03.1 - Aerodynamics – CFD Methods and Validation

LEAKAGE PREDICTION AND GEOMETRIC PARAMETER OPTIMIZATION OF STEPPED LABYRINTH SEALS BASED ON DATA-DRIVEN SURROGATE MODELS

X. Li¹, L. Yang¹, Y. Wei¹; ¹Beijing Jiaotong University, China

This study compares two surrogate-based optimization approaches for aero-engine labyrinth seals: an efficient Orthogonal Experimental Design with Response Surface Model and a data-intensive Latin Hypercube Sampling with Deep Neural Network. The trade-off between scomputational cost and predictive robustness is quantified, providing a practical decision framework for designers.

View Paper

DOI: 10.71945/icas2026_0862