Effects of Turbulence and Solidity on the Boundary Layer Development in a Low Pressure Turbine

Author(s):  
W. J. Solomon

Multiple-element surface hot-film instrumentation has been used to investigate boundary layer development in the 2 stage Low Speed Research Turbine (LSRT). Measurements from instrumentation located along the suction surface of the second stage nozzle at mid-span are presented. These results contrast the unsteady, wake-induced boundary layer transition behaviour for various turbine configurations. The boundary layer development on two new turbine blading configurations with identical design vector diagrams but substantially different loading levels are compared with a previously published result. For the conventional loading (Zweifel coefficient) designs, the boundary layer transition occurred without laminar separation. At reduced solidity, wake-induced transition started upstream of a laminar separation line and an intermittent separation bubble developed between the wake-influenced areas. A turbulence grid was installed upstream of the LSRT turbine inlet to increase the turbulence level from about 1% for clean-inlet to about 5% with the grid. The effect of turbulence on the transition onset location was smaller for the reduced solidity design than the baseline. At the high turbulence level, the amplitude of the streamwise fluctuation of the wake-induced transition onset point was reduced considerably. By clocking the first stage nozzle row relative to the second, the alignment of the wake-street from the first stage nozzle with the suction surface of the second stage nozzle was varied. At particular wake clocking alignments, the periodicity of wake induced transition was almost completely eliminated.

2004 ◽  
Vol 126 (4) ◽  
pp. 663-676 ◽  
Author(s):  
M. T. Schobeiri ◽  
B. O¨ztu¨rk

The paper experimentally studies the effects of periodic unsteady wake flow on boundary layer development, separation and reattachment along the suction surface of a low pressure turbine blade. The experimental investigations were performed on a large scale, subsonic unsteady turbine cascade research facility at the Turbomachinery Performance and Flow Research Laboratory (TPFL), Texas A&M University. The experiments were carried out at a Reynolds number of 110,000 (based on suction surface length and exit velocity) with a free-stream turbulence intensity of 1.9%. One steady and two different unsteady inlet flow conditions with the corresponding passing frequencies, wake velocities, and turbulence intensities were investigated. The reduced frequencies cover the entire operating range of LP turbines. In addition to the unsteady boundary layer measurements, blade surface measurements were performed at the same Reynolds number. The surface pressure measurements were also carried out at one steady and two periodic unsteady inlet flow conditions. The results presented in ensemble-averaged, and the contour plot forms help to understand the physics of the separation phenomenon under periodic unsteady wake flow. It was found that the suction surface displayed a strong separation bubble for these three different reduced frequencies. For each condition, the locations and the heights defining the separation bubble were determined by carefully analyzing and examining the pressure and the mean velocity profile data. The location of boundary layer separation was independent of the reduced frequency level. However, the extent of the separation was strongly dependent on the reduced frequency level. Once the unsteady wake started to penetrate into the separation bubble, the turbulent spot produced in the wake paths caused a reduction of the separation bubble height.


Author(s):  
M. T. Schobeiri ◽  
B. O¨ztu¨rk

The paper experimentally studies the effects of periodic unsteady wake flow on boundary layer development, separation and re-attachment along the suction surface of a low pressure turbine blade. The experimental investigations were performed on a large scale, subsonic unsteady turbine cascade research facility at Turbomachinery Performance and Flow Research Laboratory (TPFL), Texas A&M University. The experiments were carried out at a Reynolds number of 110,000 (based on suction surface length and exit velocity) with a free-stream turbulence intensity of 1.9%. One steady and two different unsteady inlet flow conditions with the corresponding passing frequencies, wake velocities, and turbulence intensities were investigated. The reduced frequencies cover the entire operating range of LP turbines. In addition to the unsteady boundary layer measurements, blade surface measurements were performed at the same Reynolds number. The surface pressure measurements were also carried out at one steady and two periodic unsteady inlet flow conditions. The results presented in ensemble-averaged, and the contour plot forms help to understand the physics of the separation phenomenon under periodic unsteady wake flow. It was found that the suction surface displayed a strong separation bubble for these three different reduced frequencies. For each condition, the locations and the heights defining the separation bubble were determined by carefully analyzing and examining the pressure and the mean velocity profile data. The location of boundary layer separation was independent of the reduced frequency level. However, the extent of the separation was strongly dependent on the reduced frequency level. Once the unsteady wake started to penetrate into the separation bubble, the turbulent spot produced in the wake paths caused a reduction of the separation bubble height.


1985 ◽  
Vol 107 (1) ◽  
pp. 54-59 ◽  
Author(s):  
K. Rued ◽  
S. Wittig

Heat transfer and boundary layer measurements were derived from flows over a cooled flat plate with various free-stream turbulence intensities (Tu = 1.6–11 percent), favorable pressure gradients (k = νe/ue2•due/dx = 0÷6•10−6) and cooling intensities (Tw/Te = 1.0–0.53). Special interest is directed towards the effects of the dominant parameters, including the influence on laminar to turbulent boundary layer transition. It is shown, that free-stream turbulence and pressure gradients are of primary importance. The increase of heat transfer due to wall cooling can be explained primarily by property variations as transition, and the influence of free-stream parameters are not affected.


Author(s):  
Axel Heidecke ◽  
Bernd Stoffel

This paper presents the results of a numerical investigation of a 1.5-stage low pressure turbine. The main focus of the numerical work was the prediction of the stator-2 boundary layer development under the influence of the stator stator clocking. The turbine profile used for the examination is a so called high-lift-profile and was designed for a laminar-turbulent transition over a steady separation bubble. The boundary conditions were defined by the 1.5-stage test turbine located at our laboratory, where also the measurement data was derived from. The calculations were conducted with a two-dimensional Navier-Stokes solver using a finite volume discretisation scheme. The higher level turbulence models v′2-f and the LCL-turbulence model, which are capable to predict boundary layer transition were compared with measurement data at midspan.


