Labyrinth Seal and Pocket Damper Seal High Pressure Rotordynamic Test Data

Abstract

The current centrifugal compressor design for the oil & gas market is more and more challenging, and the presence of many competitors is pushing technology towards both a casing size reduction and a rotational speed increase. The first point is leading to an increase in the number of wheels per rotor (to do the same service), and the second point is forcing to cross two or even three rotor modes (hence a higher control of rotor damping is necessary). The two points together are leading to increase the rotor “flexibility ratio” (defined as the ratio between the maximum continuous speed and the first critical speed at infinite support stiffness according to API standard, and finally the rotordynamic stability is very much challenged. The centrifugal compressor's rotordynamic stability is strongly related to the internal seals' dynamic behavior, and for this reason, the authors' company decided several years ago to develop internally a high pressure seal test rig to measure internal seals stiffness and damping. The rig is now in operation, and in a previous paper the authors described its main capabilities, the applied identification procedures, and the preliminary test results captured for a long labyrinth seal (smooth rotor, straight toothed stator) tested up to 200 bar. This paper is intended to show more data for the same long Laby with special focus on some peculiar test as negative preswirl test, single frequency versus multifrequency test, offset versus centered seal test. The negative preswirl test shows a drastic change in the effective damping (from destabilizing to stabilizing) and provides a support in favor of the selection of swirl reversal devices at seals upstream. The multifrequency excitation test approach (based on the concurrent presence of several frequencies not multiples at each other) is compared with a single frequency excitation providing similar results and thus confirming the soundness of the multiple effects linear superimposition assumption. The effect of a static offset (simulating the real position of a rotor inside an annular seal) is also investigated proving that the relevant impact is negligible within the range of eccentricity explored (10% of seal clearance). Moreover, a pocket damper seal (PDS) with the same nominal diameter, clearance, and effective length has been tested (up to 300 bar) and compared with the Laby. As expected, the PDS shows both a higher effective stiffness and damping at the same test conditions, so the promising results already collected in a previous test campaign which was performed on a smaller scale and lower pressure test rig were mostly confirmed with the only exception for the effective damping crossover frequency which was lower than expected.