Investigation on the Leakage Flow, Windage Heating and Swirl Development of Rotating Labyrinth Seal in a Compressor Stator Well
What make the labyrinth seal in a compressor stator well different from the normal labyrinth seal are the inlet and outlet rotor-stator disc cavities. Due to the presence of rotating disc cavities, the windage heating and the swirl development are remarkable, which can have a great influence on the leakage characteristic. Besides, when compressor operates at different speeds, the rotor and stator grow differently owing to centrifugal expansion and thermal expansion. Hence the tip clearance which determines the leakage mass flow changes with the varying of rotational speed and temperature in the stator well. A rotating test rig with rotational speed 8100rpm and pressure ratio range 1.05∼1.3 was designed for the test of labyrinth seal in a compressor stator well. A cantilevered structure was used to entirely collect the mass flow for an accurate measurement. To know the working tip clearance precisely, the set up tip clearance was measured with plug gauges, while the radial displacements of rotating disc and stationary casing were measured separately with two high precision laser distance sensors. The total temperatures of airflow in the stator well were measured with thermocouples to analyze the proportion of windage heating among the inlet rotating disc cavity, outlet rotating disc cavity, and labyrinth seal segment. The disc and stator casing were manufactured with non-metallic materials to reduce heat dissipation. Furthermore, the circumferential velocity of the leakage flow was measured using probes to reveal the swirl development. Two-dimensional, axisymmetric swirl flow numerical simulations were carried out to provide insight into the flow field details, total temperature variation and swirl flow development in the stator well. The numerical results of discharge coefficient, windage heating and swirl ratio were compared with the experimental data. Of particular note is, the tip clearance of numerical model at a specific rotating speed was set to be the same with the actual working clearance which was measured in the experiment. The inlet and outlet parameters corresponded with the experimental conditions also.