Flow visualization and leakage measurements of worn labyrinth seals

Abstract

A large-scale flow visualization test facility is used to conduct an experimental investigation into the leakage resistance and flow characteristics of worn labyrinth seals. Wear in labyrinth seals is a consequence of contact between the rotating and stationary components of the seal, due to thermal expansion, centrifugal growth, and rotordynamic vibration. The test facility is a large-scale, 2-D planar design, using water as the working fluid. Externally adjustable seal elements, constructed from plexiglass, simulate both straight-through and stepped labyrinth seals. Plexiglass attachments are used to simulate worn teeth tips and rub-grooves in the opposing surface. Two flow visualization methods are used to digitally record flow patterns in the labyrinth seal, and selected images are extracted and enhanced using a computer. Both worn and unworn labyrinth seals are tested over a range of Reynolds numbers, step heights, clearances, and axial offset distances. For all small clearance geometries, the throughflow "pinching" effect, which causes a high speed throughflow and increased turbulence dissipation, heavily outweighs most throughflow trajectory effects in the labyrinth chamber. At larger clearances, other effects contribute to the performance of a seal. One effect is throughflow penetration into the rub-groove cavity of a worn seal, caused by the vena contracts as the throughflow approaches the tooth clearance. If rub-groove penetration occurs, the throughflow trajectory angle into the downstream labyrinth cavity is sometimes altered. The throughflow trajectory angle is another important effect, because it determines if the throughflow directly approaches the downstream tooth, leading to carry-over, or if it approaches at a large angle, giving a larger vena contracts velocity. Another important effect is the presence of large, highspeed recirculation zones in the labyrinth cavity, which lead to greater turbulence. In nearly all cases, a geometry with worn teeth tips performs worse than its counterpart with unworn teeth tips, due to the nozzling effect of the worn teeth. Similarly, for nearly all medium and large clearance cases, a geometry with a rub-groove performed worse than its unworn counterpart. Finally, in general, a seal's leakage resistance coefficient is decreased by increasing the width of the rub-groove.