Optimization of Hole Geometry for Hot Air Anti-Icing Performance on Rotating Spinners

Optimizing hole geometry for rotating hot air anti-icing spinners is computationally expensive and requires balancing overall icing suppression against extreme local ice thickness risk. This study proposes an integrated framework combining an L9(34) orthogonal design, multi-physics simulations, a dual-branch neural-network surrogate, and Pareto-based multi-objective optimization. Row spacing, column spacing, and hole diameter were jointly optimized. Across nine configurations, the area-averaged film heating effectiveness ranged from 0.108 to 0.5988, while the maximum ice thickness reached 3.16 mm. The surrogate simultaneously predicted pointwise ice thickness and film heating effectiveness, achieving R2 values of 0.9751 and 0.8839, respectively. Minimizing maximum ice thickness and maximizing area-averaged film heating effectiveness revealed their trade-off, with the Pareto knee point yielding 0.863 mm and 0.451, respectively. The recommended configuration had a row spacing of 4.7187 mm, column spacing of 22.7646 mm, and hole diameter of 2.0639 mm. Validation under six icing conditions showed that maximum ice thickness decreased by 22.37–100% in the four baseline icing cases, and area-averaged film heating effectiveness increased by 54.45–68.01%, while compressor heat loss decreased by 12.2–16.6%. The proposed framework therefore provides an efficient method for rotating spinner anti-icing design.