Generalizing Circular Brush Seal Leakage Through a Randomly Distributed Bristle Bed

Abstract

Brush seals have established a niche in the gas-to-gas sealing against leakage in modern turbine engines. The variable nature of the brush during operation makes leakage prediction difficult. A simple semi-empirical model based on an effective brush thickness parameter has been successfully used to correlate and predict brush seal leakage in engine environments. The model was extended to correlate a range of brush densities using a physically realistic brush thickness. Later, the model was based on mean diametric brush properties for a large range of circular brush seal geometries. However, the best basis for modeling bristle, distribution was unknown. This paper proposes a solution to the distribution problem by assuming a randomly distributed bristle bed. A random distribution leads to a rectangular array model that is supported by the quality of leakage data generalization. Applying the resultant effective thickness parameter to predict brush seal performance in turbine engines is discussed.