Review of Ingress in Gas Turbines

Abstract

This review summarizes research concerned with the ingress of hot mainstream gas through the rim seals of gas turbines. It includes experimental, theoretical, and computational studies conducted by many institutions, and the ingress is classified as externally induced (EI), rotationally induced (RI), and combined ingress (CI). Although EI ingress (which is caused by the circumferential distribution of pressure created by the vanes and blades in the turbine annulus) occurs in all turbines, RI and CI ingress can be important at off-design conditions and for the inner seal of a double-seal geometry. For all three types of ingress, the equations from a simple orifice model are shown to be useful for relating the sealing effectiveness (and therefore the amount of hot gas ingested into the wheel-space of a turbine) to the sealing flow rate. In this paper, experimental data obtained from different research groups have been transformed into a consistent format and reviewed using the orifice model equations. Most of the published results for sealing effectiveness have been made using concentration measurements of a tracer gas (usually CO2) on the surface of the stator, and—for a large number of tests with single and double seals—the measured distributions of effectiveness with sealing flow rate are shown to be consistent with those predicted by the model. Although the flow through the rim seal can be treated as inviscid, the flow inside the wheel-space is controlled by the boundary layers on the rotor and stator. Using boundary-layer theory and the similarity between the transfer of mass and energy, a theoretical model has been developed to relate the adiabatic effectiveness on the rotor to the sealing effectiveness of the rim seal. Concentration measurements on the stator and infrared (IR) measurements on the rotor have confirmed that, even when ingress occurs, the sealing flow will help to protect the rotor from the effect of hot-gas ingestion. Despite the improved understanding of the “ingress problem,” there are still many unanswered questions to be addressed.