| The brush seal is emerging as a new technology to effectively control cooling and leakage flows in gas turbine engines. With their superior leakage performance, they show the potential to replace current labyrinth seals in gas turbine engines. Because the bristles slide against the rotor surface, wear at the contact becomes a major concern as it determines the life and efficiency of the seal. To optimize seal life and efficiency, an in-depth study of the factors causing the seal stiffness is needed, and a good choice of materials must be made. Although considerable research has been done on material selection and tribopairs, a brief survey reveals the lack of reliable analyses to evaluate contact loads, and to address heat transfer issues. As material pairs have been optimized for most cases, understanding and management of contact loads hold the key for further improvements in seal life. The complicated nature of bristle behavior under various combinations of pressure load and rotor interference requires computer analysis to study details that may not be available through analytical formulations. In an attempt to meet this need, this study presents a 3-D finite element model of a brush seal. The model consists of a representative bristle bundle with a backing plate and a rotor surface. Every bristle is defined by a number of 3-D quadratic beam elements. Bristles are fixed at the top nodes, while they are free to move in any direction at the tip touching the rotor surface. The model consists of 10 to 13 circumferential rows of bristles. The number of rows are based on the actual packing thickness of the seal modeled. Unlike previous analytical studies on brush seal contact loads, this work includes nonlinear frictional effects between the bristles. Frictional effects are known to drastically change the seal behavior, and are crucial in determining the contact forces.; The model applies the available published experimental data to define the boundary conditions and to prescribe the loads. Typical results indicate that when a sample uniform pressure load is applied, the overhanging portion of the bristles deflects in the flow direction. Analysis predicts a large increase in the bristle-rotor contact load when a differential pressure is applied, indicating the effect of pressure on bristle stiffening. When rotor interference is relieved under a pressure load, simulations show the presence of a hysteresis effect leaving the bristles 'hung up' due to frictional forces.; Verification of the analysis is conducted through seal wear and stiffness experiments. The limited published experimental and analytical data are also compared. Overall, the results from the model show good agreement with experimental data. Agreement with other analytical data is also satisfactory.; Taking the analysis one step further, a series of simulations are performed to obtain closed-form relations for rotor contact force, bristle tip pressure and bristle stiffness with the presence of contact friction and pressure stiffness coupling. For systematic selection of major factors contributing to the bristle tip load, two-stage designed experiments are performed. These factors are related to the contact force in nonlinear response surface model. These model relations are then verified through a series of confirmation runs. To facilitate the ease of use by design engineers, these multi-factor relations are also presented. |
