An Investigation of Rim Seal/Disk Cavity Flow and its Interaction with High Pressure Turbine Rotor Flows

Open Access
Author:
Town, Jason Edward
Graduate Program:
Aerospace Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
March 19, 2015
Committee Members:
  • Cengiz Camci, Dissertation Advisor
  • Savas Yavuzkurt, Committee Chair
  • Horacio Perez Blanco, Committee Member
  • Dennis K Mclaughlin, Committee Member
Keywords:
  • Rim Seal Cavity
  • Gas Turbine
  • Secondary Flow
  • Rotor Blade Tip Treatments
  • Rim Seal Cavity Purge
Abstract:
Ingestion and egression of hot gas path fluid into the rim seal cavity of an axial flow turbine is usually mitigated by the use of purge flow. The purge flow arrives into a disk cavity space from the final stage of the compressor and is very expensive in terms of total system efficiency. Cycle analysis for a gas turbine clearly indicates the significant efficiency penalty for the use of high pressure compressor air for cooling or disk cavity space purge action. Furthermore, the purge flow modifies flow structures inside and downstream of the cavity where the purge flow mixes with hot mainstream gasses. This thesis provides an understanding of the modes of ingestion and egression of the rim seal cavity. A review of previous studies pertaining to the effects of purge ingress/egress and methods of mitigation is provided. New testing in a simulated disk cavity space is performed at the Axial Flow Turbine Research Facility (AFTRF) using a representative, modern High Pressure Turbine (HPT) stage. A complete set of aerodynamic instrumentation is designed, assembled and validated for shakedown testing of the AFTRF. A new aerodynamic probe management system allowing for adaptive gridding in the intraspace and rotor exit frames is designed, built, and validated. This system, by moving more efficiently, increases the number of measurement locations taken in a two hour run from 336 points to 868. Using this measurement system, total pressure measurements from a Kiel probe and a Five Hole Probe downstream of the Nozzle Guide Vane (NGV) are shown to agree with each other. In the rotating frame of the AFTRF-HPT, a new 32 channel electronic pressure measurement system is implemented to measure vane and blade loadings. Vane loadings have a variation of approximately 1% from run to run and agree well with computational results. A correction for blade loading in the relative frame of reference is applied and results have a variation of 1% from run to run and agree reasonably well with predictions. Dynamic measurements of total pressure at the rotor exit, which are phase-locked with the rotor position, allow inspection of the unique flow structures generated by each passage. The next phase of testing includes varying purge rates of the rim seal cavity, modification of the rim seal cavity, changing the rotor blades to a design to control the tip leakage, and comparing experimental measurements and computational predictions of unsteady structures within the rim seal cavity. It is found that the ‘Baseline Blades’ preform less well than the ‘Tip Vortex Control (TVC) Blades’. Two tip treatments are tested, smooth and slotted, slotted tip treatments are beneficial to the ‘Baseline Blades’ and detrimental to the ‘TVC Blades’ when compared to the smooth tip treatments. Of the two rim seal cavity designs, the modified design used with the ‘TVC Blades’ with reduced radial clearance reduces the penetration and momentum deficit in the region of the hub endwall passage vortex. Experimentally measured unsteady structures within the rim seal cavity are measured using fast response aerodynamic piezoresistive transducers. They are then compared to an unsteady transient eight vane, ten blade computations. A method to extract the speed at which these structures are moving and the number of structures present is given for both the experimental and computational ways. Within the rim seal cavity 15 unsteady pressure cells moving at 77.5% of the rotor speed are experimentally measured and though the computational methods 14.5 unsteady pressure cells are moving at 81.7% rotor speed.