In non-contact annular labyrinth seals used in turbomachinery, the fluid pre-rotation in the direction of shaft rotation effectively increases the fluid velocity in the circumferential direction and could generate destabilizing fluid forces exerting on the rotor. Swirl brakes are typically employed to reduce the fluid pre-rotation at the inlet of the seal. The inlet flow separates as it follows the swirl brakes, and the ratio between the tangential component of the velocity in the seal and the velocity of the rotor surface varies consequently. Effective swirl brakes can significantly suppress the destabilizing fluid forces as it is effectively reduces the tangential velocity. In literature, it was shown that leakage rate can also be reduced by using swirl brakes with “negative-swirl”.
In this study, a labyrinth seal with inlet swirl brakes is selected as the baseline design and seal performance is evaluated using ANSYS-CFX. The design of experiments (DOE) approach is used to investigate the effects of various design variables on the seal performance. The design space consists of swirl brake’s length, width, curvature at the ends, the tilt angle, as well as the number of swirl brakes in the circumferential direction. Simple random DOE sampling method with Euclidean distances for the design matrix is used to generate the design points. The steady-state CFD simulations are then performed for each design point to analyze the performance of the swirl brakes. The quadratic polynomial fitting is used to evaluate the sensitivity of the average circumferential velocity with respect to the design variables, which gives a qualitative estimation for the performance of the swirl brakes. The results provide a better understanding of which design variables are critical and more effective in reduction of the destabilizing forces acting on the rotor, and thus will support the swirl brake design for annular pressure seals.