This paper reports numerical predictions and measurements of the flow field in a stepped labyrinth seal. The theoretical work and the experimental investigations were successfully combined to gain a comprehensive understanding of the flow patterns existing in such elements. In order to identify the influence of the honeycomb structure, a smooth stator as well as a seal configuration with a honeycomb facing mounted on the stator wall were investigated. The seal geometry is representative of typical three-step labyrinth seals of modern aero engines. The flow field was predicted using a commercial finite volume code with the standard k-ε turbulence model. The computational grid includes the basic seal geometry as well as the three-dimensional honeycomb structures.

This paper reports numerical predictions and measurements of the flow field in a stepped labyrinth seal. The theoretical work and the experimental investigations were successfully combined to gain a comprehensive understanding of the flow patterns existing in such elements. In order to identify the influence of the honeycomb structure, a smooth stator as well as a seal configuration with a honeycomb facing mounted on the stator wall were investigated. The seal geometry is representative of typical three-step labyrinth seals of modern aero engines. The flow field was predicted using a commercial finite-volume code with the standard-k-ε-turbulence model. The computational grid includes the basic seal geometry as well as the three-dimensional honeycomb structures.

with real and imaginary corrections for anomalous dispersion in the form factor of the chromium atom from D. T. Cromer and D.

The influence of a honeycomb facing on the heat transfer of a stepped labyrinth seal with geometry typical for modern jet engines was investigated. Heat transfer measurements were obtained for both a smooth stator and a stator lined with a honeycomb structure. In addition, an LDV system was used with the scaled up geometry to obtain a high local resolution of the velocity distribution in the seal. The experiments covered a wide range of pressure ratios and gap widths, typical for engine operating conditions. Local heat transfer coefficients were calculated from the measured wall and gas temperatures using a finite element code. By averaging the local values, mean heat transfer coefficients were determined and correlations for the global Nusselt numbers were derived for the stator and the rotor. The LDV results showed strong geometrical effects of the honeycomb structure on the development of the flow fields for the honeycomb seal. The distribution of the local heat transfer coefficients are compatible with the flow features identified by the LDV results and reveal a significantly reduced heat transfer with the honeycomb facing compared to the smooth facing.

The crystal structure of gismondite, a zeolite with narrow channels, has been refined using about 2200 F 0b8 gathered with a single-crystal diffractometer. All H 2 O molecules were located. The anisotropic refinement included determination of the Si/Al distribution and of site occupancies for the cation and 6 H 2 O equipoints. Although the Si and Al atoms are nearly ordered, the results suggest different arrange ments of H 2 O in otherwise identical cages. The final R w = 4.0% (conventional R = 5.5% ). 'Tphe crystal structure of natural gismondite from Hohenberg near Buehne/Westfalia was deduced by Fischer (4). The intensity data as well as the structure appeared to be suitable for a detailed refinement. Redetermination of the cell constants gave a = 9.843 ± 0.015; b = 10.023 ± 0.003; c = 10.616 ± 0.005 Α; γ = 92°25' ± 15' (first setting, space group V2 t /a). To obtain the Si/ΑΙ distribution as well as physi cally meaningful r.m.s. vibrational amplitudes, the set of Cu-K« intensities used for solving the structure had to be extended into the range of Mo-K« radiation (see Fischer, 5). Of the ca. 16,000 asymmetric reflections out side the Cu-Κα but within the Mo-K« range, about 250 were selected for measurement. After an intermediate stage of the refinement with copper data alone, structure factors and their contributions from the Si and Al atoms were computed for the Mo-Κα range. Reflections with both large F C ai and high contributions from the tetrahedral atoms were measured. Additional measurements were done with Mo-K<* radiation for about 50 reflections in the copper sphere for precise scaling. Furthermore, some 200 reflections were remeasured with Cu-K« radiation. All intensity data were gathered with a single-crystal diffractometer of the equi-inclination type (2, 6). The same crystal was used as in Ref. 4 (see also 9). The absorption correction employed a method patterned after Weber 250 Downloaded by UNIV OF CALIFORNIA SAN FRANCISCO on February 18, 2015 |