Author(s):  
Benjamin Winhart ◽  
Martin Sinkwitz ◽  
Andreas Schramm ◽  
Pascal Post ◽  
Francesca di Mare

Abstract In the proposed paper the transient interaction between periodic incoming wakes and the laminar separation bubble located on the rear suction surface of a typical, highly loaded LPT blade is investigated by means of highly resolved large-eddy simulations. An annular, large scale, 1.5-stage LPT test-rig, equipped with a modified T106 turbine blading and an upstream rotating vortex generator is considered and the numerical predictions are compared against hot film array measurements. In order to accurately assess both baseline transition and wake impact, simulations were conducted with unperturbed and periodically perturbed inflow conditions. Main mechanisms of transition and wake-boundary layer interaction are investigated utilizing a frequency-time domain analysis. Finally visualizations of the main flow structures and shear layer instabilities are provided utilizing the q-criterion as well as the finite-time Lyapunov exponent.


Author(s):  
V Michelassi

The transonic turbulent compressible flow in channels and turbine linear cascades is computed by using a Navier-Stokes solver. Turbulence effects are simulated by means of the k-ω turbulence model. A realiability constraint is introduced to improve the turbulence model performances and stability in the presence of stagnation points. In both the flow over the bump and the turbine blade, the shock induces a flow separation that affects the boundary layer development. In both cases the proposed model succeeds in predicting the flow separation. For the flow over the turbine blade a simple transition model based on integral parameters is introduced to mimic the effect of the boundary layer transition across the shock wave on the suction side. Relaminarization is also properly predicted on the pressure side, thereby allowing a good description of the boundary layer development and shock pattern.


Author(s):  
Antoine Ducoin ◽  
Jacques Andre´ Astolfi ◽  
Marie-Laure Gobert

In this paper, we investigate through an experimental approach the laminar to turbulent transition in the boundary-layer flow along a hydrofoil at a Reynolds number of 7.5 × 105, together with the vibrations of the hydrofoil induced by the transition. The latter is caused by a Laminar Separation Bubble (LSB) resulting from a laminar separation of the boundary-layer. The experiments, conducted in the hydrodynamic tunnel of the Research Institute of the French Naval Academy, are based on wall pressure and flow velocity measurements along a rigid hydrofoil, which enable a characterization of the Laminar Separation Bubble and the identification of a vortex shedding at a given frequency. Vibrations measurements are then carried out on a flexible hydrofoil in the same operating conditions. The results indicate that the boundary-layer transition induces important vibrations, whose characteristics in terms of frequency and amplitude depend on the vortex shedding frequency, and can be coupled with natural frequencies.


Author(s):  
D. Keith Walters ◽  
James H. Leylek

This paper documents computational simulations of the flow over a modern, highly-loaded turbine vane, including boundary-layer transition. Accurate prediction of transition has traditionally been difficult for commonly available RANS-based turbulence models. The present simulations used an advanced version of a three-equation eddy viscosity model recently developed and documented by the current authors. The new model is an elliptic single-point method, developed based on considerations of the universal character of pre-transitional boundary layers that have recently been published in the open literature. Simulations were performed at an engine-realistic chord Reynolds number (2.3×105) and with varying freestream turbulence intensities of 0.6, 10, and 19.5%. Detailed comparisons are made within the developing boundary layer, on both the suction and pressure surfaces, between the simulations and high-fidelity experimental measurements that have been previously documented in the open literature. Comparison of both mean velocity and Reynolds stress profiles indicates that the new model shows potential for predicting boundary layer development, including development of pre-transitional fluctuations and subsequent breakdown to turbulence.


Author(s):  
Curt H. Liebert ◽  
Raymond E. Gaugler ◽  
Herbert J. Gladden

Convection cooled turbine vane metal wall temperatures experimentally obtained in a hot cascade for a given one-vane design were compared with wall temperatures calculated with TACT1 and STAN5 computer codes which incorporated various models for predicting laminar-to-turbulent boundary layer transition. Favorable comparisons on both vane surfaces were obtained at high Reynolds number with only one of these transition models. When other models were used, temperature differences between calculated and experimental data obtained at the high Reynolds number were as much as 14 percent in the separation bubble region of the pressure surface. On the suction surface and at lower Reynolds number, predictions and data unsatisfactorily differed by as much as 22 percent. Temperature differences of this magnitude can represent orders of magnitude error in blade life prediction.


Author(s):  
Nicole M. Wolgemuth ◽  
D. Keith Walters

This study analyzes the predicted flow over a NACA 0012 airfoil at varying angles of attack and three different Reynolds numbers. The ability of three different turbulence models to predict boundary layer separation and transition behavior is investigated. Particular interest is paid to prediction of the separation bubble that develops near the leading edge of the airfoil suction surface prior to stall. The FLUENT CFD solver was used to simulate turbulent airflow over the airfoil. The three turbulence models include the standard and realizable forms of the k-ε model, available in FLUENT, as well as a recently developed transition-sensitive k-ω model that was implemented into the solver using user-defined functions. By employing the new, transition-sensitive model, computed properties of the flow field were found to be closer to experimental data than results produced by utilizing built-in turbulence models. Most importantly, the new, transition-sensitive model predicts the occurrence of the separation bubble, which the other models are unable to predict. The new model also clearly reproduces the laminar, transitional, and turbulent flow that occurs over the airfoil.


Sign in / Sign up

Export Citation Format

Share Document