Rotor purge flow cavity seals are used in gas turbine engines to prevent ingestion of the mainstream gas flow into the purge cavity. Ingestion into this cavity leads to an increase in the cavity air temperature and subsequently to the rotor disk and stator metal temperatures leading to higher thermal stresses and reduced disk and stator fatigue life. An over designed cavity seal with an excess amount of purge flow has the downside of increasing engine fuel consumption through reduced turbine efficiency. The opposite approach of strengthening the hardware to withstand the higher stress and temperatures would increase the weight of the propulsion system. Understanding how the purge flow cavity and cavity seals interact with the mainstream gas is important to producing a balanced design between weight, fuel consumption, efficiency, and fatigue life of surrounding hardware.
The main objective of this research was to perform an experimental and computational study of a one and one half stage high-pressure turbine installed at The Ohio State University Gas Turbine Laboratory Turbine Test Facility with emphasis on the rotor purge cavity. The rig housing the turbine stage incorporated many features found in a typical commercial high-pressure turbine such as a cooled high-pressure vane row with hub and shroud cooling, a downstream blade row followed by a downstream vane row, the ability to created elevated radial inlet temperature profiles using a combustor emulator, and a cooling supply line to the purge cavity. Multiple runs were performed to study the effects of cooling flows from both an aerodynamic and heat transfer perspective and incorporated instrumentation throughout the rig in order to capture time-accurate temperature, pressure, and heat flux measurements. The run matrix included cold rig configurations with no cooling flow, high-temperature uniform inlet profiles at the vane inlet for cases with and without cooling flows, and high-temperature radial inlet profiles with and without cooling flows. The computational study was performed using the Numeca FINE/Turbo code utilizing a multiple blade row model with both a steady and harmonic unsteady technique in order to simulate the physics of the experiments.
Comparisons between the data and the computational results were performed for five different operational conditions: cold inlet and no cooling flow run, an elevated radial inlet temperature profile with purge and without purge cooling, and an elevated flat inlet temperature profile with purge and without purge cooling flow. The solutions were found to match very well to the exit rake measurements, the leading edge blade profile in the rotating frame of reference, the time-averaged and time-accurate static pressure on the vane hub, the stationary cavity wall, and the rotating cavity. Time-average comparisons were shown for the thermocouples located on the blade platform and in the stationary and rotating side of the cavity. For the two radial inlet cases and for the cold inlet case, these comparisons showed very good agreement while the two flat inlet profile cases showed that the computational models general under-predicted the static temperature levels both on the rotor platform and in the cavity.