, using the NASA Glenn codes GlennHT and LEWICE3D have been carried out. This paper is an extension of previous work which did not include a particle breakup model. The simulation includes multiple blade rows separated by a mixing plane boundary condition. Results from a newly implemented particle breakup model are compared and contrasted to a simple bounce model. The focus of the investigation is centered on what happens across the fan, and then consequently what conditions result at the downstream mixing planes. Key factors from the ice crystal cloud that affect engine icing are the amount of ice that enters the core and the size of the ice particles. The analysis focusses on an operating condition (Flight 850) which produced an icing event during testing in the Propulsion Systems Lab (PSL) at NASA Glenn Research Center. For a given flow solution, several options within LEWICE3D were exercised. First, a simple ice particle bounce model (IREBOUND=1) was used for particle sizes of 5, 20, and 100 microns. Then, the newly implemented particle breakup model (IREBOUND=4) was used for particles sizes of 20 and 100 microns. Each of these cases used only a single particle size at the inlet. Initial post-processing of the results includes contour plots of collection efficiency (beta) and average particle size (diamavg). For the simple bounce cases, a distinct migration of the ice particles to the outer casing of the core mixing planes is observed for the 20 and 100 micron cases. The 5 micron, simple bounce case, remains more uniformly distributed. For the particle breakup cases, a shadow region is observed near the inner hub but a majority of the radial extent receives significant amounts of ice particles. The increase in radial extent of significant levels of ice particles can be explained by the generation of many small particle due to impact with the spinner and the fan. It is observed that for both the 20 and 100 micron cases, the particle breakup model produces average particle sizes of about 10 micron entering the tandem exit guide vanes (EGVs). When the particle breakup model was employed, for either 20 micron or 100 micron particles entering the engine, small particles with melt fraction on the order of 20% were observed at the known location of icing risk. In fact, the 20 micron breakup case, the 100 micron breakup case, and the 5 micron elastic bounce case all created similar results at the icing risk location. These observations corroborate the hypothesis that an experiment that cannot produce ice particles as large as those seen in the atmosphere, can still simulate an icing event of interest and data relevant to that condition.