This paper considers three-dimensional (3D) computational fluid dynamics (CFD) and structural modeling of brush seals, and investigates the effects of inlet swirl on the bristle pack. The model couples aerodynamic forces generated by CFD to a structural model that includes interaction between bristles. At a critical value of inlet swirl, aerodynamic forces cause circumferential slip of the upstream bristle row. In practice, this may lead to instability of the bristle pack and is consistent with anecdotal reports of seal behavior. The critical swirl velocity was reduced when the downstream pressure level was raised, keeping the same upstream total to downstream static pressure difference. This is caused by the increased dynamic head associated with the inlet swirl. Inclusion of a front plate in the seal design does not offer the intended protection to the bristle pack in highly swirling environments. This is associated with highly swirling flow impinging on the bristle tips. Fitting of roughness elements on the upstream face of the front plate could improve stability by reducing swirl of the flow impacting on the bristles. Increasing the bristle diameter and bristle stiffness does not necessarily prevent slip at higher inlet swirl velocities, but reduces the magnitude of slip of the upstream bristles.
Brush seals offer a superior sealing effectiveness compared to labyrinth seals. However, widespread use of brush seals is constrained by deleterious behaviors such as pressure-stiffening and hysteresis. For the latter, the bristles bend during the shaft incursion process and do not fully recover during the shaft retraction process. An opening gap is created, which increases seal leakage unless the pressure load drops to a certain level. In this work, analytical and numerical models based on a single bristle are proposed to capture the seal's response to shaft displacement with and without pressure loading. The models are validated using static stiffness tests at an unpressurized condition from literature. The main results show that modeling of the backing ring friction is essential to capture the bristle hang-up behavior. Shaft friction dominates at unpressurized conditions, while backing ring friction dominates at high pressure loading. An expression for shaft hang-up displacement has been derived. A sensitivity study shows that seals with shallow lay angle, short bristle length, and large bristle diameter are less prone to hang-up problems. The models developed in the present framework have been shown to qualitatively capture the pressure stiffening, hysteresis, bristle hang-up, and shaft rotation effects.