This work analyzed the performance of two foot pad designs for the compliant finger seal. The double wedge and Rayleigh step pad designs were tested numerically and experimentally. The numerical models were solved using ANSYS CFX and compared to the experimental results. A finite difference based two dimensional model of the fluid film under the pad solving the Reynolds equation coupled with the energy equation was also formulated.
The double wedge and Rayleigh step seal pad designs were analyzed using ANSYS CFX. Models matching the design drawings for the actual seal, and models with increased gap heights were created. The models consisted of one or two non-padded feet, padded feet, and the interstices between subsequent feet. Cases were run at rotational speeds of 2 krpm to 12 krpm in 2 krpm increments and at pressures of 5, 20, 25, and 30 psig, which matches the experimental cases. The small gap geometry was run as laminar compressible while the large gap geometry was run as turbulent compressible, and all were run as isothermal. The mass leakage was found to increase with increasing pressure for all models tested. For the double wedge seal and the Rayleigh step seal with the larger gap, the leakage did not change significantly with increasing rotation, but the leakage is affected by the rotation in the Rayleigh step seal with the original gap. Previously unseen compressibility effects, such as potential choking of the flow, were witnessed in the numerical models for the geometries with the larger gaps.
Both the double wedge and Rayleigh step seal pad designs were experimentally tested, and the experimental cases matched the parameters of the numerical cases. The results of the experiments on the double wedge seal indicated that as the high side pressure increased, the mass leakage remained relatively unchanged, and the mass leakage was found to decrease with increasing speed. The leakage of the Rayleigh step seal followed a similar trend as the double wedge seal as the high side pressure was increased, however, the leakage of the Rayleigh step seal increased slightly due to higher rotation. The temperature of the seal pads did not change significantly during testing for either seal pad design indicating there was no contact between the rotor and the seal.
The results of the numerical models and the experiment were compared. The mass leakage values of the experimental results were an order of magnitude higher than those of the numerical models with the small gap, and compared much better with the large gap model. The higher mass leakage results are likely due to the higher leakage due to the actual seal being warped, and the actual seal possibly having a larger gap height than the model. These issues make comparing the experimental and numeric results difficult.
A two dimensional Reynolds equation based solver was written to model the fluid film under a single foot pad. The Reynolds equation is coupled with the energy equation and solved using finite differences. The model was validated by comparing the results with those from a three dimensional Navier-Stokes based model run in ANSYS CFX. The Reynolds based model compared well with the Navier-Stokes model, both qualitatively and quantitatively. Thus, the Reynolds based model provides a fast, effective method of comparing foot pad designs.