Process Optimization for Numerical Simulation of the Manufacturing Process of Brush Seals and Comparison between Frictional and Frictionless Contact Modeling

Abstract

The aviation industry's pursuit of net-zero emissions by 2050 necessitates significant im- provements in turbomachinery efficiency. Brush seals, which offer leakage reductions of up to 90% compared to conventional labyrinth seals, are a critical enabling technology. How- ever, numerical modeling of these seals remains challenging due to the complex interaction of thousands of flexible bristles, contact friction, and manufacturing-induced pre-stress. Current state-of-the-art models often rely on idealized hexagonal bristle arrays or porous media approximations that neglect the specific mechanical history of the seal. This thesis addresses this gap by developing an automated, script-based computational framework to simulate the manufacturing process of a brush seal. The workflow integrates Python, Siemens NX, and ANSYS Workbench to generate realistic, unstructured bristle geometries and simulate the sequential clamping and axial inclination processes. A two- stage solver strategy was implemented, utilizing an implicit solver (ANSYS Mechanical) for the force-controlled clamping phase and an explicit solver (LS-DYNA) for the large- rotation inclination phase. A comparative study was conducted to determine the necessity of modeling contact fric- tion during these manufacturing stages. High-fidelity frictional simulations (p = 0.1) were compared against idealized frictionless baselines (μ = 0.0). The results reveal a dual na- ture of frictional influence. For global geometric parameters, friction is negligible; bristle tip positions agree to approximately 1% of the bristle diameter, indicating that a friction- less approximation is sufficient for defining Fluid-Structure Interaction (FSI) domains. Conversely, for structural integrity, friction is dominant. The frictional model captured a critical "wall-sticking" phenomenon at the backing plate, resulting in localized stress amplifications of 153% and contact pressure increases of over 1000% compared to the fric- tionless case. Furthermore, friction reduced the root packing density by approximately 6% due to locking effects. This work concludes that while frictionless models are computationally efficient and geo- metrically valid for CFD boundary definition, the inclusion of frictional contact mechanics is mandatory for accurate fatigue life prediction and structural safety assessment