Search

An experimental investigation of boundary layer transition in a multi-stage turbine has been completed using surface-mounted hot-film sensors. Tests were carried out using the two-stage Low Speed Research Turbine of the Aerodynamics Research Laboratory of GE Aircraft Engines. Blading in this facility models current, state-of-the-art low pressure turbine configurations. The instrumentation technique involved arrays of densely-packed hot-film sensors on the surfaces of second stage rotor and nozzle blades. The arrays were located at mid-span on both the suction and pressure surfaces. Boundary layer measurements were acquired over a complete range of relevant Reynolds numbers. Data acquisition capabilities provided means for detailed data interrogation in both time and frequency domains. Data indicate that significant regions of laminar and transitional boundary layer flow exist on the rotor and nozzle suction surfaces. Evidence of relaminarization both near the leading edge of the suction surface and along much of the pressure surface was observed. Measurements also reveal the nature of the turbulent bursts occuring within and between the wake segments convecting through the blade row. The complex character of boundary layer transition resulting from flow unsteadiness due to nozzle/nozzle, rotor/nozzle, and nozzle/rotor wake interactions are elucidated using these data. These measurements underscore the need to provide turbomachinery designers with models of boundary layer transition to facilitate accurate prediction of aerodynamic loss and heat transfer.

Experimental data from jet-engine tests have indicated that unsteady blade row interactions and separation can have a significant impact on the efficiency of low-pressure turbine stages. Measured turbine efficiencies at takeoff can be as much as two points higher than those at cruise conditions. Several recent studies have revealed that Reynolds number effects may contribute to the lower efficiencies at cruise conditions. In the current study numerical simulations have been performed to study the boundary layer development in a two-stage low-pressure turbine, and to evaluate the transition models available for low Reynolds number flows in turbomachinery. The results of the simulations have been compared with experimental data, including airfoil loadings and integral boundary layer quantities. The predicted unsteady results display similar trends to the experimental data, but significantly overestimate the amplitude of the unsteadiness. The time-averaged results show close agreement with the experimental data.

Experimental data from jet-engine tests have indicated that unsteady blade row (wake) interactions and separation can have a significant impact on the efficiency of turbine stages. The effects of these interactions can be intensified in low-pressure turbine stages because of the low Reynolds number operating environment. Measured turbine efficiencies at takeoff can be as much as two points higher than those at cruise conditions. Thus, during the last decade a significant amount of effort has been put into determining the effects of transition and turbulence on the performance of low pressure turbine stages. Experimental investigations have been performed, for example, by Hodson et al. and Halstead et al. These investigations have helped identify/clarify the roles that factors such as the Reynolds number, free stream turbulence intensity, pressure gradient and curvature have in the generation of losses. In parallel to the experimental investigations, there have been significant analytical efforts to improve the modeling of transition. Examples of such efforts include the works of Mayle and Gostelow et al. These newer models show promise of providing accurate transition predictions over a wide range of flow conditions, although they have yet to be implemented into the numerical flow analyses used by the turbine design community. Some recent computational investigations of interest include the works of Chernobrovkin and Lakshminarayana and Eulitz and Engel. The focus of the current effort has been to -use a viscous, unsteady quasi-three-dimensional Navier-Stokes analysis to study boundary layer development in a two-stage low-pressure turbine. A two-layer algebraic turbulence model, along with a natural transition model and a bubble transition model, have been used, The geometry used in the simulations has been the subject of extensive experiments. The predicted results have been compared with experimental data, including airfoil loadings and time-averaged/unsteady integral boundary layer quantities.