ScholarWorks@경상국립대학교

초록

Physical models and simulation methods continue to be developed for describing in-flight icing in aircraft, rotorcraft, and engines. This chapter presents a numerical simulation methodology and meta modeling for a rotorcraft electrothermal ice protection system. A Navier-Stokes-Fourier computational fluid dynamics code is used to simulate the flow around the air intake of a rotorcraft engine. Droplet trajectory and ice accretion codes are used to calculate the collection efficiency and ice accretion on the anti-iced surfaces of the engine air intake. The computational predictions are then validated for ice accretion on the surface of the electrothermal anti-icing system in heat-off and heat-on modes based on an icing wind tunnel study. To evaluate performance, a meta model is developed based on proper orthogonal decomposition and a general regression neural network that covers the entire envelope of icing conditions. A feasibility study is also conducted to assess the meta model, using a leave-one-out cross-validation method. The module is updated using a self-organizing map that contains k-mean clustering for the new samples in the parametric space. The practical use and high potential of the meta model for determining the optimal power and size of the heat pads are finally demonstrated. © Springer Nature Switzerland AG 2024. All rights reserved.